Note from the Editor:
The Federalist Society takes no positions on particular legal and public policy matters. Any expressions of opinion are those of the author. Whenever we publish an article that advocates for a particular position, as here, we offer links to other perspectives on the issue, including ones opposed to the position taken in the article. We also invite responses from our readers. To join the debate, please email us at email@example.com.
On September 27, 2018, the Supreme Court granted certiorari in Rimini Street, Inc. v. Oracle USA, Inc., to address “[w]hether the Copyright Act’s allowance of ‘full costs’ (17 U.S.C. § 505) to a prevailing party is limited to taxable costs under 28 U.S.C. §§ 1920 and 1821 . . . or also authorizes non-taxable costs . . . .” The case raises a textbook statutory interpretation issue, and it illuminates common pitfalls in construing statutes that are particularly tempting for textualists.
28 U.S.C. § 1920 (Section 1920) is “the general statute governing the taxation of costs in federal court.” The law allows federal courts to “tax as costs” a specific set of expenditures listed in the statute. In federal court, Section 1920 defines the “costs” awardable to a party, which are often distinct from “fees” and “expenses” that may also be awarded in some cases. For reasons discussed below, the Supreme Court has read Section 1920 to lay out the full scope of allowable costs, and “no statute will be construed as authorizing” additional costs beyond those enumerated therein “unless the statute refers explicitly” to them.
17 U.S.C. § 505 (Section 505), part of the Copyright Act of 1976 (but which dates back to 1831), provides that:
In any civil action under this title, the court in its discretion may allow the recovery of full costs by or against any party other than the United States or an officer thereof. . . . the court may also award a reasonable attorney’s fee to the prevailing party as part of the costs.
Notably for this case, this provision allows for the award of “full costs,” not just “costs.”
The question in this case is straightforward: does “full costs” in Section 505 expand the scope of awardable expenditures beyond what Section 1920 permits? The trial court in this case said “yes,” concluding that “full costs” means “all costs incurred in litigation,” not just those identified in Section 1920. Accordingly, the court awarded Oracle millions of dollars in “non-taxable costs,” including “litigation costs for expert witness fees.” The Ninth Circuit affirmed in relevant part. Following controlling circuit precedent, the panel reasoned that reading “full costs” as only those costs enumerated in Section 1920 “effectively reads the word ‘full’ out of the statute,” and that such a construction violates the canon against surplusage that requires a court to “give every word in a statute meaning.” Rimini Street challenges this reading, arguing that “full costs” only means costs awardable under Section 1920.
II. Reading Law: A “Holistic Endeavor”
At first blush, this case might seem simple, particularly to a textualist. Per the dictionary, “full” means “containing as much . . . as is possible” or “complete,” and “cost” means “the amount . . . for something” or “the outlay or expenditure . . . made to achieve an object.” Given courts’ duty “to give effect, if possible, to every . . . word of a statute,” the Ninth Circuit’s reasoning is superficially attractive: “full costs” permits recovery of “all costs incurred in litigation,” not just Section 1920 costs. From Oracle’s point of view, this is simply taking Section 505 to “mean what it says.”
As the Court’s grant of certiorari suggests, however, things are not so simple. “[S]tatutory interpretation,” of course, “is a ‘holistic endeavor,’” and “the words of a statute are not to be read in isolation.” Rather, terms in a statute must be “read together” with the rest of the law. The “plain meaning” of a given statutory provision must be ascertained not just from “the language itself,” but also from “the specific context in which that language is used, and the broader context of the statute as a whole.” This is a “fundamental principle,” not only of “statutory construction” but also “of language itself.” Thus, even reading for “plain meaning,” one cannot simply stop at the “dictionary definition of two isolated words” in trying to figure out what the law means; rather, statutory interpretation must be conducted in light of “the text and structure” of the law as a whole.
Additionally, “it is an established rule of law, that all acts in pari materia are to be taken together, as if they were one law.” Under this canon of statutory construction, statutes addressing the same subject matter generally should be read as if consisting of one law addressing the subject. Laws so related to one another thus constitute part of the “broader context” to be taken into consideration when ascertaining the plain meaning of statutory terms.
Another relevant principle is the “commonplace of statutory construction that the specific governs the general.” Accordingly, the Court will not read a general clause in one place in a way that undermines a carefully drawn statute elsewhere. This principle holds no matter how “inclusive may be the general language of a statute,” particularly when “Congress has enacted a comprehensive scheme and has deliberately targeted specific problems with specific solutions.”
In addition to the foregoing, it must be remembered that some statutes use terms of art, in which case the plain meaning—understood as the dictionary definition—is less relevant. Sometimes, for instance, Congress may use language that comes freighted with meaning due to prior constructions of that language. Other times, Congress may employ true terms of art that must be understood according to their technical, not plain, meaning. In either case, when Congress writes laws with such terms, “any attempt to break down the term into its constituent words is not apt to illuminate its meaning.”
III. The Law Means What It Says, But What Does It Say?
While “full costs” of course “means what it says,” as Oracle contends, dictionary definitions alone do not necessarily tell us what it says. Construing those words in the context of federal “costs” laws generally, and in light of the grammatical-historical background to the phrase “full costs,” it turns out that, as Rimini Street argues, “full costs” means all costs enumerated in Section 1920.
First, consider the larger statutory context. The Supreme Court has found that Section 1920 “embodies Congress’ considered choice as to the kinds of expenses that a federal court may tax as costs against the losing party.” As such, Section 1920 provides the default rule as to what costs are awardable (when costs are awardable), and that default can only be overcome by a clear indication by Congress that it intends something different in a certain case. As the Court has reasoned, if courts had “discretion to tax whatever costs may seem appropriate,” rather than just those enumerated in Section 1920, the law would “serve no role whatsoever” and be “superfluous.” By passing Section 1920, Congress “comprehensively addressed the taxation of fees for litigants’ witnesses.” “The comprehensive scope of the Act and the particularity with which it was drafted demonstrate[s] . . . that Congress meant to impose rigid controls on cost-shifting in federal courts,” controls that cannot be evaded “without plain evidence of congressional intent.”
The Court’s approach to costs statutes has been fairly consistent over time. For instance, in Henkel v. Chicago, St. Paul, Minneapolis & Omaha Railway Co., the Court addressed whether district courts had power “to allow expert witness fees” where a state statute would have permitted them in state court. The Court looked to Congress’ “[s]pecific provision as to the amounts payable and taxable as witness fees” over the years as evidence that “additional amounts paid as compensation, or fees, to expert witnesses cannot be allowed or taxed as costs in cases in the federal courts.” That is, Congress’ specific provision for particular costs means that Congress has not “extended any roving authority to the Judiciary” to award whatever costs it thinks appropriate in a given case. There must be clear statutory authorization for a court to allow additional costs beyond those enumerated in Section 1920.
Second, “costs” comes freighted with special meaning, as does “full costs.” Generally, Section 1920 defines the term “costs” in federal statutes unless the statute clearly indicates to the contrary. As such, “‘costs’ is a term of art” when it appears in fee-shifting statutes, and it refers to “the list set out in 28 U.S.C. § 1920.” Likewise, the phrase “full costs” comes loaded with meaning because of its grammatical-historical background in U.S. law. As Rimini Street points out, when the phrase “full costs” first appeared in U.S. copyright law, Congress was legislating against a background where a prevailing party might not be awarded any costs whatsoever. In response to this, Congress passed a law mandating that “full costs shall be allowed.” At the time, federal courts followed the forum state’s law with respect to costs, and so “full costs” in the copyright law effectively meant the total amount of whatever costs were allowed under state law, which could vary from state to state. In 1853, however, Congress enacted the Fee Act, “specifying in detail the nature and amount of the taxable items of cost in the federal courts” and providing that “no other compensation shall be allowed.” The Fee Act thus replaced the previously-controlling patchwork of state laws and became the relevant law defining the “full costs” awardable in federal court, including in cases involving federal copyright law. The language of “full costs” was retained over time in the copyright law, while the Fee Act laid the foundation for current Section 1920. Accordingly, “full costs” now means those costs enumerated in Section 1920, which details the total amount of costs allowable as a matter of course in federal court.
IV. But Mustn’t “Full” Mean “Full”?
Oracle, however, disagrees. In its briefing, Oracle focuses particularly on plain meaning and the canon against surplusage, but its arguments prove unavailing.
A. “Full” Is Not “Plain Evidence of Congressional Intent”
Oracle argues that the word “full” loosens the scope of “costs” to allow for recovery of expenses beyond those delineated in Section 1920. This argument runs up against Congress’ clear and specific provision for recovery of additional categories of expenses in excess of what Section 1920 allows when Congress intends for such additional expenses to be recoverable. These provisions show that Congress knows how to override the inherent limitations of Section 1920 when it so desires, meaning that language short of that should not be construed to the same effect. “[F]ull” does not provide the same “plain evidence of congressional intent” the Court has required to permit recovery of costs beyond what Section 1920 allows. “Full” is, at best, “ambiguous as to whether it includes costs beyond the scope of” Section 1920; such ambiguity cannot amount to the “clear intention” required to allow costs not enumerated in that statute.
In opposition to this conclusion, Oracle points out that Section 505 does not just allow for an award of “full costs,” but it also permits courts to “award a reasonable attorney’s fee to the prevailing party as part of the costs.” As Oracle explains, this provision for attorneys’ fees came in 1909, “after growing acceptance of the ‘American Rule’ had created some uncertainty about whether attorneys’ fees could be recovered as costs . . . .” According to Oracle, “Congress added the second sentence to the Copyright Act to clarify the continued availability of attorneys’ fees” as part of the “full costs” available under Section 505. In response to the retort that this provision would be superfluous if “full costs” already included attorneys’ fees, Oracle explains that “clarity” is not “superfluity.” That may be true, but it raises another problem for Oracle: Congress singled out attorneys’ fees to make them taxable as costs, but it did not do the same for expert witness fees, so the taxability of attorneys’ fees actually works against Oracle’s position. The 1909 amendment shows that Congress knows how to make its intention clear as to recoverable costs in Section 505 when it wants to permit costs beyond those allowed via Section 1920. Given that the circuit split at issue here is not new, Congress “could easily have” amended the statute again to indicate that expert witness fees are also recoverable as part of the costs. Congress has not done so, instead leaving Section 505 to allow only “full costs” and “a reasonable attorney’s fee.” The explicit grant of authority to award one type of fee “as part of the costs” implies that a court may not award any other type of fee beyond normally allowable costs.
B. The Limits of the Canon Against Surplusage
Oracle also argues that “full” is rendered superfluous if “full costs” simply means “the costs delineated in 28 U.S.C. § 1920.” But this argument fails for at least two reasons, plus Oracle’s reading creates a worse surplusage issue with respect to statutes enacted subsequent to the Copyright Act of 1976 that provide for recovery of “full costs” and also other kinds of litigation-related expenses.
First, assuming there is a surplusage problem, the Supreme Court has explicitly relaxed the force of the canon against surplusage in the context of costs statutes. In Marx v. General Revenue Corporation, for example, the Court rejected a surplusage argument against reading a costs statute in a certain way. The Marx Court reasoned that “[t]he canon against surplusage is not an absolute rule,” as “‘instances of surplusage are not unknown,’” and “redundancy is ‘hardly unusual’ in statutes addressing costs.” In fact, the Court here hinted that the canon is effectively of no use at all in the costs context since “a court has inherent power” to shift costs in some cases, meaning that there is “no need for Congress to specify that courts have this power” in those circumstances, making various federal statutes to this effect wholly superfluous.
Second, Oracle’s insistence on plain or ordinary meaning conflicts with its argument against surplusage. The adjective “full” quite frequently is redundant in its ordinary usage, serving only to emphasize or clarify meaning already inherently contained within the modified noun. For instance, imagine an automobile driver requesting that a passenger help defray transportation costs by buying fuel. There is no difference between asking the passenger to pay for “a tank of gas” or “a full tank of gas.” Either way, the passenger (if she is polite) is going to fill the tank. “Tank” alone means the same thing as “full tank.” The use of “full” simply emphasizes and makes clear what the word “tank,” standing alone, already indicates. Thus, the plain meaning and ordinary use of “full” is such that it is often literally superfluous. And, when “ordinary meaning would render [a] term superfluous,” the canon against surplusage “should not be used to distort [that] ordinary meaning,” a point that is especially true in “obvious instances of iteration to which lawyers . . . are particularly addicted.”
In support of its surplusage argument, Oracle points to four federal statutes—all enacted after the 1976 passage of the Copyright Act—that allow for recovery of “full costs,” arguing that Rimini Street’s reading would “render superfluous the word ‘full’ in all four.” On the contrary, those statutes support the argument that “full costs” does not mean “all costs incurred in litigation.” Take, for instance, the Digital Millennium Copyright Act of 1998, which states that a “court in its discretion may allow the recovery of full costs by or against any party and may also award a reasonable attorney’s fee to the prevailing party as part of the costs.” If “full costs” generally means all litigation-related expenses, including costs like attorneys’ fees and expert witness expenses, Congress did not need to say that the court could award “full costs” and “also . . . a reasonable attorney’s fee.” Whether or not Section 505 needed “clarifying” in the 1909 amendments, there is no reason why, nearly a century later, Congress would still be clarifying that courts may award “reasonable attorney’s fee[s] to the prevailing party as part of the costs” if “full costs” included such expenses. To the same effect is the Semiconductor Chip Protection Act of 1984, which provides that “the court in its discretion may allow the recovery of full costs, including reasonable attorneys’ fees, to the prevailing party.” The clause in this statute regarding attorneys’ fees is superfluous if “full costs” already includes them. Yet again, the Cable Communications Policy Act of 1984 provides for “the recovery of full costs, including . . . reasonable attorneys’ fees to an aggrieved party who prevails.” If “full costs” already included attorneys’ fees, the latter half of this provision is superfluous. In short, if the canon against surplusage still applies here given the risk that “multiple sentences in multiple enactments would be rendered nugatory,” the canon militates in favor of Rimini Street’s position: “full costs” does not mean “all costs incurred in litigation,” and that is why Congress has repeatedly provided for recovery of attorneys’ fees, which are additional litigation-related expenses beyond “full costs.”
The potential counterargument to this point is that the “American Rule” that each party must pay its own attorneys’ fees is an interpretive presumption that Congress must overcome with clear language, so Congress has had to provide explicitly for attorneys’ fees in addition to full costs, which otherwise generally means all litigation-related expenses besides attorneys’ fees. That is, the argument could be made that, while Congress has to provide for “attorney’s fees” in addition to “full costs” thanks to the American Rule, Congress does not have to overcome a similar presumption with respect to other litigation expenses, meaning that “full costs” generally includes all litigation-related expenses except attorneys’ fees. But of course, it is also an interpretive presumption that the term “costs” is “defined by the categories of expenses enumerated in 28 U.S.C. § 1920,” and the Court has held that “no statute will be construed as authorizing the taxation of witness fees as costs unless that statute refers explicitly to witness fees.” In other words, if clear language is needed to overcome the American Rule prohibiting an award of attorneys’ fees as costs, so too clear language is needed to overcome the presumption that witness fees are not part of awardable costs. Section 505 overcomes the American Rule by allowing for an award of attorneys’ fees, but Section 505 is silent with respect to other non-taxable litigation expenses and so does not overcome what could be dubbed the “Crawford Fitting Rule” prohibiting an award of litigation expenses beyond Section 1920 absent a clear statement allowing them.
C. The Limits of Precedent
Oracle cites Ninth Circuit precedent and cases from other circuits to support its position. But the reasoning of those cases either does not hold up or does not in fact support Oracle. There is no question that Twentieth Century Fox supports Oracle’s position. In that case, the Ninth Circuit relied on the canon against surplusage to read “full costs” as allowing for recovery of expenses beyond those delineated in Section 1920. However, the other authorities upon which Oracle relies provide little, if any, support for the position that full costs means something more than “the entire amount of allowable costs under Section 1920.”
In Coles v. Wonder, for instance, the Sixth Circuit affirmed an award under Section 505 of “$14,172.34 in non-taxable costs,” but the court focused only on the “attorney’s fee” sentence in the statute in analyzing the validity of the award. The court affirmed the district court’s entire award only after noting that it owed deference “to the discretion of the district court in the award of attorney’s fees . . . .” In fact, courts in the Sixth Circuit have explicitly rejected Oracle’s reading of Coles, holding that non-taxable costs are not awardable under Section 505. Furthermore, the Sixth Circuit has repudiated Coles to the extent it can be read as allowing costs beyond those enumerated in Section 1920.
The First Circuit in InvesSys, Inc. v. McGraw-Hill Cos. found non-taxable costs “recoverable under § 505,” but it found them recoverable as part of an “attorney’s fee” under the law, not as part of “full costs.” In fact, the InvesSys court recognized “the tendency of the courts” to treat “full costs” in Section 505 just as “costs” elsewhere, and it found that the general consensus is that non-taxable costs are not recoverable as “full costs” because Section 1920 “does not include” them. Thus, InvesSys does not support the view that non-taxable costs may be assessed as part of the “full costs” permitted by Section 505.
Finally, while the Seventh Circuit commented in dicta in Susan Wakeen Doll Co. v. Ashton-Drake Galleries that “non-taxable costs” could “come through [Section 505],” the court did not specify whether such costs would “come through” as “full costs” or as “attorney’s fee[s].” The fact that, in context, the court was addressing the district court’s award of attorneys’ fees suggests that the Seventh Circuit here may have been thinking, like the First Circuit in InvesSys, that non-taxable costs could be awarded under Section 505 as part of an award of “attorney’s fee[s].” In other words, Susan Wakeen Doll Co. does not support the idea that “non-taxable costs” may be awarded as “full costs” under Section 505.
D. The Technical Meaning of “Full Costs”
Finally, according to Oracle, “the original practice in copyright cases was for prevailing parties to receive all the costs they expended in the litigation . . . .” This was because federal courts followed state law in fashioning awards of costs and fees in copyright suits, and the states reportedly permitted taxation of all litigation expenses, not just what are now referred to as taxable costs. However, uncertainty crept into the law after an 1819 statute gave federal circuit courts original jurisdiction over copyright cases, calling into question whether state law or federal law should furnish the rule of decision regarding awardable costs in copyright actions. The Copyright Act of 1831, where “full costs” first appeared, thus “reinstated the default state rule” allowing recovery of the full gamut of litigation expenses. This argument, however, proves Rimini Street’s point: that “full costs” in Section 505 permits recovery of all that is recoverable under the governing law. Whether or not it was the case that the governing law in 1831 allowed for recovery of all litigation-related expenses as “full costs” because of the relevant state laws, Section 1920 is the governing law now, and it plainly does not.
V. Barking Up the Wrong Tree
There is one curious detail about the case, mentioned in passing by Oracle, that reinforces the foregoing. The non-taxable costs Oracle seeks to recover were purportedly incurred due to Rimini Street’s “intentional spoliation of evidence and lying under oath,” which “forced Oracle to expend an extraordinary amount of resources proving conduct” that Rimini Street committed. Federal courts have “inherent power” to address “[a]llegations of spoliation, including the destruction of evidence in pending or reasonably foreseeable litigation,” and the exercise of this “inherent power” is particularly called for where “there is no statute or rule that adequately addresses the conduct.” Courts, in fact, have invoked this “inherent authority” in similar situations where “reimbursement of . . . fees and expenses that can be traced to the spoliation” was called for “to remedy the expenses incurred” by a party. In this case, it appears the trial court invoked this inherent power, sanctioning Rimini Street for spoliation by giving an adverse inference instruction to the jury. According to Oracle, the trial court also had “ample authority to shift costs as a sanction for . . . misconduct,” but it did not, choosing instead to make an award of non-taxable costs under Section 505 that included some of the expenditures Oracle alleges were incurred because of Rimini Street’s purportedly sanctionable misconduct.
It may be, as Oracle suggests, that the trial court did not invoke its inherent power to award a monetary discovery sanction because it concluded that there was a “statute . . . that adequately addresses the conduct” under Ninth Circuit precedent: Section 505. Perhaps “costs may well have been awardable below as a sanction,” as Oracle argues, even if they were not awardable as “full costs” under Section 505. Because courts inherently possess the power to sanction, the concerns Oracle raises about chilling otherwise meritorious copyright litigation because of the potential for “irretrievably sunk” litigation costs is likely exaggerated. Even if it is not, however, the solution to this problem is not to contort the meaning of a federal statute in order to ratify a potentially acceptable (or perhaps implausible but desirable) outcome on unacceptable grounds. “If a statute needs repair, there’s a constitutionally prescribed way to do it. It’s called legislation.”
As Rimini Street illustrates well, plain meaning can be a trap for the unwary. Dictionaries alone cannot always decide questions of statutory interpretation, even under the plain meaning rule. Construing text requires reading it in its full context, taking into consideration any relevant historical or jurisprudential glosses to the text. In this case, both sides invoke fundamental principles of statutory interpretation. But, reading law here in holistic fashion, Rimini Street makes better sense of Section 505’s “full costs” when viewed in its full context.
 Order List (Sept. 27, 2018), available at https://www.supremecourt.gov/orders/courtorders/092718zr_m5n0.pdf.
 Petition for a Writ of Certiorari, Rimini Street, Inc. v. Oracle USA Inc., No. 17-1625, at i (May 31, 2018), https://www.supremecourt.gov/DocketPDF/17/17-1625/48743/20180531104056260_Rimini__Petition%20--TO%20FILE.pdf (“Petition”).
 Arlington Cent. Sch. Dist. Bd. of Educ. v. Murphy, 548 U.S. 291, 297–98 (2006) (internal quotation marks omitted). 28 U.S.C. § 1821, the other statute referenced in the question presented, delineates the “[f]ees . . . for . . . witnesses” allowed by 28 U.S.C. § 1920(3), essentially providing court witnesses with modest remuneration for actual and opportunity costs. Crawford Fitting Co. v. J.T. Gibbons, Inc., 482 U.S. 437, 440 (1987).
 These are: “(1) Fees of the clerk and marshal; (2) Fees for printed or electronically recorded transcripts necessarily obtained for use in the case; (3) Fees and disbursements for printing and witnesses; (4) Fees for exemplification and the costs of making copies of any materials where the copies are necessarily obtained for use in the case; (5) Docket fees under section 1923 of this title; [and] (6) Compensation of court appointed experts, compensation of interpreters, and salaries, fees, expenses, and costs of special interpretation services under section 1828 of this title.” 28 U.S.C. § 1920.
 See, e.g., 5 U.S.C. § 504(f) (identifying separately “costs, fees, or other expenses”). Confusion is understandable here, however, given that Section 1920 itself uses the term “fees” to refer to what it identifies “as costs.” Rimini Street refers to this distinction as a “tripartite taxonomy” in federal law. Brief for Petitioners, Rimini Street, Inc. v. Oracle USA Inc., No. 17-1625, at 20 (Nov. 13, 2018), https://www.supremecourt.gov/DocketPDF/17/17-1625/71754/20181113123325285_RIMINI%20OPENING%20BRIEF--TO%20FILE.pdf (“Rimini Brief”). However, Oracle fairly observes that this is not a consistent distinction in the U.S. Code. Brief for Respondents, Rimini Street, Inc. v. Oracle USA Inc., No. 17-1625, at 13 (Dec. 13, 2018), https://www.supremecourt.gov/DocketPDF/17/17-1625/76094/20181213142724824_17-1625bs.pdf (“Oracle Brief”).
 Murphy, 548 U.S. at 301.
 94 Pub. L. 553, 90 Stat. 2541, 2586, 17 U.S.C. § 505.
 See Humphreys & Partners Architects v. Lessard Design, 152 F. Supp. 3d 503, 524 (E.D. Va. 2015) (“Section 505 allows a court to award ‘full costs;’ it is unclear, however, what ‘full costs’ means.”).
 Oracle United States, Inc. v. Rimini Street, Inc., 209 F. Supp. 3d 1200, 1218 (D. Nev. 2016).
 Oracle USA, Inc. v. Rimini Street, Inc., 879 F.3d 948, 965–66 (9th Cir. 2018).
 Twentieth Century Fox Film Corp. v. Entm’t Distrib., 429 F.3d 869, 885 (9th Cir. 2005).
 Montclair v. Ramdsdell, 107 U.S. (17 Otto) 147, 152 (1882). See also Conn. Nat’l Bank v. Germain, 503 U.S. 249, 253 (1992) (“[C]ourts must presume that a legislature says in a statute what it means and means in a statute what it says there.”).
 Brief in Opposition, Rimini Street, Inc. v. Oracle USA Inc., No. 17-1625, at 1 (Aug. 1, 2018), https://www.supremecourt.gov/DocketPDF/17/17-1625/56170/20180801142951637_17-1625%20Oracle%20BIO%20FINAL.pdf (“Oracle Opposition”). See id. at 12 (“[F]ull means full[.]”).
 Regions Hosp. v. Shalala, 522 U.S. 448, 466 (1998) (Scalia, J., dissenting).
 Chemehuevi Tribe of Indians v. Fed. Power Comm’n, 420 U.S. 395, 403 (1975).
 Robinson v. Shell Oil Co., 519 U.S. 337, 341 (1997). See also Tyler v. Cain, 533 U.S. 656, 662 (2001) (“We do not . . . construe the meaning of statutory terms in a vacuum.”).
 Deal v. United States, 508 U.S. 129, 132 (1993).
 Bloate v. United States, 559 U.S. 196, 205 n.9 (2010). See generally Stephen Mouritsen, The Dictionary Is Not a Fortress: Definitional Fallacies and a Corpus-Based Approach to Plain Meaning, 2010 B.Y.U. L. Rev. 1915 (2010).
 United States v. Freeman, 44 U.S. (3 How.) 556, 564 (1845).
 Wachovia Bank, N.A. v. Schmidt, 546 U.S. 303, 315–16 (2006).
 See Erlenbaugh v. United States, 409 U.S. 239, 243–44 (1972) (cleaned up) (“The rule of in pari materia—like any canon of statutory construction—is a reflection of practical experience in the interpretation of statutes: a legislative body generally uses a particular word with a consistent meaning in a given context. Thus, for example, a later act can . . . be regarded as a legislative interpretation of an earlier act . . . in the sense that it aids in ascertaining the meaning of the words as used in their contemporary setting, and is therefore entitled to great weight in resolving any ambiguities and doubts. The rule is but a logical extension of the principle that individual sections of a single statute should be construed together, for it necessarily assumes that whenever Congress passes a new statute, it acts aware of all previous statutes on the same subject.”).
 Morales v. TWA, 504 U.S. 374, 384 (1992).
 Int’l Paper Co. v. Ouellette, 479 U.S. 481, 494 (1987).
 Varity Corp. v. Howe, 516 U.S. 489, 519 (1996) (Thomas, J., dissenting).
 FAA v. Cooper, 566 U.S. 284, 291–92 (2012) (noting that when Congress uses terms of art, “the ordinary meaning of the word[s]” as “defined in standard general-purpose dictionaries” is not dispositive, but rather the interpreter must take into account “the cluster of ideas” incorporated into the special term); see also Molzof v. United States, 502 U.S. 301, 306–07 (1992); cf. Yates v. United States, 354 U.S. 298, 319 (1957), overruled on other grounds by Burks v. United States, 437 U.S. 1 (1978) (“[W]e should not assume that Congress . . . used . . . words . . . in their ordinary dictionary meaning when they had already been construed as terms of art carrying a special and limited connotation.”).
 See, e.g., Hall v. Hall, 138 S. Ct. 1118, 1125 (2018) (“This is not a plain meaning case. It is instead about a term . . . with a legal lineage stretching back at least to the first federal consolidation statute, enacted by Congress in 1813. Over 125 years, this Court . . . interpreted that term to mean the joining together—but not complete merger—of constituent cases.”).
 Cooper, 566 U.S. at 291–92.
 Sullivan v. Stroop, 496 U.S. 478, 482–83 (1990).
 See Oracle Brief at 1 (“When the text of a statute is clear, judicial inquiry ends where it begins—with the text.”).
 See Fharmacy Records v. Nassar, 729 F. Supp. 2d 865, 893 (E.D. Mich. 2010) (“The majority of courts which have considered the issue have interpreted the term ‘full costs’ to include only those expenditures listed in 28 U.S.C. §§ 1821 and 1920.”).
 Crawford Fitting, 482 U.S. at 440.
 Id. at 441. See Humphreys & Partners Architects, 152 F. Supp. 3d at 524 (recognizing that Crawford Fitting “provided a general framework for considering whether a statute provides for the recovery of costs that exceed the scope of costs recoverable under Sections 1821 and 1920”).
 Crawford Fitting, 482 U.S. at 442.
 Id. at 444–45. See W. Va. Univ. Hosps. v. Casey, 499 U.S. 83, 86 (1991) (noting that these provisions “define the full extent of a federal court’s power to shift litigation costs absent express statutory authority to go further”); see also Data Gen. v. Grumman Sys. Support Corp., 825 F. Supp. 361, 367 (D. Mass. 1993) (“28 U.S.C. § 1920 defines the ‘costs’ that may be awarded under more general authority,” including “§ 505 of the Copyright Act.”).
 284 U.S. 444, 444–45 (1932).
 Id. at 446.
 Alyeska Pipeline Svc. Co. v. Wilderness Soc’y, 421 U.S. 240, 260 (1975).
 Crawford Fitting, 482 U.S. at 441.
 Murphy, 548 U.S. at 297–98 (internal quotation marks omitted). Cf. Fasa Corp. v. Playmates Toys, Inc., 108 F.3d 140, 144 (7th Cir. 1997) (“The word ‘costs’ at the conclusion of a judicial opinion is a term of art,” referring to 28 U.S.C. § 1920.).
 Petition at 16 (discussing the Act of Feb. 3, 1831, ch. 16, § 12, 4 Stat. 436, 438–39).
 See Costs in Civil Cases, 30 F. Cas. 1058, 1059 (C.C.S.D.N.Y. 1852) (“[T]he usage and practice of the circuit courts in taxing costs have uniformly been to apply the general rule prescribed in the act of September 29, 1789, namely, to fix the rate according to the fee bill of the state.”). Oracle asserts, without citation or much explanation, that “[t]he 1831 Copyright Act adopted a copyright-specific approach to cost-shifting,” and that it “did not incorporate state cost-shifting laws.” Oracle Brief at 3.
 Alyeska Pipeline Serv. Co., 421 U.S. at 252.
 The Baltimore, 75 U.S. (8 Wall.) 377, 392 (1868).
 See Petition at 16.
 See, e.g., Official Aviation Guide Co. v. Am. Aviation Assocs., 162 F.2d 541, 543 (7th Cir. 1947) (Minton, J.) (distinguishing between “ordinary costs” and “extraordinary costs of attorneys’ fees”).
 But see Stevens Linen Assocs. v. Mastercraft Corp., No. 79 Civ. 2016 (CBM), 1981 U.S. Dist. LEXIS 11046, at *7–8 (S.D.N.Y. Feb. 17, 1981) (“Plaintiff contends the word ‘full’ must be intended to expand the court’s authority to award costs. This expanded authority would allow for the award of costs beyond those taxable costs normally awarded to a prevailing party under 28 U.S.C. § 1920. . . . Plaintiff’s contention is unpersuasive . . . .”). See also NLFC, Inc. v. Devcom Mid-America, 916 F. Supp. 751, 762 (N.D. Ill. 1995) (citation omitted) (“Devcom contends that the word ‘full’ in § 505 must be read to expand the court’s authority to allow for the award of costs beyond those taxable costs normally awarded a prevailing party under § 1920. The Court disagrees with Devcom’s interpretation. Caselaw interpreting 17 U.S.C. § 116 (1970) (repealed), the precursor of § 505, has never accorded courts more discretion because of the word ‘full.’”).
 See, e.g., Casey, 499 U.S. at 89 (noting how “[a]t least 34 statutes in 10 different titles of the U.S. Code explicitly shift attorney’s fees and expert witness fees”).
 See Epic Sys. Corp. v. Lewis, 138 S. Ct. 1612, 1626 (2018); Crawford Fitting, 482 U.S. at 445; see Casey, 499 U.S. at 86 (noting that Sections 1920 and 1821 “define the full extent of a federal court’s power to shift litigation costs absent express statutory authority to go further”); cf. Fogerty v. Fantasy, Inc., 510 U.S. 517, 524 n.11 (1994) (“The 1976 Copyright [Act] did change . . . the standard for awarding costs to the prevailing party. . . . The 1976 Act changed the rule from a mandatory one to one of discretion. As the 1909 Act indicates, Congress clearly knows how to use mandatory language when it so desires.”).
 Crawford Fitting, 482 U.S. at 445. See Casey, 499 U.S. at 86 (noting that these provisions “define the full extent of a federal court’s power to shift litigation costs absent express statutory authority to go further”).
 Humphreys & Partners Architects, 152 F. Supp. 3d at 524–25.
 Crawford Fitting, 482 U.S. at 445. See BMG Rights Mgmt. (US) v. Cox Commc’ns, 234 F. Supp. 3d 760, 779 (E.D. Va. 2017) (“Although there is reasonable debate over the proper interpretation of the word ‘full’ in § 505, that term is certainly not explicit in authorizing witness fees or any other non-taxable costs. If it wanted to, Congress could have easily inserted language allowing for the recovery of any category of costs, however, it chose not to, and the Court will not implicitly read those terms into the statute.”); Tempest Publ’g, Inc. v. Hacienda Records & Recording Studio, Inc., 141 F. Supp. 3d 712, 723 (S.D. Tex. 2015) (“Although there is some evidence that Congress intended costs recoverable under § 505 to exceed those recoverable under § 1920, that evidence is not clear or explicit, as needed to conclude that the general statute controls the more specific one. Crawford Fitting, 482 U.S. at 445. When the case is close, Crawford Fitting’s standard counsels in favor of following the specific statute and applying the presumption against implied repeals. Id. Given that standard, the Supreme Court’s opinion in Marx, and the weight of circuit authority resting against a broad reading of § 505, the court concludes that the costs taxable under § 505 are limited to those enumerated in § 1920.”).
 17 U.S.C. § 505.
 Oracle Brief at 5.
 Id. at 14.
 Id. at 29.
 Casey, 499 U.S. at 95 (“Congress . . . having authorized the taxation of reasonable attorney’s fees without making any provision with respect to . . . fees of expert witnesses, must presumably have intended that they not be taxed.”) (internal quotation marks omitted).
 See Fogerty, 510 U.S. at 524 n.11 (“As the 1909 Act indicates, Congress clearly knows how to use mandatory language when it so desires.”); see also Casey, 499 U.S. at 100–01 (responding to the argument that Congress would have included expert fees in a fee-shifting statute “had it thought about it” by explaining that “[t]he facile attribution of congressional ‘forgetfulness’ cannot justify” departing from “that permissible meaning which fits most logically and comfortably into the body of both previously and subsequently enacted law”); cf. Transamerica Mortg. Advisors (tama) v. Lewis, 444 U.S. 11, 22 & n.13 (1979) (noting that “[w]hile subsequent legislation can disclose little or nothing of the intent of Congress in enacting earlier laws,” statutory amendments at least reveal “that Congress knew how to” overcome interpretive presumptions—in this case, the presumption against implied private rights of action—“when it wished to do so,” so Congress’ failure to overcome such presumptions is evidence that the presumption should still control).
 As Oracle argued in opposition to certiorari, the circuit split on this issue is “stale”—the circuit court decisions holding that “full costs” only means “taxable costs” are over fifteen years old. Oracle Opposition at 2, 14.
 Casey, 499 U.S. at 99.
 The well-established rule of expressio unius supports this conclusion. See, e.g., Jennings v. Rodriguez, 138 S. Ct. 830, 844 (2018).
 See Oracle Opposition at 18 (“Petitioners attempt to resist that conclusion by positing that ‘full’ simply means that a prevailing party can recover the entirety of the costs allowable under § 1920. But they do not and cannot explain why Congress would need to include that clarification since the same is true under § 1920 itself. Indeed, statutes often authorize recovery of ‘costs’ simpliciter, without specifying that each enumerated cost may be recovered ‘in full.’”).
 568 U.S. 371, 385 (2013).
 Id. Oracle apparently overlooked this precise point, arguing that “if ‘full costs’ really meant only those taxable costs already available under §1920, then . . . the entire grant of discretion in §505 to ‘allow the recovery of full costs’ . . . would be meaningless” because “Section 1920 already gives district courts discretion to award a prevailing party the costs the statute enumerates . . . .” Oracle Brief at 19.
 See Brief of Amici Curiae Scholars of Corpus Linguistics Supporting Petitioners, Rimini Street, Inc. v. Oracle USA Inc., No. 17-1625, at 2 (Nov. 20, 2018), https://www.supremecourt.gov/DocketPDF/17/17-1625/72865/20181120195011397_Rimini%20Street%20Amicus%20FINAL%20pdfa.pdf (“‘[F]ull’ in Section 505 should be considered a ‘delexicalized’ adjective,” which is an adjective “whose purpose is to draw attention to and underline an attribute that is already embedded in the meaning of the noun . . . .”).
 As the Corpus Linguistics amici explain, the “delexicalization” of the word “full” is quite common: a person in possession of a “deck of cards” would be presumed to have a “full deck,” for instance, so “deck” and “full deck” communicate the same point in ordinary parlance. See generally id.
 Moskal v. United States, 498 U.S. 103, 120 (1990) (Scalia, J, dissenting).
 Oracle Brief at 19; see also id. at 23.
 Oracle United States, Inc., 209 F. Supp. 3d at 1218.
 28 U.S.C. § 4001(g).
 See supra notes 56–64 and accompanying text.
 17 U.S.C. § 911(f).
 47 U.S.C. § 553(c)(2)(C); accord 47 U.S.C. § 605(e)(3)(B)(iii).
 Oracle Brief at 23.
 See, e.g., Oracle Brief at 25.
 Murphy, 548 U.S. at 301 (cleaned up) (emphasis added).
 See Oracle Opposition at 14.
 429 F.3d 869, 885 (9th Cir. 2005).
 283 F.3d 798, 803–04 (6th Cir. 2002).
 Id. (emphasis added). See R.C. Olmstead, Inc. v. CU Interface, LLC, No. 5:08CV234, 2011 U.S. Dist. LEXIS 15595, at *22–24 (N.D. Ohio Feb. 16, 2011) (awarding non-taxable costs under Section 505 per Coles by reasoning that such costs “may be subsumed within the phrase attorney’s fees” appearing in the statute).
 See, e.g., Liang v. AWG Remarketing, Inc., No. 2:14-cv-00099, 2016 U.S. Dist. LEXIS 13566, at *32–34 (S.D. Ohio Feb. 4, 2016) (noting that Coles “contains no discussion on this issue” before concluding that there was no reason to conclude that “the Sixth Circuit would adopt the rule adopted by the Ninth Circuit that non-taxable costs . . . are recoverable as costs under the Copyright Act”); Fharmacy Records, 729 F. Supp. 2d at 893 (noting that Coles “affirmed an award of non-taxable costs under § 505, without discussion,” before reviewing Crawford Fitting and holding that “the fees paid by the Defendants to their consulting experts are not recoverable against the Plaintiffs under the Copyright Act”).
 See L & W Supply Corp. v. Acuity, 475 F.3d 737, 738–41 (6th Cir. 2007) (“[W]itness fees are not recoverable as costs absent explicit statutory authority. . . . any earlier Sixth Circuit and/or any other earlier precedent is no longer controlling.”).
 Oracle Opposition at 14.
 369 F.3d 16, 22–23 (1st Cir. 2004). See also Tempest Publ’g, 141 F. Supp. 3d at 722 (recognizing that the InvesSys court “held that electronic-research costs are recoverable as part of attorney’s fees”).
 369 F.3d at 22.
 272 F.3d 441, 458 (7th Cir. 2001).
 Id. at 457–58.
 Oracle Opposition at 20.
 Id. at 21–22.
 Id. at 12.
 Rinkus Consulting Grp., Inc. v. Cammarata, 688 F. Supp. 2d 598, 611 (S.D. Tex. 2010).
 Ramos v. Swatzell, No. ED CV 12-1089-BRO (SPx), 2017 U.S. Dist. LEXIS 103014, at *48–49 (C.D. Cal. June 5, 2017). See also Process Am., Inc. v. Cynergy Holdings, LLC, No. 12 Civ.772 (BMC), 2013 U.S. Dist. LEXIS 188478, at *44 (E.D.N.Y. Sept. 23, 2013) (“[T]he Court imposes the following sanction on Process America for its spoliation: Process America is required to reimburse Cynergy for half of its costs, including attorneys’ fees and forensic expert costs, that it incurred in connection with litigating the spoliation issue.”); United States v. Philip Morris USA, Inc., 327 F. Supp. 2d 21, 26 (D.D.C. 2004) (imposing a monetary sanction to “fully reflect the reckless disregard and gross indifference displayed” by defendants “toward their discovery and document preservation obligations,” and requiring defendants “to reimburse the United States for the costs associated with . . . email destruction issues”).
 Oracle Brief at 10.
 Id. at 53.
 Oracle Opposition at 23.
 Oracle Brief at 49–51.
 Perry v. Merit Sys. Protection Bd., 137 S. Ct. 1975, 1990 (2017) (Gorsuch, J., dissenting).