“Constitution time is over.”
For academic and self-proclaimed “copyright activist” Lawrence Lessig, January 18, 2012, represents more than just the day the internet stood up to copyright law. In his view, it’s also the day “the Supreme Court shut the door, finally and firmly, on any opportunity to meaningfully challenge a copyright statute constitutionally.” For it was on that day that the Supreme Court handed down its decision in Golan v. Holder,1 or as Lessig puts it, it’s the day the Court signaled to the copyright bar: “Constitution time is over.” Lessig’s defeatist attitude is understandable. He had spent the past thirteen years representing four different plaintiffs challenging various copyright laws on constitutional grounds. Two of the cases were even heard by the Supreme Court; all four cases ended in resounding defeat for Lessig’s various constitutional arguments.2 Upset with Congress’s treatment of copyright, it’s not surprising that Lessig would turn to the courts. To get a court to strike down a copyright statute, he had to constitutionalize the issue—courts simply aren’t in the business of second-guessing legislative choices unless those choices violate the Constitution. The problem for Lessig, though, was that the Constitution itself provides few limits on Congress’s power to create copyright laws. But “few limits” does not mean “no limits.”
One of Lessig’s major arguments was that a copyright law must give way to First Amendment concerns, and it’s that particular notion that I’ll focus on in this article. While Lessig may have been defeated on this argument in that he didn’t like where the Court ultimately drew the line between copyright and free speech, as we’ll see, he was certainly vindicated in that the Court definitively recognized that copyright must in fact significantly give way to free speech interests. One culmination of Lessig’s thirteen-year-long legal battle against copyright law in the federal courts is that the Supreme Court has given us a test to use in determining whether a copyright statute has run afoul of free speech—the “traditional contours” test. And contrary to Lessig’s complaint that this test forecloses meaningful First Amendment challenges to copyright, the test instead, I think, reasonably reflects the fact that free speech and copyright promote the same public good. This sentiment is reflected in something the Court said over two decades ago: “it should not be forgotten that the Framers intended copyright itself to be the engine of free expression. By establishing a marketable right to the use of one’s expression, copyright supplies the economic incentive to create and disseminate ideas.”3
There is no doubt that copyright, in some sense, burdens free speech. The First Amendment commands that “Congress shall make no law . . . abridging the freedom of speech.” At the same time, however, the Copyright Clause allows Congress to “secur[e] for limited Times to Authors . . . the exclusive Right to their . . . Writings.” But by creating a law that secures to an author a copyright in his work, Congress has inevitably created a law that abridges free speech. Copyright, just like any other type of personal property, provides its holder with the “right to exclude others from using his property.”4 It follows then that a copyright law burdens free speech since others are not free to speak that which a copyright protects. But, under the Constitution, how much can copyright burden free speech, exactly? Thanks to Lessig’s efforts, we now have the “traditional contours” test that answers just that question.
An Amorphous Test
In Eldred v. Ashcroft,5 the Supreme Court squarely addressed a First Amendment challenge to the Copyright Term Extension Act (CTEA).6 Lessig himself participated in the oral arguments before the Court. The CTEA provided for an additional twenty years of copyright protection for most works, even those that were currently under copyright. Petitioners (Lessig’s clients) included individuals and businesses who depended on works that had fallen into the public domain for their products or services. Petitioners argued that, but for the CTEA, they would have been able to exploit works that otherwise would be in the public domain. Therefore, by extending the copyright term for these works, Congress had impermissibly abridged petitioners’ free speech rights. In light of this perceived conflict, petitioners asked the Supreme Court to find the CTEA unconstitutional under heightened First Amendment judicial scrutiny. Heightened scrutiny simply refers to how important the legislature’s ends must be, and how well those ends must fit with the means chosen to accomplish them. The greater the scrutiny, the more likely it is that a statute will be struck down as unconstitutional. So the party challenging a statute almost always argues for greater scrutiny, while the defending party, naturally, argues the opposite.
The Court declined the invitation to apply heightened scrutiny, instead agreeing with respondents that none was needed. Said the Court: “The Copyright Clause and First Amendment were adopted close in time. This proximity indicates that, in the Framers’ view, copyright’s limited monopolies are compatible with free speech principles. Indeed, copyright’s purpose is to promote the creation and publication of free expression.”7 In the Court’s view, imposition of heightened judicial scrutiny was not necessary because copyright law already contains two built-in First Amendment accommodations. The first, known as the idea/expression dichotomy,8 “distinguishes between ideas and expression and makes only the latter eligible for copyright protection.”9 The second, known as the fair use defense,10 “allows the public to use not only facts and ideas contained in a copyrighted work, but also expression itself in certain circumstances.”11 According to the Court, these built-in accommodations strike a good balance between copyright and free speech since they allow for the free flow of ideas while providing only partial protection for an author’s particular expression.
The Court explained further that the “First Amendment securely protects the freedom to make—or decline to make—one’s own speech; it bears less heavily when speakers assert the right to make other people’s speeches.”12 Petitioners’ asserted right to speak other people’s copyrighted works did raise First Amendment concerns, said the Court, but these concerns were adequately addressed by copyright’s built-in free speech protections. And then, in what can be described as an afterthought, the Supreme Court announced the “traditional contours” test: “But when, as in this case, Congress has not altered the traditional contours of copyright protection, further First Amendment scrutiny is unnecessary.”13 The “traditional contours” test, then, is a simple conditional statement: If Congress did not alter copyright’s “traditional contours” when crafting a copyright statute, then a reviewing court faced with a First Amendment challenge to that statute does not apply heightened judicial scrutiny. In other words, a court will not have to look at how important the ends are, or how closely those ends fit the means chosen to accomplish them, so long as copyright’s “traditional contours” have not been altered in creating the statute. Lower courts sorely needed a test like this to apply when faced with a free speech challenge to a copyright statute. The problem, unfortunately, was that no one knew what these “traditional contours” were because the Court didn’t really say. How could courts apply the test if they didn’t even know when the test applied? And what exactly is (or isn’t) included within these “traditional contours”?
Kahle Tests The Waters
The year after the Supreme Court opinion in Eldred was handed down, a district court in California attempted to apply the newly-minted “traditional contours” test. In Kahle v. Ashcroft,14 plaintiffs (Lessig’s clients), including the Internet Archive, were in the business of taking works that had fallen into the public domain and posting them on the Internet. They brought suit challenging the constitutionality of the Copyright Renewal Act (CRA)15 and the CTEA on First Amendment grounds. The CRA removed the renewal requirement for works created between 1964 and 1977, and it provided that the copyrights for these works would be automatically renewed for a second term. Before the CRA, these works would have only received a second term if their owners had filed for a renewal. Now, they went full-term without the owners doing anything. The CTEA, as we just saw in Eldred, added an additional twenty years of copyright protection to these now automatically-renewing works. Plaintiffs argued that, but for the CRA and the CTEA, the vast majority of the works created between 1964 and 1977 would have fallen into the public domain on January 1, 2004. Rather than a conditional copyright regime that required authors to take affirmative steps to extend the copyright term for their works, the challenged statutes created an unconditional copyright regime where copyright protections were extended automatically. This transition from a conditional to an unconditional system—in effect, from one that was “opt in” to one that was “opt out”—argued plaintiffs, altered copyright’s “traditional contours,” thereby necessitating heightened judicial scrutiny under the “traditional contours” test.
The intuitiveness of this argument is undeniable: Whereas a copyright owner once had to renew his registration to get a second term, now he had to do nothing to obtain the same result (plus an additional 20 years of protection). For parties, like plaintiffs, waiting for these works to fall into the public domain so they could use them, this change was understandably seen as an alteration of copyright’s “traditional contours.” The district court didn’t agree, and it started its analysis with the observation that the Supreme Court hadn’t actually identified the protections it considered to be within copyright’s “traditional contours.” In fact, the court noted, the phrase “traditional contours” did not appear in any other reported decision prior to its use in Eldred. The court deduced that the two concepts recognized by the Court, namely the idea/expression dichotomy and the fair use defense, related only to the scope of copyright protection. Contrasting that with the procedural steps now being challenged, the court reasoned that removing the renewal requirement did not alter the scope of the copyright protection or the copyright holder’s substantive rights. As such, the court held that the challenged statutes did not alter copyright’s “traditional contours.” For the district court, the “traditional contours” simply were not affected by changes in copyright procedures, like removing the renewal requirement.
On appeal to the Ninth Circuit,16 plaintiffs fared no better. The circuit court accepted plaintiffs’ argument that abolishing the renewal requirement would increase the copyright term for the works in question while correspondingly decreasing the number of works entering the public domain. Nonetheless, the court of appeals found that automatically renewing a copyright’s registration was qualitatively no different than extending a copyright’s term, and the Supreme Court’s holding in Eldred had in effect already answered plaintiffs’ challenge: If extending copyrights for existing works was constitutional there, then automatically renewing and thus extending copyrights for existing works was also constitutional here. So long as “traditional First Amendment safeguards such as fair use and the idea/expression dichotomy” vindicate the plaintiffs’ affected free speech interests, said the court, “extending existing copyrights . . . does not alter the traditional contours of copyright protection.”17 So for the Ninth Circuit, automatically renewing a copyright was no different than extending a copyright’s term, and plaintiffs’ clever framing of the issue fell flat. While the Supreme Court in Eldred did not define what it meant by copyright’s “traditional contours,” the Court’s decision did make clear that extending a copyright’s term leaves the “traditional contours” intact—that was the exact issue in Eldred. And once the circuit court framed the issue as one of simply extending a copyright’s term, plaintiffs’ game was over under the Court’s prior holding. Plaintiffs appealed to the Supreme Court, but the petition for certiorari was denied.18
Golan I & II
A year after the Court decided Eldred, another district court, in what would turn out to be the first step in a protracted journey back to the Supreme Court, was asked to apply the “traditional contours” test. In Golan v. Ashcroft,19 plaintiffs (Lessig’s clients) included artisans and businesses that published and performed works that were in the public domain. They sought declaratory and injunctive relief, maintaining that § 514 of the Uruguay Round Agreements Act (URAA)20 was unconstitutional. The statute restored copyright protection to foreign works whose authors had lost their copyrights due to failure to comply with certain formalities that had since been repealed. Plaintiffs argued that § 514 of the URAA altered copyright’s “traditional contours” and violated their First Amendment rights because they could no longer use certain works that had been pulled out of the public domain. The court held that plaintiffs had sufficiently distinguished the holding in Eldred so as to survive defendants’ motion to dismiss.
Plaintiffs’ victory was short-lived, though, and in a lengthy opinion the district court granted defendants’ motion for summary judgment. The court reasoned that “Congress has historically demonstrated little compunction about removing copyrightable materials from the public domain.”21 The record reflected that there were in fact many historical instances where public domain works had been granted copyright. Because of the idea/expression dichotomy, the court noted, only expressions were being restored to their authors—the ideas themselves were still public property. Considering the long string of practice for granting copyright protection to works in the public domain, the court rejected plaintiffs’ contention that copyright’s “traditional contours” had in fact been altered by doing the same thing here. The court, citing “the settled rule that private censorship via copyright enforcement does not implicate First Amendment concerns,”22 similarly rebuffed plaintiffs’ argument that having to contract for use of restored works posed too onerous a burden on their free speech rights. In the court’s opinion, while plaintiffs surely bore some free speech hardship because of § 514 of the URAA, such difficulties were an inherent feature of copyright law in general and therefore not actionable.
On appeal to the Tenth Circuit,23 plaintiffs’ luck changed. The court of appeals, after observing that the Supreme Court had not defined the “traditional contours” in Eldred, nonetheless definitively stated that “one of these traditional contours is the principle that once a work enters the public domain, no individual—not even the creator—may copyright it.”24 Moreover, the court reasoned that plaintiffs had cognizable and vested First Amendment interests in public domain works. Central to the Tenth Circuit’s analysis was the understanding that copyright’s “traditional contours” must include more than just the built-in free speech accommodations, i.e., the idea/expression dichotomy and the fair use defense. The court concluded “that the traditional contours of copyright protection include the principle that works in the public domain remain there” and that § 514 of the URAA had transgressed that “critical boundary.”25 Furthermore, the circuit court disagreed with the district court’s contention that there was a tradition of removing works from the public domain, and it characterized whatever history of the practice that did exist as the exception and not the rule. The Tenth Circuit remanded the case to the district court with instructions to subject § 514 of the URAA to heightened First Amendment scrutiny, as commanded by the “traditional contours” test.
On cross-motions for summary judgment, the district court began its analysis with the observation that § 514 of the URAA was a content-neutral regulation of speech because it could “be justified without reference to the content of the speech restricted.”26 Under the applicable standard of heightened scrutiny (here, intermediate scrutiny), the statute would be upheld only if it advanced an important governmental interest and did “not burden substantially more speech than necessary to further” that interest.27 After careful analysis, the district court granted plaintiffs’ motion for summary judgment, concluding that while the government’s interests were sufficiently important, Congress nevertheless had burdened substantially more speech than was necessary in advancing those interests. The court limited its holding to only those parties that had already been exploiting the works while they were in the public domain—the so-called reliance parties. In the court’s view, those were the only parties with vested First Amendment rights that had been contravened when copyrights were restored in the works.
The parties cross-appealed to the Tenth Circuit.28 Plaintiffs argued that § 514 of the URAA should be struck down as unconstitutional on its face, meaning that it should be found to be unconstitutional not only for the reliance parties but for everyone else as well. The defendants of course disagreed, arguing that the statute was constitutional not only as-applied to the reliance parties but for everyone in general too. On this appeal, the “traditional contours” test was not in issue—the previous circuit panel had definitively applied the test and found that the statute failed it. Under the law of the case doctrine, this circuit panel was bound to follow the prior panel’s conclusion on that point. Recall that failure to satisfy the “traditional contours” test simply means that heightened First Amendment scrutiny will be applied to the statute, so the battle on this appeal was over whether § 514 of the URAA was violative of the First Amendment under this more-stringent standard of review.
The circuit court started its de novo review by agreeing with the district court that § 514 of the URAA was a content-neutral regulation of speech, thereby calling for intermediate scrutiny. In looking at the first prong, which requires the government to assert an important or substantial interest, the court had “no difficulty in concluding that the government’s interest in securing protections abroad for American copyright holders satisfies this standard.”29 The government had introduced sufficient evidence to show that by granting copyright protection to foreign works in the public domain in the United States, foreign countries would reciprocate by granting copyright protection to American works that were in the public domain abroad. Turning to the second prong, which requires that the regulation not burden substantially more speech than is necessary to further the interest asserted, the circuit court reached the opposite conclusion than the district court. In doing so, the court of appeals found that the burdens inflicted by § 514 of the URAA on the reliance parties were congruent to the benefits afforded to American copyright holders since their works overseas would receive equal protections. The Tenth Circuit sided with defendants and reversed the district court below, holding that § 514 of the URAA was not unconstitutional under heightened scrutiny.
So to summarize: (1) the district court held that § 514 of the URAA didn’t alter copyright’s “traditional contours,” (2) the court of appeals reversed and held that it did, (3) the district court held that § 514 of the URAA didn’t pass heightened scrutiny, and (4) the court of appeals reversed and held that it did. A bit confusing, I know.
The “Traditional Contours” Test Defined
Plaintiffs petitioned for and the Supreme Court granted certiorari. In Golan v. Holder,30 the Court affirmed the Tenth Circuit below, starting with the observation that “some restriction on expression is the inherent and intended effect of every grant of copyright.”31 Despite the intrinsic conflict between the Copyright Clause and the First Amendment, said the Court, in the Framers’ view the two served the same goal of contributing to the marketplace of ideas. The Court then explained that the reference to the “traditional contours” in Eldred referred to only the idea/expression dichotomy and the fair use defense, i.e., the free speech accommodations that are built-in to copyright law. And with that simple explanation, an almost decade-long confusion about what constitutes copyright’s “traditional contours” was laid to rest. The “traditional contours” test, then, may be stated as follows: If Congress did not alter the idea/expression dichotomy or the fair use defense when crafting a copyright statute, then a reviewing court faced with a First Amendment challenge to that statute does not apply heightened judicial scrutiny.
That the Tenth Circuit the first time around had completely misunderstood the “traditional contours” test was made explicit in a footnote: “On the initial appeal in this case, the Tenth Circuit gave an unconfined reading to our reference in Eldred to ‘traditional contours of copyright.’ That reading was incorrect, as we here clarify.”32 The Court continued its analysis by stating that here, as in Eldred, there was simply no call for the heightened review that petitioners were seeking. In the Court’s view (and understanding this point is critical to understanding Golan), burdening people’s communication of an author’s protected expression simply didn’t raise the same free speech concerns that are present when the government burdens people’s communication of facts or ideas. The Court reasoned, rather simply, that since the traditional safeguards of the idea/expression dichotomy and the fair use defense had been left in place in drafting § 514 of the URAA, petitioners’ free speech interests were adequately protected. No further mitigation was necessary since the constitutional minimum requirements had been met.
The Court then turned to the argument that petitioners in this case were distinguishable from those in Eldred because they had enjoyed vested First Amendment rights in certain public domain works. Rejecting the argument that “the Constitution renders the public domain largely untouchable by Congress,” the Court accused petitioners of attempting “to achieve under the banner of the First Amendment what they could not win under the Copyright Clause.”33 The Court could find no historical or congressional practice, nor anything in the Court’s own jurisprudence, that showed that heightened scrutiny was warranted for the practice of restoring copyright protection for works that were in the public domain. The Court positively rejected petitioners’ argument that they, as members of the public using public domain works, had vested First Amendment rights in those works: “Anyone has free access to the public domain, but no one, after the copyright term has expired, acquires ownership rights in the once-protected works.”34
And with that, the Supreme Court shut down once and for all Lessig’s thirteen-year-long argument that copyright laws must give special solicitude to the First Amendment above and beyond the idea/expression dichotomy and the fair use defense.
Golan presents a great example of where simply labeling the Supreme Court’s holding as an affirmance of the court below misses the point. True, the Court affirmed the Tenth Circuit’s holding that § 514 of the URAA did not violate plaintiffs’ First Amendment rights. But the Court arrived at that destination by taking a completely different route. The better view is that, as far as the First Amendment issues are concerned, the Supreme Court in Golan completely denunciated all of the Tenth Circuit’s reasoning despite affirming its ultimate holding. The fact that the court of appeals got the right answer is entirely undercut by the fact that they were answering the wrong question.
I do want to point out as well that I think people are apt to take the holding in Golan too far. Consider the copyright statutes that we looked at—the CTEA, the CRA, and § 514 of the URAA. The CTEA extended the copyright term by twenty years for certain works. The CRA removed the renewal requirement, thus extending the copyright term for certain works. And § 514 of the URAA extended the copyright term to certain works in the public domain. See the pattern? I think the “traditional contours” test is applicable only when Congress has defined the scope of a substantive copyright right, that is, when it has “secur[ed] for limited Times to Authors . . . the exclusive Right to their . . . Writings” as the Copyright Clause expressly provides Congress may do. This is why I think Lessig goes too far when he laments that the Court has “shut the door, finally and firmly, on any opportunity to meaningfully challenge a copyright statute constitutionally.” Unless it’s a copyright law that defines the scope of the substantive right—like, say, the copyright term—I think the Court has left the door wide open for meaningful constitutional challenges. For example, I would argue that the Stop Online Piracy Act (SOPA) falls outside the “traditional contours” test since it concerns enforcement of substantive copyright rights and not the scope of the rights themselves. That means SOPA wouldn’t get a free ride when challenged on First Amendment grounds.
The holding in Golan certainly reinforces the concept that “copyright has traditionally been viewed as an exception to the First Amendment.”35 But it’s clearly not a complete exception. As I mentioned at the outset, Lessig’s notion that copyright must give way to free speech has been proved true with a vengeance—that’s exactly what the “traditional contours” test tells us. Copyright is an exception to the First Amendment, but only because it already incorporates two very important free speech safeguards. Alter either safeguard, and a copyright law’s free speech exception would have to be reexamined. Many people, no doubt, are dubious that these built-in safeguards adequately protect our free speech interests. Indeed, much has been written in the past few decades questioning precisely that. What the Court lays to rest in Golan, I think, is the dispute over whether these doubts are properly framed as constitutional issues or simply policy choices. One can reasonably believe that greater consideration for free speech is needed when it comes to copyright laws (in fact, I share that view), but what the Court has now made clear is that the First Amendment doesn’t demand it.
The last point I’ll make is that the Supreme Court is telling us in Golan that those focusing on the inherent conflict between copyright and free speech in framing their constitutional arguments are missing the forest for the trees. While the “immediate effect of our copyright law is to secure a fair return for an author’s creative labor,” the fundamental purpose of copyright is “to stimulate artistic creativity for the general public good.”36 Thus the view espoused by the Court is that copyright and free speech are not at odds with each other in a zero-sum game where a benefit to one implies a detriment to the other. The philosophy behind the Copyright Clause “is the conviction that encouragement of individual effort by personal gain is the best way to advance public welfare through the talents of authors . . . .”37 In the Court’s view, in the Framers’ view, and in my view as well, the First Amendment and the Copyright Clause are complementary provisions promoting the same goal—the public good.
Follow me on Twitter: @devlinhartline
© 2013 Devlin Hartline. Licensed under the Law Theories Public License 1.0.
- Golan v. Holder, 10-545, 2012 WL 125436 (U.S. Jan. 18, 2012). ↩
- See Eldred v. Ashcroft, 537 U.S. 186 (2003); Luck’s Music Library, Inc. v. Gonzales, 407 F.3d 1262 (D.C. Cir. 2005); Kahle v. Gonzales, 487 F.3d 697 (9th Cir. 2007), cert. denied, 128 S.Ct. 958 (2008); Golan v. Holder, 10-545, 2012 WL 125436 (U.S. Jan. 18, 2012). ↩
- Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539, 558 (1985). ↩
- Fox Film Corp. v. Doyal, 286 U.S. 123, 127 (1932); see also Metro-Goldwyn-Mayer Studios, Inc. v. Grokster, Ltd., 454 F.Supp.2d 966, 997 (C.D. Cal. 2006) (“The right to exclude is inherent in the grant of a copyright . . . .”). ↩
- Eldred v. Ashcroft, 537 U.S. 186 (2003). ↩
- Copyright Term Extension Act of 1998, Pub.L. 105–298 (amending, inter alia, 17 U.S.C. §§ 302, 304). ↩
- Eldred, 537 U.S. at 219. ↩
- The idea/expression dichotomy is codified in 17 U.S.C. § 201(b) (“In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.”). ↩
- Eldred, 537 U.S. at 219. ↩
- The fair use defense is codified in 17 U.S.C. § 107 (“Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, including such use by reproduction in copies or phonorecords or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research, is not an infringement of copyright…”). ↩
- Eldred, 537 U.S. at 219. ↩
- Id., at 221. ↩
- Id. ↩
- Kahle v. Ashcroft, 2004 WL 2663157 (N.D. Cal. Nov. 19, 2004). ↩
- Copyright Renewal Act of 1992, Pub.L. 102-307. ↩
- Kahle v. Gonzales, 487 F.3d 697 (9th Cir. 2007). ↩
- Id., at 700 (internal quotation marks omitted). ↩
- Kahle v. Gonzales, 128 S.Ct. 958 (2008) (denying cert.). ↩
- Golan v. Ashcroft, 310 F.Supp.2d 1215 (D. Colo. 2004). ↩
- Uruguay Round Agreements Act, Pub.L. 103-465. ↩
- Golan v. Gonzales, 2005 WL 914754 (D. Colo. Apr. 20, 2005). ↩
- Id., at *17. ↩
- Golan v. Gonzales, 501 F.3d 1179 (10th Cir. 2007). ↩
- Id., at 1184. ↩
- Id., at 1189. ↩
- Golan v. Holder, 611 F.Supp.2d 1165, 1170 (D. Colo. 2009). ↩
- Id., at 1172. ↩
- Golan v. Holder, 609 F.3d 1076 (10th Cir. 2010). ↩
- Id., at 1084. ↩
- Golan v. Holder, 2012 WL 125436 (U.S. Jan. 18, 2012). ↩
- Id., at *13. ↩
- Id., at fn. 29. ↩
- Id., at *15. ↩
- Id., at *16. ↩
- Jennifer E. Rothman, Liberating Copyright: Thinking Beyond Free Speech, 95 Cornell L. Rev. 463, 479 (2010). ↩
- Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156 (1975). ↩
- Mazer v. Stein, 347 U.S. 201, 219 (1954). ↩