Indiana Law Review Volume 57 2023 Number 1 ARTICLES FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE: THE SLOW DEMISE OF MIXED MOTIVES ANALYSIS SHERI LYNN JOHNSON* Justice Kavanaugh begins and ends the majority opinion in Flowers v. Mississippi1 with the same assertion: that the Court “break[s] no new legal ground.”2 Thirty-five years earlier, the Supreme Court had held in Batson v. Kentucky3 that a prosecutor’s racially motivated exercise of the peremptory challenge violated the Equal Protection Clause of the Fourteenth Amendment, and plainly that foundational holding had not changed. With respect to the question of what kind of facts count toward the establishment of discriminatory motivation in the exercise of the peremptory challenge, Justice Kavanaugh’s modesty is a little exaggerated, because there are some new tidbits in Flowers, but it is not false. Moreover, his claim is certainly accurate with respect to the question of what combination of facts, taken together, suffice to establish discriminatory motivation because there has not been—and never will be—another case like Flowers.4 But, with respect to the next question, Flowers very quietly signals a big change, one adopting an approach that prior to Flowers only one federal circuit—the Ninth—had employed.5 What should happen when the probative facts, taken together, establish a prosecutor’s racial motivation? Should a court, counterfactually, determine whether that prosecutor would have done the same thing absent racial motivation, or should the court simply reverse the conviction? Prior to Flowers, eight federal Courts of Appeal had held that Footnote 21 of Arlington Heights v. Metropolitan Housing Development Corp.6 commands the * James and Mark Flanagan Professor of Law, Cornell Law School. This Article is for Curtis Flowers, whom I was honored to represent for almost a decade. I am deeply grateful to my co-counsel and colleague, Keir Weyble, whose work with me on Flowers’s case was impeccable, indispensable, indefatigable—and unreasonably cheerful. I would also like to thank Prachee Sawant, Skyler Green, and Christine Brittain for their very able research assistance. 1. 139 S. Ct. 2228, 2235 (2019). 2. Id. at 2251. 3. 476 U.S. 79 (1986). 4. Flowers, 139 S. Ct. at 2251 (Alito, J., concurring) (“[T]his is a highly unusual case. Indeed, it is likely one of a kind.”). 5. See infra Part II.B. 6. 429 U.S. 252 (1977). 8 INDIANA LAW REVIEW [Vol. 57:7 counterfactual inquiry,7 putting up yet another barrier to the redress of covert discrimination, but the concluding language of Flowers, decided more than four decades after Arlington Heights, strongly implies the rejection of that footnote, at least for Batson claims: All that we need to decide . . . is that all of the relevant facts and circumstances taken together establish that the trial court at Flowers’ sixth trial committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not motivated in substantial part by discriminatory intent.8 Flowers’ implications for “mixed motive” analysis has not thus far been discussed in the secondary literature,9 but lower courts have begun to take notice. Since Flowers, no lower court has used “mixed motives” or “dual motivation” in its own analysis of the merits of a Batson claim, though the Fifth Circuit has, in the habeas context, employed a rule in reviewing state court decisions that is even stricter than dual motivation.10 In Colorado, the lower appellate courts are debating whether substantial motivation is the right test, a question the Colorado Supreme Court has ducked once,11 but will soon face again.12 Moreover, a petition 7. Washington v. Roberts, 846 F.3d 1283 (10th Cir. 2017); Pecor v. Walls, 56 F. App’x 723, 726 (7th Cir. 2003); Gattis v. Snyder, 278 F.3d 222, 234-35 (3d Cir. 2002); United States v. Peraza, 25 F.3d 1051 (6th Cir. 1994); United States v. Darden, 70 F.3d 1507, 1531 (8th Cir. 1995); United States v. Tokars, 95 F.3d 1520 (11th Cir. 1996); Jones v. Plaster, 57 F.3d 417, 421-22 (4th Cir. 1995); Howard v. Senkowski, 986 F.2d 24 (2d Cir. 1993). 8. Flowers v. Mississippi, 139 S. Ct. 2228, 2251 (emphasis added). 9. Although Flowers has garnered substantial commentary, see e.g., Paul Butler, Mississippi Goddamn: Flowers v. Mississippi’s Cheap Racial Justice, 2019 SUP. CT. REV. 73; Thomas Ward Frampton, What Justice Thomas Gets Rights About Batson, 72 STAN. L. REV. Online 1, 6 (2019); Dorothy E. Roberts, Supreme Court, 2018 Term—Foreword: Abolition Constitutionalism, 133 HARV. L. REV. 1, 79 (2019), none has addressed the question of dual motivation. More than a decade prior to Flowers, Professor Russell Covey made a narrower, but related argument that in the rare “smoking gun” case, “where the prosecutor candidly, or stupidly, confesses a discriminatory impulse,” permitting the state to defend the strike on other grounds “is a direct affront to basic equal protection values.” The Unbearable Lightness of Batson: Mixed Motives and Jury Selection, 66 MD. L. REV. 279, 325 (2008). I make a broader claim: Whenever a defendant establishes that a prosecutor’s peremptory strike was substantially motivated by race, whether by the prosecutor’s own admission or, as happens much more often today, through “sensitive inquiry” into all the relevant facts, the strike may not be redeemed by the citation of additional permissible reasons. More broadly, this Article further argues that other cases in the criminal justice arena should be exempted from the reach of the Arlington Heights footnote, a question that no scholar has thus far addressed. 10. See Sheppard v. Davis, 967 F.3d 458, 472 (5th Cir. 2020) (“[A] Batson claim will not succeed where the defendant fails to rebut each of the prosecutor’s legitimate reasons.”). 11. People v. Ojeda, 487 P.3d 1117,1123 (Colo. App. 2019) (adopting the “substantial motivation test”), aff’d on other grounds, 503 P.3d 856 (Colo. 2022). 12. People v. Johnson, 523 P.3d 992, (Colo. App. 2022) (adopting per se approach) cert. granted in part, People v. Johnson, 2023 WL 3587455 (Colo. 2023). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 9 for certiorari asking the Supreme Court to resolve the conflicting approaches of the lower courts was filed and denied last June;13 there will be others.14 And just beyond the horizon of jury selection claims lie other species of race discrimination claims which rejection of the Arlington Heights footnote may implicate. This Article traces the development of the dual motivation defense, argues that Flowers both does and is right to enter that defense, and then takes on the implications of its interment for other race claims in the criminal justice arena. Part I first describes the general development of purposeful discrimination doctrine in the Supreme Court up to and including Arlington Heights; it then briefly digresses to note the treatment of dual motivation in other areas of the law in order to establish that jettisoning the counterfactual inquiry that the Arlington Heights footnote demands would not be idiosyncratic. Part II traces the history of the intersection of that footnote with Batson v. Kentucky; that history is marked by a long and curious Supreme Court reticence despite widespread lower court confusion, and it culminates in the very quiet resolution in Flowers that establishment of “substantial” racial motivation in the exercise of the peremptory challenge compels reversal of the conviction. Part III then finds support for the correctness of that resolution in social and cognitive psychology literature not available when Arlington Heights was decided, literature that reveals the futility of the footnote’s inquiry, at least in the criminal justice arena. Finally, Part IV considers examples where the decision whether to apply the footnote matters, beginning with Flowers itself, hoping that the reader who likes stories better than psychology will agree that they all are best resolved by ending the inquiry once it has been determined that racial motivation substantially influenced the decision. Lest the reader have any doubt: I come to bury Caesar, not to praise him. I. “PURPOSEFUL DISCRIMINATION” AND BUT-FOR CAUSATION A. The Road to Arlington Heights A complete treatment of the history of equal protection doctrine, while certainly a worthy endeavor,15 is more than is needed here; instead, I summarize the early cases only to the extent necessary to situate purposeful discrimination doctrine in general, and the Arlington Heights footnote in particular. In the earliest equal protection cases, there was no question of racial motivation, but only the question of whether the discrimination that flowed from that motivation could be justified. Strauder v. West Virginia,16 decided by the 13. McDaniel v. California, 142 S. Ct. 2877 (2022). 14. For any given case, a grant of certiorari is unlikely, and as California’s Brief in Opposition argued, McDaniel was not a particularly good vehicle for resolving the lower court split but given the very large number of Batson claims litigated, the Supreme Court is likely to soon be presented with a straightforward presentation of the split. 15. For one comprehensive treatment, see Michael J. Klarman, An Interpretive History of Modern Equal Protection, 90 MICH. L. REV. 213 (1991). 16. 100 U.S. 303 (1879). 10 INDIANA LAW REVIEW [Vol. 57:7 Supreme Court in 1879, held that the West Virginia statute excluding Black people from jury service violated the Black defendant’s right to equal protection.17 Strauder reasoned that the law does not protect equally if “every white man is entitled to a trial by a jury selected from persons of his own race or color, . . . and a negro is not . . . ,”18 a rationale that the Court later the same year applied to race-based exclusion from grand jury service.19 Because racial discrimination was explicit on the face of the statutes, no argument about motivation was possible. Yick Wo v. Hopkins,20 decided six years later, involved an ordinance neutral on its face that was applied only to laundries owned by Chinese or Chinese Americans; the stark enforcement statistics made the inference of racial motivation so plain that it was not disputed. Or, to take the most infamous example, Louisiana did not dispute that it was Homer Plessy’s race that relegated him to a less desirable railroad car, but that social preferences justified the mandated racial segregation.21 The Court soon extended Strauder to the racially discriminatory administration of facially neutral jury discrimination laws, and in its effort to ferret out discriminatory venire selection, eventually held that a prima facie case of racial discrimination is established when the claimant shows statistically significant disparity between grand jury or venire composition and population proportions coupled with an opportunity to discriminate.22 However, it would be more than a century after Strauder before the Court recognized that discriminatory exercise of the peremptory challenge threatened the same constitutional harm that discrimination in the selection of the venire does.23 Moreover, even today, the law governing proof of discrimination in venire selection is more favorable than it is with respect to proof of discrimination in the exercise of peremptory challenges, or indeed, in any other context.24 Outside of the venire selection context, the question of what evidence sufficed to show racially discriminatory intent was largely neglected until after the Civil Rights Movement. Indeed, not until 1976, the year Arlington Heights was decided, was it clear that discriminatory intent— rather than discriminatory impact—was the trigger for strict scrutiny. How could equal protection doctrine lumber along for a century without resolving that crucial question? The short 17. Id. at 312. 18. Id. at 309. 19. Ex Parte Virginia, 100 U.S. 339, 348 (1879). 20. 118 U.S. 356 (1886). 21. Plessy v. Ferguson, 163 U.S. 537 (1896). 22. Alexander v. Louisiana, 405 U.S. 625 (1972). 23. Batson v. Kentucky, 476 U.S. 79 (1986). 24. See McCleskey v. Kemp, 481 U.S. 279, 293-94 (1987) (internal citations omitted) (“Although statistical proof normally must present a “stark” pattern to be accepted as the sole proof of discriminatory intent under the Constitution . . . ‘[b]ecause of the nature of the jury [venire]—selection task, . . . we have permitted a finding of constitutional violation even when the statistical pattern does not approach [such] extremes.’”). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 11 answer is that no Supreme Court case prior to McLaughlin v. Florida,25 decided in 1964, both articulated a rigorous standard for the review of racial classifications and actually subjected the regulation at issue to such review.26 However, once it became apparent that if a claimant established racial discrimination, the Supreme Court was likely to strike it down, discriminators had reason to conceal their motives. At about the same time that would-be discriminators risked negative legal ramifications if they admitted their motives, they also began to face a greater likelihood of social disapproval.27 Palmer v. Thompson,28 however, muddied the waters for both state officials of a mind to discriminate and claimants of a mind to assert discrimination. The city of Jackson, Mississippi had operated segregated public facilities and when faced with a federal court decision requiring their integration, desegregated all of them except the swimming pools, which the city closed. The Supreme Court responded that while the purpose behind closure may have been racial animosity, the effect on both racial groups was the same—neither could swim in public swimming pools—and consequently, the closure did not violate equal protection.29 After Palmer, some lower courts quite reasonably assumed that if the “discrimination” forbidden by the Equal Protection Clause is not established by proof of racial motive, it must be that the essence of an equal protection claim is the racially discriminatory effect.30 That is, disparate effect on a racial group constitutes the “racial discrimination” that triggers strict scrutiny. The first case to reach the Supreme Court in which the lower court had 25. 379 U.S. 184 (1964). 26. Although the opinions in both Korematsu v. U.S., 323 U.S. 214 (1944) and Hirabayashi v. U.S., 828 F.2d 591 (1987) asserted that racial classifications were presumptively invalid and subject to the most rigid scrutiny, in neither did the Court actually employ such a presumption or subject the regulation to any sort of heightened review. Conversely, although the Court clearly applied more than rational relationship scrutiny to segregated schools in Brown v. Board of Education, 347 U.S. 483 (1954), it abjured “rigid” or “strict” scrutiny language. See Klarman, supra note 15 (summarizing all of the cases). 27. Harold Sigall & Richard Page, Current Stereotypes: A Little Fading, A Little Faking, 18 J. PERSONALITY. & SOC. PSYCH. 247 (1971) (finding that social desirability affects agreement with racial stereotypes). 28. 403 U.S. 217 (1971). 29. Id. at 225. As others have observed, this analysis also neglects the fact that the effect on African Americans and whites really was not the same. Most importantly, the sequence of events made plain that the government deemed no swimming better than integrated swimming, and that whites had the option of private pools, which were largely unavailable to African Americans at the time. 30. Washington v. Davis, 426 U.S. 229, 244 (1976) (“Both before and after Palmer v. Thompson, however, various Courts of Appeals have held in several contexts, including public employment, that the substantially disproportionate racial impact of a statute or official practice standing alone and without regard to discriminatory purpose, suffices to prove racial discrimination violating the Equal Protection Clause absent some justification going substantially beyond what would be necessary to validate most other legislative classifications.”). 12 INDIANA LAW REVIEW [Vol. 57:7 adopted a focus on disparate effects was Washington v. Davis.31 Davis involved a challenge to a screening instrument for recruiting police officers.32 The plaintiffs did not allege discriminatory purpose but relied upon the discriminatory impact on African-American applicants and the lack of an established relationship between the test and job performance.33 The district court denied relief, but the Court of Appeals, borrowing the statutory standards applicable to claim under Title VII of the Civil Rights Act of 1964, held that the lack of discriminatory intent in designing and administering the test was irrelevant; the critical fact was that a greater proportion of African Americans failed the test than did whites.34 This disproportionate impact, the Court of Appeals held, was sufficient to establish a constitutional violation absent proof that the test predicted job performance, a burden which the court ruled that the police department had failed to discharge.35 The Supreme Court reversed, finding the standards of Title VII inapplicable to equal protection claims.36 Citing the distinction between de jure and de facto discrimination in school segregation cases,37 it held that “the invidious quality of a law claimed to be racially discriminatory must ultimately be traced to a racially discriminatory purpose.”38 The Court did acknowledge, referencing Palmer v. Thompson, that there were “some indications to the contrary” in its previous decisions, but asserted that “[t]o the extent that Palmer suggests a generally applicable proposition that legislative purpose is irrelevant in constitutional adjudication, our prior cases as indicated in the text are to the contrary.”39 Because the litigants had provided no evidence—or even an allegation—of discriminatory purpose, that was the end of their equal protection claim. B. Arlington Heights Village of Arlington Heights v. Metropolitan Housing Development Corp.,40 already on its way to the Supreme Court before Davis was decided, was not quite so simple.41 In 1971, in order to build 190 clustered townhouse units for low- and moderate-income tenants, the Metropolitan Housing Development Corporation 31. Id. 32. Id. at 229. 33. Id. 34. Id. at 236-37. 35. Id. at 235-38. 36. Id. at 252. 37. Id. at 240 (1976) (“The differentiating factor between De jure segregation and so-called De facto segregation . . . is Purpose or Intent to segregate.”) (quoting Keyes v. School Dist. No. 1, 413 U.S. 189, 205 (1973)). 38. Id. 39. Id. at 242, 244 n.11. 40. Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252 (1977). 41. As the opinion in Davis noted, certiorari had been granted in Arlington Heights when Davis was decided. Washington v. Davis, 426 U.S. 229, 244 n.12 (1976). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 13 (MHDC) applied to the Village of Arlington Heights, a Chicago suburb, for the rezoning of a parcel from single-family to multiple-family classification.42 When the Village denied the rezoning request, MHDC brought suit in federal court alleging that the denial was racially discriminatory and that it violated both the Equal Protection Clause of the Fourteenth Amendment and the Fair Housing Act of 1968.43 In Arlington Heights, unlike Davis, there was some evidence of discriminatory motive, though not much. The district court entered judgment for the Village, but the Seventh Circuit reversed, finding that the “ultimate effect” of the denial was racially discriminatory and that the refusal to rezone therefore violated the Fourteenth Amendment.44 One might have expected the Supreme Court to remand Arlington Heights to the circuit court for reconsideration in light of Davis,45 instead it granted full merits review and used Arlington Heights to provide guidance to lower courts on how to approach a claim of purposeful discrimination.46 The majority first instructed that “[d]etermining whether invidious discriminatory purpose was a motivating factor demands a sensitive inquiry into such circumstantial and direct evidence of intent as may be available.”47 It acknowledged that the impact of the official action provides an important starting point, and citing Yick Wo v. Hopkins48 and Gomillion v. Lightfoot,49 further noted that “[s]ometimes a clear pattern, unexplainable on grounds other than race, emerges from the effect of the state action even when the governing legislation appears neutral on its face,” but then deemed such cases both “relatively easy” and “rare.”50 Absent a pattern as stark as Gomillion’s or Yick Wo’s, “impact alone is not determinative,” and other evidence must be consulted.51 Then the Court provided a list of relevant indicators of discriminatory intent: The historical background of the decision is one evidentiary source, particularly if it reveals a series of official actions taken for invidious purposes. The specific sequence of events leading up to the challenged decision also may shed some light on the decisionmaker’s purposes. . . . . 42. Arlington Heights, 429 U.S. at 254-55. 43. Id. 44. Id. at 254. 45. Justices Marshall and Brennan, dissenting in part, objected to this departure from ordinary practice, asserting that a GVR for reconsideration in light of Davis was necessary. Id. at 271 (Marshall, J., concurring in part and dissenting in part). Justice White dissented, agreeing with Marshall and Brennan that a GVR was required, but further objecting to the majority’s elaboration on the criteria for determining the existence of discriminatory purpose. Id. at 272 (White, J., dissenting). 46. Id. at 255, 265-68 (majority opinion). 47. Id. at 266. 48. 118 U.S. 356 (1886). 49. 364 U.S. 399 (1960). 50. Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 266 (1977). 51. Id. 14 INDIANA LAW REVIEW [Vol. 57:7 Departures from the normal procedural sequence also might afford evidence that improper purposes are playing a role. Substantive departures too may be relevant, particularly if the factors usually considered important by the decisionmaker strongly favor a decision contrary to the one reached. The legislative or administrative history may be highly relevant, especially where there are contemporary statements by members of the decisionmaking body, minutes of its meetings, or reports. . . . In some extraordinary instances the members might be called to the stand at trial to testify concerning the purpose of the official action, although even then such testimony frequently will be barred by privilege.52 After noting that the foregoing list of indicia is not exhaustive, the majority opinion turns to an analysis of the evidence of discriminatory intent presented by the plaintiffs and concludes that they “simply failed to carry their burden of proving that discriminatory purpose was a motivating factor in the Village’s decision.”53 Then, ending that sentence is the footnote that contains the caveat this Article hopes to help inter: Proof that the decision by the Village was motivated in part by a racially discriminatory purpose would not necessarily have required invalidation of the challenged decision. Such proof would, however, have shifted to the Village the burden of establishing that the same decision would have resulted even had the impermissible purpose not been considered.54 C. The Rationale Behind Footnote 21 The footnote ends with a very short justification of this additional impediment to relief, and a single supporting citation. According to the majority, “[i]f this were established, the complaining party in a case of this kind no longer fairly could attribute the injury complained of to improper consideration of a discriminatory purpose.”55 Absent such a “fair” attribution of the injury, “there would be no justification for judicial interference with the challenged decision.”56 One might wonder where the imposition of this opportunity to rescue a constitutionally impermissible decision—never previously mentioned in any Supreme Court racial discrimination case—originated, or if new, why it should now be imposed. One might wonder how it was consistent with Yick Wo, which did not offer the defendant city the opportunity to demonstrate that it would have enacted a prohibition against wood laundries absent its intent to harm persons of Chinese ancestry. The footnote does not answer these questions, except by citing Mt. Healthy City School District Board of Education v. Doyle.57 52. Id. at 267-68 (internal citations omitted). 53. Id. at 270. 54. Id. at 270 n.21. 55. Id. 56. Id. 57. 429 U.S. 274 (1977). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 15 Mt. Healthy does not, however, shed much light on these questions. It is not a race case, nor even an equal protection case, and the relevant considerations seem quite different than those at stake in Arlington Heights. After an untenured teacher called a radio station to report a memorandum circulated by the school principal, the School Board advised the teacher that he would not be rehired, citing both his call to the radio station and other misconduct.58 The teacher sought damages and reinstatement for the violation of his First Amendment rights, both of which the district court granted after finding that the telephone call played a “substantial part” in the decision not to rehire, a decision affirmed by the Court of Appeals.59 The Supreme Court reversed, holding that once the district court had determined that the constitutionally protected conduct was a “substantial factor” in the Board’s decision, it “should have gone on to determine whether the Board had shown by a preponderance of the evidence that it would have reached the same decision as to [the teacher]’s reemployment even in the absence of the protected conduct.”60 Why? The Court explained that a “rule of causation which focuses solely on whether protected conduct played a part, ‘substantial’ or otherwise, in a decision not to rehire, could place an employee in a better position as a result of the exercise of constitutionally protected conduct than he would have occupied had he done nothing.”61 The Court observed that the district court’s approach “would require reinstatement in cases where a dramatic and perhaps abrasive incident is inevitably on the minds of those responsible for the decision to rehire, and does indeed play a part in that decision even if the same decision would have been reached had the incident not occurred.”62 Significantly, the Court found the need for a but-for causation requirement particularly persuasive in the tenure context where the long term consequences “are too significant for us to hold that the Board in this case would be precluded, because it considered constitutionally protected conduct in deciding not to rehire [the teacher], from attempting to prove to a trier of fact that quite apart from such conduct [his] record was such that he would not have been rehired in any event.”63 The balance of interests in vindicating rights against racial discrimination and upholding governmental decisions would seem to be quite different, or at least quite variable. One could imagine a tenure case in which racial discrimination rather than First Amendment rights were at stake, in which the Court’s Mt. Healthy rationale might seem applicable, but one could also contemplate a case like Yick Wo, where the city’s countervailing interest would seem trivial. Mt. Healthy notes another area of constitutional law as precedent for a but-for causation requirement: the attenuation of taint doctrine applicable in fruit of the poisonous tree confession cases. The balance of interests in such cases seems 58. Id. at 281-83. 59. Id. at 284, 287. 60. Id. at 287. 61. Id. at 285. 62. Id. 63. Id. 16 INDIANA LAW REVIEW [Vol. 57:7 even further removed from that at stake in racial discrimination claims, which may explain why the Arlington Heights footnote does not discuss them. However, Mt. Healthy itself had addressed whether the balance of interests are similar, reasoning that “[w]hile the type of causation on which the taint cases turn may differ somewhat from that which we apply here, those cases do suggest that the proper test to apply in the present context is one which likewise protects against the invasion of constitutional rights without commanding undesirable consequences not necessary to the assurance of those rights.”64 As I will flesh out in Parts III and IV, I think neither side of the balance is the same with respect to the cases controlled by the Arlington Heights footnote, most clearly so with respect to Batson cases; the addition of a but-for causation test dramatically diminishes the protection against the invasion of the constitutional right at stake, and the undesirable consequences avoided by that requirement are much more speculative. D. Applications of the Footnote Setting aside the application of but-for causation in Batson cases until Part III, the footnote almost disappears. In forty-five years, it has been cited in a Supreme Court majority or plurality opinion only seven times.65 None of those citations contribute much to an assessment of whether the footnote is correct or even to an understanding of how it should be applied. Only two cases involve race and neither struggles with but-for causation; Columbus v. Penick held that the city failed to prove that the racial separation in the schools would have occurred even without their concededly unlawful conduct,66 and Hunter v. Underwood held that the Alabama state constitutional provision disenfranchising persons convicted of crimes involving moral turpitude was motivated by the desire to disenfranchise African Americans, a motivation that was “beyond 64. Id. at 287. 65. Two dissents also cite the footnote, but for reasons not relevant to this discussion. The majority in School District of Omaha v. United States, 433 U.S. 667 (1977) GVR’ed the case for reconsideration in light of Arlington Heights; Justice Brennan, joined by Justice Marshall, dissented, believing the remand unnecessary because the lower court had applied the standard adopted in Arlington Heights and in so asserting, cited the footnote. Justice Marshall, dissenting in Mobile v. Bolden, contrasted the standard he would apply in Fifteenth Amendment cases with the standard of the Mt. Healthy and Arlington Heights footnote: Reallocation of the burden of proof is especially appropriate in these cases, where the challenged state action infringes the exercise of a fundamental right. The defendants would carry their burden of proof only if they showed that they considered submergence of the Negro vote a detriment, not a benefit, of the multimember systems, that they accorded minority citizens the same respect given to whites, and that they nevertheless decided to maintain the systems for legitimate reasons. City of Mobile v. Bolden, 446 U.S. 55, 137-38 (1980) (Marshall, J., dissenting). (citing Mt. Healthy and the Arlington Heights footnote). 66. Columbus Bd. of Educ. v. Penick, 443 U.S. 449, 466 (1979). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 17 peradventure . . . a ‘but-for’ motivation for the enactment.”67 Two of the other cases, like Mt. Healthy, involve First Amendment claims, and consequently the Arlington Heights footnote, though cited along with Mt. Healthy, adds nothing to their rationales.68 Another, Price Waterhouse,69 resembles Mt. Healthy in a different way; it involves an employment claim. Price Waterhouse decided that even when discriminatory motivation is established, an employer is not liable under Title VII if it can prove that had it not taken gender into account, it would have come to the same decision; the plurality did not label this a but-for causation requirement, but rather, an affirmative defense.70 As I will turn to shortly, the decision in Price Waterhouse was overturned by congressional enactment,71 and so its details are not in and of themselves important. However, the opinions flag another interesting question: What kind of evidence will satisfy the defendant’s burden? The plurality comments that “[a]s to the employer’s proof, in most cases, the employer should be able to present some objective evidence as to its probable decision in the absence of an impermissible motive.”72 Justice White, concurring, disagreed with this statement,73 asserting that there is no special burden to produce objective evidence of motivation, to which the plurality responds in a footnote that it finds “baffling” his “suggestion . . . that the employer’s own testimony as to the probable decision in the absence of discrimination is due special credence where the court has, contrary to the employer’s testimony, found that an illegitimate factor played a part in the decision.”74 Carey v. Piphus75 was factually not much like either Mt. Healthy or Arlington Heights but involved due process challenges to school discipline. After pointing out that the purpose of the Civil Rights statute was the compensation of harm flowing from violations of civil rights and citing both Mt. Healthy and the Arlington Heights footnote as analogous, the Court held that damages attributable to the discipline itself could only be awarded if failure to accord proper procedures was the cause of the discipline.76 But there was an addendum: If the deprivation of due process caused psychological harm distinct from the harm caused by the discipline, that injury was compensable even if constitutional 67. 471 U.S. 222, 232 (1985). 68. Hartman v. Moore, 547 U.S. 250 (2006) holds that absence of probable cause is a necessary element of a retaliatory selective prosecution claim because it will have high probative value in determining but-for causation and can be made mandatory with little added cost; Nieves v. Bartlett, 139 S. Ct. 1715 (2019), holds that for analogous reasons, lack of probable cause is also an element of a retaliatory arrest claim. 69. Price Waterhouse v. Hopkins, 490 U.S. 228 (1989). 70. Id. 71. Civil Rights Act of 1991, Pub. L. No. 102-166. 72. Price Waterhouse, 490 U.S. at 252. 73. Id. at 260 (White, J., concurring). 74. Id. at 252 n.14. 75. 435 U.S. 247 (1978). 76. Id. at 254-60, 266-67. 18 INDIANA LAW REVIEW [Vol. 57:7 process would have resulted in the student being found guilty of an offense and disciplined for it.77 There is only one more Supreme Court citation, this one seeming to take an even more lenient view of the state’s burden than does the footnote. Michael M. v. Superior Court involved a California law that punished only males for statutory rape, which the state defended as furthering its interest in preventing teen pregnancy.7 8 To Michael M.’s contention that the true purpose of the statute was to protect the virtue and chastity of young women—a purpose evidenced by the statute’s history—the Supreme Court responded “[w]e are satisfied not only that the prevention of illegitimate pregnancy is at least one of the ‘purposes’ of the statute, but also that the State has a strong interest in preventing such pregnancy.”79 Michael M.’s footnote seven adds: Even if the preservation of female chastity were one of the motives of the statute, and even if that motive be impermissible, petitioner’s argument must fail because ‘[i]t is a familiar practice of constitutional law that this court will not strike down an otherwise constitutional statute on the basis of an alleged illicit legislative motive.’80 The opinion then cites a number of cases, adding a “See also” cite to the Arlington Heights footnote, but without parenthetical or other explanation.81 The surprise here is that Michael M.’s language plainly overstates the rule the Arlington Heights footnote prescribes: it deems the existence of a permissible motive sufficient to defeat an equal protection claim without requiring proof that the statute would have been enacted absent the impermissible motive.82 If this recitation of the Arlington Heights Footnote 21 progeny leaves the reader uncertain as to the vitality or interpretation of the footnote, then she may have sympathy with the lower courts that attempted to discern whether and how to apply the footnote to Batson cases, which were almost the only racial discrimination cases in which they faced such questions. Before I turn to Batson, and the application of the footnote in those cases, I digress briefly to note that other approaches to intent and but-for causation are possible. E. Other Approaches to Intent and But-For Causation This section makes no claim that but-for causation requirements or affirmative defenses that function like but-for cause requirements8 3 are 77. Id. at 247. 78. 450 U.S. 464 (1981). 79. Id. at 472. 80. Id. at 472 n.7. 81. Id. 82. Id. 83. There is some dispute as to whether it is more appropriate to see but-for causation as an element of a discrimination claim or as an affirmative defense to such a claim. I do not distinguish between those arguments because, as I will argue, I think both are mistaken in the criminal justice 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 19 uncommon, but only that they are not universal. As noted above, in 1989 the Supreme Court decided Price Waterhouse,84 holding that once a Title VII plaintiff has shown that a forbidden motivation influenced an employment decision, the burden shifts to the government or the employer to show that the adverse decision would have been made absent the forbidden motivation.85 This, of course, comports with the approach of the Arlington Heights footnote. However, in 1991 Congress amended Title VII to say that except as otherwise provided, an unlawful employment practice is established when the complaining party demonstrates that the race, color, religion, sex or national origin was a motivating factor for an employment practice—whether or not other factors also motivated the practice.86 So the revised Title VII jettisons the idea of that an employer can rescue a discriminatorily motivated decision by saying it would have made the same decision were its motives pure.87 The reader may think: That example shows little, because it is a statute. Congress can impose or decline to impose whatever burdens it chooses. But another example where but-for causation is disputed arises in Establishment Clause jurisprudence. Although the rule of Lemon v. Kurtzman88—now abandoned89—had exceptions, it generally prohibits governmental actions taken for a religious purpose (as well as some forms of governmental action that have large effects on religion or create undue entanglement between church and state).90 But the language the opinions use to describe the religious purpose test varied quite a bit, sometimes suggesting but- for causation and other times appearing to reject it. Lemon itself said only that “[t]he statute must have a secular purpose.”91 This sounds like the state need only show the existence of some secular purpose (much as Michael M. implies), not that the legislation would have been enacted were that secular purpose the only motivation. In contrast, Edwards v. Aguillard92 seemed to require much more of context, which is not to say that they are equivalent or equally desirable in other contexts. 84. See Price Waterhouse v. Hopkins, 490 U.S. 228 (1989). 85. See id. at 250. 86. 42 U.S.C. § 2000e-2 (1991). 87. Although the Civil Rights Act of 1991, which amends Title VII of the Civil Rights Act of 1964, supplants Price Waterhouse’s determination that mixed motives completely shield an employer, it does not render them irrelevant. If a plaintiff proves that an illicit motivation (race, sex, color, national origin, or religion) was the basis for the adverse employment action, the defendant faces liability regardless of whether or not he has a “same decision defense.” But if a defendant can establish the defense, remedies are limited to declaratory and injunctive relief and attorney’s fees. Civil Rights Act of 1991 §§ 107(a)-(b), codified at 42 U.S.C.A. §§ 2000e-2(m), 2000e-5(g)(2)(B). 88. 403 U.S. 602 (1971). 89. The Supreme Court did not explicitly overturn Lemon in Kennedy v. Bremerton School District, 142 S. Ct. 2407 (2022), but it instructed lower courts to replace Lemon with a new standard for evaluating religious actions in a public school. Justice Sotomayor’s dissent convincingly asserted that Kennedy had overturned Lemon. 90. Lemon, 403 U.S. at 612-13. 91. Id. 92. 482 U.S. 578, 594 (1987). 20 INDIANA LAW REVIEW [Vol. 57:7 the state. According to the Edwards majority, “[b]ecause the primary purpose of the Creationism Act is to advance a particular religious belief, the Act endorses religion in violation of the First Amendment,”93 this characterization of the Establishment Clause seems to preclude resuscitating the statute by reference to a secular purpose even if the secular purpose would have been sufficient to assure passage of the statute absent the impermissible purpose. In Edwards, Justice Scalia dissented, objecting that a secular purpose is sufficient to sustain the statute.94 Thus, the Establishment Clause dispute (in which the argument of this Article has no stake) also demonstrates the plausibility of either incorporating or rejecting a but-for cause requirement. Another rejection of but-for causation worth noting is more thought experiment than example. Suppose that a defendant charged with arson says: “I did intend to burn down the building, but really, I would have lit the match in order to see in the dark even if I had not wanted to burn down the building. That would have resulted in the building burning down anyway.” Or, “Yes, I picked up the wallet the man dropped on the subway intending to permanently deprive him of its contents, but if I hadn’t intended to do that, I would have picked it up anyway because I am a neat freak, and I hate to see trash on the subway. If I had picked it up for that reason, he would have lost his money anyway because there is no identification in the wallet, and I could not have given it back.” Now we might not believe these claimed alternative motivations, but even if we did, it seems laughable that such a “dual motivation” would make any difference in the defendant’s criminal liability. The criminal law looks at the defendant’s worst motivation to establish or calibrate his criminal responsibility, regardless of whether the action he took might have been taken even absent that motivation. The relationship of prophylactic rules to but-for causation requirements is also instructive. Why is a statement obtained without Miranda warnings inadmissible? Not because it is always the product of compulsion, but because it often is, and Miranda warnings are designed to dispel such compulsion when it exists. Consequently, Miranda v. Arizona95 did not remand the case to the Arizona courts to give the state the opportunity to show that Ernesto Miranda would have waived his rights had he been informed of them. Similarly, the deprivation of counsel at a line-up results in the suppression of any identification that occurs at the line-up, rather than an inquiry into whether something the lawyer might have done at the line-up would have either prevented the identification or impeached it. One defense of these prophylactic rules might be that they are justified by the difficulty of the but-for inquiry, or the infrequency with which the state could disprove but-for causation of harm to the defendant, or the likelihood that such but-for determinations would be tainted by the desire to obtain the evidence. That may be so, but it only makes the point that but-for requirements are not always worth the candle, and that whether to permit the state to disprove but-for causation should depend upon a weighing of the costs and 93. Id. at 593 (emphasis added). 94. Id. at 619 (Scalia, J., dissenting). 95. 384 U.S. 436 (1966). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 21 benefits. I offer these examples to show that the election of a but-for-cause requirement is not inevitable. If regardless of context, the law virtually always answers questions of a particular kind in the same way, then choosing to answer such a question in a different way in a single context requires a strong justification. Here, however, no such broad generalization exists; the law does not take a uniform approach to the question of whether to focus solely on the most—or least—invidious reason in assessing responsibility for the harm that ensues from an action taken for multiple reasons. Nor, more broadly, does it always require but-for causation before imposing a remedy. I turn now to recounting the application of the Arlington Heights footnote in the Batson line of cases, and then will come back to the merits of its application in that context (and in other criminal justice settings). II. BATSON AND THE ARLINGTON HEIGHTS FOOTNOTE A. The Road to Flowers As noted in Part I, one of the earliest applications of the Equal Protection Clause of the Fourteenth Amendment was in the area of venire selection in 1880.96 Though the Supreme Court became increasingly vigilant in policing exclusion from the jury venire, it did not address the legitimacy of exclusion by peremptory challenge until 1965, when Swain v. Alabama9 7 held that the prosecutor’s racially motivated strikes of all six African Americans from the jury of an African-American defendant did not violate equal protection. According to the unanimous Swain court, a generalization that an African-American juror is more likely to be partial to an African-American defendant is permissible; Swain’s only caveat was that if a defendant could prove that the prosecutor struck African-American jurors in every case, regardless of the crime, race of the defendant, or race of the victim, then he had violated the Equal Protection Clause.98 Not surprisingly, “Swain claims” that a prosecutor had engaged in such pervasive exclusion virtually always failed.99 A decade and an African-American justice later, the Supreme Court reversed course, and in Batson v. Kentucky,100 held that a prosecutor violates equal protection norms when he or she exercises even a single peremptory challenge for a racial reason. Batson instructs trial judges on the proper procedure for adjudicating an allegation of such discrimination: First the defendant must be permitted to establish a prima facie case of discrimination; next the prosecutor 96. Strauder v. West Virginia, 100 U.S. 303 (1879). 97. 380 U.S. 202 (1965). 98. Id. at 223. 99. Sheri Lynn Johnson, Black Innocence and the White Jury, 83 MICH. L. REV. 1611, 1658 n. 240 (1985). 100. Batson v. Kentucky, 476 U.S. 79, 92-93 (describing Swain’s burden of proof as “crippling”). 22 INDIANA LAW REVIEW [Vol. 57:7 must be permitted to offer race-neutral explanations for his or her strikes; and finally, the court must determine whether those stated reasons were pretextual.101 Batson quotes Arlington Heights for the proposition that in “deciding if the defendant has carried his burden of persuasion, a court must undertake ‘a sensitive inquiry into such circumstantial and direct evidence of intent as may be available.’”102 However, Batson makes no reference to the Arlington Heights footnote and concomitantly, at no point imposes a fourth-step determination of whether the strike, if determined to be racially motivated, would have been made absent the racial motivation. The Court soon extended Batson’s reach, first to white defendant/African American juror cases,103 then to civil cases,104 and finally to peremptory strikes by defense counsel,105 but, as many critics complained, seemed uninterested in enforcement of Batson’s command.106 Hernandez v. New York 107 held that striking Latino jurors because they spoke Spanish was a race-neutral reason which a judge could find was not a pretext for racial discrimination—despite the fact, omitted from the Supreme Court’s opinion, that the prosecutor had only asked Latino jurors if they spoke Spanish.108 Purkett v. Elem109 was equally unhelpful, reversing an Eighth Circuit grant of relief and criticizing it for requiring even a “minimally persuasive” reason at the second step. Some lower courts took Batson seriously, but for twenty years, those that did not faced no sanction from the Supreme Court.110 Until Miller-El111 Miller-El took two trips to the Supreme Court before his conviction was reversed for Batson error,112 but the second time around, the Supreme Court determined that considering all the relevant circumstances, the only reasonable conclusion was the prosecutor’s exercise of the peremptory 101. Id. at 80. 102. Id. at 93 (quoting Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 266 (1977)). 103. Powers v. Ohio, 499 U.S. 400 (1991). 104. Edmonson v. Leesville Concrete, 500 U.S. 614 (1991). 105. Georgia v. McCollum, 505 U.S. 42 (1992). 106. See Miller-El v. Dretke, 545 U.S. 231, 268-69 (Breyer, J., concurring) (noting a sampling of the critical literature). For similar more recent judicial criticism, see Taylor v. Jordan, 10 F.4th 625, 642 (6th Cir. 2021) (Moore, J., dissenting) (also sampling the critical literature). 107. 500 U.S. 352 (1991). 108. People v. Hernandez, 552 N.E.2d 621, 622 (N.Y.1990) (Kaye, J., dissenting) (“[T]here is no indication that any other members of the panel were also asked if they spoke Spanish.”). 109. 514 U.S. 765 (1995). 110. See Miller-El, 5425 U.S. at 268-69 (Breyer, J., concurring). 111. Miller-El v. Dretke, 545 U.S. 231 (2005). 112. Interestingly, three of the four Batson merits wins—Miller El, Snyder, and Flowers—required two Supreme Court certiorari grants, perhaps suggesting the degree of resistance in some lower courts, or perhaps suggesting the Supreme Court’s unwillingness to reverse for Batson error except when faced with open recalcitrance. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 23 challenge was racially motivated.113 The Miller-El majority observed that “[i]f any facially neutral reason sufficed to answer a Batson challenge, then Batson would not amount to much more than [Swain],114 and proceeded to identify factors “bear[ing] upon the issue of racial animosity.”1 15 Interestingly, the Court did not reference the general Arlington Heights list of factors probative of discriminatory intent, but provided a context-specific list: the strength of the prima facie case;116 “side-by-side comparisons of some black venire panelists who were struck and white panelists allowed to serve;”117 failure to voir dire on the reasons purportedly grounding a strike;118 “how reasonable, or how improbable, the explanations are . . . and [] whether the proffered rationale has some basis in accepted trial strategy;”119 “contrasting voir dire questions posed respectively to black and nonblack panel members;”120 mischaracterization of the evidence;121 and a history of racial discrimination by the prosecuting office.122 Miller-El, like Batson, cited Arlington Heights123 but not Footnote 21, and like Batson, said nothing about providing the state an opportunity to show that it would have struck the same jurors even absent racial motivation. Snyder v. Louisiana,124 first granted, vacated, and remanded for reconsideration in light of Miller-El, and then granted certiorari after the Louisiana Supreme Court affirmed its prior decision, suggested that Miller-El was not a one-off. The Supreme Court carefully scrutinized two proffered reasons for striking prospective juror Brooks; it declined to assume that the trial court relied upon the first reason—demeanor—when the trial court itself only cited the second reason and based upon the voir dire and comparisons to seated white jurors, found the second reason implausible.125 More importantly for our purposes, the Supreme Court took up the question of but-for causation, albeit only to set it aside. Citing Hunter v. Underwood (though not the Arlington Heights footnote), the opinion acknowledges that “in other circumstances . . . once it is shown that a discriminatory intent was a 113. Miller-El, 545 U.S. at 265-66. 114. Id. at 239-40. 115. Snyder v. Louisiana, 552 U.S. 472, 478 (2008). 116. Miller-El, 545 U.S. at 240. 117. Id. at 241. 118. Id. at 244. 119. Id. at 247. 120. Id. at 255. 121. Id. at 244. 122. Id. at 263. Miller-El also noted that “the prosecution’s decision to seek a jury shuffle when a predominant number of African–Americans were seated in the front of the panel, along with its decision to delay a formal objection to the defense’s shuffle until after the new racial composition was revealed” raised a suspicion that the State sought to exclude African–Americans from the jury. Id. at 254. 123. Id. at 239. 124. 552 U.S. 472 (2008). 125. Id. at 479-80. 24 INDIANA LAW REVIEW [Vol. 57:7 substantial or motivating factor in an action taken by a state actor, the burden shifts to the party defending the action to show that this factor was not determinative.”126 Then, however, the opinion (correctly) points out that the Court had never previously applied this rule in a Batson case, and asserts that “we need not decide here whether that standard governs in this context.” Why not? For present purposes, it is enough to recognize that a peremptory strike shown to have been motivated in substantial part by discriminatory intent could not be sustained based on any lesser showing by the prosecution. And in light of the circumstances here—including absence of anything in the record showing that the trial judge credited the claim that Mr. Brooks was nervous, the prosecution’s description of both of its proffered explanations as “main concern[s],” . . . and the adverse inference noted above—the record does not show that the prosecution would have pre-emptively challenged Mr. Brooks based on his nervousness alone. Nor is there any realistic possibility that this subtle question of causation could be profitably explored further on remand at this late date, more than a decade after petitioner’s trial.127 Thus, the Snyder Court both declined to decide whether the but-for causation requirement applies to a Batson claim and hinted that the state could not meet that requirement were it applicable. As I will summarize shortly, this treatment provided little guidance for lower courts that were struggling with the but-for question. Despite disarray in the lower courts, the Supreme Court said nothing for another eight years. Then Foster v. Chatman,128 decided in 2016, again noted the question, but again failed to resolve it. After reviewing the evidence of discriminatory intent, including comparisons between struck Black jurors and seated white jurors, “shifting explanations, the misrepresentations of the record, and the persistent focus on race in the prosecution’s file” the Foster majority concluded: “[c]onsidering all of the circumstantial evidence that ‘bear[s] upon the issue of racial animosity,’ we are left with the firm conviction that the strikes of [two African American jurors] were ‘motivated in substantial part by discriminatory intent.’”129 So far, so good. But the quoted language comes from Snyder, and at the end of citation to Snyder, buried in a footnote, comes what little the Court is willing to say about shifting the burden of proof: In Snyder, we noted that we had not previously allowed the prosecution to show that “a discriminatory intent [that] was a substantial or motivating factor” behind a strike was nevertheless not “determinative” to the prosecution’s decision to exercise the strike. The State does not raise such an argument here and so, as in Snyder, we need not decide the 126. Id. at 485 (citing Hunter v. Underwood, 471 U.S. 222, 228 (1985)). 127. Id. at 485-86. 128. 578 U.S. 488 (2016) (emphasis added). 129. Id. at 1754 (citing Snyder v. Louisiana, 552 U.S. 472, 478, 485 (2008)). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 25 availability of such a defense.130 B. Confusion in the Lower Courts Until Batson very few cases cited the Arlington Heights footnote, either in the Supreme Court, as discussed above, or in the lower courts. Despite what might seem like both doctrinal and practical importance, Footnote 21 almost disappeared until it surfaced in lower court Batson cases; such courts by and large had not been required to consider the footnote in racial discrimination claims prior to Batson mainly because racial motivation was so rarely established in those cases. But in the early Batson cases, establishing racial motivation was sometimes easy because prosecutors admitted it, generally without being aware that they had done so. A prosecutor might say, for example, “I struck her because she has teenage sons, and the defendant is a Black teenage boy,” which plainly is not race-neutral as Batson’s second step requires. For a prosecutor trained under the rule of Swain, it sounded fine, and then, when confronted with the fact that she had cited a race-based reason, the prosecutor would respond: “Yes, but I also struck her because she was weak on the death penalty.” Other times a prosecutor might cite two reasons in the same breath: “I struck him because he and the defendant are both black men in their 20s, and also because he lives in a crime-prone neighborhood.”131 In either situation the trial court (and the reviewing court when the strike was sustained) would have to decide what to do with one impermissible and one permissible reason. The lower courts labeled this question in various ways—”mixed motives,” “dual motivation,” or “but-for causation”—and responded in various ways. But before I turn to the different approaches, it is important to notice that this question of dual motivation can and does occur even in cases where no explicitly race-based reason is proffered; this matters because in more recent cases, explicit reference to race as a reason is much less common. Suppose a prosecutor says: “I struck her because she was friends with the defendant’s aunt, because her occupation begins with the letter “p,” and because she has a brother who was charged with the same crime as the defendant in this case.” Further suppose that the first reason is not true. Now the prosecutor has cited one false reason, one implausible reason, and one seemingly legitimate reason. On that basis, a judge might find the first two reasons pretextual, and that they establish racial motivation, but then face the contention that the last reason is legitimate and that absent racial motivation, the strike would have been made for the legitimate reason alone. Indeed, Snyder and Foster, which explicitly put off the question of mixed motives, as well as Flowers, are all cases where multiple race-neutral 130. Id. at 1754 n.6. 131. See, e.g., Walker v. Girdich, 410 F.3d 120, 121-23 (2d Cir. 2005) (finding that despite giving several reasons for striking an African-American juror, the prosecutor “focused upon Mr. Jones’s race,” when she stated, “‘[o]kay, one of the main things I had a problem with was that this is an individual who was a Black man with no kids and no family,’” and holding that the prosecutor’s reason “was not race-neutral”) (internal citation and emphasis omitted). 26 INDIANA LAW REVIEW [Vol. 57:7 reasons were proffered, some obviously pretextual and others not. Currently, there are three principal responses to the counterfactual contention that a challenged strike, even if racially motivated, would have been made absent illicit motivation: the “mixed motives” or “dual motivation” approach; the “taint,” or “per se” approach; and the “substantial motivation” approach. 1. Dual Motivation.—The “dual motivation” or “mixed motives” approach gives the state the opportunity to convince the court that the prosecutor would have made the strike even absent the racial motivation established by the defendant. The Second Circuit pioneered this application of but-for causation to Batson in Howard v. Senkowski.132 Howard acknowledged that Batson itself did not impose any sort of dual motivation analysis, but after extensive discussion of Arlington Heights and Footnote 21, concluded that: Batson challenges may be brought by defendants who can show that racial discrimination was a substantial part of the motivation for a prosecutor’s peremptory challenges, leaving to the prosecutor the affirmative defense of showing that the same challenges would have been exercised for race-neutral reasons in the absence of such partially improper motivation.133 According to the Second Circuit, “in concluding that dual motivation analysis applies to a Batson challenge, we do no more than apply that analysis precisely as previously enunciated by the Supreme Court in prior dual motivation cases such as Arlington Heights and Price Waterhouse.134 Since Howard, the Second Circuit has regularly employed “dual motivation” analysis,135 and the Third,136 Fourth,137 Fifth,138 Sixth,139 Seventh,140 Eighth,141 and Eleventh Circuits,142 as well 132. 986 F.2d 24 (2d Cir. 1993). 133. Id. at 30. 134. Id. 135. See e.g., U.S. v. Taylor, 92 F.3d 1313 (2d Cir. 1996); United States v. Brown, 352 F.3d 654 (2d Cir. 2003); Wells v. Ricks, 2008 WL 506294 (S.D.N.Y. 2008); United States v. Douglas, 525 F.3d 225 (2d Cir. 2008); Meikle v. Dzurenda, 2009 WL 413157 (D. Conn. 2009); Perez v. Miller, 2009 WL 1044629 (E.D.N.Y. 2009); and Alston v. Cahill, 2012 WL 3288923 (D. Conn. 2012). 136. Gattis v. Snyder, 278 F.3d 222 (3d Cir. 2002), a habeas case, cited Howard as support for the proposition that a mixed motive approach was not contrary to Supreme Court precedent, and then later adopted a “mixed motive” analysis for application in direct appeal cases. United States v. DeJesus, 347 F.3d 500 (3d Cir. 2003). 137. Jones v. Plaster, 57 F.3d 417, 421-22 (4th Cir. 1995). 138. Haynes v. Thaler, 438 F. App’x 324 (5th Cir. 2011). 139. United States v. Peraza, 25 F.3d 1051 (6th Cir. 1994). 140. Pecor v. Walls, 56 F. App’x 723, 726 (7th Cir. 2003). 141. United States v. Darden, 70 F.3d 1507, 1531 (8th Cir. 1995). 142. The Eleventh Circuit first cited Howard with approval in a habeas case, Wallace v. Morrison, 87 F.3d at 1274-75, and then adopted it for application in direct appeal cases in United States v. Tokars, 95 F.3d 1520 (11th Cir. 1996). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 27 state courts in Illinois143 and Texas,144 have followed its lead.145 2. Taint.—The most defendant-friendly approach was first advocated by Justice Marshall, dissenting from denial of certiorari in Wilkerson v. Texas: [A] ‘neutral’ explanation for challenging an Afro-American juror means just what it says—that the explanation must not be tainted by any impermissible factors. Requiring anything less undermines an already underprotective means of safeguarding the integrity of the criminal jury selection process.146 Justice Marshall distinguished the Batson inquiry from employment discrimination claims on the ground that, while a company’s hiring criteria can be used to test an employer’s assertion that a Black job applicant would not have been hired even absent a discriminatory motive, meaningful review of a prosecutor’s assertion that he would have struck the juror absent racial motivation is not possible.147 At least eight state courts have been persuaded by this reasoning.148 3. Substantial Motivation.—Although no federal circuit has adopted the taint 143. People v. Hudson, 745 N.E.2d 1246 (2001). 144. Guzman v. State, 85 S.W.3d 242 (Tex. Crim. App. 2002). 145. Even more generous to the state than the dual motivation approach is the “sole motivation” approach, which upholds a strike unless the proponent was motivated “solely” by race; it does not ask the state to disprove but-for causation. The Second Circuit explicitly rejected this approach in Howard, and it is rare. Tennessee employs this standard, which the Sixth Circuit has upheld in the federal habeas context as not an unreasonable application of the relevant clearly established Supreme Court law (as is required to grant relief under 28 U.S.C. § 2254(d)). Akins v. Easterling, 648 F.3d 380, 392 (6th Cir. 2011). At one time Kansas also applied the sole motivation standard, which the Tenth Circuit also upheld in the habeas context, while noting that Kansas law no longer applies that test. Washington v. Roberts, 846 F3d. 1283, 1287 (10th Cir. 2017). 146. 493 U.S. 924, 928 (1989) (Marshall, J., dissenting from denial of certiorari). 147. Id. at 927. 148. See, e.g., People v. Douglas, 232 Cal. Rptr. 3d 305 (Ct. App. 2018) (reversing lower court based on discriminatory taint in jury selection); State v. Lucas 18 P.3d 160, 163 (Ariz. App. 2001) (“any consideration of a discriminatory factor directly conflicts with the purpose of Batson and taints the entire jury selection process”); Payton v. Kearse, 495 S.E.2d 205, 210 (S.C. 1998) (“Once a discriminatory reason has been uncovered—either inherent or pretextual—this reason taints the entire jury selection procedure”); Riley v. Commonwealth (1995) 21 Va.App. 464 S.E.2d 508, 510 (“The fact that the Commonwealth [also] used age to identify which women to strike does not overcome the constitutional infirmity”); Rector v. State 444 S.E.2d 862, 865 (Ga. App. 1994) (trial court erred in ruling a “purportedly race-neutral explanation[ ] cured the element of the stereotypical reasoning employed by the State’s attorney”); Ex parte Sockwell 675 So.2d 38, 41 (Ala. 1995); McCray v. State 738 So.2d 911, 914 (Ala. Crim. App. 1998) (following Sockwell; “in Alabama, a race-neutral reason . . . will not ‘cancel out’ a race-based reason”); State v. King 572 N.W.2d 530, 535-36 (Wis. App. 1997) (“[W]here the challenged party admits reliance on a prohibited discriminatory characteristic . . . a response that other factors were also used is [in]sufficient rebuttal under the second prong of Batson.”). 28 INDIANA LAW REVIEW [Vol. 57:7 approach, the Ninth Circuit has rejected dual motivation and adopted a “substantial motivation” approach, which is closely related to the taint approach, albeit one that asks slightly more of the defendant. Cook v. LaMarque explicitly rejects a mixed motive/dual motivation approach, instead “limit[ing] . . . inquiry to whether the prosecutor was motivated in substantial part by discriminatory intent.”1 49 Thus, the difference between Justice Marshall’s taint approach and the substantial motivation approach lies in those cases where the racial motivation, though established, is not “substantial;” it seems unlikely that there would be many such cases. Cook defends its decision to not follow the circuits that adopted a mixed motives approach by claiming that it is following Snyder v. Louisiana, which according to Cook, “declin[ed] to adopt mixed-motives analysis for Batson cases.”150 Unfortunately, Snyder did not really do that. Rather, Snyder first noted that “in other circumstances, we have held that, once it is shown that a discriminatory intent was a substantial or motivating factor in an action taken by a state actor, the burden shifts to the party defending the action to show that this factor was not determinative,” then observed that the Court has “not previously applied this rule in a Batson case,” and finally, deemed it unnecessary to decide “whether that standard governs in this context.”151 Why was it unnecessary? Because “a peremptory strike shown to have been motivated in substantial part by discriminatory intent could not be sustained based on any lesser showing by the prosecution.”152 Here, “lesser showing” means any showing less than the burden-shifting showing required in the other cases where the Court has permitted the state to show that the forbidden motive “was not determinative.”153 Snyder then sets forth the case-specific reasons that Louisiana would be unable to make that “not determinative” showing.154 Thus, Snyder did not, contrary to the Ninth Circuit, reject the applicability of the Arlington Heights footnote to Batson cases, and no court other than the Ninth Circuit has concluded that it did. Snyder did, however, as other courts have overlooked, flag that the importation of but-for inquiry into Batson cases wrought by the dual motivation approach might be wrong. And as discussed above, so does Foster v. Chatman, which brings us, finally, to Flowers. B. Flowers and Arlington Heights Footnote 21 This Article began with Justice Kavanagh’s assertion that Flowers “breaks no new ground.” As I address below, that is not true and should not be true with respect to the applicability of the Arlington Heights footnote to Batson cases. However, Flowers also makes several other noteworthy, though perhaps not 149. 593 F.3d 810, 815 (9th Cir. 2010) (emphasis added). 150. Id. at 814. 151. Snyder v. Louisiana, 552 U.S. 472, 485 (2008). 152. Id. 153. Id. 154. Id. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 29 “groundbreaking,” contributions to jury selection law. Justice Kavanagh’s long introduction firmly situating Batson in both Equal Protection Clause and criminal justice history makes plain that despite what Justice Thomas might wish,155 Batson is here to stay. Flowers also revives reliance on the defendant’s interest in racially neutral jury selection, an interest which disappeared in Powers v. Ohio156 in favor of a seemingly exclusive focus on the rights of excluded jurors. More specifically, the opinion supplements Miller-El’s list of the indicia of discrimination. It notes the race of the prosecutor and the victims, thus suggesting the relevance of the racial identity of various actors to the Batson analysis, offers a very broad description of what counts as a relevant history of discrimination,157 and chastises the judge for not “sufficiently” accounting for the history.158 The opinion also expands Miller-El’s “disparate questioning” factor to include disparate investigation159 and stresses the importance of disparate inquiry as a factor in assessing racial motivation.160 It deems implausible one of the reasons 155. See Flowers v. Mississippi, 139 S. Ct. 2228, 2269 (2019) (Thomas, J, dissenting) (“That rule [Batson] was suspect when it was announced, and I am even less confident of it today.”). 156. 476 U.S. 79 (1986). 157. The Mississippi Supreme Court’s assertion that the prosecutor’s personal history of discrimination was unlike the office policy of discrimination the Court had deemed relevant in Miller-El was plainly wrong; it is more probative, and it is no surprise that Flowers notes and corrects that error. But the opinion also endorses the relevance of the total number of Black jurors the prosecutor struck over the course of six trials, and the fact that in the two trials where more than one Black juror was permitted to serve. It points out that Batson permits the use of the same kind of historical evidence that Swain had allowed and deems relevant a broad swath of historical evidence: “the State’s discriminatory peremptory strikes from past trials in the jurisdiction.” Flowers v. Mississippi, 139 S. Ct. 2228, 2245 (2019). 158. Id. at 2246. 159. As Mississippi Supreme Court Justice King—who dissented from the state court opinion Flowers reversed—has protested, the Mississippi Supreme Court has refused to take note of this aspect of Flowers: I raise serious concerns over the trial court’s refusal to allow the defense to examine which jurors the State had investigated, and its reliance on unproved outside information. Disparate investigation of black and white potential jurors is an indication of pretext. The defense cannot hope to make a record of disparate investigation when the trial court refuses to allow it to access the investigation to make that record. Eubanks v. State, 291 So. 3d 309, 328 (2020) (King, J., dissenting) (internal quotation marks omitted). 160. Flowers, 139 S. Ct. at 2248 (“[B]y asking a lot of questions of the black prospective jurors or conducting additional inquiry into their backgrounds, a prosecutor can try to find some pretextual reason—any reason—that the prosecutor can later articulate to justify what is in reality a racially motivated strike. And by not doing the same for white prospective jurors, by not asking white prospective jurors those same questions, the prosecutor can try to distort the record so as to thereby avoid being accused of treating black and white jurors differently. Disparity in questioning and investigation can produce a record that says little about white prospective jurors and is therefore resistant to characteristic-by-characteristic comparisons of struck black prospective jurors and seated 30 INDIANA LAW REVIEW [Vol. 57:7 that distinguished a struck Black juror from seated white jurors and points out characteristics of the struck juror that should have made her attractive to the prosecution. Finally, the opinion vehemently insists upon viewing all of the evidence of racial motivation cumulatively, rather than considering it juror by juror.161 These additions have been noted by several lower courts,162 notwithstanding Justice Kavanagh’s “we break no new ground” disclaimers. My point in noting these additions to the discriminatory purpose tool kit is that Justice Kavanagh’s claim that Flowers breaks no new legal ground is not strictly accurate even with respect to the smaller question of what counts in establishing racial motivation.163 And with respect to the next question—what happens when racial motivation is established—Flowers works a much larger change. Perhaps one might defend Kavanagh’s assertion by saying that it was Snyder that “broke” the new ground, but no matter. Whether Flowers accomplishes what Snyder and Foster only foreshadowed or completes what Snyder began, it should end the controversy with which the lower courts have struggled. Recall that Snyder first pointed out that in other circumstances “once it is shown that a discriminatory intent was a substantial or motivating factor in an action taken by a state actor, the burden shifts to the party defending the action to show that this factor was not determinative,”164 but then observed that the white jurors. Prosecutors can decline to seek what they do not want to find about white prospective jurors.”). 161. Id. at 2250 (“As we see it, the overall context here requires skepticism of the State’s strike of Carolyn Wright. We must examine the Wright strike in light of the history of the State’s use of peremptory strikes in the prior trials, the State’s decision to strike five out of six black prospective jurors at Flowers’ sixth trial, and the State’s vastly disparate questioning of black and white prospective jurors during jury selection at the sixth trial. We cannot just look away. Nor can we focus on the Wright strike in isolation.”). 162. See, e.g., Ervin v. Davis, 12 F.4th 1102, 1108 (9th Cir. 2021) (“Given the Court’s recent guidance in Flowers, and under the unique circumstances of this case, we believe that the district court is in the best position to evaluate the Flowers factors anew [and] therefore vacate the judgment and remand to the district court so it can evaluate Ervin’s Batson claims in light of the Supreme Court’s guidance in Flowers); Jones v. Broomfield, 562 F. Supp. 3d 652, 678 ((C.D. Cal. 2021) (“Here, the Court considers only that evidence that was before the trial court, but notes that five of the six types of evidence referred to in Flowers apply here”). 163. At least two lower courts have, after extensive discussion of Flowers, cited it as a compelling reversal of the case before them. Ervin, 12 F.4th at 1102; State v. Kirk, 145 N.E.3d 1092 (Ohio App. 2019); State v. Alexander, 851 S.E.2d 411 (N.C. App. 2020) (noting that “our review of Defendant’s appeal is controlled by [Flowers and a state court decision that relies on Flowers] not available to the lower court at the time of trial”). See also State v. Hobbs, 841 S.E.2d 492 (N.C. App. 2020) (relying heavily on Flowers in its consideration of the history of discrimination in the county); State v. Andujar, 254 A.3d 606, 628 (noting and applying Flowers’ disapproval of disparate investigation of minority jurors); but see Willacy v. State, 314 So. 3d 246, 247 (Fla. 2021) (“Flowers did not establish a new constitutional right that has been held to apply retroactively”). 164. Snyder v. Louisiana, 552 U.S. 472, 485 (2008) (citing Hunter v. Underwood, 471 U.S. 222, 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 31 Court had never applied this rule in a Batson case, and finally declined to determine “whether that standard governs in this context.”165 Snyder justified delaying resolution of the question because the record itself did not reflect that the prosecution would have pre-emptively challenged the juror based on a permissible purpose alone, and because “this subtle question of causation” could not profitably be explored so long after Snyder’s trial.166 Foster167 again flagged the question of whether the state is entitled to an opportunity to disprove but-for causation, and like Snyder, failed to resolve it, this time more briefly defending “not decid[ing] the availability of such a defense” by pointing to the State’s failure to raise the issue.168 Flowers, however, with respect to the question of the availability of the defense, is much simpler: All that we need to decide . . . is that all of the relevant facts and circumstances taken together establish that the trial court at Flowers’ sixth trial committed clear error in concluding that the State’s peremptory strike of black prospective juror Carolyn Wright was not motivated in substantial part by discriminatory intent.169 No comment on what was not decided, no explanatory footnote, no reason for not reaching the question of but-for causation. “All” that needs to be decided is that it was clear error to determine that the peremptory strike of Carolyn Wright “was not motivated in substantial part by discriminatory intent” because proof of substantial discriminatory motivation in the Batson context does not lead to a Footnote 21 inquiry but is all that needs to be determined to compel reversal.170 No lower court has explicitly addressed whether Flowers ends the controversy over the applicability of a Footnote 21 inquiry to Batson cases, but the import of that single sentence has not gone unheeded. Two circuits that previously employed dual motivation analyses now cite the “motivated in substantial part by discriminatory intent” language of Flowers171 and the Ninth Circuit has reaffirmed its rejection of dual motivation analysis in an opinion that cites Flowers,172 as has the Arizona Supreme Court.173 An Ohio intermediate 228 (1985)). 165. Id. 166. Id. at 485-86. 167. Foster v. Chatman, 578U.S. 488 (2016). 168. Id. at 513 n.6. 169. Flowers v. Mississippi, 139 S. Ct. 2228, 2251 (2019) (emphasis added). 170. Id. 171. United States v. Adams, 996 F.3d 514, 520 (8th Cir. 2021); United States v. Iron Crow, 970 F.3d 1003, 1007 (8th Cir. 2020); Hunt v. Sunquist, 822 F. App’x 468 (6th Cir. 2020); Meirs v. Ottawa Cty., 821 F. App’x 445, 459 (6th Cir. 2020). See also King v. Warden, Ga. Diagnostic Prison, 2020 WL 423344 at *31 (SD Ga. 2020) (citing Flowers generally and quoting the “substantial motivation” language from Foster and Snyder). 172. Walker v. Davis, 822 F. App’x 549 (9th Cir. 2020); see also, Infante v. Martel, 953 F.3d 560, 566 (9th Cir. 2020) (quoting without discussing the “substantially motivated” language from 32 INDIANA LAW REVIEW [Vol. 57:7 appellate court has, citing Flowers, held that the establishment of racial motivation precludes any reliance on race-neutral reasons.174 The Colorado Supreme Court ducked the question last year after granting certiorari to review a lower court’s decision to apply an “unprecedented” substantial motivating factor test by affirming the decision on other grounds,175 but already faces another lower appellate court decision refusing to apply mixed-motive analysis.176 Since Flowers, no lower court has used “dual motivation” in its own analysis of the merits of a Batson claim, although the Fifth Circuit has, in the habeas context, employed a rule in reviewing state court decisions that is even stricter than dual motivation.177 Flowers). 173. State v. Porter, 460 P.3d 1276, 1282 (Ariz. App. 2020), vacated, 491 P.3d 1100 (Ariz. 2021) (“Arizona law provides that one non-race-neutral reason for a strike will taint any other neutral reason for the strike”). 174. State v. Saunders, 162 N.E.3d 959 (Ohio App. 2020). 175. People v. Ojeda, 487 P.3d 1117,1123 (Ct. App. 2019) (adopting the “substantial motivation test”), aff’d on other grounds, 503 P.3d 856 (Colo. 2022). 176. People v. Johnson, 523 P.3d 992 (Ct. App. 2022) (adopting a per se approach), cert. granted in part, No. 22SC852, 2023 Colo. LEXIS *458 (Colo. May 22, 2023). 177. The Fifth Circuit has, since Flowers, reiterated what appears to be an even harsher rule in its review of state court decisions, holding in a habeas case that “a Batson claim will not succeed where the defendant fails to rebut each of the prosecutor’s legitimate reasons.” Sheppard v. Davis, 967 F.3d 458, 472 (5th Cir. 2020). Since Flowers, one federal district court did hold that a state court decision that appears to employ dual motivation analysis is not “contrary to” or an “unreasonable application of” Supreme Court precedent, and thereby cannot ground federal habeas relief, but went out of its way to opine that the state court decision was nonetheless wrong in so doing. Porter v. Coyne-Fague, 528 F. Supp. 3d 2, 7 (D.R.I. 2021), rev’d and remanded, 35 F.4th 68 (1st Cir. 2022). The First Circuit reversed, citing the “substantial motivation” language of Flowers, and punting on the question of dual motivation: The Supreme Court has left open the possibility that the respondent might be able to prevail by showing that the prosecution’s discriminatory intent was “‘a substantial or motivating factor’ behind a strike” but “was nevertheless not ‘determinative’ to the prosecution’s decision to exercise the strike.” Foster, 578 U.S. at 513 n.6, 136 S.Ct. 1737 (quoting Snyder, 552 U.S. at 485, 128 S.Ct. 1203). We need not address this possibility because, here, as in Foster, the respondent has advanced no such argument. See id. We add, moreover, that the record does not establish that showing on its own; and there is no “realistic possibility that this subtle question of causation could be profitably explored further on remand at this late date.” Snyder, 552 U.S. at 485-86, 128 S. Ct. 1203. Porter v. Coyne-Fague, 35 F.4th 68, 82 n.6 (1st Cir. 2022). See also, Dixon v. State, 485 P.3d 1254, 1257 n.2 (Nev. 2021) (declining the state’s invitation to apply dual motivation analysis but on the narrow reason that no gender-neutral reason was proffered and without commenting on the continued validity of dual motivation analysis). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 33 III. IN DEFENSE OF FLOWERS This brings us to whether the Court was right to—quietly—disavow the counterfactual inquiry Footnote 21 authorizes, at least in the resolution of Batson claims. Neither Snyder nor Foster explain why the Court hadn’t previously applied the Footnote 21 inquiry in Batson cases and Flowers does not discuss why the majority concluded that “all we need to decide” is whether the strike of Carolyn Wright was substantially motivated by race. Put sympathetically, the argument against rejecting the Footnote 21 inquiry boils down to: A claimant should not get a windfall but should be put in the same position she would have been had the wrongdoer not misbehaved. Stated so generally, that sounds fair, but at least with respect to established racial discrimination in the criminal justice system, it is misguided. The problem is, how should we know whether to believe the prosecutor who says she would have done the same thing absent racial motivation? After all, she did entertain the impermissible motivation and did deny that she entertained it. We might be worried that her next assertion is unlikely to be truthful, and/or that we are unlikely to be able to discern its truthfulness. As the Price Waterhouse plurality observed in the employment context, any “suggestion . . . that the [discriminator’s] own testimony as to the probable decision in the absence of discrimination is due special credence where the court has, contrary to the [discriminator’s] testimony, found that an illegitimate factor played a part in the decision . . .” is counterintuitive.178 But at least in the employment context, as the plurality also observed, “the employer should be able to present some objective evidence as to its probable decision in the absence of an impermissible motive.”179 A comparable safeguard, however, is not present in Batson mixed motive determinations. Thus, trying to resolve the value of but-for inquiries in general is not fruitful; there is no inevitable answer to the question of whether dual motivation matters, which is why, as sketched in Part I.E., there are some areas of the law that do not follow the no-windfall approach. How the law treats mixed motives varies with the context, and that variation means that the “no windfall” approach might or might not be best. At least three factors bear on its desirability in a particular context: the frequency of the wrongful motivation; the ease or difficulty of discerning what would have occurred absent the wrongful motivation; and the relative cost of “windfalls” that accrue to the state when inquiry into but-for causation is permitted and to the claimant if inquiry ends when wrongful motivation is established. I discuss each consideration in turn, but first offer a quick summary: Significant empirical evidence–all developed since Arlington Heights was decided–supports the conclusion there is a lot of racial discrimination, and that it is virtually impossible to discern what a person who acted with racial motivation would have done absent that motivation. Moreover, the cost of conducting a Footnote 21 inquiry is greater than the cost imposed by 178. Price Waterhouse v. Hopkins, 490 U.S. 228, 295 n.14 (1989). 179. Id. at 252. 34 INDIANA LAW REVIEW [Vol. 57:7 its forbearance. A. The Frequency of Discrimination A study of federal cases in the 1980s–when we might expect to find more discrimination than we would today–revealed that very few constitutional race discrimination claims won, only about one per district per year.180 That low number is why, prior to Batson, there are virtually no Arlington Heights footnote cases; if claimants overwhelmingly lose on round one, the state never needs to make the round two footnote argument. That courts were not remedying very much racial discrimination by government factors could be because there wasn’t much discrimination, or it could be because the Washington v. Davis purposeful discrimination standard is almost impossible to satisfy once state actors are on notice that they should not reveal discriminatory motivation. At the time the Arlington Heights footnote was written, we did not have a very good way to decide between those two alternatives. But now we do. A brief history of the literature on racial bias is necessary to contrast what might have been believed about prejudice and discrimination circa Arlington Heights, and what social scientists have now established. The first well-known study of racial bias in the United States, Gunnar Myrdal’s An American Dilemma,181 compiled sociological data on race relations. Even prior to its publication in 1944, less well-known racial bias researchers were administering scales that measured the prevalence of stereotyping and antipathy. Those researchers found that large numbers of respondents would, without much hesitation, acknowledge their agreement with gross racial stereotypes,182 and the violent reaction to the Civil Right movement made plain that significant numbers of people who held such biases were willing to act on them. In the 1960s and 70s, however, studies showed substantial decreases in the number of respondents reporting racially biased attitudes.183 Racial discrimination—at least in its most obvious forms—had been legally condemned, and racial bias was rapidly becoming socially unacceptable.1 84 So, one might wonder whether bias had 180. See Theodore Eisenberg & Sheri Lynn Johnson, The Effects of Intent: Do We Know How Legal Standards Work?, 76 CORNELL L. REV. 1151, 1174 (1991) (analyzing all federal appellate and district court published opinions from June 7, 1976 to February 6, 1988 that cited either Washington v. Davis or Arlington Heights and eliminating those cases that lacked a constitutional race-based intent claim). 181. GUNNAR MYRDAL, AN AMERICAN DILEMMA: THE NEGRO PROBLEM AND MODERN DEMOCRACY (1944). 182. For the earliest such study, see Daniel Katz & Kenneth Bealy, Racial Stereotypes of One- Hundred College Students, 28 J. ABNORMAL & SOC. PSYCH. 282 (1933). 183. Sigall & Page, supra note 27; Marvin Karlins, Thomas Coffian, & Gary Walters, On the Fading of Social Stereotypes: Studies in Three Generations of College Students, 13 J. PERSONALITY & SOC. PSYCH. 1-16 (1969). 184. Sigall & Page, supra note 27. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 35 precipitously declined or simply had gone underground.185 Neither Washington v. Davis nor Arlington Heights seemed very concerned about covert discrimination, though shortly after they were decided legal commentators began protesting both that requiring proof of racial motivation would leave much racial discrimination undetected,186 and that requiring proof of racial motivation left longstanding racial inequality unaddressed.187 But whatever one might have conjectured about the prevalence of bias at the time Davis and Arlington Heights were written, since then the evidence of pervasive racial bias has grown enormously. It is no longer possible to credibly assert that racial bias and discrimination had largely disappeared; although there is ample evidence that blatant forms of racial bigotry were declining, a large body of literature makes plain that subtle and covert forms are pervasive.188 A variety of subfields—cognitive, social, and neuro psychology—all document the rising number of individuals whose ambivalent racial attitudes led them to deny their own prejudice while expressing it indirectly, covertly, and often unconsciously.189 Moreover, the evidence that racial bias continues to infect determinations of guilt and sentencing is particularly compelling; this evidence comes both from archival studies and from new laboratory studies.190 Mock jury studies were a first attempt at circumventing reluctance to self- report racial bias; by holding constant all the facts presented to subjects except the defendant’s race (and/or the victim’s race), psychologists could attribute differences in conviction rates and sentences to racial bias. Then they discovered another way to track bias that is not within conscious awareness or control: the 185. Id. 186. See, e.g., Theodore Eisenberg, Disproportionate Impact and Illicit Motive: Theories of Constitutional Adjudication, 52 N.Y.U. L. REV. 36, 114-16 (1977); Kenneth L. Karst, The Costs of Motive-Centered Inquiry, 15 SAN DIEGO L. REV. 1163 (1978); Michael J. Perry, The Disproportionate Impact Theory of Racial Discrimination, 125 U. PA. L. REV. 540, 548-49 (1977); Robert G. Schwemm, From Washington to Arlington Heights and Beyond Discriminatory Purpose in Equal Protection Litigation, 1977 U. ILL. L. REV. 961, 1001; Larry G. Simon, Racially Prejudiced Governmental Actions: A Motivation Theory of the Constitutional Ban Against Racial Discrimination, 15 SAN DIEGO L. REV. 1041 (1978). 187. See, e.g., Alan D. Freeman, Legitimizing Racial Discrimination Through Antidiscrimination Law: A Critical Review of Supreme Court Doctrine, 62 MINN. L. REV. 1049 (1978); see also David A. Strauss, Discriminatory Intent and the Taming of Brown, 56 U. CHI. L. REV. 935, 41-46 (1989) (reviewing various broader conceptions of discrimination and citing their proponents). 188. See J. F. DOVIDIO, PREJUDICE, DISCRIMINATION, AND RACISM (S.L. GAERTNER ED., 1986) (reviewing the literature). For an early and very influential comment on the legal implications of that literature, see Charles R. Lawrence III, The Id the Ego, and Equal Protection: Reckoning with Unconscious Racism, 39 STAN. L. REV. 317 (1987). 189. See Sheri Lynn Johnson, Unconscious Racism and the Criminal Law, 73 CORNELL L. REV. 1016, 1027-29 (1988) (reviewing the literature and noting the variation in terminology). 190. See Sheri Lynn Johnson, Black Innocence and the White Jury, 83 MICH. L. REV. 1611 (1985) (reviewing archival and mock jury studies). 36 INDIANA LAW REVIEW [Vol. 57:7 Implicit Associations Test, or “IAT.” IATs compare the speed with which subjects can pair words from different categories; their underlying premise is that if a person “automatically” associates one thing with another, she can more quickly pair them with each other.191 By this measure, 70 to 80 percent white Americans implicitly, automatically associate white with good and Black with bad. 1 9 2 Even legal training does not make a difference; judges, law students, and defense lawyers also exhibit these automatic reactions.193 A number of criticisms have been leveled at IATs, but they have been shown to predict discriminatory behavior better than self-report measures,194 and they have been correlated with decisions ranging from cardiologist diagnoses195 to employment decisions196 to amygdala activation.197 Moreover, IATs are not the only measure of automatic reactions that reveals racial bias. In “shooter studies,” researchers show subjects the image of a person who either has a gun or a tool in his hand and the subjects must decide very quickly whether to shoot the image. Most people shoot black targets too often, mistaking tools for guns.198 Conversely, they shoot too few white targets, mistaking guns for tools.199 Researchers studying retaliation have also uncovered racial bias. They ask subjects to perform a difficult task for either a white or a Black experimenter, and then, in a task that is purportedly unrelated, they ask the subject to assign punishment to a person of unspecified race for a described transgression. The assigned penalty when the experimenter is Black is significantly higher than when 191. Although there is a paper and pencil version of the test, typically the subject sits at a computer and sees a series of words and images. She is instructed to press the “I” key when she sees either a “good” word (such as happy or pleasant) or a white man’s face, or to press the “E” key when sees either a “bad” word (such as dangerous or vomit) or a Black man’s face. Then the process is reversed, and she is asked to press “I” for either good words or Black faces and “E” for either bad words or white faces. (For some subjects, the order of the pairings is reversed.) The computer records response times to each stimulus, and at the end of the test, calculates a score. 192. Brian A. Nosek et al., Harvesting Implicit Group Attitudes and Beliefs From a Demonstration Website, 6 GRP. DYNAMICS 101 (2002). 193. Theodore Eisenberg & Sheri Lynn Johnson, Implicit Racial Attitudes of Death Penalty Lawyers, 53 DEPAUL L. REV. 1539 (2004). 194. Anthony Greenwald et al., Understanding and Using the Implicit Association Test: III. Meta-Analysis of Predictive Validity, 97 J. PERSONALITY & SOC. PSYCH. 17 (2009). 195. Chloë FitzGerald & Samia Hurst, Implicit Bias in Healthcare Professionals: A Systematic Review, 18.1 BMC MED. ETHICS 19 (2017). 196. Jonathan C. Ziegert & Paul J. Hanges, Employment Discrimination: The Role of Implicit Attitudes, Motivation, and a Climate for Racial Bias, 90 J. APPLIED PSYCH. 553 (2005). 197. E. A. Phelps et al., Amygdala Activation Predicts Performance on Indirect Measures of Racial Bias, 12 J. COG. NEUROSCI. 729 (2000). 198. THOMAS GILOVICH ET AL., SOCIAL PSYCHOLOGY 432 (W.W. Norton 2006) (reviewing the relevant literature); see also Jerry Kang, Trojan Horses of Race, 118 HARV. L. REV. 1489, 1525-28 (2005) (reviewing shooter studies and noting that the race of the participants had no effect on shooter bias). 199. Kang, supra note 198, at 1525-28. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 37 the experimenter is white, despite the fact that the penalty is not being directed at the experimenter. We can interpret these findings two ways: The subject is angrier at being frustrated by the Black experimenter, or the subject who was frustrated by a Black experimenter is imagining that the person he is penalizing is also Black.200 Either way, frustration with a Black person leads to greater punitiveness. Additional evidence bearing on the frequency of racially skewed cognitive processing has been amassed by neuropsychologists, who look at the activation of various brain areas. When annoyance or anger is triggered in many white people, its activation is both greater and lasts longer when the source of the annoyance is Black than when he is white.201 Similarly, the activation of fear lasts longer when the source is Black.202 When white people attempt to identify another white person, the fusiform region of the brain is activated, but for many white people, that area of the brain is not activated at all when attempting to identify a person of another race; one might even say that the brain is not reacting to an other-race face as human. Subjects with high implicit racial bias register higher levels of emotion when viewing unfamiliar African-American faces than when viewing unfamiliar white faces,203 and when faced with a disagreement with a person of another race, the area of the brain that attempts to mediate conflict is less active than when faced with a disagreement with a person of the same race.204 What is particularly noteworthy about the results of brain activation studies is that differential neurological responses based on race, like implicit associations, are neither very easily detected, nor very amenable to change. B. Untangling Mixed Motives Thus, race influences many decisions in many ways for many people, sometimes consciously but often subconsciously, and it is more likely to do so in settings where stereotypes are triggered. Can a judge reliably determine whether a prosecutor whose strike of a Black juror was “substantially” motivated by race would have, absent racial motivation, struck that juror? Here too, empirical data developed since Arlington Heights is illuminating. 200. Mark Chen & John H. Bargh, Nonconscious Behavioral Confirmation Processes: The Self- Fulfilling Consequences of Automatic Stereotype Activation, 33 J. EXPERIMENTAL SOC. PSYCH. 541, 547 (1997). 201. William A. Cunningham et al., Separable Neural Components in the Processing of Black and White Faces, 15 PSYCH. SCI. 806, 809 (2004). 202. Andreas Olsson et al., The Role of Social Groups in Persistence of Learned Fear, 309 SCI. 785, 785-86 (2005). 203. See generally John F. Dovidio et al., On the Nature of Prejudice: Automatic and Controlled Processes, 33 J. EXPERIMENTAL SOC. PSYCH. 510, 511-18 (1997) (demonstrating that implicit negative racial attitudes among whites may be unconscious and automatic); Elizabeth A. Phelps et al., Performance on Indirect Measures of Race Evaluation Predicts Amygdala Activation, 12 J. COG. NEUROSCI. 729, 730-33 (2000). 204. Cunningham et al., supra note 201, at 811. 38 INDIANA LAW REVIEW [Vol. 57:7 The cognitive psychology findings cited below are relevant for their contribution to the evidence that substantial racial discrimination is occurring, but they are even more probative of the question of whether the Footnote 21 inquiry is feasible. One important cognitive processing distortion is the halo effect.205 If I love you or think highly of you, I am less likely to notice negative things about you and more likely to notice good things. On the other hand, if I dislike you or think you are stupid, the opposite is true. Thus, if two people hear the same competent-but-not exceptional presentation, the one who admires the presenter will focus on the better parts, while the one who does not is likely to notice and remember the weaker aspects. Or to take a more pertinent example, when a prosecutor has negative beliefs about a juror based on her race, he is likely to interpret other information—if it is at all ambiguous—in a much more negative way than if he were interpreting information about a person of his own race. Even if the prosecutor honestly says that he believes the fact that the juror is on disability is an independent and sufficient reason for striking him, it may be that his view of the juror’s disability is an example of the halo effect. A related phenomenon that also distorts cognition is attribution bias. When we observe a person doing something bad, do we attribute that action to his character or to the situation? Conversely, if we observe a good deed, do we give the actor credit for being a good person, or do we conclude that circumstances caused the good behavior? In general, people tend to attribute their own bad behavior to surrounding circumstances, and their own good behavior to their moral worthiness.105 But, when others are concerned, people are more likely to do the reverse: attribute good behavior to the circumstances and bad behavior to character. Race exacerbates this tendency.206 Thus, if a juror is determining whether to sentence a defendant to death, when the defendant is a person of another race the juror is more likely to attribute all of the defendant’s bad acts to his character (rather than the situation, his mental illness, or his upbringing) and to rationalize that any good acts he has done were due to his fear that he would be punished for bad behavior. Similarly, a prosecutor whose strike was “substantially” motivated by race might cite, as an alternative and sufficient reason for the strike, the juror’s lateness as evidence of disrespect for the criminal process. When she does so, however, she may be sincere, yet inaccurate; because the juror is Black, the prosecutor was more likely to infer that he was late because he was lazy rather than because he was caught in traffic, and more likely to infer that he was polite only because he was intimidated by the judge rather than because he is respectful of the judicial system. A third relevant concept from cognitive psychology is schema accessibility. Schemas are cognitive frameworks or concepts that help organize and interpret information in our environment, but often more than one schema could be applied to the information in front of us; which one we employ at a particular time depends on salience and priming, which makes particular schema more or less 205. See GILOVICH ET AL., supra note 198. 206. SUSAN T. FISKE & SHELLEY E. TAYLOR, SOCIAL COGNITION 80-81 (McGraw-Hill 1991). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 39 accessible.207 To take a mundane example, when one of my students sees me, she might be interpreting information about me through schemas relating to professors, to capital defense lawyers, to mothers, to tall people, and so on. Which schema will be accessible to her depends in part on priming; if she sees me in the grocery store with teenagers in tow, she is more likely to think about the mother schema than the capital defense lawyer schema, but if she sees a photo of me in front of the Supreme Court, she is more likely to think about capital defense lawyers. However, salience also matters; if she sees me in a group of colleagues, the law professor schema might be triggered, but if all of those colleagues are male, my gender may stand out, and therefore trigger a woman or mother schema. More pertinently, in criminal trial settings, both salience and priming are likely to trigger racial schema. Jurors are more likely to have racial schemas about the defendant triggered if the defendant is the only person in the courtroom (or the only person at either counsel table) who is a person of color, and also more likely to have racial schema triggered because the information about the crime is likely to make accessible schema associating race and criminality. And for prosecutors, racial schema relating to potential jurors are likely to be salient when few of the potential jurors are people of color, and more likely to be primed by the race of the defendant and/or information about the crime the prosecutor is discussing with the potential juror. Thus, throughout the voir dire process, when organizing information a prosecutor has about jurors, racial schema are likely to play a significant role. Taken together, halo effects, attribution bias, and schema accessibility all operate to make a prosecutor who is influenced by a juror’s race believe that she has race-neutral reasons for her strike when in fact those reasons are dependent on race. Moreover, two other kinds of cognitive processing distortions make her protest–even if she is not dissembling–that she would have struck the juror absent racial motivation. The first is belief persistence.208 After an individual forms a belief, she becomes reluctant to revise it, even in the face of contrary evidence. If a juror thinks that Black people are dumb, and that’s why so many fail to graduate from high school, evidence that a defendant’s school was chaotic and dangerous is unlikely to make that juror think that the school is responsible for the defendant’s failure to graduate. If a prosecutor thinks that Black people stick together, he is unlikely to believe a Black juror who says the race of the defendant and victim won’t affect her decision-making, and quite likely will “see” an untrustworthy demeanor when the juror states her impartiality. More generally, if a prosecutor forms the belief–based on race–that a juror is undesirable for the prosecution, and then is told that his reason for striking a juror is impermissible, he is unlikely to change his mind about the juror’s desirability. The second and related phenomenon is confirmation bias.209 Suppose you formed a strong belief that Donald Trump is going to be the Republican candidate 207. GILOVICH ET AL., supra note 198, at 408; FISKE & TAYLOR, supra note 206. 208. FISKE & TAYLOR, supra note 206, at 150-51. 209. ZIVA KUNDA, SOCIAL COGNITION: MAKING SENSE OF PEOPLE 159 (MIT Press 1999). 40 INDIANA LAW REVIEW [Vol. 57:7 for President again in 2024. Whether you formed that belief because you hope to see him President, or because you dread that possibility, if you see a newscast of him giving a speech to which the audience responded positively, you will remember that speech, remember the audience’s reaction as more enthusiastic than it was, and emphasize the significance of that audience, all because it confirms your prior belief. In contrast, should you see a speech where audience applause for Trump is lukewarm, you may forget that you saw it, remember the audience as less inattentive than it was, or dismiss that audience as less important than audiences whose reactions were enthusiastic. With race, people who consciously or unconsciously hold negative stereotypes about racial minorities attend more to stereotype-consistent information, remember new information that supports their stereotypes more than information that undermines them, remember the stereotype-consistent information in an exaggerated form and stereotype-inconsistent information in a diminished form if they remember it all, and emphasize the importance of stereotype-consistent information.210 Thus, confirmation bias may lead a juror, whether or not he consciously thinks about the defendant “he’s Black and less than human,” to hear that he beat his girlfriend, remember that testimony when deciding whether to sentence him to death, and deem it important in assessing his character, while forgetting testimony that the defendant himself was raped as a child, and rejecting as unimportant testimony that he rescued a prison guard who was under attack by a psychotic inmate. Likewise, it will lead the prosecutor who believes Black jurors are bad for the prosecution to selectively remember and emphasize facts about a particular Black juror that are consistent with his prior belief while forgetting or downplaying facts that would seem to make that juror attractive to the prosecution. Although a prosecutor could be made aware of these phenomena (and should be), there is no way she could measure their impact on her own decision-making, assuming she wanted to do so; even psychologists don’t know how to do that. And there is little reason to think that a prosecutor who was substantially motivated by race in the exercise of her peremptory challenge would even want to do so. D. The Cost of “Windfalls” Both the frequency of biased decision-making and the difficulty of establishing purposeful discrimination support an inference that removing the counterfactual Footnote 21 inquiry will produce very few “windfalls” to the criminal defendant who has actually managed to prove purposeful discrimination. Beyond that, the frequency with which racially biased motives are inextricably entwined with permissible motives suggests that the Footnote 21 inquiry cannot reliably ascertain when racial motivation was determinative, and therefore will not accurately separate appropriate compensation for harm from windfalls. Moreover, even in cases in which it would be possible to ascertain that the racial 210. Id. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 41 motivation was not a but-for cause of the strike, it is important to note that in the Batson context, the size of any “windfall” is small. If, per Flowers, the trial judge correctly determines the existence of racial motivation, and disallows the racially motivated strike without performing a Footnote 21 inquiry, even if such an inquiry would have correctly determined that the racial motivation was not determinative, the “windfall” cost is very small; because challenges for cause have already weeded out jurors who are demonstrably partial or otherwise unqualified, the only loss to the prosecution is a juror who might be less favorably disposed toward the prosecution’s case. Or, if the trial judge fails to perform the “sensitive inquiry” required under Batsons’ third step, and an appellate court, doing so, concludes that racial motivation has been established, and then, per Flowers, reverses the conviction without performing a Footnote 21 inquiry, it might well have done so even if its but-for inquiry were theoretically available; as Snyder illustrates, an appellate court often would be unable to perform such an inquiry based on the record before it, and remand for a retrospective hearing on but-for causation might be unproductive years after the strike was made. And finally, at the very most, the only “windfall” the defendant gets is a new trial (during which he will doubtless be incarcerated), not an acquittal. More broadly, although windfall arguments have much less persuasive power with respect to Batson claims than they do with respect to employment discrimination claims, wrongful discharge (or failure to hire) is wrong because it deprives the individual of employment, and if it did not actually deprive him of employment, it needs no remedy. In contrast, while purposeful exclusion of members of a racial group from the jury deprives a particular prospective juror of the opportunity to serve only if he would otherwise have been selected, racially motivated exclusion also threatens a central tenet of our criminal justice system and compromises the integrity of the trial process, thereby harming the defendant and the society regardless of the final composition of the jury.211 But even for those who do not subscribe to the view that no “windfall” is ever involved in the jury selection context, it should be apparent that the size of any windfall in the Batson context is much smaller than in the employment context, where failure to consider whether racial motivation was determinative would sometimes mean awarding tenure, a job, or backpay to someone who is unqualified. It is important to remember that Mt. Healthy—the progenitor of Footnote 21—expressed concern that absent an opportunity to rebut but-for causation, the school district might be saddled with a terrible teacher for life;212 those fears are inapplicable in 211. See State v. Lucas 18 P.3d 160, 163 (Ariz. App. 2001) (“any consideration of a discriminatory factor directly conflicts with the purpose of Batson and taints the entire jury selection process”); Payton v. Kearse, 495 S.E.2d 205, 210 (S.C. 1998) (“Once a discriminatory reason has been uncovered—either inherent or pretextual—this reason taints the entire jury selection procedure”). 212. Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 286 (1977). (“The long- term consequences of an award of tenure are of great moment both to the employee and to the employer. They are too significant for us to hold that the Board in this case would be precluded, 42 INDIANA LAW REVIEW [Vol. 57:7 the Batson context, and virtually every criminal justice context. IV. IN PRAISE OF FLOWERS As the introduction made explicit, I come to bury Arlington Heights, not to praise it, and not surprisingly, I really come to praise Flowers. Part III defended Flowers by asserting that what Footnote 21 asks is rarely possible, at least in the context of racially motivated decisions in the criminal justice system, and that getting rid of it is not very costly. This last section speaks to those who like stories better than they like data and praises Flowers for the cases in which, if it takes root, Flowers will permit the avoidance of terrible costs created by Footnote 21. I begin with a Batson case but will consider other criminal justice cases as well, for I think the principle is the same: “All” that need be decided is that a decision was “substantially” motivated by race. A. Curtis Flowers It is hard for me to know what to say about Curtis Flowers. Given the two state court briefs, two petitions for certiorari, the Supreme Court brief and reply on merits filed by my co-counsel and I, the problem is not a lack of words, but saying something that does more than rehash the litigation. “The only plausible interpretation of all of the evidence viewed cumulatively is that Doug Evans began jury selection in Flowers’ [sixth trial] with an unconstitutional end in mind: To seat as few African American jurors as he could.”213 Seven members of the Supreme Court agreed with that assertion, convinced by the 41 African American jurors Evans struck over the course of six trials, his multiple misrepresentations of the record, his disparate questioning and investigation of Black jurors, and comparisons of struck Black jurors to white jurors Evans was willing to accept.214 Thus, his “substantial” racial motivation in the exercise of his peremptory challenges has been conclusively determined by the Supreme Court. But Evans’ expression of racial animus towards Curtis Flowers went beyond the African-American jurors he struck. He pursued a case against Flowers that any objective prosecutor would have rejected, one that lacked a plausible motive, probative physical evidence, and a coherent timeline.215 He recruited three lying informants to claim that Flowers had confessed to them; when the first recanted, he recruited a second, and when the second recanted, he recruited a third (who also ultimately recanted).216 He dismissed evidence pointing toward a more likely because it considered constitutionally protected conduct in deciding not to rehire Doyle, from attempting to prove to a trier of fact that quite apart from such conduct Doyle’s record was such that he would not have been rehired in any event.”). 213. Transcript of Oral Arguments at *3, Flowers v. Mississippi 139 S. Ct. 2228 (2019) (No. 17-9572). 214. Flowers v. Mississippi, 139 S. Ct. 2228, 2235 (2019). 215. See Brief for the Petitioner, Flowers v. Mississippi, 139 S. Ct. U.S. 2228 (2019) (No. 17- 9572). 216. Id. at *61 n.3. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 43 perpetrator and concealed it from the defense, along with evidence impeaching his star witness.217 He violated established Mississippi law in an attempt to obtain multiple opportunities to prosecute Flowers for the same offense, probably as insurance against the risk that even an all-white jury might not convict on the weak evidence he was presenting.218 He lied about several crucial facts in closing argument to cover the holes in his case, and when reversed for doing so, at a subsequent trial repeated those lies dressed in slightly different clothes.2 1 9 He attempted to prosecute two African-American jurors for hanging the jury in Flowers’ fifth trial, a move likely to induce fear of jury service and fear of the consequences of voting to acquit in prospective African-American jurors.220 After the Supreme Court decision reversing Flowers’s conviction, Evans agreed to recuse himself, and the Mississippi Attorney General’s Office agreed to take over the prosecution. Fortuitously, that office had just changed parties. New (Republican) Attorney General Lynn Fitch reviewed the case and only a little more than a year after the State of Mississippi was defending its right to execute Curtis Flowers, she moved to dismiss with prejudice all charges against him.2 2 1 And less than two years after the Supreme Court decision, Flowers was awarded $500,000 by the State of Mississippi for his wrongful conviction and twenty-some years on death row.222 Upon learning of this award, Flowers issued a statement that bespeaks his character, one that focuses not on the wrong done him, but on the support he received: Today, I am finally free from the injustice that left me locked in a box for nearly twenty-three years . . . I’ve been asked if I ever thought this day would come. I have been blessed with a family that never gave up on me 217. Id. at *3-14. 218. Id. 219. Id. 220. Brief of Appellant at *122-23, Flowers v. Mississippi, 139 S. Ct. 2228, No. 2010-DP-1348- SCT (2013) (These facts are all documented with record cites in the direct appeal Brief of Appellant, with the exception of the recantation of the third informant, which occurred several years after the sixth trial.); see David Mann, Curtis Flowers’ Prosecution Witness Says Testimony Was “All Make Believe,” CLARION LEDGER, June 10, 2018, https://www.clarionledger.com/story/news/2018/06/10/ curtis-flowers-case-prosecution-witness-says-he-lied/685575002/) [https://perma.cc/8C6J-5MTW]. 221. After Twenty-three Years, Mississippi Drops All Charges Against Curtis Flowers, ABA DEATH PENALTY REPRESENTATION PROJECT (Oct. 28, 2020), https://www.americanbar.org/groups/ committees/death_penalty_representation/project_press/2020/fall-2020/curtis-flowers-all-charges- dropped/ [https://perma.cc/8C6J-5MTW]. 222. Curtis Flowers to be Awarded $500,000 by Mississippi After Spending 23 Years on Death Row, Enduring Six Trials for Crime He Didn’t Commit, ATLANTA BLACK STAR (Mar. 5, 2021), https://atlantablackstar.com/2021/03/05/curtis-flowers-to-be-awarded-500000-by-mississippi-after- s p e n d in g -2 3 -ye a r s-on -dea t h -r o w -e n d u r in g -s ix -t r ia ls -f o r -c r im e -h e -d id n t -c o m m it / [https://perma.cc/6EAN-TJQD]. 44 INDIANA LAW REVIEW [Vol. 57:7 and with them by my side, I knew it would.223 The point of recounting Evans’ misconduct is not to argue that most prosecutors who violate Batson have racial hatred in their hearts. Most do not. The point is that racial motivation—whether prompted by animosity or stereotypes of what African Americans are like—is not expressed in just one act but will be expressed when opportunity permits it to be expressed. If a court finds that a prosecutor was motivated by his belief that Black jurors are bad jurors, it cannot trust that prosecutor’s assertion that he would have struck a Black juror absent racial motivation, because that prosecutor’s racial motivation did not disappear when the judge announced that the defendant had established it. “Dual motivation” analysis, which allows a prosecutor to resuscitate a racially motivated challenge by supplying another motivation for that challenge, assumes that the racial motivation will not influence his statement of an alternative motivation; there is no basis for such an assumption. The point in tracing Curtis Flowers’ post-Supreme Court exoneration is not just that he is innocent. It is that the Attorney General’s Office had a stake in seeing him as guilty, and a stake in seeing Evans as innocent, despite compelling evidence to the contrary on both issues. Once it was committed to upholding his conviction and seeking his execution, it was blinded to evidence that readily convinced the new Attorney General to dismiss the charges and further convinced a Mississippi state court judge to award him the highest compensation for wrongful conviction allowed under state law. Attribution bias, confirmation bias, and halo effects ride again. If the entire office could be so blinded by its past commitment to Flowers’ guilt and Evans’ innocence, we should not imagine that a prosecutor committed to striking a Black juror because she is Black will be able to put that motivation aside and neutrally assess whether there are also race- neutral reasons that would have led to the strike. And the point in quoting Curtis Flowers? The point is that he is a gentle, innocent man, and despite the fact that the purpose of this Article is not to establish his innocence, my own motivation to do so persists, and I use the opportunity to advance that aim. I do not talk about Flowers’ innocence when I speak to the grocery store clerk or my teenager’s high school counselor, but when I am thinking about Batson, I think about Curtis Flowers, and when I think about Curtis Flowers, I want everyone to be convinced of the wrongfulness of his conviction. Motivation—for prosecutors as well as everyone else—is often like that. B. Johnny Bennett Johnny Bennett is an African American man who was accused of a brutal capital crime in South Carolina.2 2 4 His prosecutor did three things that I think reflect racial motivation. First, at the sentencing phase he elicited the testimony 223. Id. 224. Bennet v. Stirling, 842 F.3d 319 (4th Cir. 2016). 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 45 of the victim of another crime Bennett committed.225 Presenting testimony of other “aggravating” offenses is both permissible and expected. Even soliciting testimony concerning how that offense traumatized the victim is not unusual, but what this victim said was quite unusual. He told the jury about a dream in which “Black Indians” were chasing him.226 Since Bennett is not indigenous, it is hard to see what relevance that dream had to Bennett, and even harder to see what relevance it had in determining whether he should be sentenced to death. The second thing the prosecutor did that had racial resonance was to inquire about the woman Johnny Bennett was dating, referring to her not by name, but as “the blond-headed lady,”227 and then referring to his sexual relationship with her no less than seven times during closing argument. And the third thing he did was to refer to Johnny Bennett as “King Kong,”228 and use other forms of subhuman imagery to describe him. Moreover, Bennett was tried twice, and in the first trial, where some of the jurors were African American, the prosecutor did not ask the first witness about his “Black Indian” dream, did not ask the second witness about the “blonde-headed lady,” and did not refer to Bennett as “King Kong;” he only made these racially loaded remarks in the second trial, when his audience was all- white.229 Defense counsel objected to these comments, and those objections were repeated on appeal.2 3 0 Nonetheless, the South Carolina Supreme Court affirmed Bennett’s conviction and death sentence. It began by acknowledging that the remarks had racial connotations, but held that such connotations are not dispositive: As a starting point, we recognize that the terms “blond[e] lady” and “King Kong” could have racial connotations. However, this court’s jurisprudence does not prohibit the use of terms with racial meanings, nor does our case law stand for the proposition that arguments or evidence in a case must be void of racial allusions. Instead, this court has recognized that it is impermissible to use race to “inflame the passions or prejudices” of the jury. Accordingly, our inquiry focuses on how these terms were used in Appellant’s case.231 Then the Court rationalized that the “blond[e] lady” remark “was not made to inflame the passions or prejudices of the jury:” Because “[i]f the State had sought to make the race of Appellant’s former lover an issue in this case, we believe they would have elicited evidence to this effect while examining their witness who 225. Id. at 326. 226. Id. (“Indians were chasing me trying to kill me, and the thing that I thought was they were black . . . [T]here might have been a link. You know, that I was remembering something about trying to get away from someone.”). 227. Id. 228. Id. at 321. 229. Id. at 327. 230. State v. Bennett, 632 S.E.2d 281, 285 (S.C. 2006). 231. Id. at 288. 46 INDIANA LAW REVIEW [Vol. 57:7 testified extensively about the affair” instead of occurring during the prosecutor’s cross-examination of a defense witness.232 The court’s explanation of why the use of “King Kong” was not an appeal to the prejudice of the jury is more remarkable: The comment referred to Appellant’s immense size, strength, and the destructiveness of his previous crimes. In this case, the trial court properly determined that Appellant’s size and strength were probative of the aggravating circumstance of physical torture, which the court charged to the jury. In this regard, the Solicitor’s use of the term “King Kong” was not suggestive of a giant black gorilla who abducts a white woman, but rather, descriptive of Appellant’s size and strength as they related to his past crimes.233 It is true that the prosecutor could have brought out the race of Bennett’s girlfriend with a different witness. It is true that Johnny Bennett is a big, strong man, and that the crime of which he was accused was brutal. But by asking only whether the racial allusion fits some facts that are race-neutral, the South Carolina Supreme Court missed the point. The right question is, why did he say anything about a blonde girlfriend and why did he pick “King Kong” as the descriptor? Eventually, that is the question the Fourth Circuit asked. It affirmed the district court’s grant of the writ of habeas corpus after finding the state court’s determination that the remarks were not racially motivated to be unreasonable: The prosecutor’s comments were poorly disguised appeals to racial prejudice. It is impossible to divorce the prosecutor’s “King Kong” remark, “caveman” label, and other descriptions of a black capital defendant from their odious historical context. And in context, the prosecutor’s comments mined a vein of historical prejudice against African Americans, who have been appallingly disparaged as primates or members of a subhuman species in some lesser state of evolution.234 Considering all the racial allusions together, as well as the fact that the prosecutor only made them when the jury was entirely white, the Fourth Circuit concluded: “Race was a recurrent theme throughout the capital sentencing proceeding, a theme designed to implant both racial fears and prejudices in the mind of the jury by playing upon ancient staples of racial disparagement and discrimination.”235 Once a court sees that there was a racial allusion, and from the surrounding facts, that a choice was made to employ that allusion, that ought to be the end of the matter. “All” that a court then needs to decide is that the remark was “substantially” motivated by race, which is “all” the Fourth Circuit did.236 232. Id. 233. Id. 234. Bennett v. Stirling, 842 F.3d 319, 327 (4th Cir. 2016). 235. Id. 236. Id. (This condemnation of racial motivation did not occur until after round two in the state courts, which was another sort of dual motivation case, though this one focused on a juror’s motivation. Post-conviction counsel interviewed jurors who were willing to talk, and one cooperative juror when asked why he thought Johnny Bennett committed the murder, responded: “Because he 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 47 D. Curtis Osborne The next example is Curtis Osborne.237 It involves racial bias on the part of defense counsel and does not, in my judgment, end so well as Johnny Bennett’s story. Osborne killed two people in a bizarre incident in rural Georgia that was completely unlike anything he had ever done previously. Osborne had been a hard-working, devoted young father, when out of the blue he shot the driver and passenger in a car in which he was riding, running the car off the road while Osborne was in it.238 His capital case was assigned to Johnny Mostiler, a lawyer with an enormous caseload.239 Mostiler never had Osborne evaluated by a mental health professional and never presented any evidence about what caused the drastic change in Osborne’s behavior, though he did put on evidence about how Osborne had been a hard worker since the age of 13, and supported his mother, the sort of lay testimony that requires little preparation.240 Osborne was sentenced to death.241 State post-conviction counsel did some investigation and discovered both a short-term explanation for the crime—cocaine hallucinosis242—and a longer-term mental health mitigation story that explains why Osborne had turned to cocaine at the time of the murders.243 Osborne also claimed that trial counsel never conveyed to him the prosecutor’s offer to plea bargain the case for a life sentence.244 Trial counsel, Mostiler, defended his failure to hire any mental health expert as due to the lack of any reason to suspect drug use or mental illness and further claimed that he had informed Osborne of the plea offer, who rejected it.245 The state court deemed trial counsel’s decision not to seek expert evaluation reasonable and credited his claim that he had conveyed the plea offer to Osborne; was just a dumb n*****.” The lawyers called this juror as a witness at a postconviction relief hearing, fully expecting that he would deny what he said. Instead, he admitted, “I did say it.” That, following Flowers, should have been the end of it; again, “all” that needs to be established is substantial racial motivation, and the juror’s remark clarifying remark “[t]hat’s just the way I think of people who are in that category, who really don’t do anything” seems to cement the existence of racial motivation. But then the prosecutor asked, “[s]o when you decided to sentence him to death, that didn’t influence you, right?” and the juror responded “[r]ight.” After hearing this testimony the post-conviction court denied relief on the grounds that the juror was not racially biased at the time of the sentencing, and the state supreme court denied certiorari.). 237. Osborne v. Terry, 466 F.3d 1298 (11th Cir. 2006), cert. denied, 552 U.S. 841 (2007). 238. Id. at 1302-03. 239. Id. at 1315. 240. Id. at 1311-13. 241. Osborne v. State, 430 S.E.2d 576, 577 (Ga. 1993). 242. Osborne, 466 F.3d at 1311. 243. Id. at 1311-13. 244. Id. at 1316. 245. Id. at 1311-12. 48 INDIANA LAW REVIEW [Vol. 57:7 it therefore denied Osborne’s ineffective assistance of counsel claims.246 Then as the case reached federal court, Osborne’s lawyer was told by another one of his clients, Gerald Huey, that he knew something he must reveal: That Johnny Mostiler said to him more than once, referring to Osborne, “That little n***** deserves the chair.”247 Huey is white, and was represented by Johnny Mostiler at the same time as Osborne was. Huey further swore: I recall Mr. Mostiller (sic) telling me that I wouldn’t believe the amount of money he was going to spend on my case. He said he was going to hire a private investigator and get expert witnesses. He said the money he would spend on m[e] was going to be a lot more than he would spend on Mr. Osborne because “that little n[*****] deserves the chair.”248 By the time Huey made this statement, Mostiler had died, but the State never alleged that Huey was lying, or was influenced by affection for Osborne, with whom he had previously had unpleasant interactions.249 Indeed, there was never any dispute that Mostiler had in fact used this epithet to refer to Osborne, and as it turned out, a published opinion in another case reported the allegation of another Black client that Mostiler had addressed him with the same slur, an allegation to which Mostiler responded by saying he did not “recall” whether he had used the slur. I was asked to argue this case in the Eleventh Circuit after habeas was denied in federal district court and did so naively believing I would convince the panel that the establishment of racial motivation should end the matter. But the panel concluded otherwise: Even if the affidavit correctly recounts Mostiler’s statements to Huey, it does not establish that Mostiler failed to convey the plea offer to Osborne. Moreover, Osborne presents no other evidence to support his claim that Mostiler’s alleged racial animosity affected his representation.250 After certiorari was denied, a big push for clemency began. Former President Jimmy Carter sent a letter to the clemency board members asking them to grant clemency,251 as did a United States Attorney General who grew up in Georgia.252 A former justice of the Georgia Supreme Court appeared before the clemency board and asked the board to commute Osborne’s sentence.253 But the board did 246. Id. at 1314-15. 247. Id. at 1316. 248. Id. 249. Id. 250. Id. at 1318. 251. See David Von Drehle, If Your Lawyer Wants You Executed, TIME MAG. (June 2, 2008), https://content.time.com/time/nation/article/0,8599,1811174,00.html [https://perma.cc/9UYA- MV2F]. 252. Id. 253. See Greg Bluestein, Georgia Executes Man Who Killed 2 in 1990, ACCESSWDUN (June 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 49 not commute his sentence, and the only explanation it offered for denial was that the crime was horrible.254 Osborne was executed.255 How could the Eleventh Circuit trust that a lawyer who expressed extreme racially-motivated disregard for Osborne’s welfare was truthful in his claim that he conveyed a life offer to Osborne? Especially a lawyer who said that he thought Osborne did not “deserve” a life sentence. How could the Eleventh Circuit say that Osborne presented no other evidence to support the claim that racial animosity “affected Mostiler’s representation,” when he stated to Huey that “the money he would spend on [Huey] was going to be a lot more than he would spend on Mr. Osborne because ‘that little n[*****] deserves the chair?’”256 The Eleventh Circuit acknowledged defense counsel’s racial motivation yet credited his assertions that other motivations determined his actions.257 As the Price Waterhouse plurality observed in the employment context, any “suggestion . . . that the [discriminator’s] own testimony as to the probable decision in the absence of discrimination is due special credence where the court has, contrary to the [discriminator’s] testimony, found that an illegitimate factor played a part in the decision . . .” is counterintuitive.258 And appalling. E. Theodore Kelly The next story I want to tell, that of Theodore Kelly,259 is one in which the deciding court explicitly rejected counterfactual speculation. McCleskey v. Kemp held that statistical showings of racial discrimination in the administration of the death penalty standing alone did not establish racial motivation and therefore could not ground relief.260 Kelly’s is the only post-McCleskey claim of racially discriminatory administration of the death penalty that has prevailed. This is not to say that few studies were done, or that the studies failed to find evidence of racial motivation; on the contrary, there have been many studies, and as the Government Accounting Office’s metanalysis concluded, overwhelmingly those studies do find evidence of racial discrimination.261 One such study, done in Spartanburg, South Carolina, formed the basis of a selective prosecution claim 4, 2008), https://accesswdun.com/article/2008/6/210545 [https://perma.cc/8FR2-LBLB] (Norman Fletcher, the ex-chief justice of the Georgia Supreme Court, showed up in person Monday to ask for leniency. It was the first time he had made such an appearance, but he said he was drawn in by the ``extraordinary nature’‘ of the case.”). 254. Id. 255. Id. 256. Osborne v. Terry, 466 F.3d 1298, 1316 (11th Cir. 2006), cert. denied, 552 U.S. 841 (2007). 257. Id. at 1318. 258. Price Waterhouse v. Hopkins, 490 U.S. 228, 295 n.14 (1989). 259. State v. Kelly, 502 S.E.2d 99 (S.C. 1998). 260. McCleskey v. Kemp, 481 U.S. 279 (1987). 261. U.S. Gov’t Accountability Off., Death Penalty Sentencing: Research Indicates Pattern of Racial Disparities, (1990), https://www.gao.gov/products/t-ggd-90-37 [https://perma.cc/ZF4M- 5WMA]. 50 INDIANA LAW REVIEW [Vol. 57:7 that lost in state court despite the fact that solicitor of Spartanburg had never sought death in a black victim case and had sought death in about half of the death-eligible white victim cases.262 After the study had been presented in court in another case, the Spartanburg solicitor sought and obtained a death sentence for Theodore Kelly.263 Kelly’s victim was Black and in the course of post-conviction hearings an assistant solicitor admitted: I told [Solicitor Gossett] that I felt like the black community would be upset . . . if we did not seek the death penalty because there were two black victims in this case . . . The only mention that was ever made of race was when I said that I felt like if we did not seek the death penalty . . . the black community would be upset because we are seeking the death penalty in the (Andre) Rosemond case for the murder of two white people.264 The court, granting relief, dismissed the State’s argument that it would have sought death absent the racial motivation: While the witness adamantly denies that race was the reason the State sought the death penalty in this case, his own words admit that race was, in fact, a consideration, and thus a factor in the decision to seek the death penalty. Race can never be a factor, in any degree whatsoever, in the decision to seek death.265 After ruling that the consideration of race could never be harmless, the Kelly court observed that the State’s claim, that the decision to seek death was not solely motivated by race, was in fact a false one, belied by the timing of that decision.266 The State’s Footnote 21 argument may have been disingenuous, or it may have been spawned by some combination of the cognitive distortions discussed in Part III. Either explanation, however, underlines the cost of Footnote 21-type inquiries, for in most cases, objective facts that disprove the state’s claim of mixed motives will not be available. 262. See Sheri Lynn Johnson, Litigating for Racial Fairness after McCleskey v. Kemp, 39 COLUM. HUM. RTS. L. REV. 178, 181-82 (2007) (noting a death sentence study showing that during Solicitor Holman Gossett’s time as a prosecutor in Spartanburg, he sought death in 50% of the 52 death-eligible white victim cases and in 0% of the 19 death-eligible Black victim cases). 263. State v. Kelly, 502 S.E.2d 99 (S.C. 1998). 264. Kelly v. Catoe, Case No. 99 CP-42-1174 at *38 (S.C. 7th Cir. 2003) (unpublished order granting relief on file with author). 265. Id. at 38; see also John H. Blume, Sheri Johnson, Emily C. Paavola & Keir M. Weyble, When Lightning Strikes Back: South Carolina’s Return to the Unconstitutional Standardless Capital Sentencing Regime of the Pre-Furman Era, 4 CHARLESTON L. REV. 479, 516 (2010). 266. Kelly, No. 99 CP-42-1174 at *38-41. 2023] FLOWERS FOR THE ARLINGTON HEIGHTS FOOTNOTE 51 CONCLUSION Justice Marshall, writing before any of the social science evidence reported in Part III was published, was right that permitting a dual motivation defense to a racially motivated strike “undermines an already under-protective means of safeguarding the integrity of the criminal jury selection process” undermine the slim protection Batson affords.267 Flowers, albeit belatedly, was right to reject the application of Footnote 21 to Batson cases. The lower courts are right to read the “All that we need to decide” sentence as sufficient reason to abandon mixed- motive analysis.268 And now? I am hopeful that Flowers’ rejection of Footnote 21 may spread to other criminal procedure claims; the reasons for abandoning its counterfactual, unreliable inquiry apply wherever stereotypes and implicit bias hold sway, which is, as I have spent my career exploring, virtually all criminal justice decisions. But for just this paragraph I want to go beyond the cases and settings I know. I have to notice that in employment cases, too, powerful stereotypes are omnipresent. Suppose an employer has said: “She was tardy, just like a lot of Mexican people are,” and upon being asked about that comment argues “Well, I didn’t fire her because she’s Mexican, it’s because she was tardy.” I am doubtful that a factfinder can judge whether that is true, in part because there is little reason to think the employer himself knows whether it is true. Or consider the issue in Arlington Heights itself: If racial motivation had been established and mixed with other motivations about poor families and multiple-family dwellings, those motivations, I think, would have been inextricably entwined due to associations between race and poverty. Further, I suspect it is not just jury selection cases, or capital cases, or criminal justice cases, or employment cases, or housing cases. Isn’t any setting that triggers actionable racial motivation also likely to be a setting where racial stereotypes and biases—about how people of color act, what they believe, or what they are worth—are both prevalent and deeply, inextricably embedded? The Arlington Heights footnote is wrong. Ashes to ashes, dust to dust. 267. Wilkerson v. Texas, 493 U.S. 924, 928 (1989) (Marshall, J., dissenting from the denial of certiorari). 268. Flowers v. Mississippi, 139 S. Ct. 2228, 2251 (2019).