Chapman v. California/Dissent Harlan
Mr. Justice HARLAN, dissenting.
The Court today holds that the harmlessness of a trial error in a state criminal prosecution, such error resulting from the allowance of prosecutorial comment barred by the Fourteenth Amendment, must be determined under a 'necessary rule' of federal law. The Court imposes a revised version of the standard utilized in Fahy v. State of Connecticut, 375 U.S. 85, 84 S.Ct. 229, on state appellate courts, not because the Constitution requires that particular standard, but because the Court prefers it.
My understanding of our federal system, and my view of the rationale and function of harmless-error rules and their status under the Fourteenth Amendment, lead me to a very different conclusion. I would hold that a state appellate court's reasonable application of a constitutionally proper state harmless-error rule to sustain a state conviction constitutes an independent and adequate state ground of judgment. Believing this to be the situation here, I would dismiss the writ. Viator v. Stone, 336 U.S. 948, 69 S.Ct. 882, 93 L.Ed. 1104.
The key to the Court's opinion can, I think, be found in its statement that it cannot 'leave to the States the formulation of the authoritative laws, rules, and remedies designed to protect people from infractions by the States of federally guaranteed rights,' and that 'in the absence of appropriate congressional action 'the Court must fashion protective rules. The harmless-error rule now established flows from what is seemingly regarded as a power inherent in the Court's constitutional responsibilities rather than from the Constitution itself. The Court appears to acknowledge that other harmless-error formulations would be constitutionally permissible. It certainly indicates that Congress, for example, could impose a different formulation. 
I regard the Court's assumption of what amounts to a general supervisory power over the trial of federal constitutional issues in state courts as a startling constitutional development that is wholly out of keeping with our federal system and completely unsupported by the Fourteenth Amendment where the source of such a power must be found. The Fourteenth Amendment guarantees individuals against invasions by the States of fundamental invasions by the States of Connecticut, 302 U.S. 319, 58 S.Ct. 149, 82 L.Ed. 288 and under more recent decisions of this Court some of the specifics of the Bill of Rights as well. See, e.g., in the context of this case, Malloy v. Hogan, 378 U.S. 1, 84 S.Ct. 1489, 12 L.Ed.2d 653; Griffin v. State of California, 380 U.S. 609, 85 S.Ct. 1229. It thus serves as a limitation on the actions of the States, and lodges in this Court the same power over state 'laws, rules, and remedies' as the Court has always had over the 'laws, rules, and remedies' created by Congress. This power was classically described by Chief Justice Marshall in Marbury v. Madison, 1 Cranch 137, 178, 2 L.Ed. 60:
'So if a law be in opposition to the constitution; if both the law and the constitution apply to a particular case, so that the court must either decide that case conformably to the law, disregarding the constitution; or conformably to the constitution, disregarding the law; the court must determine which of these conflicting rules governs the case. * * *'
Nothing in the Fourteenth Amendment purports to give federal courts supervisory powers, in the affirmative sense of McNabb v. United States, 318 U.S. 332, 63 S.Ct. 608, 87 L.Ed. 819, over state courts. See id., at 340-341, 63 S.Ct., at 612-613. Moreover, where the constitutional power described by Marshall has been invoked, the Court has always been especially reluctant to interfere with state procedural practices. See Spencer v. Texas, 385 U.S. 554, 87 S.Ct. 648, 17 L.Ed.2d 606. From the beginning of the federal Union, state courts have had power to decide issues of federal law and to formulate 'authoritative laws, rules, and remedies' for the trial of those issues. The primary responsibility for the trial of state criminal cases still rests upon the States, and the only constitutional limitation upon these trials is that the laws, rules, and remedies applied must meet constitutional requirements. If they do not, this Court may hold them invalid. The Court has no power, however, to declare which of many admittedly constitutional alternatives a State may choose.  To impose uniform national requirements when alternatives are constitutionally permissible would destroy that opportunity for broad experimentation which is the genius of our federal system.
Even assuming that the Court has the power to fashion remedies and procedures binding on state courts for the protection of particular constitutional rights, I could not agree that a general harmless-error rule falls into that category. The harmless-error rules now utilized by all the States and in the federal judicial system are the product of judicial reform early in this century. Previously most American appellate courts, concerned about the harshness of criminal penalties, followed the rule imposed on English courts through the efforts of Baron Parke, and held that any error of substance required a reversal of conviction. See Orfield, Criminal Appeals in America 190. The reform movement, led by authorities like Roscoe Pound and Learned Hand, resulted in allowing courts to discontinue using reversal as a 'necessary' remedy for particular errors and 'to substitute judgment for the automatic application of rules * * *.' 4 Barron, Federal Practice and Procedure § 2571, at 438. This Court summarized the need for that development in the leading case of Koteakos v. United States, 328 U.S. 750, 759, 66 S.Ct. 1239, 1245, 90 L.Ed. 1557:
's 269 (a federal harmless error provision) and similar state legislation grew out of widespread and deep conviction over the general course of appellate review in American criminal causes. This was shortly, as one trial judge put it after § 269 had become law, that courts of review, 'tower above the trials of criminal cases as impregnable citadels of technicality.' * * * (C)riminal trial became a game for sowing reversible error in the record.'
Holding, as is done today, that a special harmless-error rule is a necessary remedy for a particular kind of error revives the unfortunate idea that appellate courts must act on particular errors rather than decide on reversal by an evaluation of the entire proceeding to determine whether the cause as a whole has been determined according to properly applicable law. In this case, California has recognized the impropriety of the trial comment here involved, and has given clear direction to state trial courts for the future. Certainly this is the appropriate remedy for the constitutional error committed. The challenged decision has no direct relation to federal constitutional provisions, rather it is an analysis of the question whetehr this admittedly improper comment had any significant impact on the outcome of the trial. In Kotteckos, supra, this Court described the 'material factors' in harmless-error determinations as 'the character of the proceeding, what is at stake upon its outcome, and the relation of the error asserted to casting the balance for decision on the case as a whole * * *.' Id., at 762, 66 S.Ct., at 1246. None of these factors has any relation to substantive constitutional provisions, and I think the Court errs in conceiving of an application of harmless-error rules as a remedy designed to safeguard particular constitutional rights.  It seems clear to me that harmless-error rules concern, instead, the fundamental integrity of the judicial proceedings as a whole.
As indicated above, I am of the opinion that the validity of a challenged state harmless-error rule itself is a federal constitutional question. Harmless-error rules may, as the Court says, 'work very unfair and mischievous results.' And just concern can be expressed over the possibility that state harmless-error decisions may result in the dilution of new constitutional doctrines because of state hostility to them. However, the record is barren of any showing that the California courts, which have been in the vanguard in the development of individual safeguards in criminal trials,  are using their harmless-error rule to destroy or dilute constitutional guarantees. If the contrary were the case and the harmless-error rule itself were shown to have resulted in a course of convictions significantly influenced by constitutionally impermissible factors, I think it clear that constitutional due process could not countenance the continued application of the rule.  And individual applications of a permissible rule would still be subject to scrutiny as to the tenability of the independent and adequate state ground. See Thompson v. City of Louisville, 362 U.S. 199, 80 S.Ct. 624, 4 L.Ed.2d 654; Terre Haute & Indianapolis Railroad Co. v. State of Indiana ex rel. Ketcham, 194 U.S. 579, 24 S.Ct. 767, 48 L.Ed. 1124; Note, The Untenable Non-federal Ground in the Supreme Court, 74 Harv.L.Rev. 1375.
I thus see no need for this new constitutional doctrine.  Decision of this case should turn instead on the answers to two questions: Is the California harmless-error provision consistent with the guarantee of fundamental fairness embodied in the Due Process Clause of the Fourteenth Amendment? See Palko v. Connecticut, supra. Was its application in this instance by the California Supreme Court a reasonable one or was the rule applied arbitrarily to evade the underlying constitutional mandate of fundamental fairness? These issues will now be considered.
The California harmless-error rule, is incorporated in that State's constitution. It was first adopted by a vote of the people in 1911 and readopted as part of the revised constitution in 1966. While its language allows reversal only where there has been a 'miscarriage of justice,' a long course of judicial decisions has shaped the rule in a manner which cannot be ignored. California courts will not allow a conviction based upon an improperly obtained confession to stand. See, e.g., People v. Dorado, 62 Cal.2d 338, 42 Cal.Rptr. 169, 398 P.2d 361; People v. Sears, 62 Cal.2d 737, 44 Cal.Rptr. 330, 401 P.2d 938. Nor will the fact that sufficient evidence to support the conviction is present absent the tainted evidence preclude a reversal. See, e.g., People v. Patubo, 9 Cal.2d 537, 71 P.2d 270, 113 A.L.R. 1303; People v. Mahoney, 201 Cal. 618, 258 P. 607. And reversal will be required when the tainted evidence is introduced in intentional violation of constitutional standards. See People v. Sarazzawski, 27 Cal.2d 7, 161 P.2d 934. Thus the California rule and the 'federal rule' today declared applicable to state adjudication are parallel in these special instances  and their divergence, if any, arises from the general formulation found in the opinions of the California Supreme Court.
In People v. Watson, 46 Cal.2d 818, 299 P.2d 243, the California Supreme Court undertook a general discussion of the application of the state harmless-error rule. It declared that the 'final test' was 'the 'opinion' of the reviewing court, in the sense of its belief or conviction, as to the effect of the error; and that ordinarily where the result appears just, and it further appears that such result would have been reached if the error had not been committed, a reversal will not be ordered.' Reversal would be required only when 'it is reasonably probable that a result more favorable to the appealing party would have been reached,' and this judgment 'must necessarily be based upon reasonable probabilities rather than upon mere possibilities; otherwise the entire purpose of the constitutional provision would be defeated.' 46 Cal.2d, at 835-837, 299 P.2d, at 254-255. This formulation may sound somewhat different from that announced today, but on closer analysis the distinction between probability and possibility becomes essentially esoteric. In fact, California courts have at times equated the California standard with the standard utilized by this Court in Fahy v. State of Connecticut, supra. See, e.g., People v. Jacobson, 63 Cal.2d 319, 331, 46 Cal.Rptr. 515, 523, 405 P.2d 555, 563.
Similarly members of this Court have used a variety of verbal formulae in deciding questions of harmless error in federal cases, ranging from today's 'reasonable doubt' standard to the ability to 'say with fair assurance * * * that the jury was not substantially swayed * * *.' Fiswick v. United States, 329 U.S. 211, 218, 67 S.Ct. 224, 228, 91 L.Ed. 196. And the circuit courts have been equally varied in their expressions. United States v. Brown, 79 F.2d 321; United States v. Feinberg, 2 Cir., 140 F.2d 592; United States v. McMaster, 6 Cir., 343 F.2d 176.
Against this background the California rule can hardly be said to be out of keeping with fundamental fairness, and I see no reason for striking it down on its face as a violation of the guarantee of 'due process.' 
A summary of the evidence introduced against the petitioners and events of the trial will make it apparent that the application of the California rule in this case was not an unreasonable one. California courts have not hesitated to declare that comment has caused a miscarriage of justice when that conclusion has been warranted by the circumstances, see, e.g., People v. Keller, 234 Cal.App.2d 395, 44 Cal.Rptr. 432; People v. Sigal, 235 Cal.App.2d 449, 45 Cal.Rptr. 481, but the posture of this case minimized the possible impact of the comment.
Petitioners were tried for the murder of a night club bartender in the course of a robbery of the club. The State established that petitioners were the last customers remaining in the club on the night of the murder. Three people with descriptions matching those of Chapman, Teale, and the victim were seen leaving the club together. The club had been ransacked and its condition indicated that the victim had been forced out of it. He was later shot from close range with a .22-caliber weapon and left beside a country road. It was shown that Chapman had purchased a similar weapon five days before the murder and this weapon was in Teale's possession when he was arrested. Blood matching the type of the victim was found on the floormat of the vehicle in which Chapman and Teale had been traveling. Other scientific testimony established that the victim had been in petitioners' car. Blood (untypable) was found on Chapman's clothes, and blood matching the victim's was found on her shoes. Similar evidence connected Teale with the murder.
After his arrest Teale made admissions, amounting almost to a full confession, to a fellow prisoner and these were introduced against him. The jury was cautioned to disregard them as against Chapman. Petitioners pleaded not guilty, but offered no defense on the merits. The only defense witness was a Dr. Sheuerman who was called by Chapman in an effort to establish a defense of lack of capacity to form the requisite intent because of 'disassociative reaction.'
The prosecutor's comment on petitioners' failure to explain away or challenge the evidence presented against them was admittedly extensive.  The California Supreme Court found it harmless error for a number of reasons. First the court noted the convincing and unchallenged evidence presented by the State. It next observed that the jurors were certain to take notice of petitioners' silence whether or not there was comment since the evidence itself cried for an explanation. I think this point crucial, since it seems to me that this Court has confused the impact of petitioners' silence on the jury with the impact of the prosecution's comment upon that silence. The added impact of that comment would seem marginal in a case of this type where the jury must inevitably look to petitioners for an explanation of the innuendo of the real evidence and in Teale's case of his damaging admissions. Finally the California Supreme Court noted that Chapman, against whom the evidence was less strong, had keyed her defense to evidence of her mental defect, a subject upon which the comment had not touched. From this discriminating analysis it was concluded that another result was not 'reasonably probable' absent the erroneous comments.
I cannot see how this resolution can be thought other than a reasonable, and therefore constitutional, application of the California harmless-error rule.
When we consider how little is empirically known about the workings of a jury, see Kalven & Zeisel, The American Jury, passim, it seems to me highly inappropriate for this Court to presume to take upon itself the power to pass directly on the correctness of impact evaluations coming from 50 different jurisdictions. Juries must invariably react differently to particular items of evidence because of local predispositions and experience factors. The state courts, manned by local judges aware of and in touch with the special factors affecting local criminal trials, seem the best, and the constitutionally required, final authority for ruling on the effect of the admission of inadmissible evidence in state criminal proceedings, absent the application of a fundamentally unfair rule, or any unreasonable application of a proper rule manifesting a purpose to defeat federal constitutional rights. Once it appears that neither of these factors is present in a state harmless-constitutional-error decision, federal judicial responsibility should be at an end. This decision, however, encompasses much more. It imposes on this Court, in cases coming here directly from state courts, and on the lower federal courts, in cases arising on habeas corpus, the duty of determining for themselves whether a constitutional error was harmless. In all but insubstantial instances, this will entail a de novo assessment of the entire state trial record.
For one who believes that among the constitutional values which contribute to the preservation of our free society none ranks higher than the principles of federalism, and that this Court's responsibility for keeping such principles intact is no less than its responsibility for maintaining particular constitutional rights, the doctrine announced today is a most disturbing one. It cuts sharply into the finality of state criminal processes; it bids fair to place an unnecessary substantial burden of work on the federal courts; and it opens the door to further excursions by the federal judiciary into state judicial domains. I venture to hope that as time goes on this new doctrine, even in its present manifestation, will be found to have been strictly contained, still more that it will not be pushed to its logical extremes.
I respectfully dissent.
^1 For myself, I intimate no view on congressional power with respect to state courts in this regard.
^2 Cases in which lower federal courts, acting under the authority of the Fourteenth Amendment, as expanded by this Court's decision in Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506, have promulgated their own reapportionment plans may superficially be thought to support such a power. E.g., Reynolds v. State Election Board, D.C., 233 F.Supp. 323. But such cases are quite apart from the present one because they arise from a situation where some positive constitutional action is a necessity and thus require the exercise of special equity powers. Here the ordinary remedy of striking down unconstitutional harmless-error rules and applications is sufficient to deal with any problem that may arise. There is no necessity for a State to have a harmless-error rule at all.
^3 The Court indeed recognizes, as does my Brother STEWART in his concurring opinion, that errors of constitutional dimension can be harmless, a proposition supported by ample precedent. See Snyder v. Commonwealth of Massachusetts, 291 U.S. 97, 54 S.Ct. 330, 78 L.Ed. 674; Motes v. United States, 178 U.S. 458, 20 S.Ct. 993, 44 L.Ed. 1150; Haines v. United States, 9 Cir., 188 F.2d 546; United States v. Donnelly, 7 Cir., 179 F.2d 227. Prsumably all errors in the federal courts will continue to be evaluated under the single standard of 28 U.S.C. § 2111 as interpreted today. Certainly there is nothing in the substantive provisions of the Bill of Rights which suggests any standard for assessing the impact of their violation.
^5 It is clear enough that this is not the rationale that the Court is employing. The Court would leave California free to apply its harmless-error rule to errors of state law and must thus consider the rule itself consistent with constitutional due process. This leaves the anomalous situation where the impact of a particular piece of evidence is to be assessed by a different 'constitutional' standard depending only on whether state law or federal constitutional law barred its admittance.
^7 Some special limitations on harmless error have always been respected by this Court and seem to me essential to the fundamental fairness guaranteed by the Due Process Clauses of the Fifth and Fourteenth Amendments. These limitations stem from what I perceive as two distinct considerations. The first is a recognition that particular types of error have an effect which is so devastating or inherently indeterminate that as a matter of law they cannot reasonably be found harmless. E.g., Payne v. State of Arkansas, 356 U.S. 560, 78 S.Ct. 844 (confessions); see Fahy v. State of Connecticut, supra, at 95, 84 S.Ct. at 234 (dissenting opinion of Harlan, J.); cf. Bollenbach v. United States, 326 U.S. 607, 66 S.Ct. 402 (independently sufficient evidence). The second is a recognition that certain types of official misbehavior require reversal simply because society cannot tolerate giving final effct to a judgment tainted with such intentional misconduct. E.g., Berger v. United States, 295 U.S. 78, 55 S.Ct. 629, 79 L.Ed. 1314 (prosecutorial misconduct). Although they have never been viewed in this light, I would see violations of Gideon v. Wainwright, 372 U.S. 335, 83 S.Ct. 792, 9 L.Ed.2d 799, as falling in the first category, and violations of Tumey v. Ohio, 273 U.S. 510, 47 S.Ct. 437, as falling in the second. However, as I understand my Brother STEWART's opinion concurring in the result, he would read all such limitations into the content of the Due Process Clause and limit the application of harmless-error rules with respect to constitutional errors to an undefined category of instances. I think it preferable to resolve these special problems from an analysis of the nature of the error involved rather than by an attempt to discover limitations in the policy underlying the substantive constitutional provisions. The latter course seems to me to blur analysis and lead to distinction by fiat among equally specific constitutional guarantees.
^9 The decision in Griffin v. State of California, 380 U.S. 609, 85 S.Ct. 1229, was not announced until after the trial of the case. Hence the trial was conducted according to what was, at the time, constitutional California law. No implication of prosecutorial misconduct can be drawn from these circumstances.