Stewart v. United States, 366 U.S. 1 (1961)

Argued: February 21, 1961
Decided: April 24, 1961

U.S. Supreme Court

Stewart v. United States, 366 U.S. 1 (1961)

Stewart v. United States

No. 143

Argued February 21, 1961

Decided April 24, 1961

366 U.S. 1



U.S. Supreme Court

Stewart v. United States, 366 U.S. 1 (1961) Stewart v. United States

No. 143

Argued February 21, 1961

Decided April 24, 1961

366 U.S. 1




Petitioner was tried three times in a federal court for murder. At the first two trials, he did not testify in his own defense, but he did so at the third trial, at which the main issue was whether or not he was insane when the offense was committed. On cross-examination, the prosecutor alluded to the two earlier trials and asked, "This is the first time you have gone on the stand, isn't it, Willie?" Petitioner's counsel moved for a mistrial on the ground that it was prejudicial to inform the jury of petitioner's failure to take the stand in his previous trials. The motion was denied, and petitioner was convicted.

Held: the question was prejudicial; the error was not harmless; a mistrial should have been granted; and the judgment affirming the conviction is reversed. Pp. 366 U. S. 2-10.

107 U.S.App.D.C. 159, 275 F.2d 617, reversed.

Page 366 U. S. 2

MR. JUSTICE BLACK delivered the opinion of the Court.

The Fifth Amendment to the United States Constitution provides in unequivocal terms that no person may "be compelled in any criminal case to be a witness against himself." To protect this right, Congress has declared that the failure of a defendant to testify in his own defense "shall not create any presumption against him." [Footnote 1] Ordinarily, the effectuation of this protection is a relatively simple matter -- if the defendant chooses not to take the stand, no comment or argument about his failure to testify is permitted. [Footnote 2] But where for any reason it becomes necessary to try a particular charge more than one time, a more complicated problem may be presented. For a defendant may choose to remain silent at his first trial and then decide to take the stand at a subsequent trial. When this occurs, questions arise as to the propriety of comment or argument in the second trial based upon the defendant's failure to take the stand at his previous trial. This case turns upon such a question.

Petitioner has been tried three times in the District Court for the District of Columbia upon an indictment charging that he had committed first-degree murder under a felony-murder statute. [Footnote 3] In all three trials, petitioner's

Page 366 U. S. 3

chief defense has been insanity but, on each occasion, the jury has rejected this defense and returned a verdict of guilty upon which the District of Columbia's mandatory death sentence has been imposed. [Footnote 4] After the first two trials, in which petitioner did not testify, the convictions and death sentences were set aside on the basis of trial errors that the Court of Appeals found had prevented a proper consideration of the case by the jury. [Footnote 5] At the third trial, in an apparent effort to bolster the contention of insanity, petitioner was placed upon the stand and asked a number of questions by defense counsel -- a maneuver obviously made for the purpose of giving the jury an opportunity directly to observe the functioning of petitioner's mental processes in the hope that such an exhibition would persuade them that his memory and mental comprehension were defective. Petitioner's responses to these questions were aptly described by the court below as "gibberish without meaning." [Footnote 6]

Page 366 U. S. 4

Upon cross-examination, the prosecutor attempted without noticeable success to demonstrate that these irrational answers were given by petitioner in furtherance of his plan to feign a mental weakness that did not exist. To this end, the prosecutor asked petitioner a number of questions about statements petitioner had allegedly made subsequent to his arrest, apparently in the hope that one of these questions would surprise petitioner and provoke a sensible response. When petitioner continued to talk in the same manner that he had used upon direct examination, the prosecutor concluded his cross-examination with the following remarks in the form of questions: "Willie, you were tried on two other occasions." And, "This is the first time you have gone on the stand, isn't it, Willie?" [Footnote 7]

The defense moved immediately for a mistrial on the ground that it was highly prejudicial for the prosecutor to inform the jury of the defendant's failure to take the stand in his previous trials. The prosecutor defended his actions on the ground that this "is a fact that the Jury is entitled to know." The trial judge agreed with the prosecutor, denied the motion for a mistrial, and the trial proceeded, culminating in the third verdict of guilty and death sentence. On appeal, the case was heard by

Page 366 U. S. 5

all nine members of the Court of Appeals sitting en banc, and was affirmed by a 5-4 vote [Footnote 8] -- the majority concluding that the issue was controlled by the decision of this Court in Raffel v. United States [Footnote 9] and the minority concluding that the issue was controlled by our decision in Grunewald v. United States. [Footnote 10] We granted certiorari to consider whether it was error for the trial court to deny the motion for a mistrial under the circumstances. [Footnote 11]

In this Court, the Government concedes that the question put to the defendant about his prior failures to testify cannot be justified under Raffel, Grunewald, or any other of this Court's prior decisions. This concession, which we accept as proper, rests upon the Government's recognition of the fact that in no case has this Court intimated that there is such a basic inconsistency between silence at one trial and taking the stand at a subsequent trial that the fact of prior silence can be used to impeach any testimony which a defendant elects to give at a later trial. The Raffel case, relied upon by the majority below, involved a situation in which Raffel had sat silent at his first trial in the face of testimony by a government agent

Page 366 U. S. 6

that Raffel had previously made admissions pointing to his guilt. On a second trial, Raffel took the stand and denied the truth of this same testimony offered by the same witness. Under these circumstances, this Court held that Raffel's silence at the first trial could be shown in order to discredit his testimony at the second trial on the theory that the silence itself constituted an admission as to the truth of the agent's testimony. The result was that Raffel's silence at the first trial was held properly admitted to impeach the specific testimony he offered at the second trial. Here, on the other hand, the defendant's entire "testimony" comprised nothing more than "gibberish without meaning," with the result that there was no specific testimony to impeach. Any attempt to impeach this defendant as a witness could therefore have related only to his demeanor on the stand, and, indeed, the majority below expressly rested its conclusion upon the view that the prosecution had the right, under Raffel, to test the genuineness of this sort of "demeanor evidence" by questions as to why it was not offered at previous trials. [Footnote 12] But if Raffel could properly be read as standing for this proposition, such questions would be permissible in every instance, for, whenever a witness takes the stand, he necessarily puts the genuineness of his demeanor into issue. [Footnote 13] The Government quite properly concedes that

Page 366 U. S. 7

this cannot be the law, since it would conflict with the precise holding of this Court in the Grunewald case. [Footnote 14]

Despite this concession, however, the Government persists in the contention that petitioner's conviction should be upheld, arguing that the error committed was harmless, and could not have affected the jury's verdict. This argument is rested upon three grounds: first, that the jury may not even have heard the improper question; secondly, that, even if the jury did hear the question, it may not have inferred that petitioner in fact did not testify at his previous trial; and, finally that, even if the jury did infer that petitioner did not testify previously, no inference adverse to petitioner would have been drawn from this fact. The first two of these grounds can be quickly disposed of. We can think of no justification for ignoring the part of a record showing error on a mere conjecture that the jury might not have heard the testimony that part of the record represents. Nor do we believe it reasonable to argue that the jury trying this case would not have inferred that this defendant had failed to testify in his prior trials when the prosecutor asked, "This is the first time you have gone on the stand, isn't it, Willie?" Indeed, the recognition that such an inference will in all likelihood be drawn from leading questions of this kind lies at the root of the long established rule that such questions may not properly be put unless the inference, if drawn, would be factually true. [Footnote 15] Thus, the Government's argument that

Page 366 U. S. 8

the error was harmless must stand or fall upon the third ground it urges -- that the jury's awareness of petitioner's failure to take the stand at his previous trials would not have prejudiced the consideration of his case. The disposition of this contention requires the statement of a few more of the relevant facts of the case.

In connection with the defense of insanity, petitioner had introduced evidence of both mental disease and mental defect, as those terms are applied in the relevant law of the District of Columbia. [Footnote 16] On the mental disease issue, the testimony was that petitioner was suffering from manic depressive psychosis, a disease which the record shows tends to fluctuate considerably in its manifestations from time to time. On the mental defect issue, the defense introduced evidence that petitioner had an intelligence level in the moronic class. The case went to the jury on both of these points, the jury being directed to acquit if it found the homicide to have been the product either of mental disease or mental defect. [Footnote 17] Petitioner's "testimony" thus raised at least two different issues in the minds of the jury: first, whether petitioner was simply

Page 366 U. S. 9

feigning this testimony, and, secondly, whether, if not, petitioner's condition at the time of his third trial fairly represented his condition at the time of the act charged in the indictment. [Footnote 18]

We think it apparent that the jury's awareness of petitioner's failure to testify at his first two trials could have affected its deliberations on either or both of these issues. Thus, the jury might well have thought it likely that petitioner elected to feign this "testimony" out of desperation brought on by his failure to gain acquittal without it in the two previous trials. Similarly, even if the jury believed petitioner's "testimony" was genuine, it might have thought that petitioner's condition was caused by a mental disease, and concluded that it is unlikely that a disease that had manifested itself only one out of three times for exhibition at trial was active at the occasion of the homicide. Or, on the same assumption, it might have thought that petitioner's failure to exhibit himself at the previous trials indicated that the condition manifested at this trial was the result of a worsening in his mental condition since those trials and, consequently, also since the commission of the acts charged in the indictment. There may be other ways in which the jury might have used the information improperly given it by the prosecution -- we have mentioned more than enough already, however, to satisfy ourselves that the Government's contention that the error was harmless must be rejected.

The Government's final contention is that, even if the error was prejudicial, the conviction should be allowed

Page 366 U. S. 10

to stand on the theory that the error was not sufficiently prejudicial to warrant the granting of a mistrial, and the defense made no request for cautionary instructions. One answer to this argument is to be found in the Government's own brief. For, in its argument regarding the possibility that the jury may not have been aware of the improper question, the Government stresses the fact that the question was not emphasized by any reference to it in the instructions to the jury. During the course of this argument the Government expressly recognizes that the danger of the situation would have been increased by a cautionary instruction in that such an instruction would have again brought the jury's attention to petitioner's prior failures to testify. Plainly, the defense was under no obligation to take such a risk. The motion for a mistrial was entirely appropriate, and, indeed, necessary to protect the interests of petitioner. [Footnote 19]

We thus conclude that this conviction and sentence against petitioner cannot stand. In doing so, we agree with the point made by the Government in its brief -- that it is regrettable when the concurrent findings of 36 jurors are not sufficient finally to terminate a case. But, under our system, a man is entitled to the findings of 12 jurors on evidence fairly and properly presented to them. Petitioner may not be deprived of his life until that right is accorded him. That right was denied here by the prosecutor's improper questions.


Page 366 U. S. 11

[Footnote 1]

"In trial of all persons charged with the commission of offenses against the United States and in all proceedings in courts martial and courts of inquiry in any State, District, Possession or Territory, the person charged shall at his own request, be a competent witness. His failure to make such request shall not create any presumption against him."

62 Stat. 833, 18 U.S.C. § 3481.

[Footnote 2]

Wilson v. United States, 149 U. S. 60.

[Footnote 3]

"Whoever, being of sound memory and discretion, kills another purposely, either of deliberate and premeditated malice or by means of poison, or in perpetrating or attempting to perpetrate any offense punishable by imprisonment in the penitentiary, or without purpose so to do kills another in perpetrating or in attempting to perpetrate any arson, as defined in section 22-401 or 22-402 of this Code, rape, mayhem, robbery, or kidnapping, or in perpetrating or in attempting to perpetrate any housebreaking while armed with or using a dangerous weapon, is guilty of murder in the first degree."

District of Columbia Code, § 22-2401. (Emphasis supplied.)

[Footnote 4]

Section 22-2404 of the District of Columbia Code provides: "The punishment of murder in the first degree shall be death by electrocution."

[Footnote 5]

The first conviction was set aside because of erroneous instructions on the defense of insanity. 94 U.S.App.D.C. 293, 214 F.2d 879. The second conviction was set aside because of improper argument by the prosecutor. 101 U.S.App.D.C. 51, 247 F.2d 42.

[Footnote 6]

107 U.S.App.D.C. 159, 160, 275 F.2d 617, 618. The following excerpt from petitioner's testimony is entirely typical:

"Q. Who is your lawyer?"

"A. Well, I mean, I am my own lawyer, as far as my concern."

"Q. Have I been representing you here the last couple days?"

"A. As far as I am concerned, you all look the same to me."

"Q. Do you know what is going on in this courtroom the last couple days?"

"A. I ain't asked about what is going on. It is up to you go on and describe yourself. I mean, don't ask me. As far as I am just sitting here."

"Q. Did you ever hear the name Harry Honigman (the man with whose murder petitioner was charged) before?"

"A. I haven't."

"Q. Do you know you are charged with first degree murder?"

"A. As far as I am concerned, I ain't charged with nothing."

"Q. What is first degree murder; do you know?"

"A. I don't know."

[Footnote 7]

The record reveals the following exchange at the conclusion of the cross-examination of petitioner by the prosecutor, a Mr. Smithson:

"Q. Willie, you were tried on two other occasions. "

"A. Well, I don't care how many occasions, how many case -- you say case. I was a case man once in a time."

"Q. This is the first time you have gone on the stand, isn't it, Willie?"

"A. What?"

"Q. This is the first time you have gone on the stand, isn't it, Willie?"

"A. I am always the stand; I am everything, I done told you."

"Mr. Smithson: That is all."

[Footnote 8]

107 U.S.App.D.C. 159, 275 F.2d 617.

[Footnote 9]

271 U. S. 271 U.S. 494.

[Footnote 10]

353 U. S. 353 U.S. 391.

[Footnote 11]

363 U.S. 818. The petition for certiorari also raised objections based upon other alleged errors during the course of the trial. In view of our disposition of the primary issue, and because the actions complained of may not arise at any subsequent trial, we find it unnecessary to pass upon these other objections.

[Footnote 12]

Thus, the majority reasoned:

"The logical and permissible first step under Raffel v. United States, supra, was to have him say whether he had previously testified in order to lay the groundwork for developing an inconsistency inherent in the difference in his 'demeanor-evidence' in the two trials."

107 U.S.App.D.C. 159, 167, 275 F.2d 617, 625.

[Footnote 13]

This is so because the defendant's credibility is in issue whenever he testifies. If the failure to testify at a previous trial were to amount to evidence that testimony at a subsequent trial was feigned or perjurious, the fact of failure to testify would always be admissible.

[Footnote 14]

The holding in Grunewald was that the defendant's answers to certain questions were not inconsistent with his previous reliance upon the Fifth Amendment to excuse a refusal to answer those very same questions. Since defendant's testimony placed his credibility in issue, the necessary implication of that holding is that his prior refusal to testify could not be used to impeach his general credibility.

[Footnote 15]

III Wigmore, Evidence (3d ed.), § 780. Wigmore quotes Chitty, Practice of the Law, 2d ed., III, 901, for the proposition:

"It is an established rule, as regards cross-examination, that a counsel has no right, even in order to detect or catch a witness in a falsity, falsely to assume or pretend that the witness had previously sworn or stated differently to the fact, or that a matter had previously been proved when it had not."

This Court has previously recognized that principle. Berger v. United States, 295 U. S. 78, 295 U. S. 84.

[Footnote 16]

The difference between the terms "disease" and "defect" was explained in the charge to the jury in the following manner:

"We use 'disease' in the sense of a condition which is considered capable of either improving or deteriorating. We the 'defect' in the sense of a condition which is not considered capable of either improving or deteriorating, and which may be either congenital or the result of injury, or the residual effect of a physical or mental disease."

[Footnote 17]

These instructions stemmed from the test of criminal responsibility that prevails in the District of Columbia under the decision of the Court of Appeals in Durham v. United States, 94 U.S.App.D.C. 228, 214 F.2d 862.

[Footnote 18]

This second issue arises from the fact that the jury was not here trying the question whether petitioner was mentally competent to stand trial. Under the District of Columbia practice, that question is decided in a separate proceeding. See District of Columbia Code, § 24-301.

[Footnote 19]

Johnson v. United States, 318 U. S. 189, relied upon by the Government, does not sustain its argument on this point. There, the defense made no objection at all, choosing instead to rest its chances upon the verdict of the jury. Petitioner here made no such choice, for he has repeatedly pressed his right to a mistrial, in the District Court, in the Court of Appeals, and here.


The result which the Court draws from its account of the trial seems not unreasonable. But, by force of what the Court does not relate, there is such disparity between its account and the almost nine hundred pages of the trial transcript that, in fairness, the Court's opinion hardly conveys what took place before the jury and what must, therefore, rationally be evaluated in attributing any influence on the jury's verdict to the questions which the Government now concedes were improperly asked.

"In reviewing criminal cases, it is particularly important for appellate courts to relive the whole trial imaginatively, and not to extract from episodes in isolation abstract questions of evidence and procedure. To turn a criminal appeal into a quest for error no more promotes the ends of justice than to acquiesce in low standards of criminal prosecution."

Johnson v. United States, 318 U. S. 189, 318 U. S. 202 (concurring opinion).

What emerges from the transcript at the outset is that Willie Lee Stewart's killing of Harry Honikman was practically never in issue. The testimony of two eyewitnesses who positively identified Stewart as the killer [Footnote 2/1] was not seriously challenged. A third witness had examined in Stewart's hands, shortly before the killing, the gun which unimpugned ballastic evidence established fired the lethal shots. The testimony of a fingerprint expert, also unimpugned, linked Stewart to the killing. Nowhere in their opening or closing statements did experienced defense counsel ask the jury to doubt that

Page 366 U. S. 12

Stewart was the killer: the whole of the defense was that Stewart was not responsible, because insane.

Insanity was not merely, as the Court says, Stewart's chief defense; it was his defense. His lawyer put it aptly: "[The prosecutor] knows as well as I, as anybody in this courtroom, the only defense we have is insanity." [Footnote 2/2] Thus, there is not involved in this case the danger that the jury, being told as laymen of the defendant's previous failure to testify in his own behalf, reasoned that, if Stewart did not do the acts with which he was charged, he would have said so. Here, those acts were not contested. If prejudice is not to be blindly assumed, but to be discovered in the record, it must be discovered by some more subtle train of associations.

Stewart's trial took the major part of six court days: twelve calendar days. The Government's opening case, presenting the testimony of the eyewitnesses, fingerprint and ballistic experts, arresting and investigating officers, etc. -- ten witnesses in all [Footnote 2/3] -- consumed a day and a half. Thereafter, beginning on the second court day and running into the third, the defense put in the testimony of a series of witnesses -- Stewart's cousin, landlady, friend, sister, employer, wife, neighbor, sister-in-law -- all of whom recounted episodes of Stewart's behavior tending to show his unsoundness of mind. [Footnote 2/4] These episodes spanned the period of his life from early childhood until the time of the killing, and they painted what, to say the least, is a bizarre portrait.

Page 366 U. S. 13

If the jury believed them, they believed, inter alia: (1) that Stewart, as a child, threw all his food on the floor, ran away from school, tore his clothes off, cut them up, roamed the house at night; (2) that Stewart's aunts and brother were of unsound mind, in that they would often sit with saliva running out of their mouths and would never say anything; (3) that Stewart, as an adult, once shot at his wife, and sat on his wife and beat her while she was pregnant; (4) that he once punched a hole in a low ceiling with his fist for no apparent reason, and, on another occasion, threw all the food out of his refrigerator and beat the refrigerator door so hard with his first that he broke it; (5) that he locked his children out of the family's room in cold weather; that he threatened to throw one of his children, while a baby, out of the window and threatened to throw another into a burning stove; that he would have done both if not forcibly prevented; (6) that he insisted on pushing through a boarded front door and jumping in and out of the house at a time when the porch was under repair; that he once jumped out of a window; that he threw his nephew's toy piano out of a window; (7) that he attempted to have sexual relations with his sister-in-law in her husband's presence; (8) that, having been told by his employer that he would get a requested pay raise, he kicked down a brick wall that he had been constructing. Following this testimony, defense counsel read to the jury portions of Stewart's military record, revealing that a medical discharge had been recommended for Stewart after a fight with another soldier, largely on the basis of tests taken at that time which placed Stewart's intelligence in the feeble-minded range.

On the third trial day, the defendant took the stand and was examined and cross-examined briefly. His testimony occupies fifteen pages of the 885-page

Page 366 U. S. 14

trial minutes. Let this sample of it give its quality of meaninglessness:

"Q. What is your wife's name, Willie?"

"A. You should ask her that. As far as I am concerned, I don't have no wife. I don't consider I have any; therefore, I can't say what her name is."

"Q. Have you ever been married?"

"A. I wouldn't say married."

"Q. What do you mean you wouldn't say married?"

"A. Well, as far as I concerned, nobody is married, as far as my way of understanding."

"Q. Do you have any children?"

"A. I don't consider -- I have none. She say I have some. I don't have none. If she say I have some, I guess I have to leave it to her. As far as my concern, I don't have none, and I don't want none."

"Q. Do you know where you are now?"

"A. Looking at you, as far as I know."

"Q. What is my name?"

"A. I don't know."

"Q. Who is your lawyer?"

"A. Well, I mean, I am my own lawyer, as far as my concern."

On his direct examination, Stewart testified that he did not know what kind of a building he was in, that he had never shot nobody, but that the white folks told him he was supposed to kill; that he considered himself master, as far as the killing situation; that he was the monkey, the monkey with the tail; that he still remained to see that monkey with the tail; that he had been told to kill -- his mind tells him to kill -- and he was always going to kill until he conquered; that the good man upstairs say so; that he had talked to God and God told him to conquer everybody, that he was the master;

Page 366 U. S. 15

hated everybody; counsel shouldn't ask him no more. The brief cross-examination proceeded in the same vein. The prosecutor's questions, designed less to elicit any information from the witness than to call forth some revealingly intelligent response, some sign of memory or understanding, which would show that Stewart's apparently grave mental estrangement was a pose, evoked only wild and unresponsive answers. The cross-examination closed on the following dialogue:

"Q. You can see me, can't you, Willie?"

"A. Sure. You can see me, too, can't you? We see one another. I am going to be the master and you ain't going to stop me and nobody else."

"Q. Tell me, Willie, do you know a Dr. Williams?"

"A. Dr. Williams?"

"Q. Yes, E. Y. Williams."

"A. Why you keep asking me? If I told you once, I told you a hundred time, I am my own doctor. Why you keep asking me the same question over and over again. I told you I am my own doctor."

"Q. Do you know a Deputy Marshall by the name of Ballinger?"

"A. I am my own marshal. I am everything. That takes care of the whole question. I am everything. Everything you ask me, I am talking to me, I am it."

"Q. Willie, you were tried on two other occasions."

"A. Well, I don't care how many occasions, how many case -- you say case. I was a case man once in a time."

"Q. This is the first time you have gone on the stand, isn't it, Willie?"

"A. What?"

"Q. This is the first time you have gone on the stand, isn't it, Willie? "

Page 366 U. S. 16

"A. I am always the stand; I am everything, I done told you."

"Mr. Smithson [the prosecutor]: That is all."

"The Witness: You and nobody else going ever stop me."

"The Court: Mr. Carey [defense counsel], anything further?"

"Mr. Carey: That is all."

Defense counsel immediately moved for a mistrial, which was denied. The defense then qualified Dr. E. Y. Williams, a psychiatrist, as an expert witness. Responsive to hypothetical questions predicated upon Stewart's army record, the various instances of odd behavior testified to by the previous lay witnesses, and the circumstances of Honikman's killing, Dr. Williams gave his professional opinion that Stewart was, at the time of the killing, suffering from both a mental defect and a mental disease. He explained in detail the psychiatric significance of Stewart's intelligence quotient of sixty-five, a rating which, he told the jury, would characterize Stewart as a moron. He further typified Stewart's mental disease as manic-depressive psychosis and, by the use of a blackboard, diagrammed and described the cyclic character of that disease. He testified that his own examination of the defendant in 1953 had yielded insufficient personal history to base a diagnosis, but that he had examined Stewart on several occasions since that time and found nothing which would change his opinion that Stewart was a manic-depressive psychotic. Dr. Williams was cross-examined at length on the afternoon of the third and the morning of the fourth days of the trial.

The remaining three trial days were taken up, in large part, by the testimony of seven government witnesses put forward to rebut Stewart's defense of insanity. Two psychiatric experts testified that they had examined

Page 366 U. S. 17

Stewart shortly after the killing in 1953 and found no mental defect or disease. A neighbor and friend of Stewart's who had known him for six years and seen him regularly during at least three years preceding 1953 testified that, on the basis of Stewart's conduct in his presence, he believed that Stewart was normal. An attendant at Saint Elizabeths Hospital, where Stewart had been committed during late 1957 and early 1958, described Stewart's behavior there as that of a model patient who had caused no specific trouble, gotten along with others, played cards and checkers, been seen with a Bible, etc. A police lieutenant at the District of Columbia jail similarly related Stewart's activities at the jail over the four years between the killing and the present trial. Through this witness, there were put in evidence as exhibits portions of the jail file tending to show that Stewart had signed certain forms, made certain written requests, and sent numerous letters to his wife and sister-in-law. A third psychiatric expert, who had examined Stewart early in 1958, testified that he found no evidence of mental disease, and did not regard Stewart as a mental defective. A fourth testified, on the basis of two examinations made in 1958, that the defendant was not a manic-depressive psychotic. Both of these psychiatrists agreed that Stewart was malingering at the time of their examinations.

It is unnecessary to describe in greater detail here the testimony of these seven government witnesses. All were cross-examined, two of the experts at considerable length. On the sixth trial day, counsel for the Government and for the defense addressed the jury. Neither in these exhaustive closing statements nor in the court's extended charge was any reference made to the two questions, asked several days before and, in effect, unanswered, which are now assigned as prejudicial error. The jury retired, deliberated, and found the defendant guilty.

Page 366 U. S. 18

On the totality of this record, with solicitous regard for the heavy obligation which rests upon us in a capital case, I cannot but conclude that the prosecutor's questions concerning Stewart's prior failures to testify are of that class of errors "which do not affect the substantial rights of the parties," and which, therefore, this Court, by virtue of an Act of Congress, is under duty to disregard. 40 Stat. 1181 (1919), in its present form 63 Stat. 105, 28 U.S.C. § 2111. This is so in light of a number of considerations, none of which, viewed in isolation, might be determinative, but whose sum -- in the whole context of the trial -- convinces me that the Court's conjectures of prejudice are chimerical.

First, Stewart never intelligibly answered the questions. The jury was not told, and did not know as a fact, that he had not previously taken the stand. The Court now finds that the jury may nevertheless have inferred the information from the leading form of the prosecutor's questions. But this conclusion should not be reached merely on the basis of the broad generalization that "such an inference will in all likelihood be drawn from leading questions of this kind." Such an abstraction does not get us to the heart of the question before us. That question, in one aspect, is whether it is likely that this jury, in the circumstances of this case, drew the inference from this leading question. It is not only not likely, but overwhelmingly unlikely.

The question was not pressed or persisted in by the prosecutor so as to concentrate the jury's attention on it as an assertion of fact. It was once repeated -- when Stewart asked "What?" -- and then dropped. It was asked in a setting in which it is not to be assumed, because most improbable, that the jury took in and paid heed to the content of the prosecutor's questions as such, particularly the one now so inflated in importance. On the

Page 366 U. S. 19

stand was a witness who had just testified that he was the master and the monkey with the tail, and that he had been told by God to conquer and kill. His responses appeared raving and incoherent. The only significance of his testimony, of course, was his demeanor, and it was upon the manner and character of his responses, not upon the subjects inquired into, that the jury can plausibly be supposed to have focused. The offending question followed a series of others -- "You can see me, can't you, Willie?" " . . . Willie, do you know a Dr. Williams?" "Do you know a Deputy Marshal by the name of Ballinger?" -- which had absolutely no significance of content, except insofar as they prodded the witness to respond. There is no reason to think that the jury could have regarded the questions concerning previous failure to testify any differently, or attributed special significance to them. In any event, assuming that the jury were given to pondering subtle inferences in the face of this manifest madman, they could have learned no more from the prosecutor's questions than what Stewart's own counsel had already elicited. The jury knew that this defendant had been tried before, because testimony from prior trials had been read to them. Yet defense counsel asked Stewart on direct examination: "Have you ever taken an oath?" and Stewart answered: "Not that I knows of."

Even had the jurors not been absorbed by the eye-catching spectacle of Stewart on the stand, and even had the unanswered questions been answered, the inference attributed to the jury by the Court would hardly have been a probable one. For the prejudice which the Court conceives does not arise from the simple knowledge that Stewart had not previously testified. It arises only upon the supposition that the jury indulged conjectures concerning the reasons for his not testifying, and upon the

Page 366 U. S. 20

further supposition that, in the course of those conjectures, it rejected alternatives favorable to the defense -- for example, that Stewart, being insane, capriciously refused to go on the stand -- and fixed on the explanation that Stewart was sane at the time of the earlier trials. Perhaps, were there nothing else in this case, this chain of suppositions might be entertainable. But the weakness of its links is one more factor making it implausible to find prejudice here.

Finally, these two concededly impermissible questions -- more accurately, a single question once repeated at the witness' request -- must be viewed in the perspective of the proceedings as a whole. Asked and left unanswered on the third day of a six-day trial at which eighteen witnesses testified and the testimony of eight more was read to the jury, the questions were never again adverted to. They had been preceded by a series of what the jury cannot but have found startling accounts of Stewart's behavior, were contemporaneous with a glaring display of the symptoms of madness, and were followed by a two-day battle of expert witnesses -- one accoutered with blackboard and chalk -- all addressed to the question of Stewart's sanity. It weaves solidities out of gossamer assumptions to attribute to fleeting and argumentative implications of fact in a leading question an impact so ponderous as to discredit and reverse a jury's verdict in the context of a record that impressively carries the contrary meaning. The jury was not left to pick at such threads in order to weave the cords of its verdict. On both sides -- by both the prosecution and the defense -- strong, heavy cables were furnished it. To suppose that, even if noticed when asked and made the occasion of implausible deductions, these questions amounted to more than a whisper drowned in the compulsion of ear-resounding testimony, seems to me a striking example of pursuing a quest for error.

Page 366 U. S. 21

More than a half-century ago, William H. Taft, reflecting his wide experience even before he became Chief Justice, laid this charge at the door of the courts:

". . . The . . . disposition on the part of the courts to think that every provision of every rule of law in favor of the defendant is one to be strictly enforced, and even widened in its effect in the interest of the liberty of the citizen, has led courts of appeal to a degree of refinement in upholding technicalities in favor of defendants, and in reversing convictions that render one who has had practical knowledge of the trial of criminal cases most impatient."

". . . When a court of highest authority in this country thus interposes a bare technicality between a defendant and his just conviction, it is not too much to charge some of the laxity in our administration of the criminal law to a proneness on the part of courts of last resort to find error and to reverse judgments of conviction. [Footnote 2/5]"

I am convinced that today's decision falls within these weighty strictures. To explain the jury's rejection of Stewart's sole defense of insanity, with its consequent finding of guilt, on the ground, as a matter of assumption, that the jury was influenced by the two questions on which the verdict is reversed here is to show less respect for the jury system than do the opponents of the system. [Footnote 2/6] One does not have to accept all the encomia which opinions of this Court have showered on the jury's functions and values not to attribute fecklessness to the twelve men and women chosen to sit in this murder case. To make

Page 366 U. S. 22

such attribution is to be unconsciously betrayed, as sophisticates sometimes are, into a depreciation of the capacities of the run of men. I dissent from the judgment of the Court.

[Footnote 2/1]

Honikman's daughter took the stand and testified at the trial. A transcription of her mother's testimony at a previous trial, corroborating the daughter's account of the killing, was read to the jury.

[Footnote 2/2]

This remark was made at the bench, out of the hearing of the jury.

[Footnote 2/3]

In addition to the testimony of Mrs. Honikman, that of two other witnesses was read to the jury. The remaining seven appeared at this trial.

[Footnote 2/4]

Three of these eight witnesses took the stand. In the case of the other five, excerpts from their testimony at prior trials were read.

[Footnote 2/5]

Taft, The Administration of Criminal Law, 15 Yale L.J. 1, 15 (1905).

[Footnote 2/6]

See, e.g., Frank, Courts on Trial (1949), cc. VIII, IX.

MR. JUSTICE CLARK, with whom MR. JUSTICE WHITTAKER joins, dissenting.

It may be that Willie Lee Stewart "had an intelligence level in the moronic class," but he can laugh up his sleeve today, for he has again made a laughingstock of the law. This makes the third jury verdict of guilt -- each with a mandatory death penalty -- that has been set aside since 1953. It was in that year that Willie walked into Harry Honikman's little grocery store here in Washington, bought a bag of potato chips and a soft drink, consumed them in the store, ordered another bottle of soda, and then pulled out a pistol and killed Honikman right before the eyes of his wife and young daughter. The verdict is now set aside because of some hypotheticals as to what the jury might have inferred from a single question asked Willie as to whether he had testified at his other trials. In my view, none of these conjectures is sufficiently persuasive to be said to cast doubt on the validity of the jury's determination. Let us first review the setting of the fatal question in the trial.

The jury heard evidence for six days and from some 26 witnesses. The printed record here, which is only partial, consists of 400 pages. Willie Stewart's "gibberish" comprises nine pages, representing perhaps some 20 minutes of testimony. It came during the third day of the trial. Mr. Carey, Willie's counsel, had placed him on the stand. He had asked on direct examination, "Have you ever taken an oath?" Willie replied, "Not that I knows of." Willie was also asked by his counsel, "Did you ever stand trial before this trial for the murder of Harry Honikman?" He answered, "Well, you talk. You

Page 366 U. S. 23

just go ahead and explain yourself. Have you ever stand trial? Go ahead. Don't ask me. I don't know." Mr. Carey had not represented Willie on the other trials. Carey then asked, "Were you ever tried for first degree murder before this time?" And Willie replied, "I ain't never been tried. I ain't never been tried." With these openings made by Carey, the Government, on cross-examination, asked the same questions. No issue is made of the examination relating to the fact of prior trials. Then came the question which has brought on this reversal: "This is the first time you have gone on the stand, isn't it, Willie?" There was no objection. Willie answered, "What?" And the Government's counsel again asked the same question in identical words. Still there was no objection. Willie answered: "I am always the stand; I am everything, I done told you." Thereafter Willie was excused as a witness, whereupon his counsel approached the bench and made his motion for mistrial. He asked for no curative instruction. Counsel had set his trap, lain in wait, and was now demanding all or nothing. The demand for a mistrial was denied.

A government witness then testified that, on the very night of the murder, Willie was playing cards, that he exhibited the pistol used in the slaying to one of the players, that he left the card game before the hour of the murder, and that he returned to the card game after the hour of the murder and continued playing cards until about 2 a.m. This witness testified, "he [Willie] seemed normal to me." This was followed by testimony of an aide at St. Elizabeths Hospital and a guard at the District jail as to his conduct all during the period after his arrest up until a few weeks before his third trial. All said that he was perfectly normal; that he talked freely and understood the conversation; that he used a Bible and a dictionary, played bid whist and checkers, and was a "model" patient or prisoner. His jail file revealed that

Page 366 U. S. 24

he mailed letters to his wife and sister-in-law, both of whom testified in his behalf, during April, October, and November, 1953; July, August, September, and October, 1954; October, November, and December, 1955; January, February, and March, 1956; and October, November, and December, 1957; and forwarded his wife $10 on each of two occasions, once in 1954 and the other in 1955. On several occasions, he sent memo requests for conferences with jail officials. He asked for work to pass the time while in the District jail, and actually put in many hours working day-in and day-out during the time of his custody. He first did cleaning, then plumbing, and finally was continually engaged in painting cell blocks throughout the jail. In 1957, his son was ill, and he requested permission, which was granted, to visit him in custody. These witnesses all related that Willie "acted normal" during this period. In fact, his only expert witness, a psychiatrist, testified that he could not decide in June, 1953, when he examined Willie whether or not he was suffering from a mental disease. However, he stated that, after talking with Willie's sister-in-law and hearing the story of Willie's background, he decided that Willie suffered a manic-depressive psychosis. The three government psychiatrists, two of whom examined him in March, 1953, found him "perfectly normal." He answered their questions freely, went through various tests cooperatively, and was found to be in "average normal range of intelligence." Each agreed that Willie was later malingering, i.e., feigning mental illness. This began shortly before his third trial. In addition, Willie had served two enlistments in the Army before 1953. On discharge, he was found "illiterate but mentally adequate."

In the light of this testimony, I find the hypotheses of the Court, with due deference, entirely unrealistic, if not

Page 366 U. S. 25

completely absurd. The crucial date was the time of the killing, 1953, not the date of the third trial, 1958. Despite this and the uncontradicted evidence, detailed above, of Willie's normality all during the period 1953-1958, the Court assumes that, from the asking of the question by the prosecutor, the jury believed that Willie had not testified in the two prior trials, and therefore the jury "might" have inferred that (1) Willie "elected to feign this "testimony" (gibberish) out of desperation brought on by his failure to gain acquittal" previously; or (2) the jury "might have thought" Willie suffered from a mental disease but "concluded that it is unlikely that a disease that had manifested itself only one out of three times for exhibition at trial was active at the occasion of the homicide"; or (3) the jury "might have thought" that the condition was worsening as indicated by his action at the trial.

In the first place, it seems to me a violent assumption to say that the jury believed, solely from the Government's question on cross-examination, that Willie had not testified at the prior trials, especially since he had already testified in response to a query from his own counsel on direct examination that he had never been under oath. Moreover, in opening up the issue of prior trials, the defense counsel was obviously trying to leave the impression with the jury that they had not concluded in guilty verdicts. When he received answers such as "you talk" -- "You just go ahead and explain" -- "Don't ask me," he repeated the question. And the government counsel got like answers to his questions: "I don't care how many occasions," etc. And the answer to the question found prejudicial was first a "What?" and, upon its repetition, "I am always the stand." Using the majority's speculative approach, it is the more likely that the jury thought from those questions that the previous trials resulted in hung

Page 366 U. S. 26

juries and never speculated upon the nice distinctions the Court makes as to Willie's demeanor. * The uncontradicted evidence was that he was a faker. They needed no inference to so conclude. Discounting the speculative effects of his own counsel's question on oaths, and the Government's question on testifying, his answers themselves might well have led the jury to believe that he did testify on the previous trials. In any event, a simple instruction to the jury to consider this trial alone, to strike our of its minds and give no consideration whatever to any reference to a former trial or to any event or thing that might or might not have happened there, would have certainly been sufficient. But Willie did not ask for this. He wanted "all or none," and the Court is giving him "all." But, returning to the hypotheses, whether or not Willie "elected" to feign his testimony was not the question. The jury's concern was whether he did feign it, and the uncontradicted testimony was that he did so. Secondly, the only testimony as to Willie's activity on the very night of the killing was that of the card player. He stated that Willie "seemed normal to me." How the jury might infer from the prosecutor's question that Willie had a mental disease but it was inactive at the time of the murder is beyond me. Every witness testified to the contrary -- save one psychiatrist -- and even he said that his examination of Willie was inconclusive. The jury knew it had been five years since the killing, and that both lay and medical evidence -- uncontradicted -- was that Willie was normal during all that period. Lastly, as to the disease worsening, that possibility had no relevancy to the condition in 1953 at the time of the killing.

Page 366 U. S. 27

I might add that, as I read the Government's brief, it conceded only that the question asked Willie "was of but negligible importance to the government's case." The sole issue, it said, was whether the question was prejudicial. This does not license the Court to find other and further concessions as to the Raffel and Grunewald cases. Nor do I find the Government contending, in its point that no prejudice resulted from the question, that "the jury may not even have heard the improper question." To so state its attitude makes the Government appear ridiculous. Its true position was that one could not assume, as the Court does, that "the jury noted and focused attention on a question given so little emphasis that it was overlooked by the trial judge." I add that, in the light of the long trial, the uncontradicted evidence as to Willie's malingering, and the fact that the question was never mentioned again during the remaining three days of the trial, the jury did not need, nor as a matter of relevancy was it able, to go through the mental gymnastics the Court supposes.

I note that the Court does adopt one point made by the Government. It says "that it is regrettable when the concurrent findings of 36 jurors are not sufficient finally to terminate a case." I, too, agree with that, but, in view of the Court's approach, I would add that its regret is tempered by its willingness to indulge in such hypothesizing as to effectively remove from our law the concept of harmless error in capital cases.

* If there was any impression relating to Willie's failure to take the stand in prior trials, it was surely due to the questioning by his own counsel on the issue of oaths.