Markets, Rights, and Discrimination by Customers

Discrimination by Customers is motivated by a puzzle.1 Many firms are prohibited from engaging in certain discriminatory practices. For example, in deciding which customers to serve, or which job applicants to hire. Katherine Bartlett and Mitu Gulati see the reason for these prohibitions as straightforward: The prohibited forms of discrimination are harmful.2 But, the authors believe, discrimination by customers is similarly harmful.3 And yet we don’t currently prohibit it. The puzzle is why. Why do we prohibit, we think legitimately, harmful discrimination by employers and sellers, while allowing harmful discrimination by customers?4 In seeking an answer to this, the authors investigate what they take to be the two most promising justifications for this asymmetric treatment. Namely, (1) an argument that appeals to efficacy and (2) an argument that appeals to ideals of privacy and autonomy. They end up defending the latter.5 They believe that the direct regulation of customer discrimination would unjustifiably impinge upon consumers’ privacy and autonomy – what they call the “basic liberty interests” of citizens.6 Indeed, while they never fully commit to the position, they at least suggest that these consumer privacy and autonomy interests are constitutionally enshrined. But notwithstanding these liberty interests, Bartlett and Gulati believe that at least some forms of customer discrimination can be regulated, albeit indirectly, through the placement of additional antidiscrimination duties on the firms that facilitate that customer discrimination.7 More specifically, they “propose that entities that already have a legal obligation not to discriminate . . . also should have an explicit obligation to curtail and not to facilitate discrimination by their customers, and to refrain from discrimination when they, themselves, act in the role of customer.”8

Bartlett and Gulati’s framing of the puzzle is useful and most will agree with the background assumptions that motivate it. Government regulation of private (i.e., non-government) discrimination can be legitimate and the current regulation of firm discrimination is an example of legitimate regulation of this sort. Discrimination by customers looks similarly harmful to the discrimination by firms that we legitimately proscribe and so something must explain our asymmetric treatment. And while pragmatic and efficacy concerns may have some explanatory force, they do not seem to fully capture what’s going on. Liberty, Autonomy, and Privacy—those are important and seem imperiled by laws prohibiting consumer discrimination, and so their protection provides a more plausible explanation of the asymmetry. The upshot being that, to the extent we want to further protect market actors from discrimination, we must regulate around these consumer interests. But while all of this is intuitively appealing, Bartlett and Gulati’s account faces a number of problems. I offer some concerns in Section I on the efficacy argument that the authors ultimately reject. In Section II, I explain that if the authors mean to suggest that there is a constitutional right for customers to engage in harmful discriminatory conduct, then their proposal for redressing the relevant harms indirectly, by imposing additional antidiscrimination duties on firms, might turn out to be constitutionally problematic. Moreover, even if Bartlett and Gulati disavow any constitutional basis for these consumer “basic liberty interests,” so long as they remain committed to the position that the government cannot directly prohibit consumer discrimination without violating them, their proposal of indirect regulation is in tension with the more general principle that if the end is prohibited, one is prohibited from trying to effectuate that end through indirect means as much as through direct ones. Bartlett and Gulati may reject that principle but given our general commitment to it, they need to provide an account of why it should be rejected. My primary critical challenge is in Section III. There, I criticize the authors’ autonomy- and privacy-based rationale for specially protecting discrimination by customers. I argue that this rationale needs to be supplemented by a more expansive account of how the relevant constitutional rights transfer over from non-commercial contexts into commercial ones.

I.     Efficacy and Antidiscrimination Law

Bartlett and Gulati find it prima facie plausible that reasons of efficacy explain why we prohibit harmful discrimination by firms but not by customers.9 Firms are easier to regulate, for one thing.10 They are also typically the cheapest cost avoiders.11 Firms also have independent reasons to avoid discrimination, the authors say, since discrimination involves operational inefficiencies that work against profitability.12 Bartlett and Gulati also appeal to exposure theory to suggest that the regulation of discrimination by firms reduces customer discrimination indirectly, because it exposes the customer to members of discriminated-against groups, and thereby tends to reduce the customer’s “biases toward these groups, thus altering the preferences that drive customer discrimination.”13

Ultimately, however, the authors find problems with an efficacy-based explanation for the differing treatment of discrimination by customers and firms. First, the authors seem to reject the argument that efficacy should determine our antidiscrimination regime. As they see it, discrimination by customers is harmful, both in individual instances and in its cumulative effect, and we have reason to directly regulate this harmful conduct regardless of whether this is the most efficient antidiscrimination policy strategy overall.14 The authors also raise two objections internal to the efficacy explanation for our focus on firms. First, they claim that direct regulation of customer discrimination may be especially important for reducing customer discrimination because customers are more likely than firms to be motivated by discriminatory biases and less likely to resist those biases due to countervailing motives (e.g., a profit motive).15 Second, and most significantly, an appeal to considerations of efficacy, to defend an antidiscrimination regime focused on firms, mistakenly assumes that firms can most effectively reduce discrimination by refraining from it themselves, when in fact the causal story is more complex.16 Customer discrimination affects business practices, and business practices influence customers’ discriminatory biases in turn. For the authors, an exclusive focus on firms “ignores the complex interaction between business practices and customer preferences.”17 And this undermines any efficacy-based justification for regulating discrimination by firms while leaving customer discrimination unregulated.18

Before getting to my more fundamental criticism of Bartlett and Gulati’s efficacy analysis, I have two concerns with their understanding of how discrimination works and how it can be lessened. To start, the authors appeal to Gary Becker’s economic model of discrimination, which treats discrimination as a matter of individual taste, to support the claim that discrimination is inefficient, and that firms therefore have a profit motive to avoid it.19 While they admit that this sort of modeling is oversimplified, others (with whom I agree) argue that Becker’s model is not oversimplified but false.20 I am also skeptical of Bartlett and Gulati’s claim that forcing customers to interact with members of discriminated-against groups who are in service positions, which is achieved by prohibiting firms from discriminating against employees, will lessen discrimination. Lessening wrongful discrimination almost certainly requires integration, but of a more systematic kind.21 According to the contact hypothesis, for intergroup interaction to reduce prejudice, the contact must, among other things, occur between individuals with equal status and be sufficiently frequent to lead to personal acquaintances and informal interaction.22 These forms of contact generally are not present in the modern relationship between service workers and customers, to put it mildly. Perhaps the authors’ idea will work when customers interact with high-status service providers (e.g., lawyers and doctors), but that does not make up the majority of customer-service worker interactions most of us face.23 The mischaracterization of discrimination’s mechanics––what it is, how it works, and how contact among members of various groups can (and cannot) help eliminate it––renders unpersuasive their efficacy analysis.

My larger concern, though, is more fundamental. Although in the end Bartlett and Gulati reject an efficacy-based argument for the differing treatment of discrimination by customers and firms, their discussion of this argument exhibits an explanatory reticence that runs through the entirety of Discrimination by Customers. In order to assess the efficacy of our current antidiscrimination regime, we must first specify at least two things: the purpose (or purposes) of the relevant law (or laws) and the limits within which the legislators were willing to pursue that purpose (or purposes). The need for the former is perhaps obvious: We cannot assess the efficacy of an antidiscrimination policy without knowing precisely what the antidiscrimination law aims to achieve. As for the latter specification, the Court in Rodriguez put it well:

No legislation pursues its purposes at all costs. Deciding what competing values will or will not be sacrificed to the achievement of a particular objective is the very essence of legislative choice––and it frustrates rather than effectuates legislative intent simplistically to assume that whatever furthers the statute’s primary objective must be the law.24

Even if we know a law’s objectives, we cannot assess its efficacy without knowing the limits placed on its pursuit of them.25  Bartlett and Gulati, however, do not provide a clear account of either of these things. We never get an account of the limits within which our antidiscrimination goals are pursued.26

As for what Bartlett and Gulati think the goals of antidiscrimination law are, it is unclear. They suggest in passing that the aim is to “stop discrimination,”27 but I would be surprised if the authors thought, as a descriptive matter, this is what our current laws aim at. Outside the market individuals and associations have wide discretion to discriminate and I suspect many think the law does not and ought not have the goal of trying to curtail it. As for intentional discrimination that the law allows firms to engage in within the market—e.g., against employees through the bona fide occupational requirement (BFOQ) exception and against contractors, given Title VII’s limited coverage—those gaps can plausibly be understood as either the result of the rough-and-tumble of realpolitik (and thus consistent with a goal of “stopping discrimination”) or instead as reflecting the limited aims our antidiscrimination laws have in the first instance. At another point the authors state that “the aim of anti-discrimination law is to change how individuals act, not what they believe.”28 But this description is also too vague and over-inclusive to be of any help in evaluating whether our current regime is efficient.

Setting aside the descriptive, if the authors are interested in whether the current system efficiently limits the types of discrimination we ought to proscribe, they should first clarify that they are talking in the normative register. But once clarified, a two-part normative inquiry remains: which types of discrimination can we legitimately deter or remedy and, of those, which ought our laws actually target.

One approach to answering the underlying normative question would focus on harming and wronging. In schematic terms, the idea would be that the person who is discriminated against is harmed or wronged in some contexts (e.g., a woman denied employment on grounds of her sex), but not in other contexts (e.g., a woman overlooked as a romantic partner by a gay man).29 But giving an account of precisely what makes discrimination wrongful or harmful—beyond this schematic, preliminary sketch—is no easy task, and the debates around this issue are extensive.30 Another related approach would be to focus on legitimacy. Regardless of whether a given form of discriminatory conduct is wrongful, it might be illegitimate for the state to regulate it. For any theory of the legitimacy conditions of legal regulation one might adopt, there would be theoretical complications to contend with.31

My general point is that while Bartlett and Gulati purport to be examining and ultimately challenging the efficacy explanation for the asymmetries in antidiscrimination law, no such examination is possible and their challenge to the efficacy explanation is more assertion than argument. To the extent their project is descriptive (i.e., is our current regime efficacious), we cannot know until we fix what our antidiscrimination laws are trying to achieve and within what limits. If the project is normative, we must conduct a broader survey of the normative terrain that antidiscrimination law governs—the kind of survey that would allow us to articulate precisely which ends we ought to be aiming to efficaciously realize—and then consider those aims alongside other values we seek to protect that come into conflict with them. Until this further work is done, the efficacy argument has a hole at its center.

II.     Regulating Customer Discrimination Indirectly

Where the efficacy argument falls short, Bartlett and Gulati think an argument appealing to reasons of privacy and autonomy will explain why the law only directly regulates discrimination by firms.32 In essence, customers have a privacy- and autonomy-based right to engage in discriminatory acts, irrespective of the harm those acts cause.33 On this basis the authors oppose any direct prohibition of customer discrimination, and instead favor an indirect approach to curtailing the relevant harms. They propose that “entities that already have a legal obligation not to discriminate” (e.g., employers of a certain size, educational institutions, and entities subject to public accommodations laws) should also “have an explicit obligation to curtail and not to facilitate discrimination by their customers, and to refrain from discrimination when they, themselves, act in the role of a customer.”34 In this section, I’ll discuss three worries about this proposal for the indirect regulation of harmful discriminatory conduct performed by customers. The first two stem from their suggestion that while direct prohibitions on consumer discrimination would violate consumer privacy- and autonomy-based rights, indirect prohibitions on the same conduct would not. The third concerns the legitimacy of placing additional burdens on firms.

To start, it is unclear what the authors take their citations to various constitutional cases to be doing. Do they think the customer autonomy ‘rights’ they point to are merely consumer interests that are normatively deserving of protection, where the case citations merely gesture at the content of those values? Or, instead, do they think these customer rights are actually constitutionally protected rights, where the cases are cited to show the sources of that constitutional protection?35 If it is the later (i.e., they think the customers have a constitutional right to discriminate), then the authors’ proposal to indirectly regulate exercises of this right will be constitutionally vulnerable.36 Consider, for instance, how indirect burdens on people’s speech rights and reproductive rights are widely understood as constitutionally problematic.37 As the Court has said of free speech rights, they are protected against both “heavy-handed frontal attack, but also from being stifled by more subtle governmental interference.”38 As I will explain in Section III, I am unconvinced that people do in fact have a constitutional right to engage in discriminatory conduct within the market, as customers. But if they do, then stifling the exercise of that right by subtle regulatory interference is likely to be illicit for much the same reasons that apply in the free speech and reproductive rights cases.

Indeed, if customers have a constitutional right to discriminate, then even some current antidiscrimination laws seem like they are open to constitutional challenge, and the narrow bona fide occupational qualification (“BFOQ”) exceptions that are sometimes used to satisfy the discriminatory preferences of customers look like a constitutional requirement rather than a mere pragmatic concession.39

Elsewhere I have discussed another problem with simultaneously assuming a consumer’s constitutional right to discriminate while also arguing for curbing its exercise through indirect methods.40 In short, while the state may be free to decide whether to prohibit the buying or selling (or both) of particular goods and services, the choice to prohibit only one side of that transaction seems permissible only if the state could have prohibited the other side as well.41 Lee Fennell has noted this issue in the context of fair housing law: It seems problematic to prohibit discriminatory collateral search behavior—e.g., the behavior of third-parties who assist homeseekers in engaging in discrimination when finding a property—if the underlying discriminatory conduct of the homeseeker is constitutionally protected.42

It is surprising that the authors do not acknowledge this concern, given that the type of argument sketched above has been successfully used by the on-demand economy firms they intend to target with their proposal. Consider, in which the Ninth Circuit had to decide, inter alia, whether an online company violated fair housing laws in prompting users who were seeking co-tenants to enter their racial preferences for potential roommates.43 The Ninth Circuit invoked the doctrine of constitutional avoidance to hold that the relevant fair housing laws did not prohibit consumers from racially discriminating in the selection of roommates and thus—crucially for Bartlett and Gulati—that it wasn’t unlawful for Roommate (the firm) to encourage those roommates to engage in racial discrimination.44 Put another way, Roommate was able to successfully avoid liability for its own policy by arguing that users had a constitutional right to discriminate in the relevant ways.45 The vague invocation of consumer rights to discriminate also makes appearances in BFOQ cases.46 Given that Airbnb was able to successfully compel arbitration in its own race discrimination case, we do not yet know what substantive arguments they will call upon.47 However, given the result in Roommate, it seems likely that they too will attempt to avoid liability by appealing to the protection of a purported customer constitutional right to discriminate.

Here is a third concern about Bartlett and Gulati’s proposal to indirectly regulate customer discrimination, by regulating firms that facilitate such discrimination. Although this proposal might not entail a major departure from the current state of antidiscrimination law, the authors are clear that they would seek to impose additional antidiscrimination duties on firms. Therefore, even if we assume that indirect controls on consumer discrimination can avoid autonomy- and privacy-based concerns, we will need some explanation of why firms should be seen as the legitimate bearers of antidiscrimination duties generally, and what the limits to those duties should be. Addressing this issue requires us to think more generally about what firms are and what sorts of norms they should be subject to. This may depend in part on one’s theory of the firm. If one subscribes to the view that firms are just artificial entities then there may be no in-principle limits to the antidiscrimination duties that a state can impose on them.48 But there are other theories of the nature of firms—the real entity theory, the nexus of contracts theory, and the collaboration theory—on which there will be limits to the extent to which firms can be legitimately saddled with these antidiscrimination duties, given the interests of the firm itself or of those who organize, operate, work within, or control it.49 For instance, under some theories of the firm, firms might be understood as both agents of and subscribers of justice.50 To the extent that firms are agents of justice—given how, in economic systems like ours, they control access to highly valuable goods and opportunities—we may think it appropriate to impose special antidiscrimination duties on them. However, to the extent they are subscribers of justice, firms will also have some claim to be treated justly, where this might mean there are limits to the antidiscrimination duties we can place on them. How to reconcile their dual agent-subscriber role and determine how they ought to be treated under different theories of the firm is no easy feat. But Bartlett and Gulati cannot just pile additional burdens on firms without taking a position on these issues.51

III.     The Customer’s Right to Discriminate

The pivotal thesis in Bartlett and Gulati’s account, again, is that customers have a privacy- and autonomy-based right to engage in discriminatory conduct. This right, they suggest, is constitutional.52 In order to defend this latter claim the authors appeal to First Amendment rights of expression and association, citing Rotary Club and NAACP,53 along with an amalgamation of constitutional rights pertaining to decisional and spatial privacy, drawing on Lawrence, Griswold, and Skinner.54 The contours of current antidiscrimination law can be understood, they suggest, as an expression of the governing principles in these cases.55 The solution to our initial puzzle can be roughly summarized as follows, then: Although we directly prohibit discrimination by firms, we should not directly prohibit discriminatory conduct by customers, because constitutional rights to individual privacy and individual autonomy disallow any direct legal intervention in discriminatory actions performed by private individuals.

One initial worry about this way of ascertaining the scope of privacy rights is that it operates at too abstract a level, and thus pays insufficient attention to how laws protecting individual privacy actually function.56 Setting that worry aside, though, I will focus my criticisms here on one of the crucial premises that underwrites Bartlett and Gulati’s reasoning, namely, the claim that “[i]t seems appropriate to grant [the] same freedom” one has outside of the market “in market transactions.”57 Let’s call this the parity premise, i.e., the idea that people’s constitutional rights apply much the same, regardless of whether they are acting in the market, as customers, or acting outside the commercial sphere. The importance of this premise to the authors’ view is evident given the cases they appeal to. Both Rotary Club and Lawrence, for example, tackle questions about the requirements of privacy and autonomy in relation to voluntary associations that bear no evident connection to commercial acts or practices. But despite the fact that these cases bear no relation to markets, the authors interpret them as showing that the constitution carries an expectation of “a privacy and autonomy right of individuals to make their own market decisions”—indeed, an expectation that’s “deep,” “strong,” and “broadly shared.”58 This reading relies on the parity premise.

The clearest argument the authors provide in defense of the parity premise comes when they argue that the state should not be allowed to prohibit dating websites from giving users tools to filter potential partners based on race. “[T]he freedom to make one’s own choices with respect to one’s sexuality, family, and procreation” they say, “is one of this nation’s most protected constitutional rights.”59 And therefore, they continue, “If individuals can apply their own selective criteria in choosing with whom to dance in a bar––hardly something one can imagine being the subject of state regulation––it seems reasonable that they also be able to do so when they access a commercial dating site.”60 In essence, their argument is that if the Constitution protects a certain type of discriminatory choice in non-commercial interactions, as in a person racially discriminating in their choice of dance partner, then it should protect similar choices in commercial contexts, as in someone racially discriminating in interactions on a dating website. But before I get to my more substantive concern, notice that even if we grant the parity premise there is a questionable inference here. Although it’s true that people are free to choose to dance with whomever they like at a bar, this does not mean bars are permitted to facilitate racially discriminatory choices via any means whatsoever. The dating website’s racial filter is arguably more like a bar that sets up segregated dance floors and allows patrons to select the ethnicity of their dance partner in that way. Even accepting arguendo that cases like Lawrence and Rotary Club mean that people have a constitutional right to racially discriminate qua customers does not yet entail anything about whether firms always are or should be protected in assisting exercises of that right.61

At any rate, I am skeptical that cases like Lawrence and Rotary Club in fact support the key claim that people have a constitutional right to racially discriminate qua customers. Take Lawrence.62 To support its holding—striking down as unconstitutional a law criminalizing sodomy—the Court adopted two arguments from Justice Stevens’ dissent in Bowers. First, the state cannot prohibit a practice merely because the majority sees it as immoral. Second, people’s decisions “concerning the intimacies of their physical relationship” are “a form of ‘liberty’ protected by the due process clause of the Fourteenth Amendment.”63 The question for us is what these arguments or principles entail in relation to whether the direct regulation of discriminatory conduct by customers infringes on the individual’s privacy and autonomy rights.

It seems to me that the principles in Lawrence provide little support for Bartlett and Gulati’s position. One reason for this is that those who endorse the regulation of customer discrimination think this is merited not because such discrimination is immoral, but because it is harmful. Indeed, the authors grant this claim in their framing of the puzzle. Lawrence says that the state cannot prohibit mere immorality, but the case for regulating customer discrimination, just like the case for regulating discrimination by firms, has never rested on the charge that it is merely immoral.64

The more difficult question in interpreting Lawrence’s implications pertains to the scope of the liberty protected by substantive due process. Given the ongoing debate and uncertainty about the scope of this liberty, there is room for reasonable doubt here about whether overtly commercial acts—acts of buying, selling, tendering, contracting, etc.—are encompassed by substantive due process.65 And such doubts are reinforced when we inspect the substance of the reasoning in Lawrence. In striking down prohibitions on sodomy, the Court saw it as noteworthy that the case did “not involve public conduct or prostitution.”66 The apparent suggestion is that the applicability of the due process clause in this context owes to the fact that the statute in question was regulating unpaid sex acts among private partners, as opposed to sex acts that were part of a commercial transaction. What this implies, then, is that the anti-sodomy statute in Lawrence violated substantive due process (at least in part) because it prohibited sex acts in non-commercial contexts, as opposed to sex acts between sex workers and their clients.

In light of the reasoning that underpins Lawrence, then, the natural conclusion is that protected liberty interests in sex outside of market contexts do not automatically or necessarily transfer over when people are acting in the guise of customer. This is at odds with the authors’ reliance on the parity principle. They say if people are protected in performing an act outside the market, they are also protected in performing that act in the market.67 But commercial interactions seem to modify the application of people’s constitutional rights to privacy and autonomy, such that there is no simple parity, as a descriptive matter, in which acts are protected by those rights when people move between market and non-market contexts. 

Why, then, do Bartlett and Gulati see Lawrence as implying a privacy- and autonomy-based right for individuals to make whatever market decisions they like? Here is a line of reasoning that one could follow to try to defend that interpretation.

The first step is to argue that, contrary to what is suggested by the remarks about prostitution as I have explained them above, Lawrence in fact should cast doubt on the constitutional permissibility of bans on prostitution.68 On this view, if people’s decisions about their intimate physical relationships are protected by the Fourteenth Amendment’s due process clause, then these decisions are protected regardless of whether they involve any kind of commercial transaction. 69

The second step is to argue that what is true for decisions about sex also applies to various other decisions that involve a significant exercise of personal discretion in one’s voluntary associations. If you are free to have consensual sex however and with whomever you like (regardless of whether money changes hands), then you are free to converse, fraternize, dance, shake hands, exchange goods, or spend time, however and with whomever you like (again, regardless of whether money changes hands). And this means that customers have a right to engage in potentially harmful forms of discrimination. If you do not want to be operated on by a female surgeon, or receive treatment from a gay masseuse, or if you want to exclude certain racial groups when selecting an (otherwise anonymous) sperm donor in your fertility treatment, that is your constitutional right.

This seems to be roughly what Bartlett and Gulati have in mind. They see cases like Lawrence as standing for something far broader than the right to choose one’s sexual partner without the law’s intrusion. Instead, they see Lawrence as standing for something like a right to “decisional privacy,” which they think is infringed upon by laws directly prohibiting discriminatory conduct on the part of customers.

The problem with all this is descriptive, and potentially normative. As a descriptive matter, the introduction of market activity does matter for determining the scope and contours of rights. Across a range of different areas of law—lying,70 fair use,71 and sexual harassment,72 to name a few—people’s constitutional rights are routinely limited and constrained in a variety of ways when they are exercised in the market. So, when Bartlett and Gulati say it seems reasonable to grant the same freedoms one has outside the market to market transactions, they are endorsing a more radical view than they acknowledge.73 In terms of general normative theory, defenders of this view will need to subscribe to some kind of sweeping economic libertarianism. They will need to deny that there is anything structurally distinctive about interactions and decisions in market contexts, and they will need to characterize the acts people perform when acting as workers, managers, contractors, vendors, customers, etc. as normatively equivalent to—that is to say, safeguarded by the same schedule of rights as—the acts that people perform in private and informal social intercourse.

Of course there are arguments available to those who want to defend that kind of radical libertarianism as a matter of normative theory.74 However, Bartlett and Gulati are not well-placed to avail themselves of those arguments. First, they are meant to be grounding their justification for the protection of discrimination by customers in constitutional principles. The radical libertarian position I’ve sketched here would overturn a wide swathe of settled constitutional doctrine. Second, the libertarian position does not only undermine the legal regulation of discrimination by customers, it also undermines the justificatory framework that underpins the regulation of labor and markets in general, including regulations on discriminatory actions of firms.75 But Bartlett and Gulati’s whole theoretical approach to this issue is bounded by their explicit endorsement of the legitimacy of imposing antidiscrimination duties on firms.

What is needed, in order to defend an asymmetry in the imposition of antidiscrimination duties on firms and customers, is a much more finely calibrated account of when and how individual rights are modified when people move from acting in non-market to market contexts.76 The claim cannot just be that we grant the very same freedoms one has outside in the market in market transactions. Rather, for Bartlett and Gulati, it has to be something like this: People bring their full schedule of constitutional rights with them when they act in the market as customers, but when they act in the market in other roles (e.g., as employees in firms) the application of those rights is significantly curtailed.

But why should we accept such a view? It would be premature to simply assert that this kind of view is utterly untenable as a reading of the principles underlying cases like Lawrence. My point is that getting from Lawrence to this kind of conclusion requires a more detailed and ambitious reading of the reasoning involved in Lawrence. More than merely observing its (uncontroversial) implications regarding individual privacy in non-commercial sexual contexts.

The puzzle motivating Discrimination by Customers is whether the asymmetric treatment of consumer and firm discrimination is justified. The intuition that prohibiting customer discrimination directly may violate autonomy or privacy is merely a starting point. To build on this we need a comprehensive account of what is, and what ought to be, the legal significance of the market on the contours of the rights, obligations, and interests of actors operating within it. Citations to cases like Lawrence pump moral intuitions about the rightness or wrongness of the outcomes in those cases in a way that invites us to quickly accept as true (or false) Bartlett and Gulati’s assertion that constitutional rights (or values) are at stake and that the laws in question risk violating them. But this sort of intuition-pumping surreptitiously elides the very inquiry Bartlett and Gulati set out to undertake.


  1. [1]. Katharine T. Bartlett & Mitu Gulati, Discrimination by Customers, 102 Iowa L. Rev. 223, 226 (2016).

  2. [2]. Id.

  3. [3]. Id. at 230.

  4. [4]. Bartlett and Gulati do not see all customer discrimination as harmful. Some customer discrimination they see as harmless, or even desirable, e.g., customer-led campaigns aimed at helping black-owned businesses. See id. at 242. Nonetheless, they maintain that much discriminatory conduct by customers is harmful. Ultimately, they divide this conduct into two categories: (i) that which is harmful but must be tolerated because of privacy and autonomy rights, and (ii) that which is liable to regulation in principle because it does not as seriously impinge upon privacy and autonomy rights. See id. at 242–45.

  5. [5]. Id. at 249.

  6. [6]. Id. at 227.

  7. [7]. Bartlett and Gulati’s work may be situated within both the second and third wave of theoretical work on antidiscrimination law. The first waved focused on justifications for affirmative action, the second on the proper scope of antidiscrimination law, and the third (and current) on when and why discrimination is objectionable. For discussion of these waves see John Gardner, Lippert-Rasmussen, Kasper. Born Free and Equal? A Philosophical Inquiry into the Nature of Discrimination, 125 Ethics 1204, 1204–05 (2015) (book review).

  8. [8]. Bartlett & Gulati, supra note 1, at 249.

  9. [9]. See id. at 228–30.

  10. [10]. See id. at 228.

  11. [11]. See id. at 228–29.

  12. [12]. See id. at 229.

  13. [13]. Id.

  14. [14]. See id. at 230. Latent here may be some intuitions about desert and the law’s role in holding people accountable for the harms they inflict.

  15. [15]. See id. Given they assert this within their efficacy analysis, I take it Bartlett and Gulati must think this dynamic means a focus on firms is therefore not efficient.

  16. [16]. Id. at 230–31.

  17. [17]. Id. at 231.

  18. [18]. See id. at 238. It is unclear why recognition of this feedback loop cannot, in principle, be squared with the kind of efficacy-based argument the authors are critiquing. A legal regime whose architects were fully aware of the effects of this feedback loop might seek to prohibit firms from enacting policies to facilitate customer discrimination in order to address the very problem that the feedback loop contributes to. In short, it seems at least possible for a firm-only antidiscrimination regime to be sensitive to the effects of the feedback loop the authors describe. (Note: the hypothetical I describe above is different to our current antidiscrimination regime. Title VII does not allow firms to cater to customers’ discriminatory preferences in general, but it does allow some forms of customer discrimination to influence the actions of firms. See42 U.S.C. § 2000e–2(e)(1) (2012) (Title VII’s “bona fide occupational qualification” (BFOQ) exception does not apply to race); 29 C.F.R.
    § 1604.2(a)(1)(iii) (2016) (stating that, in general, the BFOQ exception will not cover “[t]he refusal to hire an individual because of the preferences of coworkers, the employer, clients or customers”) (emphasis added)).

  19. [19]. See Bartlett & Gulati, supra note 1, at 229.

  20. [20]. See generally, e.g., Richard H. McAdams, Cooperation and Conflict: The Economics of Group Status Production and Race Discrimination, 108 Harv. L. Rev. 1003 (1995) (questioning the validity of Becker’s model based on a study of intra-group relations and discrimination).

  21. [21]. See generally Elizabeth Anderson, The Imperative of Integration (2010) (arguing for the need for robust racial integration of this sort).

  22. [22]. See generally Gordon W. Allport, The Nature of Prejudice (1954) (finding that intergroup contact reduces prejudice only when that contact is coupled with equal status, common goals, intergroup cooperation, and social sanction of the contact via the support of law, authorities, or custom).

  23. [23]. Though my skepticism here ends up supporting Bartlett and Gulati’s ultimate conclusion that an exclusive focus on firm discrimination may not be efficient.

  24. [24]. Rodriguez v. United States, 480 U.S. 522, 525–26 (1987) (emphasis omitted).

  25. [25]. And with regards to “how far” the legislators are willing to go to pursue their objectives, “you can’t tease that out of a purpose.” Frank H. Easterbrook, The Absence of Method in Statutory Interpretation, 84 U. Chi. L. Rev. 81, 81 (2017) (emphasis omitted).

  26. [26]. We might understand their autonomy and privacy concerns as themselves creating some of the limits within which we pursue the goals of antidiscrimination law. If that is right, then the efficacy analysis should be conducted in light of those concerns.

  27. [27]. See Bartlett & Gulati, supra note 1, at 238 (concluding that the efficacy argument for regulating only firms assumes, inter alia, that “firms have the capacity to stop discrimination” on their own).

  28. [28]. Id. This claim also seems in tension with the authors’ earlier analysis. When discussing the efficacy arguments, Bartlett and Gulati speculate that focusing on firms might make sense because it will result in customers being forced to interact with people they would otherwise avoid. Invoking the exposure theory, the authors think it is plausible that these interactions will result in a lessening of customers’ discriminatory preferences. See id. at 229 (explaining the efficacy rationale). This makes it sound like the authors think one of the goals of antidiscrimination law is to change what people believe by changing the material conditions under which their preferences are formed. They would not be alone in worrying about laws that seek to shape individuals’ beliefs. See Seana Valentine Shiffrin, What is Really Wrong With Compelled Association?, 99 Nw. U. L. Rev. 839 (2005) (arguing the application of antidiscrimination laws to private associations can infringe on mental autonomy). But see infra Part II (detailing my concerns with this argument); Heather M. Whitney, The Autonomy Defense of Private Discrimination (Feb. 24, 2017) (unpublished manuscript), (explaining and critiquing how autonomy is used in antidiscrimination scholarship).

  29. [29]. One could contest the substantive claims embedded in these examples. The point of the examples is just to give a rough indication of the contours of the relevant distinction.

  30. [30]. For a sampling of this literature see, e.g., Tarunabh Khaitan, A Theory of Discrimination Law 117–39 (2015); Kasper Lippert-Rasmussen, Born Free and Equal? A Philosophical Inquiry into the Nature of Discrimination 103–92 (2014) (discussing, in Part II, the wrongness of discrimination); Philosophical Foundations of Discrimination Law (Deborah Hellman & Sophia Moreau eds., 2013); Larry Alexander, What Makes Wrongful Discrimination Wrong? Biases, Preferences, Stereotypes, and Proxies, 141 U. Pa. L. Rev. 149, 218 (1992).

  31. [31]. There are some accounts of legitimacy that re-inscribe a version of the harm principle, by taking the view that the law’s legitimate purpose is to prevent harm to others. But there are also justice-based accounts of legitimacy, on which the law’s purpose is to ensure corrective or distributive justice, or both, irrespective of whether this involves preventing harms. See John Gardner, Discrimination as Injustice, 16 Oxford J. L. Stud. 353, 365 (1996). For more on this see generally Heather M. Whitney, The Regulation of Discrimination by Individuals in the Market, 2017 U. Chi. Legal F. (forthcoming 2017) (on file with author) (discussing some of these complications).

  32. [32]. Bartlett & Gulati, supra note 1, at 238–39.

  33. [33]. Id. at 238.

  34. [34]. Id. 249.

  35. [35]. See, e.g., id. at 227 (“Society may regulate firms and not customers because it believes that limiting customer choice would infringe on the basic liberty interests of its citizens in ways not present in the regulation of firms. The concern with such potential infringement is not only that it potentially violates strong constitutional values ... .”) (emphasis added); id. at 239 (noting the “constitutional importance” of consumer privacy and autonomy).

  36. [36]. I focus here on the constitutional case but my concern about indirect regulation remains even if Bartlett and Gulati think the consumer right not constitutionally protected. In general, if something is prohibited—here they think that ‘something’ is prohibiting consumers from discriminating—then the law forbids doing that thing through both direct and indirect means.

  37. [37]. For speech see, e.g., Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786 (2011) (striking down a law banning the sale of violent video games to minors, absent parental approval, because the law violates the First Amendment rights of the children buyers); Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 71 (1963) (striking down a system that had the de facto effect of prohibiting the distribution of certain materials on the grounds that it was an unconstitutional infringement on the speech rights of, inter alia, readers). See also Seth F. Kreimer, Censorship by Proxy: The First Amendment, Internet Intermediaries, and the Problem of the Weakest Link, 155 U. Pa. L. Rev. 11 (2006) (arguing that the First Amendment rights of downstream users can be used to protect against the indirect censorship of those users, enacted via the direct censorship of internet intermediaries). As for abortion see, e.g., Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833 (1992) (discussing indirect burdens on abortion rights); Planned Parenthood of Wis., Inc. v. Schimel, 806 F.3d 908 (7th Cir. 2015) (holding Wisconsin law requiring abortion providers to obtain admitting privileges at nearby hospitals an unconstitutional unburden on women’s right to an abortion).

  38. [38]. Bates v. City of Little Rock, 361 U.S. 516, 523 (1960).

  39. [39]. See, e.g., 29 C.F.R. § 1604.2(a)(1)(iii) (2016) (finding that “[t]he refusal to hire an individual because of the preferences of coworkers, the employer, clients or customers” does not qualify as a legitimate BFOQ unless the refusal is in regard to sex “[w]here [sex] is necessary for the purpose of authenticity or genuineness”). Of course, some do argue that current antidiscrimination laws are unconstitutional. For instance, that some sexual harassment laws violate the First Amendment rights of workers. See generally Eugene Volokh,Freedom of Speech and Workplace Harassment,39 UCLA L. Rev. 1791, 1846 (1992) (arguing that the state cannot legitimately regulate all workplace speech contributing to a hostile environment but instead only “directed speech––speech that is aimed at a particular employee because of her race, sex, religion, or national origin”); id. (comparing “directed speech” to “[u]ndirected speech,” which is “speech between other employees that is overheard by the offended employee, or printed material [like pornographic or bigoted posters], intended to communicate to the other employees in general”).

  40. [40]. See Whitney, supra note 31, (manuscript at 4–6).

  41. [41]. See id. That said, I can see a conceptual way to find it legitimate to prohibit the sale but not the purchase of certain forms of discrimination. Borrowing Hohfeld’s taxonomy of legal concepts, one posits: (1) individuals qua consumers have a privilege to discriminate against all others within the market; (2) workers have a claim right to not be discriminated against by their employers (which entails that firms qua employers have a correlative duty to not discriminate against their workers); (3) the strength of (2) is such that it outweighs both any liability the firm qua seller might have toward consumers, to aid consumer discrimination, and outweighs any objections consumers can make against laws that recognize and enforce firms’ duties to their employees that have the effect of limiting the ability of consumers to engage in their privileged discrimination; and (4) individuals qua consumers have no duty to aid employers in the performance of (2). If one accepts (1)-(4), then we can see one way prohibiting firms qua employers from selling discrimination might be permissible (at least when the type of discrimination sold entails discriminating against employees) while directly prohibiting consumers from trying to buy that discrimination is not. As readers no doubt see, this setup also provides a solution to Bartlett and Gulati’s original puzzle. But the above simply lays out the suppositions one would need in order to explain why the sale but not purchase of certain forms of discrimination is acceptable. Bartlett and Gulati would then need to defend against reasonable alternatives this allocation and relative weighing of duties, rights, and entitlements. This setup would also commit Bartlett and Gulati to a position they seem unwilling to take. Namely, that while our treatment of firm and customer discrimination appears asymmetric, because both seem normatively on par as similarly harmful, it is not: firm discrimination is harmful while consumer discrimination is not. Why would they be forced into this position? If we adopt a Millian framework and find that a harm, in contrast to a mere offense, occurs when someone fails to treat another in the way that other is entitled to be treated, the relevant correlatives of the above
    (1)-(3) are: (5) firms harm employees when they discriminate against them (because the firm is violating those workers right to not be discriminated against by firms, given (2)), and (6) consumers do not harm employees when they discriminate against them (because workers have no right to not be discriminated against by customers, given (1 and 4)). I develop this much more elsewhere. See generally Whitney, supra note 28 (laying bare the presuppositions undergirding claims that regulation of certain forms of discrimination violates the autonomy of the would-be discriminator). For more on Hohfeld’s taxonomy see generally Wesley Newcomb Hohfeld, Fundamental Legal Conceptions as Applied in Judicial Reasoning, 26 Yale L.J. 710 (1917); Wesley Newcomb Hohfeld, Some Fundamental Legal Conceptions as Applied to Judicial Reasoning, 23 Yale L.J. 16 (1913).

  42. [42]. Lee Anne Fennell, Searching for Fair Housing, 97 B.U. L. Rev. 349, 389 (2017) (finding that while “§ 1982 does not offer an especially useful or attractive stand-alone vehicle for reaching homeseeking” it nevertheless “buttresses (and saves from a key criticism) the use of other provisions within the FHA, including the prohibition on discriminatory advertisements and statements, to reach certain collateral actions undertaken by homeseekers or by those assisting them in the course of conducing a housing search”).

  43. [43]. Fair Hous. Council of San Fernando Valley v., LLC, 666 F.3d 1216 (9th Cir. 2012).

  44. [44]. Id. at 1223.

  45. [45]. See, e.g., Appellant, LLC’s Third Brief on Cross-Appeal at 51,, LLC v. Fair Hous. Council of San Fernando Valley, Nos. 09-55272, 09-55875, 09-55969, 2010 WL 2751575 (9th Cir. May 14, 2010) (“The injunction impermissibly inhibits the rights of homeseekers to state and act on preferences, however, by forbidding Roommate from making available to them formatted questions or matching using even voluntary responses. Denying such users of the choice of focusing their search based on the characteristics of those with a home to share is not only unwarranted, it is unconstitutional.”). Given the Ninth Circuit invoked constitutional avoidance it is possible that prohibiting both discrimination in roommate selection and its facilitation is constitutional (that we cannot say so definitively is one of the problems with the canon). But what we can say is that “those who invoke the doctrine must believe that the alternative is a serious likelihood that the statute will be held unconstitutional.” Almendarez-Torres v. United States, 523 U.S. 224, 238 (1998). This serious likelihood, especially in conjunction with the quasi-precedential effect constitutional avoidance is known to sometimes have, is enough to raise problems for Bartlett and Gulati’s proposal of indirect regulation. See Neal Kumar Katyal & Thomas P. Schmidt, Active Avoidance: The Modern Supreme Court and Legal Change, 128 Harv. L. Rev. 2109, 2135–37 (2015) (discussing this effect and how it allows later courts to avoid engaging in robust analysis of the original issue).

  46. [46]. See Whitney, supra note 31, (manuscript at 17–29).

  47. [47]. See Selden v. Airbnb, Inc., No. 16-CV-00933, 2016 WL 6476934, at *9 (D.D.C. Nov. 1, 2016), motion to certify appeal denied, No. 16-CV-933, 2016 WL 7373776 (D.D.C. Dec. 19, 2016).

  48. [48]. The artificial entity theory, also known as the concession theory, posits that because corporations are creatures of statute, owing their existence to a government grant of rights and privileges, the state has the power to limit the scope of permissible corporate behavior. On this view, I could imagine an argument that the state is free to set whatever requirements, vis-à-vis antidiscrimination law, that it wants and, because corporations have no right to operate, there can be no principled objection to these. The artificial entity theory is not in vogue with the Court today.

  49. [49]. For a brief but clear overview of the artificial entity theory see Eric C. Chaffee, The Origins of Corporate Social Responsibility, 85 Cin. L. Rev. 347, 355–58 (2017).

  50. [50]. See Gardner, supra note 31, at 363.

  51. [51]. I do not mean to imply that the duties we have to others in virtue of our relationships and roles should always be conceived of as net burdens or “bad.” While friendship or family relations come with (or consist of) substantial duties, many would say these relationships and duties make life better. In contract, we often accept duties because, all things considered, the arrangement is to our benefit. My point is simply that there are important questions about what duties we can legitimately impose that need to be addressed.

  52. [52]. Or at minimum concerns constitutional values. See supra note 36 and accompanying text.

  53. [53]. See Bartlett & Gulati, supra note 1, at 238.

  54. [54]. See id. at 243.

  55. [55]. See id. at 238–39. Boy Scouts of America v. Dale, 530 U.S. 640 (2000) is arguably the clearest example of constitutional rights limiting the reach of antidiscrimination law, but the authors do not mention it. Instead, in addition to Rotary Club, Bartlett and Gulati look at the application of antidiscrimination law to the sale and rental of housing and suggest that Mrs. Murphy’s exception is constitutionally required. See Bartlett & Gulati, supra note 1, at 239. While the literature appears sparse, at least one author has argued against this view. See James D. Walsh, Note, Reaching Mrs. Murphy: A Call for Repeal of the Mrs. Murphy Exemption to the Fair Housing Act, 34 Harv. C.R.-C.L. L. Rev. 605 (1999). See also Elizabeth F. Emens, Intimate Discrimination: The State’s Role in the Accidents of Sex and Love, 122 Harv. L. Rev. 1307, 1379 n.332 (“[W]e might read an explicit carve-out in federal housing law—the so-called Mrs. Murphy exception for owner-occupied dwellings of no more than five families—as a concession to fears of miscegenation, though it is usually framed in terms of privacy and freedom of association.”).

  56. [56]. In related work, I examine the consumer privacy rights judges use to defend BFOQ exceptions to Title VII antidiscrimination duties. I argue that the bases for invoking these privacy interests are shakier than those who do so acknowledge. See Whitney, supra note 31, manuscript at 17–29. See also Amy Kapczynski, Same-Sex Privacy and the Limits of Antidiscrimination Law, 112 Yale L.J. 1257, 1273–74 (2003) (“Same-sex privacy doctrine cannot be defended with recourse to the law of privacy. In many cases, the third parties in question have no relevant privacy rights. Even where they do have such rights, it is not self-evident that there is or ought to be a link between those privacy rights and sex.”).

  57. [57]. Bartlett & Gulati, supra note 1, at 243. We see this view today, in cases like Hobby Lobby and the erosion of the commercial speech doctrine. See, e.g., Matal v. Tam, 137 S. Ct. 1744, 1767 (2017) (Kennedy, J. concurring) (“‘Commercial speech is no exception,’ the Court has explained, to the principle that the First Amendment ‘requires heightened scrutiny whenever the government creates a regulation of speech because of disagreement with the message it conveys.’”) (citing Sorrell v. IMS Health Inc., 564 U.S. 552, 566 (2011)); id. at 1769 (Thomas, J., concurring) (“I continue to believe that when the government seeks to restrict truthful speech in order to suppress the ideas it conveys, strict scrutiny is appropriate, whether or not the speech in question may be characterized as ‘commercial.’”) (citations omitted). We arguably also see something like this view in Craig. v. Masterpiece Cakeshop, Inc., 370 P.3d 272 (Colo. Ct. App. 2015), cert. granted sub nom. Masterpiece Cakeshop, Ltd. v. Colorado Civil Rights Commission, 85 U.S.L.W. 3593 (U.S. June 26, 2017) (No. 16-111). The question presented is whether Colorado’s public accommodation law, which requires the petitioner to provide a wedding cake for a same-sex couple, unconstitutionally violates his free speech or free exercise rights under the First Amendment.

  58. [58]. Bartlett & Gulati, supra note 1, at 239 (emphasis added).

  59. [59]. Id. at 243.

  60. [60]. Id.

  61. [61]. As I discussed above, see supra note 41 and accompanying text, this will depend on the contours, scope, and relative strength of the right, entitlements, and duties we place on different actors in virtue of their different roles and relations.

  62. [62]. Lawrence v. Texas, 539 U.S. 558 (2003).

  63. [63]. Lawrence, 539 U.S. at 577–78 (citing Bowers v. Hardwick, 478 U.S. 186, 216 (1986) (Stevens, J., dissenting)).

  64. [64]. See Whitney, supra note 28, (manuscript at 9–13).

  65. [65]. The complication here is access to contraception, which is often purchased and/or mediated by a medical relationship that occurs within the market. See generally Griswold v. Connecticut, 381 U.S. 479 (1965) (striking down as an unconstitutional violation of marital privacy rights a statute that prohibited both the sale and use of contraceptives).

  66. [66]. Lawrence, 539 U.S. at 578. It is noteworthy that the Court mentioned both the absence of public conduct and prostitution. As Bartlett and Gulati point out, the public/private line is contested and in flux. See Bartlett & Gulati, supra note 1, at 239–41. Prostitution could be conceived as falling on either side of that line. The Court does not take a position in that debate: regardless of where prostitution falls, the Court thought it significant that sex for money (sex in the market) was not at issue when it articulated the relevant liberty interest.

  67. [67]. See Bartlett & Gulati, supra note 1, at 243.

  68. [68]. Some litigants have argued (unsuccessfully) precisely this. See, e.g., State v. Freitag, 130 P.3d 544, 546 (Ariz. Ct. App. 2006) (“We thus join other state courts that have specifically rejected any constitutionally protected fundamental liberty or privacy interest in soliciting or engaging in prostitution.”); People v. Williams, 811 N.E.2d 1197 (Ill. App. Ct. 2004). The Lawrence dissent also took this position. See Lawrence, 539 U.S. at 590 (Scalia, J., dissenting) (“State laws against ... prostitution ... are likewise sustainable only in light of Bowers’ validation of laws based on moral choices. Every single one of these laws is called into question by today’s decision.”). The dissent takes this position by assuming what Justice Stevens first argued in Bowers—that anti-sodomy laws do not prevent harm, but instead only prevent what some find immoral—and extending it to laws against prostitution. But that understanding of the basis for anti-prostitution laws—that they merely prohibit immoral but not harmful conduct—is highly controversial to say the least.

  69. [69]. There is a hidden further issue here, and how one comes down on it has implications for how Lawrence speaks to prohibitions on customer discrimination. Accepting arguendo that, post-Lawrence, prostitution cannot be prohibited, there is the separate question of whether prostitution can be regulated at all. There are some who see sex work like other forms of dangerous or exploitable labor and think it can and should be regulated as such; this is what Adrienne Davis calls the “erotic assimilationist” view. Adrienne D. Davis, Regulating Sex Work: Erotic Assimilationism, Erotic Exceptionalism, and the Challenge of Intimate Labor, 103 Calif. L. Rev. 1195, 1201 (2015). If one is an erotic assimilationist and finds Lawrence to entail the decriminalization of prostitution, then one believes that the government is freer to regulate the exercise of the constitutionally protected liberty interest when exercised inside the market than when exercised outside it. Someone with this view might think, for instance, that while the government cannot regulate how many hours a person has sex in their private lives, the government can regulate how many hours and at what price a person can have sex for money. In contrast, one might be an erotic exceptionalist, and think that sex work, because it involves sex, is an exceptional form of labor and thus no government regulation is permissible. See id. at 1220–21. Bartlett and Gulati seem committed to an exceptionalist position when they say “[i]t seems appropriate to grant” the same freedoms one has outside the market “in market transactions” to explain why racial filters on dating websites must be permitted yet seem assimilationist when they argue for indirect regulations of customer discrimination through the placement of additional antidiscrimination duties on firms in other contexts. Bartlett & Gulati, supra note 1, at 243.

  70. [70]. See United States v. Alvarez, 567 U.S. 709, 723 (2012) (“Where false claims are made to effect a fraud or secure moneys or other valuable considerations, say offers of employment, it is well established that the Government may restrict speech without affronting the First Amendment.”).

  71. [71]. The Court’s views on the extent that commerciality weighs against a finding of fair use has fluctuated over time. See Sony Corp. of Am. v. Universal City Studios, 464 U.S. 417, 449 (1984) (stating that “a commercial or profit-making purpose” was “presumptively ... unfair”). The Court walked this back somewhat ten years later. See, e.g., Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 579 (1994) (“[T]he more transformative the new work, the less will be the significance of other factors, like commercialism, that may weigh against a finding of fair use.”); Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 539, 562 (1985) (when copying is commercial that “tends to weigh against a finding of fair use”) (emphasis added). Even today, though, commerciality is used to argue strongly against fair use determinations. See Opening Brief and Addendum for Oracle America, Inc. at 27–28, Oracle America, Inc., v. Google, Inc., No.
    17-11118 (Fed. Cir. Feb. 10, 2017).

  72. [72]. See generally Volokh, supra note 39 (discussing the limits on free speech in the workplace).

  73. [73]. Though perhaps this position is no longer so radical. The Court has had an increasingly ambivalent relationship to the market/non-market distinction, at least when it comes to the First Amendment. See supra text accompanying note 56. For more on the relevance of the market/non-market distinction and the Court’s changing attitude towards it, see generally Whitney, supra note 31.

  74. [74]. Within the more libertarian tradition this position has more recently been adopted by Jason Brennan and Peter Martin Jaworski. See generally Jason Brennan & Peter M. Jaworksi, Markets Without Limits 10 (2016) (“Our view on the scope of the market can be summarized as follows: Markets without Limits: If you may do it for free, then you may do it for money.”); Jason Brennan & Peter Martin Jaworski, Markets Without Symbolic Limits, 125 Ethics 1053 (2015) (arguing against semiotically justified limits on what can be bought and sold). As discussed above, within the context of sex work, this is the position of erotic exceptionalists. See supra text accompanying note 68. Numerous philosophers have argued against this position. See, e.g., Debra Satz, Why Some Things Should Not Be For Sale: The Moral Limits of Markets (2012), Margaret Jane Radin, Contested Commodities132–36 (2001), Elizabeth Anderson, Value in Ethics and Economics150–58 (1995) (discussing prostitution).

  75. [75]. See generally Samuel R. Bagenstos, The Unrelenting Libertarian Challenge to Public Accommodation Law, 66 Stan. L. Rev. 1205 (2014) (discussing this phenomenon).

  76. [76]. See supra text accompanying note 40 for a sketch of how this might work and the problems it poses for Bartlett and Gulati.


Ph.D. Candidate in Philosophy, New York University. Visiting Researcher, Harvard Law School (Spring 2017); Bigelow Fellow and Lecturer in Law, University of Chicago Law School (2014–2016).

For more helpful feedback than anyone could reasonably ask for, my thanks and gratitude to Robert Simpson. Additional thanks to Will Baude, Ryan Doerfler, Tienmu Ma, Richard McAdams, Alex Prescott-Couch, and the editors of the Iowa Law Review.