Ernest Benjamin Smith v. W. J. Estelle, Jr., Director, Texas Dept. Of Corrections

GOLDBERG, Circuit Judge: Ernest Benjamin Smith, Jr., the appellee in this case, participated in an armed robbery in which a person was killed. He was indicted for capital murder and convicted by a jury in Dallas County, Texas. Texas law provides that when a defendant is convicted of a capital crime, the court must immediately hold a hearing, before the same jury, to decide whether the defendant is to be sentenced to life imprisonment or put to death. The jury answers “yes” or *696 “no” to each of three questions; if it answers all the questions in the affirmative, the judge must impose a death sentence. The second question 1 — the only one involved in this appeal — is whether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society. Tex.Code Crim.Proc.Ann. art. 37.071 (Vernon Supp.1978). At the sentencing hearing in Smith’s case, a psychiatrist who had examined Smith testified for the state and strongly urged that Smith would commit acts of violence and would constitute a continuing threat to society. Smith was sentenced to death. After appealing through the state courts he petitioned for federal habeas corpus. Judge Porter of the United States District Court for the Northern District of Texas held that several particular aspects of the psychiatrist’s testimony violated Smith’s constitutional rights. He did not disturb Smith’s conviction but he ordered that the death sentence not be carried out. The state appeals. We affirm. I. Smith and Howie Ray Robinson robbed a convenience store in Dallas on September 28, 1973. Both carried guns. During the robbery, the clerk in the store “reached for something.” Smith called out to Robinson; one prosecution witness testified that Smith said “Watch out, Howie,” but another asserted that Smith had later recalled saying, “Get him, Howie.” Robinson shot and killed the clerk. Smith never fired his gun. There was testimony that Smith knew it was defective and could not be fired; there was also testimony that he had attempted unsuccessfully to fire it during the robbery. Smith and Robinson were both arrested and indicted for capital murder. Approximately six weeks later, while Smith was awaiting trial, the state trial judge asked the prosecutor to have Smith examined by a psychiatrist named James P. Grigson. The judge wanted Dr. Grigson to decide whether Smith was competent to stand trial. No one had ever questioned Smith’s competence; the trial judge ordered the examination entirely on his own motion. He later testified that he always ordered such an examination in a capital case. Dr. Grigson spoke to Smith for ninety minutes. Smith was never told that Dr. Grigson was assessing anything but his competence to stand trial, and he cooperated fully. Dr. Grigson did conclude that Smith was competent to stand trial, but he filed no formal report; instead he wrote a letter notifying the trial judge of his conclusion. Judge Porter found that Smith’s attorneys were never told that Dr. Grigson had examined their client. 2 At the beginning of the sentencing hearing, after Smith had been tried and convicted, the prosecution announced that it was resting “subject to reopening.” Smith presented three witnesses — his stepmother, his aunt, and the owner of the gun he used in the robbery — and rested. The prosecution then called Dr. Grigson. By this time, the defense attorneys had seen, in the court’s file of the case, Dr. Grigson’s letter advising the trial judge that Smith was competent, but they did not know that Dr. Grigson was connected to the case in any other way. In fact, before the trial the defense attorneys had obtained an order requiring the prosecution to disclose all the witnesses it planned to use in its case in chief. The state presented a list of witnesses, and the defense successfully moved that the prosecution be prohibited from presenting the testimony of any witnesses not on *697 the list. Dr. Grigson’s name was not on the list. When Dr. Grigson was called at the sentencing hearing, the defense objected and asked that the jury be excused. Defense counsel then conducted a voir dire examination of Dr. Grigson, in which he testified that he had never told the defense attorneys about his interview with their client. He acknowledged, however, that he had discussed his conclusions with the prosecution. He testified that at some time between his examination of Smith and the trial, one of the prosecutors had told him that he might be called to testify against Smith. Then five days before the sentencing hearing, at the very beginning of the trial, the prosecution told him that his testimony would be needed within the week. App. 122. At this point in the voir dire defense counsel moved that Dr. Grigson not be allowed to testify because his name had not been included on the prosecution’s witness list. The trial judge denied the motion. The trial judge also agreed that the jury could be told that it was he who had asked Dr. Grigson to examine Smith. The jury was so informed but was never told that the judge had appointed Dr. Grigson only to determine Smith’s competence to stand trial, not to decide if he was prone to commit acts of violence. Dr. Grigson’s testimony was extremely damaging to the defendant. He described his extensive qualifications and detailed the five phases of his ninety-minute interview with Smith. He had had no other contact with Smith or with any of Smith’s relatives, but on the basis of the interview he concluded that Smith was “a sociopathic personality ... a very severe sociopath.” App. 135, 136. He was “[absolutely . convinced that [Smith] is a severe sociopath on the far end of the sociopathic scale.” App. 141. Sociopaths, he said, “do not have a conscience that most of us develop at an early age.” App. 135. “They will tell the truth if it serves their purpose. If it’s harmful, whatever distortion of the truth is necessary is what they will use.” App. 136. And then: Q No conscience, no remorse, no guilt feelings? A No. He has none. Q Now, Dr. Grigson, what is your prognosis in this case? A Oh, he will continue his previous behavior — that which he has done in the past. He will again do it in the future. Q All right. So, were he released into society, I take it, then, you would not expect his behavior to differ from what it has been? A No. If anything it would only get worse. Q Now, Dr. Grigson, this sociopathic personality that Smith has — is this a condition that will improve with time? I guess — what I’m asking you is this: Is this a stage that he is passing through that he will grow out of? A No. This is not what you would consider a stage. This is a way of life. Just as you work every day, well his personality comes out in his behavior, but it is not a stage he is going through. It’s only something he will continue. Q You said get worse? A Yes. Q If it can? A Right. A . [I]t’s my opinion that really, Mr. Smith does not have any regard for another human being’s property or for their life, regardless of who it may be. This is what makes him such a very severe sociopath. He has complete disregard for another human being’s life. Q Dr. Grigson, does the field of medicine — perhaps psychiatric treatment and/or hospitalization of some sort — does that have anything beneficial to offer someone like Ernest Benjamin Smith? A We don’t have anything in medicine or psychiatry that in any way at all modifies or changes this behavior. We don’t have it. There is no treatment, no medicine. Nothing that’s going to change this behavior. *698 Q Now, Dr. Grigson, do you have an opinion as to whether or not there is a probability that the Defendant, Ernest Benjamin Smith, will commit criminal acts of violence that will constitute a continuing threat to society? A Yes, sir. I have an opinion as to that. Q And what is that opinion? A That certainly Mr. Smith is going to go ahead and commit other similar or same criminal acts if given the opportunity to do so. Q Now, Dr. Grigson, I believe you have stated that this man has no remorse or sorrow for what he has done? A No. He has none. App. 136-39. The prosecution called no other witnesses at the sentencing hearing. The jury answered “yes” to all three questions and Smith was sentenced to death. The Texas Court of Criminal Appeals affirmed Smith’s conviction and sentence. Smith v. State, 540 S.W.2d 693 (Tex.Cr.App. 1976). The United States Supreme Court, over the dissent of three justices, denied certiorari. 430 U.S. 922, 97 S.Ct. 1341, 51 L.Ed.2d 601 (1977). On federal habeas corpus, Judge Porter held that Dr. Grigson’s testimony violated Smith’s sixth and fourteenth amendment right to the effective assistance of counsel, his eighth amendment right to present mitigating circumstances, and his fifth amendment right not to be forced to incriminate himself. Smith v. Estelle, 445 F.Supp. 647 (N.D.Tex.1977). Judge Porter specified several rules that, he said, must be followed in the future to prevent defense attorneys from being unfairly surprised with psychiatrists’ reports about the dangerousness of their clients. 3 He held that the results of a psychiatric examination of the defendant’s dangerousness could not be admitted at the sentencing phase of a capital trial unless the defendant had been advised, before the examination, that he had a right to remain silent. Our approach is slightly different from Judge Porter’s, but we agree with him that Texas violated Smith’s constitutional rights, both in the way it introduced Dr. Grigson’s evidence and in the way it procured it. We shall address these two matters in turn. II. In Gardner v. Florida, 430 U.S. 349, 97 S.Ct. 1197, 51 L.Ed.2d 393 (1977), the Supreme Court reversed a death sentence because the judge who imposed it acted partly on the basis of information that was not disclosed to the defendant or his attorneys. The Court did not deny that in an ordinary case such a procedure might be acceptable. See id. at 357, 97 S.Ct. 1197, 1204 (plurality opinion). But “five Members of the Court have now expressly recognized that death is a different kind of punishment from any other which may be imposed in this country.” Id. For that reason the Gardner Court, balancing the benefits of withholding the information against the costs, found that the advantages of confidentiality to the state — obtaining information more easily, avoiding delay, and, if the defendant was not executed, preventing damage to his rehabilitation — were easily outweighed by the damage to “the interest in reliability.” Id. at 358-60, 97 S.Ct. 1197. 4 This damage *699 was caused by denying the “opportunity for . explanation or argument by defense counsel,” and for “counsel to challenge the accuracy or materiality of [the] information.” Id. at 356, 362, 97 S.Ct. at 1204. Gardner requires us to set aside Smith’s death sentence. The defect which the Supreme Court identified in Gardner — defense counsel’s inability to challenge or answer the evidence on which the death sentence was based, see id. — is conspicuous here as well. As a result Smith’s sentencing hearing was at least as unreliable as the proceedings in Gardner. And the justifications for the unreliable procedures were far weaker here. Surprise can be as effective as secrecy in preventing effective cross-examination, in denying the “opportunity for [defense] counsel to challenge the accuracy or materiality of” evidence, Gardner v. Florida, 430 U.S. at 357, 97 S.Ct. at 1204, and in foreclosing “that debate between adversaries [which] is often essential to the truth-seeking function of trials,” id. at 360, 97 S.Ct. at 1206. Smith’s attorneys were manifestly surprised by Dr. Grigson’s testimony. They were never notified, according to Judge Porter, the trial judge had ordered a psychiatrist to interview their client. 5 When he had completed his interview, Dr. Grigson filed no report. No copy of the letter stating his conclusions was ever sent to the defense attorneys. 6 And despite the trial court’s order that the state’s attorney list “all the State’s witnesses whom he in good faith expects to use at the trial of the State’s case in chief, both during the guilt or innocence state [sic] and the punishment stage,” App. 97, the prosecution did not list Dr. Grigson’s name. The consequences of this surprise were devastating. 7 Thus the *701 trial judge’s informality and the prosecution’s tactics combined to give Smith’s attorneys no chance to prepare an effective response to Dr. Grigson’s testimony or to impeach it in any significant way. Surprise in this ease impaired “the interest in reliability,” id. at 359, 97 S.Ct. 1197 as much as secrecy did in Gardner, 8 Indeed, the sentencing judge in Gardner relied upon a presentence report, prepared by the Florida Parole and Probation Commission, see Fla. R.Crim.Proc. 3.710-3.713, under conditions that presumably gave some assurance that the reported information would be accurate. Here we deal with testimony offered by one party and not effectively cross-examined by the other; it carries no assurance of reliability whatever. Similarly, in Gardner, the failure to disclose the information had a recognized, substantial justification; under the federal rules, for example, information in a presentence report can be kept from a defendant under certain circumstances. See Fed.R. Crim.P. 32(c)(3). And in some cases, procedures which allow the prosecution to surprise the defense may have a similarly *702 strong justification. 9 In this case, however, the price of avoiding surprise was, at most, the insignificant cost of more regular and formal procedures. Requiring prosecutors and judges to exercise greater care than is shown in this record will scarcely burden them unduly; the surprise in this case resulted, in the first instance, from several minor irregularities that could have been easily corrected. The state trial judge— whom we do not fault for trying to assure himself that the defendant was competent to stand trial — might have ordered the psychiatric examination directly, instead of having a prosecuting attorney contact Dr. Grigson. The trial judge or the prosecutors might have formally notified the defense attorneys of Dr. Grigson’s examination. Or the prosecution might have listed Dr. Grigson’s name as a prospective witness. Even if we assume that their failure to list Dr. Grigson was inadvertent, Judge Porter’s holding would follow a fortiori from Gardner; the gains from informality and relaxed procedures cannot possibly outweigh the risk that the state may execute a person who would not have been sentenced to death if the jury had had full and “accurate sentencing information” — “an indispensable prerequisite to a reasonable determination of whether- a defendant shall live or die.” Gregg v. Georgia, 428 U.S. 153, 190, 96 S.Ct. 2909, 2933, 49 L.Ed.2d 859 (1976) (plurality opinion). 10 But Judge Porter found that the prosecutors intentionally omitted Dr. Grigson’s name from the witness list. In other words, they intended to surprise the defense attorneys and leave them, at best, not fully prepared. The state claims that Judge Porter’s finding is clearly erroneous but it is plainly supported by the record. Dr. Grigson testified that a prosecuting attorney had alerted him, some time before the trial, that he might be called as a witness. At the beginning of the week-long trial, the same prosecutor asked him to stand ready because the prosecution “would probably need him to testify this week.” App. 122. Then at the start of the sentencing hearing, the prosecution, which was entitled to present witnesses first, rested, “subject to the right to reopen.” It is difficult to resist the conclusions that the prosecution was attempting to lay a foundation for the argument — ultimately accepted by the Texas Court of Criminal Appeals, see Smith v. State, 540 S.W.2d 693, 699 (Tex.Cr.App.1976), and apparently by the trial judge — that Dr. Grigson was a rebuttal witness, not a witness to be used on the case in chief, so his omission from the witness list did not violate the trial judge’s order. Whatever the merits of this view, 11 the prosecution’s unusual actions at the sentencing hearing only support Judge Porter’s conclusion that the prosecutors were deliberately attempting to surprise the defense *703 attorneys and to force them to examine Dr. Grigson while they were unprepared. 12 This is irresponsible conduct. It lacks even the meager benefits some might credit to informality. We have reversed a conviction in a non-capital case where the government’s similarly inexcusable surprise tactics left defense counsel unable to cross-examine meaningfully. See Riggs v. United States, 280 F.2d 750, 753-54 (5th Cir. 1960). And the Supreme Court has commented that even a defendant has no right to treat a criminal trial as “a poker game in which players enjoy an absolute right always to conceal their cards until played.” Williams v. Florida, 399 U.S. 78, 82, 90 S.Ct. 1893, 1896, 26 L.Ed.2d 446 (1970). At the sentencing phase of a capital trial, Gardner and other cases, see e. g., Woodson v. North Carolina, 428 U.S. 280, 304-305, 96 S.Ct. 2978, 49 L.Ed.2d 944 (1976) (plurality opinion), demand extraordinarily fair and reliable procedures. “[T]he penalty of death is qualitatively different from a sentence of imprisonment, however long. . . . Because of that qualitative difference, there is a corresponding difference in the need for reliability in the determination that death is the appropriate punishment in a specific case.” Woodson v. North Carolina, 428 U.S. 280, 305, 96 S.Ct. 2978, 2991, 49 L.Ed.2d 944 (1976) (plurality opinion); see Gardner v. Florida, 430 U.S. 349, 357-59, 97 S.Ct. 1197, 51 L.Ed.2d 393 (plurality opinion); id. at 363, 97 S.Ct. 1197 (White, J., concurring). To some extent, at least, a state’s decision to kill a person must be insulated from the vagaries of the criminal process. “It is of vital importance to the defendant and to the community that any decision to impose the death sentence be, and appear to be, based on reason rather than caprice or emotion.” Gardner v. Florida, 430 U.S. 349, 358, 97 S.Ct. 1197, 1205, 51 L.Ed.2d 393 (1977) (plurality opinion). For these reasons, the death sentence imposed on Smith cannot stand. III. No one ever told Smith that Dr. Grigson’s examination concerned more than Smith’s competence to stand trial. No one ever revealed that their ninety-minute interview would be used to gather the evidence that would send Smith to his death. These facts are undisputed, and they prompted Judge Porter to hold that the state’s using Dr. Grigson’s testimony would have violated Smith’s fifth amendment right not to incriminate himself even if there had been no surprise. More specifically, Judge Porter ruled that “if the State or Court seeks to have the Defendant examined by the psychiatric expert on the issue of dangerousness, the Defendant must be advised he has a right to remain silent. If the Defendant indicates that he wishes to exercise that right, he may not be questioned' by the psychiatrist for the purpose of determining dangerousness.” Smith v. Estelle, 445 F.Supp. 647, 664 (N.D.Tex.1977). Before we decide whether such warnings are required, however, we must consider the logically prior issue of whether Smith even had a right to refuse to be examined by Dr. Grigson. 13 Ordinarily *704 there would be no question; under the fifth amendment a criminal defendant of course cannot be forced to discuss his alleged crime with anyone who is able to use his statements as evidence against him at his trial. But the state insists that when the defendant is facing jury deciding whether to impose a death sentence, this rule does not apply, at least so long as the interrogator is a psychiatrist. The state argues first that the evidence Smith gave Grigson is not “testimonial” and therefore not within the fifth amendment privilege. See Livingston v. State, 542 S.W.2d 655 (Tex.Cr.App.), cert. denied 431 U.S. 933, 97 S.Ct. 2642, 53 L.Ed.2d 250 (1977). The Supreme Court has held that a criminal defendant may be compelled to wear a piece of clothing in order to show that it fits him, Holt v. United States, 218 U.S. 245, 31 S.Ct. 2, 54 L.Ed. 1021 (1910), to give a blood sample, Schmerber v. California, 384 U.S. 757, 86 S.Ct. 1826, 16 L.Ed.2d 908 (1966), or a handwriting, Gilbert v. California, 388 U.S. 263, 87 S.Ct. 1951, 18 L.Ed.2d 1178 (1967), or voice exemplar, United States v. Dionisio, 410 U.S. 1, 93 S.Ct. 764, 35 L.Ed.2d 67 (1973), or to speak so that a witness may hear his voice quality or tone, United States v. Wade, 388 U.S. 218, 222-23, 87 S.Ct. 1926, 18 L.Ed.2d 1149 (1967). In these cases, the Court said, evidence given by the defendant was not “rélat[ed] to some communicative act,” Schmerber v. California, 384 U.S. at 765, 86 S.Ct. 1826, and was not “used . . . for the testimonial or communicative content of what was to be said,” United States v. Dionisio, 410 U.S. at 7, 93 S.Ct. at 768. Had Dr. Grigson drawn his conclusion from Smith’s manner or deportment, his attention span or facial expressions, a strong argument might be made that he gathered only evidence like that involved in these cases. If Dr. Grigson had been analyzing only the patterns of the defendant’s speech, his grammar, organization, logical coherence, and similar qualities, the question would be closer but arguably the fifth amendment would still not apply. But as we have said, Dr. Grigson’s diagnosis of Smith rested principally on his conclusion that Smith showed no remorse; obviously Dr. Grigson drew that conclusion from the content of Smith’s statements to him. Indeed, Dr. Grigson testified, on cross-examination, that the most important basis of his diagnosis was the account of the crime which Smith gave him during their interview. App. 142-43. Dr. Grigson accepted that account as accurate and drew his conclusions accordingly, App. 192; he also based his diagnosis on comments Smith made and failed to make while he was recounting the crime, App. 142-43. Plainly, then, Dr. Grigson — and therefore the prosecution, when it called him as a witness— used the content, not the non-testimonial aspects of Smith’s statements. The state also argues that United States v. Cohen, 530 F.2d 43 (5th Cir. 1976), entitles it to force Smith and other capital defendants to undergo a psychiatric examination designed to decide whether they are likely to commit violent crimes in the future. In Cohen, we held that a defendant can be required “to submit to an examination by psychiatrists selected by the government to determine his sanity at the time of the commission of the crime.” Id. at 47 (emphasis added). But far from supporting the state’s position, Cohen furnishes excellent arguments against it; our decision in Cohen rested on two premises, and neither applies when the compelled psychiatric examination concerns not the defendant’s sanity but his propensity to be violent in the future. The first premise was that a psychiatrist’s conclusions about á defendant’s sanity rested on the non-testimonial aspects of the defendant’s conduct during the examination, see 530 F.2d at 48; for the reasons we have just given, Dr. Grigson’s conclusions about the defendant’s future dangerousness were based almost entirely on the defendant’s statements. Indeed at times they appeared to be not just the fruits of Smith’s statements but summaries of them. 14 The second and more important *705 premise of Cohen was that if a defendant raises insanity as a defense, and introduces psychiatric testimony, “the government will seldom have a satisfactory method of meeting defendant’s proof on the issue of sanity except by the testimony of a psychiatrist it selects . . . who has had the opportunity to form a reliable opinion by examining the accused.” Id. at 48; see United States v. Albright, 388 F.2d 719, 724-25 (4th Cir. 1968). At most, this rationale of Cohen implies that a defendant who introduces the results of a psychiatric examination at the sentencing phase of a capital trial must submit to an examination by a psychiatrist nominated by the state. 15 But it is undisputed that Smith never used psychiatric evidence and never suggested that he might. This second premise is therefore also inapposite. In Cohen, the defendant himself raised the insanity defense and intended to present psychiatric evidence of his own; a psychiatric examination of such a defendant is utterly unlike an examination initiated by the state or the court, designed to determine not the defendant’s sanity but his future dangerousness, and offered in an affirmative effort to have the defendant sentenced to death, rather than as a response to the defendant’s own psychiatric evidence. For these reasons, Cohen does not suggest that evidence from a compelled psychiatric examination can be used by the state in its efforts to show the defendant’s future dangerousness. Preventing the prosecution from relying on compelled psychiatric examinations will scarcely disable it from proving the defendant’s dangerousness in appropriate cases. Many defendants may consent to examinations in the hope of being exonerated or of escaping a death sentence. As we have said, we leave open the possibility that a defendant who wishes to use psychiatric evidence in his own behalf can be precluded from using it unless he is willing to be examined by a psychiatrist nominated by the state. Moreover, the inquiry demanded by Texas’s second question is not obviously a matter for psychiatrists. 16 There seems to be a consensus among psychiatrists that any long-term predictions of future violence must be very tentative if, indeed, they can be made at all. See, e.g., American Psychiatric Association Task Force on Clinical Aspects of the Violent Individual, Clinical Aspects of the Violent Individual 23, 33 (1974); note 7 supra. Indeed, many in the psychiatric community apparently believe that nothing about a psychiatrist’s training or skill qualifies him to predict how dangerous a person will be in the distant future. See note 7 supra. Dr. Grigson evidently thinks otherwise, but even he conceded that the “sociopathic personality disorder” be diagnosed in Smith “is not an illness or a sickness. It’s simply a descriptive term that does describe an individual that has certain characteristics.” App. 135. Moreover, when Dr. Grigson explained how he had concluded that Smith was a sociopath, he often reasoned in ways that manifestly required no expertise. 17 *706 More important, however, is the legal meaning which the United States Supreme Court has, in effect, given to Texas’s second question. In Jurek v. Texas, 428 U.S. 262, 96 S.Ct. 2950, 49 L.Ed.2d 929 (1976), the Court refused to hold Texas’s capital sentencing statute unconstitutional on its face. Responding to arguments that the second question asked for a prediction that experts could not make, the Court said: [Prediction of future criminal conduct is an essential element in many of the decisions rendered throughout our criminal justice system. The decision whether to admit a defendant to bail, for instance, must often turn on a judge’s prediction of the defendant’s future conduct. And any sentencing authority must predict a convicted person’s probable future conduct when it engages in the process of determining what punishment to impose. For those sentenced to prison, these same predictions must be made by parole authorities. The task that a Texas jury must perform in answering the statutory question in issue is thus basically no different from the task performed countless times each day throughout the American system of criminal justice. Id. at 275-76, 96 S.Ct. at 2957-2958. While the Court did not disapprove the use of psychiatric testimony, quintessential^ then, see id. at 273, 96 S.Ct. 2950, it did say that the second question was constitutionally acceptable precisely because it could be resolved by non-experts. Moreover, the Court clearly held that the Texas procedure would have been unconstitutional if the second question did not present issues to which a psychiatrist’s expertise was unquestionably irrelevant. In Woodson v. North Carolina, 428 U.S. 280, 96 S.Ct. 2978, 49 L.Ed.2d 944 (1976) and Roberts v. Louisiana, 428 U.S. 325, 96 S.Ct. 3001, 49 L.Ed.2d 974 (1976), decided the same day as Jurek, the Court held that a capital sentencing procedure would violate the eighth amendment if it did not give a defendant an opportunity to present at least some mitigating circumstances. Texas’s procedure did not explicitly provide such an opportunity, but the Court concluded that Texas “will interpret this second question so as to allow a defendant to bring to the jury’s attention whatever mitigating circumstances he may be able to show.” Jurek v. Texas, 428 U.S. 262, 272, 96 S.Ct. 2950, 2956, 49 L.Ed.2d 929 (1976). On this understanding, the Court upheld the Texas procedure. Id. at 272-74, 96 S.Ct. 2950. In order to be constitutional, then, Texas’s second question must reach far beyond a psychiatrist’s expertise to issues that are quintessential the domain of lay persons. See Witherspoon v. Illinois, 391 U.S. 510, 519, 88 S.Ct. 1770, 1775, 20 L.Ed.2d 776 (1968) (“[A] jury that must choose between life imprisonment and capital punishment can do little more — and must do nothing less — than express the conscience of the community on the ultimate question of life or death.”) 18 *707 Finally, any attempt to compel defendants to submit to a psychiatric inquiry into their dangerousness is likely to be erratic and capricious. Such an examination evidently cannot succeed unless the defendant cooperates; Dr. Grigson for example, testified that Smith had cooperated with him. No one has suggested that a psychiatrist examining a defendant for future dangerousness can simply observe him, as a psychiatrist concerned with insanity might. Nor can a defendant be physically forced to cooperate with an examining psychiatrist, as he can, for example, with a doctor extracting a blood sample or a policeman who is fingerprinting him. Civil contempt is unlikely to move a defendant who is contemplating the sentencing phase of a capital trial. A defendant can be barred from pleading insanity if he did not submit to an examination by the state’s psychiatrist or psychologist; as we have said, we leave open the possibility that a defendant might be precluded from presenting psychiatric evidence of his own if he does not agree to be examined by a psychiatrist chosen by the state. But no such sanction is available against a capital defendant who, like Smith, proposes to use no such evidence. The state might, of course, be permitted to comment on the defendant’s refusal to be examined by a psychiatrist. But it is not clear what the jury is to infer; surely in most cases it would be unwarranted to conclude that the defendant knew the clinical details of his psychiatric state and did not want them discovered. And simply mentioning the defendant’s refusal to be examined and allowing the jury to draw its own conclusions might clash with the Supreme Court’s insistence that capital sentencing procedures be unusually reliable, see Part II supra. In any event, it is unlikely that the threat of such a remark by the prosecution will induce any well-informed defendant to submit to an examination by a psychiatrist who habitually testifies for the state and who is likely to give an emphatic affirmative answer to the precise question before the jury. The principal significance of allowing compelled examination on the question of dangerousness, then, will be in cases like Smith’s. If the state is entitled to compel a defendant to submit to an examination, it can, in an effort to gain the defendant’s *708 cooperation, mislead him or indeed lie to him about the significance of the examination; it can take advantage of his ignorance or lack of understanding. It can coerce him in any way that does not make his statements less useful to the interrogating psychiatrist. Psychological pressure, sharp practices, and deceit are likely to be, in effect, the means of compelling examinations. These tactics are inherently discriminatory. A knowledgeable defendant, or one with vigilant attorneys, will either simply refuse to submit to an examination or will bargain with the state to have the examination conducted by a psychiatrist who is more likely to favor the defense. Only defendants who do not know better will allow themselves to be examined by psychiatrists antecedently favorable to the state. We have every reason, therefore, to give effect to the apparent command of the fifth amendment and to hold that a defendant may not be compelled to speak to a psychiatrist who can use his statements against him at the sentencing phase of a capital trial. 19 If a state wishes to prove a defendant’s propensity to commit future crimes of violence by using evidence gathered at a psychiatric examination, the defendant must voluntarily consent to the examination. It follows that Judge Porter was correct when he held that if a defendant indicates that he wishes to remain silent, “he may not be questioned by the psychiatrist for the purpose of determining dangerousness.” Judge Porter also held that the defendant must be warned that he had a right to remain silent; since Smith was in custody when he was interviewed, this holding, too, was correct. Compare Schnekloth v. Bustamonte, 412 U.S. 218, 232, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973) with Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966). We also agree with Judge Porter’s holding that a defendant has no constitutional right to have an attorney present during a psychiatric evaluation of his dangerousness. Here Judge Porter followed United States v. Cohen, 530 F.2d 43 (5th Cir. 1976), which held that there was no right to have an attorney present when the examination was to decide if the defendant was sane. See id. at 48. Judge Porter reasoned, as we had in Cohen, that an attorney present during the psychiatric interview could contribute little and might seriously disrupt the examination. See id. In this he was correct. But Judge Porter overlooked the role that an attorney might have played in helping a client like Smith decide whether he wished to submit to an examination. This is a vitally important decision, literally a life or death matter. It is a difficult decision even for an attorney; it requires a knowledge of what other evidence is available, of the particular psychiatrist’s biases and predilections, of possible alternative strategies at the sentencing hearing. For a lay defendant, who is likely to have no idea of the vagaries of expert testimony and its possible role in a capital trial, and who may well find it difficult to understand, even if he is told, whether a psychiatrist is examining his competence, his sanity, his long-term dangerousness for purposes of sentencing, his short-term dangerousness for purposes of *709 civil commitment, his mental health for purposes of treatment, or some other thing, it is a hopelessly difficult decision. There is no reason to force the defendant to make it without “the guiding hand of counsel.” Powell v. Alabama, 287 U.S. 45, 57, 53 S.Ct. 55, 77 L.Ed. 158 (1933). The Supreme Court has spoken of the “vital need” for counsel “at the pretrial stage,” and has said that “whatever else it may mean, the right to counsel granted by the Sixth and Fourteenth Amendments means at least that a person is entitled to the help of a lawyer at or after the time that judicial proceedings have been initiated against him — ‘whether by way of formal charge, preliminary hearing, indictment, information, or arraignment.’ ” Brewer v. Williams, 430 U.S. 387, 398, 97 S.Ct. 1232, 1239, 51 L.Ed.2d 423 (1977), quoting Kirby v. Illinois, 406 U.S. 682, 689, 92 S.Ct. 1877, 32 L.Ed.2d 411 (1972). Smith, of course, had been indicted when Dr. Grigson examined him; indeed, an attorney had been appointed to represent him. And the Supreme Court has repeatedly held that a defendant is entitled to the assistance of counsel when he faces decisions that may have a crucial effect on his trial. See, e. g. Coleman v. Alabama, 399 U.S. 1, 90 S.Ct. 1999, 26 L.Ed.2d 387 (1970); Hamilton v. Alabama, 368 U.S. 52, 82 S.Ct. 157, 7 L.Ed.2d 114 (1961). See also Powell v. Alabama, 287 U.S. 45, 57, 53 S.Ct. 55, 77 L.Ed. 158 (1933). We therefore hold that at the sentencing phase of a capital trial, Texas may not use evidence based on a psychiatric examination of the defendant unless the defendant was warned, before the examination, that he had a right to remain silent; was allowed to terminate the examination when he wished; and was assisted by counsel in deciding whether to submit to the examination. Smith was denied these rights; for this reason, too, his death sentence must be set aside. 20 The judgment of the district court is AFFIRMED. Footnotes [1] The other two questions are: Whether the conduct of the defendant that caused the death of the deceased was committed deliberately and with the reasonable expectation that the death of the deceased or another would result; [I]f raised by the evidence, whether the conduct of the defendant in killing the deceased was unreasonable in response to the provocation, if any, by the deceased. Tex.Code Crim.Proc.Ann. art. 37.071 (Vernon Supp.1978). [2] See note 5 infra. [3] In the future to insure the right to effective cross examination and the right to effectively defend the case through expert rebuttal testimony, which the due process clause guarantees, defense counsel must be notified that a psychiatric examination will be held on the issue of “dangerousness” the results of which may be used at the penalty phase of the trial. Defense counsel prior to trial, must have meaningful access to the report, now required by state statute, prepared by the psychiatrist stating his findings and conclusions. If the trial court wishes to appoint a psychiatric expert on the issue of dangerousness the order of appointment should so reflect. The testimony of a court appointed psychiatrist and his report would then be equally available to both defense and prosecution for presentation to the jury. However, both sides should have a meaningful right to obtain additional psychiatric experts to either supplement or challenge the conclusions of the court appointed expert. [4] The principles of Gardner seem to be derived from both the eighth amendment and the due process clause of the fourteenth amendment. The Gardner plurality rested its opinion squarely on the due process clause. See 430 U.S. at *699 351, 97 S.Ct. 1197. But Justice White concurred in the judgment because he thought the eighth amendment compelled the result. See id. at 363, 97 S.Ct. 1197 (White, J., concurring). And Gardner was clearly adumbrated by Woodson v. North Carolina, 428 U.S. 280, 305, 96 S.Ct. 2978, 49 L.Ed.2d 944 (1976) (plurality opinion), which rested on the eighth amendment, and perhaps by discussions of the eighth amendment in Gregg v. Georgia, 428 U.S. 153, 190, 96 S.Ct. 2909, 49 L.Ed.2d 859 (1976) (plurality opinion) and Jurek v. Texas, 428 U.S. 262, 271, 96 S.Ct. 2950, 49 L.Ed.2d 929 (1976) (plurality opinion). [5] Both defense attorneys swore in affidavits that they were never notified that Dr. Grigson had examined their client. The state trial judge, in an affidavit, said “[a]s best I recall, I informed” one of the defense attorneys that Dr. Grigson had been appointed. Judge Porter held no evidentiary hearing but found that the defense attorneys had never been notified. On this appeal the state protests that the state judge’s affidavit conflicted with the attorneys’, and that Judge Porter could not properly resolve the conflict without an evidentiary hearing. But we need not decide whether the state is correct, because it is undisputed, first, that no formal notification was given; second, that so far as the judge knew, the only purpose of Dr. Grigson’s examination was to help decide if Smith was competent; and third, that the defense attorneys had no notice, and no reason to believe, that Dr. Grigson would ever be called to testify. [6] By happenstance, they saw a copy of the letter while inspecting the court’s files in the case, but the letter reflected that Dr. Grigson was appointed to evaluate the defendant’s competence to stand trial. Since the defense attorneys did not claim that Smith was incompetent, they naturally did not expect Dr. Grigson to testify. [7] The Supreme Court in Gardner did not attempt to detail the ways in which defense counsel might have explained or challenged the information that was kept confidential, cf. Davis v. Alabama, 596 F.2d 1214, 1221-23 (5th Cir. 1979) (in some circumstances, attorneys’ failure to prepare adequate defense vitiates state conviction only if it prejudiced defendant). But in this case it is clear that Smith’s defense attorneys — whose actual efforts to impeach Dr. Grigson were at best ineffectual, see, e. g., App. 139-41, 146-48, and sometimes damaging to their client, see, e.g., App. 141-43 might, with even minimal preparation, have met Dr. Grigson’s testimony in several ways. Had they been able to do even rudimentary research, they would have learned that a Task Force of the American Psychiatric Association concluded that predictions like Dr. Grigson’s “are fundamentally of very low reliability.” Psychiatric expertise in the prediction of “dangerousness” is not established and clinicians should avoid conclusory judgments in this regard. American Psychiatric Association Task Force on Clinical Aspects of the Violent Individual, Clinical Aspects of the Violent Individual 23, 33 (1974). The American Psychiatric Association itself, in an amicus brief filed in this case, advised us that Dr. Grigson’s “kind of inquiry *700 about long-term future violence essentially does not involve medical analysis, and is not within the realm of established psychiatric expertise.” Amicus Br. 10-11. According to one student of the literature, the most that any psychiatrist can predict is that a person “is within a group which, as a whole, poses a comparative greater risk of dangerous behavior than the general population.” Dix, the Death Penalty, “Dangerousness,” Psychiatric Testimony, and Professional Ethics, 5 Am.J.Crim.L. 151, 198 (1977). The APA Task Force agreed. “Predictions of ‘dangerousness’ are judgments of a ‘relative risk’ sort, statements of comparative probabilities that are usually quite low.” Task Force Report at 27. And, as Judge Porter detailed in his opinion, even such tentative predictions as these are not very reliable. Smith v. Estelle, 445 F.Supp. 647, 656-57 (N.D.Tex. 1977). Dr. Grigson, of course, was not so restrained. He made a “prognosis” that Smith “will continue his previous behavior — that which he has done in the past. He will again do it in the future . . . . If anything it will only get worse . . . . It’s only something he will continue . . . Nothing [is] going to change this behavior . . . [C]ertainly Mr. Smith is going to go ahead and commit other similar or same criminal acts if given the opportunity to do so.” App. 136-39 (emphasis added). Even a moderately well prepared defense attorney would have had some success in impeaching Dr. Grigson’s testimony or offering an opposing view. Moreover, Dr. Grigson reached his conclusions entirely, see App. 149, on the basis of a single ninety minute “mental status examination” of the defendant. The state, summarizing Dr. Grigson’s testimony, explains that “[f]rom this mental status examination, [Dr. Grigson] was able to draw opinions not only as to [Smith’s] present competency, but also as to his entire mental and emotional composition.” Appellant’s Br. at 26. One standard textbook flatly asserts that it is medically impossible to draw such conclusions on the basis of a mental status examination. Sands, Psychiatric History and Mental Status, in A. Freedman & H. Kaplan, eds., Comprehensive Textbook of Psychiatry 499 (1977). In addition, if the notions of sociopathy used by Dr. Grigson have any meaning at all — matter of some dispute in the literature, see, e. g. E. Sutherland & D. Cressey, Criminology 159-60 (8th ed. 1970) — then before a person may be labelled a sociopath, several elements must be found that Dr. Grigson did not identify in Smith. The American Psychiatric Association’s Diagnostic Manual defines “antisocial personality” in this way: This term is reserved for individuals who are basically unsocialized and whose behavior pattern brings them repeatedly into conflict with society. They are incapable of significant loyalty to individuals, groups, or social values. They are grossly selfish, callous, irresponsible, impulsive, and unable to feel guilt or to learn from experience and punishment. Frustration tolerance is low. They tend to blame others or offer plausible rationalizations for their behavior. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders 43 (2d ed. 1968). There was no evidence that Smith was prone to impulsive behavior, that he had a low “frustration tolerance,” or that he tended to blame others or to rationalize his failures. And there was clear evidence that Smith had not come “repeatedly into conflict with society”; he had no history of antisocial behavior, and his only previous criminal conviction was for possessing less than a matchbox full of marijuana. One commentator, discussing Dr. Grigson’s testimony in this case, concluded that “the views he expressed stood — generously analyzed — at the fringes of professional acceptability” and that Dr. Grigson “operate[d] at the brink of quackery.” Dix, supra at 172. Whatever the merit of that view, much of this material, easily available, might have been used to impeach or answer Dr. Grigson’s conclusions if Smith’s attorneys had had some reasonable time to prepare. To be sure, as the state points out, Smith’s defense attorneys did make some effort to suggest tnat Dr. Grigson was biased toward the prosecution. Here again, however, their lack of preparation made them ineffective. Smith’s lawyer could ask only broad, argumentative questions. See, e.g., App. 139 (“Dr. Grigson, you’re kind of the hatchet man down here for the District Attorney’s Office, aren’t you?”); App. 154 (“Did you ever meet a person you didn’t think was a sociopath?”). Again, the most basic research would have disclosed some evidence of bias that might have been used— and under Texas law, see, e.g., Castro v. State, 562 S.W.2d 252 (Tex.Cr.App.1978); Evans v. State, 519 S.W.2d 868 (Tex.Cr.App.1975); Smith v. State, 516 S.W.2d 415 (Tex.Cr.App.1974); Wood v. State, 486 S.W.2d 359 (Tex.Cr.App.1972), could have been used — to impeach Dr. Grigson. For example, even in reported, appellate cases, Dr. Grigson has repeatedly testified for the state, see, e.g., Bruce v. Estelle, 536 F.2d 1051, 1054-56 (5th Cir. 1976); Chambers v. State, 568 S.W.2d 313, 324-327 (Tex.Cr.App.1978); L. L. S. v. State, 565 S.W.2d 252, 256-57 (Tex.Cr.App.1978); Hughes v. State, 562 S.W.2d 857, 863-64 (Tex.Cr.App.1978); Jackson v. State, 551 S.W.2d 351, 359 (Tex.Cr.App.1977) (dissenting opinion); Graham v. State, 546 S.W.2d 605, 609 (Tex.Cr.App.1977); Moore v. State, 542 S.W.2d 664, 676 (Tex.Cr.App.1976); Livingston v. State, 542 S.W.2d 655, 661 (Tex.Cr.App.1976); Gholson v. State, *701 542 S.W.2d 395, 400-01 (Tex.Cr.App.1976); Hicks v. State, 525 S.W.2d 177, 180 (Tex.Cr.App.1975); Hurd v. State, 513 S.W.2d 936, 944 (Tex.Cr.App.1974); Culley v. State, 505 S.W.2d 567, 569 (Tex.Cr.App.1974); Armstrong v. State, 502 S.W.2d 731, 735 ’(Tex.Cr.App.1974); Boss v. State, 489 S.W.2d 580, 582 (Tex.Cr.App.1972); Blankenship v. State, 432 S.W.2d 945, 946-47 (Tex.Cr.App.1968). He has not appeared in the report of any case as a witness for the defense. On many occasions he has declared that a person he examined was a sociopath or was otherwise likely to commit crimes in the future. See, e.g., Bruce v. Estelle, 536 F.2d 1051, 1054-56 (5th Cir. 1976); Chambers v. State, 568 S.W.2d 313, 324-327 (Tex.Cr.App.1978); L. L. S. v. State, 565 S.W.2d 252, 256-57 (Tex.Cr.App.1978); Hughes v. State, 562 S.W.2d 857, 863-64 (Tex.Cr.App.1978); Jackson v. State, 551 S.W.2d 351, 359 (Tex.Cr.App.1977) (dissenting opinion); Moore v. State, 542 S.W.2d 664, 676 (Tex.Cr.App.1976); Livingston v. State, 542 S.W.2d 655, 661 (Tex.Cr.App.1976); Gholson v. State, 542 S.W.2d 395, 400-01 (Tex.Cr.App.1976); Hurd v. State, 513 S.W.2d 936, 944 (Tex.Cr.App.1974); Armstrong v. State, 502 S.W.2d 731, 735 (Tex.Cr.App.1974). Frequently he reached this conclusion after he was assigned to examine only for competence or sanity. See, e.g., Bruce v. Estelle, 536 F.2d 1051, 1054-56 (5th Cir. 1976); Livingston v. State, 542 S.W.2d 655, 661 (Tex.Cr.App.1976); Gholson v. State, 542 S.W.2d 395, 400-01 (Tex.Cr.App.1976); Hurd v. State, 513 S.W.2d 936, 944 (Tex.Cr.App.1974); Armstrong v. State, 502 S.W.2d 731, 735 (Tex.Cr.App.1974). We dealt with Dr. Grigson’s testimony in Bruce v. Estelle, 536 F.2d 1051 (5th Cir. 1976). There a habeas corpus petitioner claimed that he had been incompetent to stand trial. Dr. Grigson testified that the petitioner had been competent and was in fact a sociopath, whose apparent symptoms of incompetence were actually part of an attempt to manipulate others. The federal district court relied on Dr. Grigson’s testimony and denied relief. We reversed, holding that the reliance on Dr. Grigson’s testimony was clear error in view of other psychiatrists’ views and their bases. [8] The state seems to suggest that Smith cannot claim unfair surprise because his attorneys did not attempt to gain more time to prepare by moving for a continuance. At best, this amounts to a claim that Smith forfeited his right to object to the surprise because he did not raise his objection in the proper form. There are abundant reasons to reject this position. First, the state points us to no rule of Texas law saying that moving for a continuance is the only way to object to surprise. Even if there were such a rule, the gross disparity between the actual cross-examination conducted by Smith’s attorneys, see App. 139-52, 153-54, and the cross-examination they might have conducted if they had requested and received time to prepare, see note 7 supra, suggest that we should be prepared to excuse the defense attorneys’ procedural default in order to avoid “a miscarriage of justice.” Wainwright v. Sykes, 433 U.S. 72, 97 S.Ct. 2497, 2508, 53 L.Ed.2d 594 (1977). See generally Jurek v. Estelle, 593 F.2d 672, 680-85 (1979), reh. en banc granted, 597 F.2d 590 (1979). See also Smith v. Estelle, 445 F.Supp. 647, 658 n. 16 (suggesting that if admitting Dr. Grigson’s testimony was not error, defense counsel provided ineffective assistance in rebutting it). Even if we cannot excuse the forfeiture in this way, however, Judge Porter held that the state had waived any claims that Smith committed a procedural default, see Smith v. Estelle, 445 F.Supp. 647, 659 (N.D.Tex.1977), and as we say, see note 19 infra, we are not inclined to disturb his decision. More fundamentally, however, the defense attorneys’ failure to move for a continuance does not distinguish this case from Gardner. Here Smith’s attorneys at least objected to Dr. Grigson’s appearing as a witness; in Gardner, the attorneys never asked the trial judge to disclose the confidential information. The Supreme Court invalidated the death sentence nonetheless. [9] For example, concealing a witness’s identity from the defense may, in certain circumstances, be the only way to prevent the witness from being intimidated. [10] The appellant insists that Smith did not exhaust his state remedies on this issue. This is a mystifying claim. In its answer to Smith’s federal habeas petition the state expressly abjured any contention that Smith had failed to exhaust. App. 45-46. In its reply brief on this appeal the state concedes that Smith’s petition for federal habeas corpus exactly duplicated the allegations in his state appeal and' state petition for habeas corpus. Appellant’s Reply Br. at 2. On direct appeal, Smith had claimed that Dr. Grigson’s surprise testimony denied him a fair trial, see App. 159-60; Smith v. State, 540 S.W.2d 693, 699 (Tex.Cr.App.1977). In his state habeas petition he asserted: The introduction of the testimony of Dr. Grigson at petitioner’s trial violated his privilege against self-incrimination and his rights to fair notice and the effective assistance of counsel secured by the Fifth, Sixth and Fourteenth Amendments to the Constitution of the United States and said that Dr. Grigson’s surprise appearance “prevented petitioner’s trial counsel from effectively defending against the charge that there existed a probability that petitioner would commit criminal acts of violence that would constitute a continuing threat to society.” App. 220. That Smith did not anticipate the precise arguments Judge Porter used in support of his decision to grant relief obviously does not mean he failed to exhaust his remedies. [11] The defense presented no psychiatric evidence or expert testimony of any kind, so it is ^ not clear what Dr. Grigson was rebutting. In any event, the prosecution rested its case in brief subject to the right to reopen — that is, presumably, to reopen its case in chief — and said nothing about rebuttal evidence. In this way the state avoided having to explain exactly whom or what Dr. Grigson’s testimony was offered to refute. [12] The state’s explanation, in its brief on this appeal, is that the prosecutors rested because they knew Dr. Grigson would not be available to testify until later in the penalty trial. This suggestion lacks support in the record and, in any event, is wholly immaterial; it does not make Dr. Grigson a rebuttal witness, see note 11 supra, and obviously it does not excuse the state’s failure to list him. [13] If Texas retried Smith, Part II of this opinion, standing alone, would permit it to use the same testimony from Dr. Grigson, based on the same examination of Smith, provided it gave adequate notice to the defense attorneys. As Judge Porter’s opinion thoroughly shows, however, there is a serious question about whether Dr. Grigson’s testimony was obtained by violating Smith’s fifth and sixth amendment rights; if it was, of course, it is not admissible in any event. Under Texas law, evidence like Dr. Grigson’s in this case is apparently admissible, see, e.g., Livingston v. State, 542 S.W.2d 655 (Tex.Cr.App.), cert. denied, 431 U.S. 933, 97 S.Ct. 2642, 53 L.Ed.2d 250 (1977), so if Smith were convicted or sentenced on the basis of that evidence another federal collateral proceeding, raising the fifth and sixth amendment issues, would be inevitable. Since the issues have been briefed and argued, and were decided by Judge Porter, we see no reason not to resolve them here. [14] The crucial portions of Dr. Grigson’s testimony were his statements that Smith showed “no guilt feelings, no remorse, no sorrow,” and the like; these may be not just conclusions *705 drawn from Smith’s statements but actually reports or summaries of the statements themselves. In either case, of course, the fifth amendment would require Dr. Grigson’s evidence to be suppressed if Smith’s statements were not voluntary. [15] Judge Porter imposed such a rule. See Smith v. Estelle, 445 F.Supp. 647, 663 (N.D. Tex. 1977). Because Smith did not attempt to use any psychiatric evidence, however, we need not decide whether Judge Porter’s quite plausible view is correct. [16] In this way, too, the inquiry under Texas’s second question differs sharply from the issue raised by an insanity defense. This Circuit’s law on insanity is, in relevant part, A person is not responsible for criminal conduct if at the time of such conduct as a result of mental disease or defect he lacks substantial capacity either to appreciate the wrongfulness of his conduct or to conform his conduct to the requirements of law. Blake v. United States, 407 F.2d 908, 916 (5th Cir. 1969) (en banc). The very language of this definition — “mental disease or defect,” “substantial capacity,” and the like — point to the role of the psychiatrist or psychologist. [17] Q Would you say, being a psychiatrist, that a person could have been frightened or in a daze or state of shock, and that would have caused him to act in such a way as Ernest Benjamin Smith did? A Well, certainly. Yeah. I think if I had shot somebody I would certainly be in a daze *706 or state of shock. But, you know, I think that if I were in that sort of situation I would probably be frightened and run like the devil. I wouldn’t step over the body to get the money. I would want to get out of there. But an individual that is involved in, say, a goal directed fashion certainly is not functioning in a daze or shock type state. Q Would it make any difference in your evaluation of Ernest Benjamin Smith if you thought or believed that he felt that he was not responsible for Mr. Moon’s death? Would that make any difference with regard to his remorse? A But, you know, he was. He at least stated he picked out the store. You know, he chose it. He was in there, so he was responsible. So, I feel like that — you know, certainly he should feel remorse. App. 143 — 44, 151-52. [18] In fact, the logic of the Court’s decision in Jurek suggests that the second question must be interpreted to mandate an even broader inquiry to which a psychiatrist’s expert knowledge is even less relevant. The Court held that the death penalty would be unconstitutional if it were “so totally without penological justification that it resulted] in the gratuitous infliction of suffering.” Gregg v. Georgia, 428 U.S. 153, 96 S.Ct. 2909, 2929-30, 49 L.Ed.2d 859 (1976). But, the Court said, capital punishment as Texas administers it “is not without justification and thus is not unconstitutionally severe,” id. at 187, 96 S.Ct. at 2931, see Jurek v. Texas, 428 U.S. 262, 274, 96 S.Ct. 2950, 49 L.Ed.2d 929 *707 (1976). The Court identified two justifications, “retribution and deterrence of capital crimes by prospective offenders.” Gregg v. Georgia, 428 U.S. at 183, 96 S.Ct. at 2929. But it is not clear why killing offenders who are likely to “commit criminal acts of violence that would constitute a continuing threat to society” is more likely to deter others in the future, see The Supreme Court, 1975 Term, 90 Harv.L.Rev. 56, 71-72 (1976), than an equal number of random executions — a system that the Court presumably would not tolerate, see Furman v. Georgia, 408 U.S. 238, 92 S.Ct. 2726, 33 L.Ed.2d 346 (1972). Executing particularly “dangerous” criminals would serve to prevent them from committing crimes in the future, but the Court declined to rely on this sort of prevention as a justification for capital punishment, Gregg v. Georgia, 428 U.S. at 183 n. 28, 96 S.Ct. 2909, and the reason is apparent: a desire to prevent recidivism does not justify a legislature’s decision to kill an offender rather than to imprison him for life with no possibility of parole, and the Court has made it clear that the unique severity of capital punishment makes it necessary to justify that choice. See, e.g., id.; Furman v. Georgia, 408 U.S. 238, 312-13, 92 S.Ct. 2726, 33 L.Ed.2d 346 (1972) (White, J„ concurring). It follows, then, that Texas’s. second question must be seen as an effort to promote the goal of retribution, or “an expression of society’s moral outrage at particularly offensive conduct.” Gregg v. Georgia, 428 U.S. at 183, 96 S.Ct. at 2930. The Court explained these notions by saying that “the decision that capital punishment may be the appropriate sanction in extreme cases is an expression of the community’s belief that certain crimes are themselves so grievous an affront to humanity that the only adequate response may be the penalty of death.” Id. at 184, 96 S.Ct. at 2930. The Court must have been saying, then, that the justification for imposing the death sentence on particularly dangerous offenders is that it allows the jury to express its belief, on behalf of the community, that killing the defendant is “the only adequate response.” If this is the task which Texas’s second question presents to the jury, then the jury must, in answering that question, simply ask itself whether the defendant deserves to die. A psychiatrist’s diagnosis and predictions are apparently relevant to this horrifying inquiry, but only peripherally, and forcing the government to rely on lay evidence — at least when the defendant himself is not relying on experts — will not significantly hinder it. [19] The state asserts that Smith forfeited his fifth and sixth amendment claims by not raising them when he objected to Dr. Grigson’s testimony at the sentencing phase. There are three sufficient answers. First, in the habeas proceeding in federal district court, the state did not raise this argument in the pleading or at any time until it moved for a new trial. Judge Porter held that this constituted a waiver, see Smith v. Estelle, 445 F.Supp. 647, 659 (N.D. Tex. 1977), and we are not disposed to disagree. See LaRoche v. Wainwright, 599 F.2d 722, 724 (1979). Second, Texas courts interpreted the fifth and sixth amendments to permit testimony like Dr. Grigson’s to be admitted, see, e. g., Livingston v. State, 542 S.W.2d 655, 661-62 (Tex.Cr.App.1976), cert. denied, 431 U.S. 933, 97 S.Ct. 2642, 53 L.Ed.2d 250 (1977); we have held that the apparent futility of objecting to an alleged constitutional violation excuses a failure to object. See Rummel v. Estelle, 587 F.2d 651, 653-54 (5th Cir. 1978) (en banc), cert. granted, - U.S. -, 99 S.Ct. 2403, 60 L.Ed.2d 1064 (1979). Third, Smith’s objection was, after all, essentially surprise, see note 8 supra; his counsel can scarcely be faulted for failing to enumerate all of the many constitutional rights that the state violated when it unexpectedly presented Dr. Grigson’s testimony. [20] In this appeal, Smith did not allege any errors that would undermine his conviction, and we understand Judge Porter to have held simply that Smith cannot be executed. We affirm his judgment on that understanding. We of course leave to state authorities any questions that arise about the appropriate way to proceed when the state cannot legally execute a defendant whom it has sentenced to death.