Vitruvius wrote that architecture must possess strength, utility, and beauty.1 Two thousand years later, can political communities in the United States ensure that architecture is not just strong and useful but also beautiful? Recently, in Burns v. Town of Palm Beach,2 the Eleventh Circuit addressed whether the First Amendment protects residential architecture from aesthetic regulation.3 Though the court declined to extend First Amendment coverage to the plaintiff’s proposed home, both the majority and the dissent relied on expressive conduct doctrine to evaluate the architectural expression claim. Expressive conduct, however, is a poor fit for architecture, and future courts should either develop a standard tailored to architecture’s unique character or decline to extend First Amendment coverage.
In 2013, Donald Burns, a telecom executive and resident of Palm Beach, Florida, decided to make a lifestyle change.4 He would demolish his traditional Bermuda Style mansion and replace it with a modern International Style one twice its size.5 The modern structure would communicate his newfound philosophy of “lifestyle simplicity”6 and the message that he was “unique and different”7 from his neighbors. Between 2014 and 2015, Burns applied to the Palm Beach Town Council for permission to begin construction.8 The neighbors objected. The proposed structure, they argued, was too large and too dissimilar from the neighborhood aesthetic.9 After hearing witnesses for and against, the Town Council voted unanimously to defer its decision until the project was reviewed by the town’s architectural commission (ARCOM).10
ARCOM was created to preserve and foster the beauty of Palm Beach.11 The Palm Beach municipal code — recognizing that beautiful communities result from the “deliberate search for beauty”12 and that the “essential foundation” of such beauty is “harmony”13 — tasks the seven-member14 ARCOM with evaluating proposed construction projects according to ten criteria, including excessive similarity and dissimilarity.15 Over three meetings, ARCOM twice deferred issuing a decision so that Burns could modify his proposal in light of the concerns raised.16 He reduced the mansion’s size, increased the landscaping to better shield the mansion from public view, and reduced the amount of glass.17 It was not enough. In September 2016, ARCOM voted 5–2 to deny Burns’s application, citing the proposed structure’s lack of harmony with and dissimilarity from neighboring structures.18
Burns sued the Town in the U.S. District Court for the Southern District of Florida,19 alleging that two ARCOM-related ordinances — one outlining ARCOM’s purpose and the other ARCOM’s evaluative criteria (“the Ordinances”)20 — infringed upon protected First Amendment architectural expression, both facially and as applied.21 The Town, in response, filed motions for dismissal and summary judgment.22
The district court granted summary judgment for the Town,23 relying on a Report and Recommendation from a magistrate judge.24 The magistrate judge, noting the absence of case law on the First Amendment status of architecture, had begun by considering the appropriate test for architectural expression claims.25 Both parties suggested the magistrate judge evaluate the claim using some version of the Spence v. Washington26 test, which was developed to determine when conduct is sufficiently expressive to trigger First Amendment coverage.27 Under Spence, conduct receives First Amendment coverage if (1) the actor intended to communicate a message and (2) that message was likely to be understood by reasonable viewers.28
But architecture, the magistrate judge concluded, is an object, not conduct.29 And so, the magistrate judge looked to the Second Circuit’s expressive merchandise cases. Those cases, most notably Mastrovincenzo v. City of New York,30 use a predominant purpose test to determine whether an object is “predominantly expressive” and thus entitled to First Amendment coverage.31 Blending Spence with Mastrovincenzo, the magistrate judge fashioned a modified architectural Spence test, requiring (1) that an owner intend to communicate a message, (2) that the structure’s predominant purpose be communicative, and (3) a great likelihood that a reasonable viewer would understand the structure as predominantly communicative.32
Burns’s proposed mansion failed this test. Burns had intent, as all parties conceded, but the structure’s predominant purpose was functional (i.e., residential), not artistic, and the reasonable viewer would not understand the proposed mansion as predominantly communicative.33 Accordingly, the magistrate judge recommended summary judgment for the Town.34 Finding the magistrate judge’s analysis “well-reasoned and correct,” the district court adopted the Report and Recommendation and granted summary judgment for the Town.35
The Eleventh Circuit affirmed.36 Writing for the majority, Judge Luck37 held that Burns’s proposed mansion was not sufficiently expressive to trigger First Amendment coverage under Spence.38 Unlike the magistrate judge and the district court, however, Judge Luck declined to apply the modified architectural Spence test. For Judge Luck, Burns’s proposed mansion could not even satisfy the ostensibly easier-to-meet pure Spence test, so it was unnecessary to consider the lower court’s modified version.39
As before, Spence prong one was satisfied: all agreed that Burns had intent to communicate a message.40 But Judge Luck found that Burns’s proposed mansion failed Spence prong two for at least two reasons. First, the reasonable viewer would not understand Burns’s proposed mansion as communicating some message because the reasonable viewer could not see Burns’s proposed mansion, obscured as it would be by trees and landscaping.41 Second, there was no evidence that the reasonable viewer — even if one existed — would understand Burns’s proposed mansion as anything other than a “really big house.”42 As a result, Burns’s proposed mansion failed to trigger First Amendment coverage. Responding to the dissent, Judge Luck noted the narrowness of this holding. The court determined only that Burns’s residential architecture was not expressive conduct, not that residential architecture could never be expressive conduct.43
In dissent, Judge Marcus criticized the majority for failing to extend First Amendment protection to Burns’s proposed mansion.44 As a preliminary matter, Judge Marcus would have held that architecture is art.45 Nevertheless, Judge Marcus agreed with the majority that Spence was the appropriate standard for assessing whether a structure is expressive enough to qualify as architectural art.46 Unlike the majority, however, Judge Marcus concluded that Burns’s mansion did satisfy Spence prong two. First, Judge Marcus argued that the “reasonable viewer exists here.”47 Burns’s proposed mansion, contrary to the majority, would have been visible to the public, and even if it was not, the reasonable viewer could be an invited guest.48 Second, Judge Marcus found that the reasonable viewer would likely understand Burns’s mansion as communicating some message, particularly when considered in context.49 As a result, First Amendment coverage was triggered.50
Having concluded that Burns’s mansion was expressive conduct under Spence, the dissent’s next task was to determine the level of scrutiny with which to evaluate the ARCOM Ordinances. Judge Marcus decided the Ordinances were content based — assessments of architectural harmony necessarily turn on the content, that is, architectural style, of proposed structures — and that strict scrutiny was therefore appropriate.51 ARCOM failed strict scrutiny. The Town’s only asserted interest, Judge Marcus argued, was in “aesthetic uniformity,”52 and aesthetics alone could not justify content-based restrictions on Burns’s First Amendment freedom of expression.53
Though Burns’s First Amendment claim failed, the majority and the dissent both relied on expressive conduct doctrine to evaluate the architectural claim. Expressive conduct, however, is a poor fit for architecture, and insofar as architecture is art, it should be understood as a sui generis category of art, one that demands either a First Amendment standard tailored to its unique character or no coverage at all.
The First Amendment encompasses areas of expression beyond written or spoken words.54 One such area, as noted in Burns, is expressive conduct. In general, conduct, whether walking down the street or driving to work, is not covered under the First Amendment.55 Some conduct, however, is deemed so expressive as to merit First Amendment coverage. The Spence test, applied by both the majority and the dissent, has emerged as the tool for determining when conduct is sufficiently expressive to merit coverage.56
The Spence test, however, is a poor fit for architecture, as well as for art more generally. To understand the problem, consider how the Spence analysis might proceed for John Constable’s The Hay Wain. Painted in 1821, The Hay Wain is regarded as a masterpiece of English landscape painting.57 It also, like Burns’s proposed mansion, exists in an area of expression beyond written or spoken words. And so, the Burns opinions would suggest that Spence is applicable. The application of Spence, however, creates conceptual confusion and anomalous results. Regarding prong one, it is not clear that Constable had a subjective intent to communicate a particularized message. Some artists do (e.g., political cartoonists), but many do not, and it is unclear why intent to communicate a message should be dispositive in the context of art.58 Regarding prong two, it is even less clear that the reasonable viewer would be likely to understand The Hay Wain as communicating some message. Application of Spence thus creates an anomaly: Spence would plausibly deny First Amendment coverage for The Hay Wain and other artistic objects even though the Supreme Court has assured us that art is protected under the First Amendment.
Art is presumably protected under the First Amendment, but the Supreme Court has never explained how or to what extent.59 Instead, the Supreme Court has assumed, usually in passing, that art is “unquestionably”60 (and probably strongly) protected.61 Left unexplained is how to determine what constitutes “art.” What is clear is that the Spence test, contrary to the dissent, is a poor candidate for evaluating whether something qualifies as “art.” Neither prong seems to track the core attributes of art, whatever they may be.
Architecture is art. The dissent, as well as legal commentators, are almost certainly correct in that conclusion.62 If architecture is art, however, and art receives strong First Amendment protection, then the question becomes what structures qualify as architecture. Some have argued for a broad conception of architecture that would, for example, treat all single-family houses as fully protected First Amendment expression.63 Others, including the dissent, find the broad view implausible — conceptually and pragmatically — and thus need some mechanism for distinguishing buildings (unprotected) from architecture (protected).64
The mechanisms adopted — expressive conduct or otherwise — have at least two additional shortcomings, apart from poor fit. First, they would involve the courts in line drawing between different styles. In Burns, for example, the dissent concluded that Burns’s International Style mansion was art (perhaps universally but certainly in Palm Beach) while seeming to suggest that the surrounding Georgian and Regency mansions were not.65 Second, and more importantly, the practical effect of distinguishing some houses as buildings (unprotected) and others as architecture (protected) is to privilege those structures that are increasingly removed from the conventional house, which is presumably just a building (unprotected). In practice, the more bizarre, unusual, or un-house-like a house becomes, the more likely it would be to merit First Amendment protection. Both the dissent and legal commentators acknowledge as much, with some arguing that the protection of “iconoclastic” houses is the whole point.66 The First Amendment, however, is not limited to the protection of iconoclasm and neither should the constitutional protections for architecture, whatever they may be.
Though architecture is art, architecture is distinctive among the arts. For this reason, even if a structure qualifies as “architecture,” it should not receive the same heightened protection that presumably attaches to other arts. At minimum, any First Amendment standard should take into account at least three distinguishing features of architecture.67 First, architecture is functional, not incidentally but essentially. Commentators, including the dissent, often acknowledge architecture’s functionality but counter that all art is functional; all art can be put to use.68 That argument confuses incidental attributes with essential ones. One might be able to live inside some sculptures, but the ability to do so is incidental to sculpture. One must be able to live inside residential architecture, as that is essential. Architecture’s functionality therefore complicates the task of distinguishing genuine architectural expression (e.g., harmony with nature) from functional architectural preferences (e.g., more natural light), as both are necessarily related and easily reframed in terms of the other.
Second, architecture is permanently localized. Most art can be realized in an infinite number of locations, whether by performance or transport or reproduction.69 Architecture cannot. Architecture exists in one place, forms the built environment of that place and no other, and persists in the community after the artist-owner departs. The community thus has a stronger, more particularized, and more concrete interest in preventing bad architecture than in preventing, for example, bad poetry.70 No other art presents such localized and extreme burdens.
Third, architecture is always a public imposition. Art is generally enjoyed at the private level as a matter of personal choice.71 Architecture, by contrast, is the movie that must be seen day in and day out for decades. Moreover, the publicness of architecture suggests the inadequacy of the dissent’s notion that communities must endure “art that we hate” to enjoy “art that we love.”72 Architecture we hate prevents us from enjoying architecture we love, as architecture’s publicness shapes the context in which other architecture is viewed and appreciated.
In sum, architecture is essentially functional, localized, and public, all of which have implications for its First Amendment status. As the Supreme Court acknowledged in Southeastern Promotions, Ltd. v. Conrad,73 each “medium of expression . . . must be assessed for First Amendment purposes by standards suited to it, for each may present its own problems.”74 Architecture poses its own problems.
As one of the first cases to engage with those problems, Burns will likely have important implications going forward. Future courts, however, should consider Conrad’s exhortation and either develop a First Amendment standard tailored to architecture’s unique character or decline to extend First Amendment coverage. Whatever consensus emerges, the First Amendment does not and should not impede the authority of political communities to ensure that their architecture is not only strong and useful but beautiful as well.