concurring in part and dissenting in part:
We are unanimous in our decision today to abandon the formulation of criminal responsibility adopted eighteen years ago in Durham v. United States, 94 U.S.App.D.C. 228, 214 F.2d 862 (1954). We held there that a person is not responsible for a criminal act if the act was the product of mental disease or mental defect. In place of the Durham jury instruction, juries will now be instructed in terms of the American Law Institute test that a person is not responsible for a criminal act if 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. But the adoption of this new test is largely an anticlimax, for even though Durham’s language survived until today's decision, the significant differences between our approach and the approach of the ALI test vanished many years ago. As described in Judge Leventhal’s scholarly opinion, the ALI test may make possible an improvement in the adjudication of the responsibility issue. But on the whole I fear that the change made by the Court today is primarily one of form rather than of substance.1
Durham was designed to throw open the windows of the defense and ventilate a musty doctrine with all of the inf ormation acquired during a century’s study of the intricacies of human behavior. It fueled a long and instructive debate which uncovered a vast range of perplexing and previously hidden questions. And the decision helped to move the question of responsibility from the realm of esotérica into the forefront of the critical issues of the criminal law.
While Durham never suffered a shortage of critics, virtually all of them missed what I consider the crucial failure which emerged in its operation. The critics assumed that our ruling would generate far-reaching changes, and they questioned or condemned the changes they foresaw. In fact, for all our efforts to push the defense onto a new and more meaningful track, Durham actually produced very little change at all. The first few years’ experience dispelled any illusion that the decision would alter fundamentally the operation of the defense.
By its adoption of the American Law Institute test of criminal responsibility, this Court today repudiates none of the objectives of Durham, see pages 1030-1031, infra, but embraces a new test in the hope that it will succeed where Durham failed. The principal question before us, therefore, is whether the adoption of the ALI test is responsive to the lessons we learned from our efforts to implement Durham, and whether it offers any promise of resolving the difficulties that have always troubled us. The analysis must begin with a clear understanding of the reasons why Durham failed to achieve its objectives.
Plainly, we did not fail for want of trying. Durham reformulated the responsibility test in the hope that new and more useful information would be presented to the jury. We acted largely in response to the plea of behavioral *1011scientists that they did not want to decide ultimate questions of law and morality, but wanted only an opportunity to report their findin’gs as scientific investigators without the need to force those findings through -the prism of M’Naghten.2 3 See pages 1015-1016, infra. By removing the obstacles to the presentation of those findings, Durham challenged the experts to provide the information they had long promised. We expected, perhaps naively, that the presentation of this new information would permit — indeed, require — the jury to undertake a much broader inquiry and to rely less on the ultimate conclusions of the experts. But it quickly became apparent that while our decision produced some expansion of the inquiry, it did not do nearly enough to eliminate the experts’ stranglehold on the process. Even after Durham counsel for both sides often sought to present the issue to the jury in “simplified” form by eliciting from the experts little more than conelusory yes-or-no answers to the questions, “Was the accused suffering from a mental disease or defect ? ” “Was his act the product of that disease or defect?” And so the experts continued, on the whole, to speak in conelusory terms which inevitably included but concealed their underlying value judgments, and their own views as to the appropriate legal outcome. The use of conelusory psychiatric labels often provided an aura of certainty which made it difficult to discern the inadequacies of the examination on which the expert testimony was based, and the limitations of psychiatric knowledge generally. See pages 1017-' 1018 and n. 21, infra. The experts were able to retain their influence in part because of the manner in which Durham was construed. The term “mental disease or mental defect” was saddled with an unintended and astringent medical meaning. And the “productivity” requirement was perversely viewed as a locked door which could only be opened by an expert’s key. But most important, the Court failed to deal with crucial practical obstacles that operate under any formulation o.f the test to impede the flow of information to the jury.
The first of these difficulties was the subject of our 1962 decision in McDonald v. United States, 114 U.S.App.D.C. 120, 312 F.2d 847 (1962) (en banc), where we attempted to rescue the term “mental disease or defect” from the grip of the expert witnesses. The definition of mental disease adopted in McDonald3 rendered our test, in almost every significant respect, identical to the ALI test. Yet McDonald, no less than Durham, left the power of the experts intact. Expert witnesses still testify in misleading and conelusory terms about the medical or psychiatric definitions of mental disease. Since the Court today grafts McDonald onto the ALI test, this decision provides no new answers to this aspect of the problem. In fact, the Court makes clear that the new test rests squarely on a “medical model,” thereby enhancing the power of the experts. See pages 1027-1030, infra.
The second source of difficulty concerns the productivity requirement — the albatross of the Durham decision. This Court’s frustration with the conelusory expert testimony on the issue of productivity culminated in our decision in Washington v. United States, 129 U.S.App.D.C. 29, 390 F.2d 444 (1967), which barred such testimony altogether. And yet, in the face of our prohibition, the experts have continued to testify in conelusory terms, as the records in Brawner and dozens of other cases attest. A reiteration of our ban will not be effective, and I join the Court’s holding that the issue of productivity must henceforth be eliminated from the instructions to the *1012jury. But it should be clear that the ALI test comprises its own variant of the productivity requirement. And, as I will point out below, the Court’s discussion of that aspect of the ALI test carries the strong implication that the albatross is with us still. In my view, we can prevent encroachments on the jury’s function only by adopting an instruction that candidly describes the jury’s power and responsibility. Since we have no simple, scientific formula that will provide a clear-cut answer to every case, we have no choice, in my opinion, but to tell the truth: that the jury, not the experts, must judge the defendant’s blameworthiness; that a calibrated, easily-applied standard is not yet available to guide that decision; and that the jury must resolve the question with reference to its own understanding of community concepts of blameworthiness. See pages 1030-1034, infra.
The third source of difficulty — and to my mind the paramount cause of Durham’s failure — is the cluster of practical obstacles that stand in the way of the full disclosure of information that Durham hoped to secure. See pages 1034-1039, infra. Here too the Court’s decision sheds no new light. For no matter how felicitous its phrasing, a responsibility test cannot, singlehanded, overcome these practical obstacles. Neither Durham nor Brawner lets slip our well-guarded secret that the great majority of responsibility cases concern indigents, not affluent defendants with easy access to legal and psychiatric assistance. In a long line of cases we have been asked to confront difficult questions concerning the right to an adequate psychiatric examination, the right to psychiatric assistance in the preparation of the defense, the right to counsel at various stages of the process, the role and responsibility of a government expert who testifies on behalf of an indigent defendant, the burden of proof, the right to treatment during postacquittal hospitalization, and many more. If the promise of Durham has not been fulfilled, the primary explanation lies in our answers, or lack of answers, to those questions. I fear that it can fairly be said of Braiuner, just as it should be said of Durham, that while the generals are designing an inspiring new insignia for the standard, the battle is being lost in the trenches. In fact, our obligation to confront the practical problems now is greater than it was in 1954, if only because our efforts to implement Durham have brought many of these problems to first light.
If Brawner is not responsive to the problems uncovered by Durham and makes no fundamental change in the operation of the defense,4 why has the Court bothered to do anything at all? Perhaps the decision rests on an unstated assumption that change is futile because we lack enough information about human behavior to make possible a meaningful use of the defense, or because we are unwilling or unable to act upon the information that is already at hand.5 Adoption of the uncontroversial ALI test might then be explained as an attempt to discourage the fruitless expenditure of time and energy on an effort doomed to failure. If that is indeed the Court’s intention, then this decision will undoubtedly prove a great disappointment. The adoption of this test, or any other new test, is sure to touch off a flurry of litigation in which counsel will call upon us not only to face the under*1013lying moral questions, but also to pour some concrete meaning into the ambiguous language of the ALI test. True, the adoption of this new test does not foreclose the possibility of further change and development that will respond directly to the central problems of the defense. But the decision does seem to me an important signal of the Court’s current attitude. It is an attitude sharply at odds with the spirit of experimentation, inquiry, and confrontation that have characterized so much of our work in this field. Brawner offered us an opportunity to explore the most difficult questions— to what end do we maintain the defense ? and how can we facilitate a meaningful use of the defense by all defendants, including indigents who must rely on the government for expert assistance? If the Court’s decision today rests on the belief that nothing is wrong which cannot be cured by fixing a new label to our test, then eighteen years’ experience has surely been wastd.
TABLE OF CONTENTS
I. The Trial Record...........1013
II. The Origins and Development of the Durham Rule of Criminal Responsibility ...............1014 -1017
III. The Need to Abandon the Durham-McDonald Test ........1017 -1021
IV. The Court’s Articulated Reasons for Replacing Durham-McDonald With ALI-McDonald........1021 -1030
A. The Interest of Uniformity ....................1021 -1030
B. The Need to Depart from the Product Formulation . .. 1021 -1022
V. The Advantages of a Rule that Instructs the Jury to Acquit the Defendant if He Cannot Justly Be Held Responsible........... 1022
VI. Practical Problems of the Defense and the Disposition of This Case......................1034 -1039
VII. Conclusion ................1039
I. THE TRIAL RECORD
The facts underlying this appeal are simple. After spending an afternoon drinking wine, appellant went to a party at the apartment of three friends. During the evening several fights broke out. Appellant was hit in the jaw and knocked down; he left the apartment immediately. During the next hour he was seen by several friends, who described him as dizzy, unclear of speech and bleeding from the mouth. He refused to go to a hospital for medical attention, and told his friends he had been jumped. Pounding on a mailbox with a fist, he said that someone — perhaps himself— was going to die. Returning to the building in which the party had been held, appellant fired five shots through the closed door of the apartment. Two of the shots struck and killed one of the occupants. Appellant was arrested nearby shortly afterward.
On his own motion appellant was committed to St. Elizabeths Hospital for observation. The standard commitment order asked the Hospital to report on’ both his competence to stand trial and his responsibility at the time of the act charged.6 With respect to responsibility, the Hospital was asked “whether the defendant, at the time of the alleged criminal offense, committed on or about September 8, 1967, was suffering from a mental disease, or defect which substantially affected his mental or emotional processes and substantially impaired his behavior controls, and if so, whether his criminal act was the product of his mental condition . . . . ”7 The Hospital’s summary report stated that appellant was competent for trial; that he was mentally ill at the time of the act; and *1014that the act was not causally related to his illness.
At trial, four expert witnesses from the staff of the Hospital testified on the issue of responsibility. All four agreed that appellant was mentally ill at the time of his unlawful act. They used various labels,8 but in general they agreed that he had an organic brain pathology and an associated explosive personality disorder. The organic damage was indicated by a history of epileptic seizures, an abnormal electroencephalogram test, and a pattern of responses to psychological tests characteristic of persons with organic impairment. The explosive personality disorder was indicated by psychological testing and by psychiatric interviews and observations.
All four experts went into commendable detail in describing the nature of appellant’s disorder and its effect on his behavior. Each expert in turn stated that appellant’s disorder was manifested in an inability to deal with provocation.9 Appellant was said to have severe .feelings of inadequacy,10 and to be easily threatened. He would respond to a felt threat without control; his behavior would be explosive, and out of proportion to the situation.
The only conflict in the expert testimony arose in the course of the prohibited inquiry into productivity. Dr. Stammeyer and Dr. Hamman testified that in their view appellant’s unlawful act was the product of his explosive epileptoid personality disorder. Dr. Weickhardt and Dr. Platkin, on the other hand, testified that appellant's act in shooting through the closed door of an apartment was not the product of his illness but rather the product of a normal desire to retaliate for a broken jaw. That is, even if appellant had not been ill, he would have retaliated in this way. Dr. Platkin’s notes in the Hospital records describe the act as “a more or less legitimate response to a situation in which he had been severely injured in a fight and was very vindictive.” Dr. Platkin testified that “I think I would, too, under the same circumstances want to get even with somebody who broke my jaw.”
II. THE ORIGINS AND DEVELOPMENT OF THE DURHAM RULE OF CRIMINAL RESPONSIBILITY
Eighteen years ago this Court formulated a new test o,f criminal responsibility for the District of Columbia. In Durham v. United States we held that “an accused is not criminally responsible if his unlawful act was the product of mental disease or mental defect.”11 The Durham test was formulated in response to the widespread complaints of psychiatrists that under traditional tests of responsibility the law asked them to go beyond their professional competence; the traditional tests seemed to saddle them with the job of deciding which defendants were guilty and which ones should be excused for lack of criminal responsibility. The M’Naghten Rule and the various “irresistible impulse” tests attempted to define the types of illness that the law would regard as exculpatory, and asked the psychiatrist whether the *1015particular defendant suffered from such an illness at the time of his unlawful act. Thus under M’Naghten the psychiatrist was asked whether the defendant could tell right from wrong, and under the irresistible impulse test he was asked whether the defendant had the capacity to refrain from doing the unlawful act.12 The psychiatrist’s answer was the whole answer to the question of responsibility. The only function of the jury was to choose which expert to believe in case of a conflict.
Psychiatrists vociferously . criticized that approach to the problem of criminal responsibility on the ground that it did not correspond to the state of psychiatric knowledge. In their view few if any persons could be said to be totally lacking in the capacity to distinguish right from wrong or to control their actions. At the same time psychiatrists believed that they could provide extensive insights into other aspects of behavior that were highly relevant to the problem of responsibility.13 Since the traditional tests were deemed too narrow to allow consideration of such insights, many psychiatrists sought to include them under the cover of psychiatric labels and legal conclusions. The expert treated a neutral scientific question about the defendant’s mental condition as one addressed to the legal significance of that condition.14 He was often allowed to do so because the factfinder was happy to be relieved of a difficulty and troubling task, namely, deciding whether the defendant’s illness was severe enough to excuse him. The psychiatrist performed *1016this task simply by fitting, or refusing to fit, the defendant into one of the categories of exculpatory mental illness. But even if psychiatric diagnosis is sufficiently flexible to permit this kind of manipulation,15 it does not provide a satisfactory solution from a legal standpoint. The difficulty of deciding the issue of responsibility may not be avoided by turning it over entirely to the experts.
One solution might have been to formulate a new and broader definition of “exculpatory mental illness,” in order, to retain M’Naghten’s goal of offering the psychiatrist a limited role, i. e., ascertaining whether the defendant suffered from such an illness. But we found it impossible to formulate a general definition which would eliminate the need to evaluate blameworthiness in each individual case. Our problem, therefore, was to make it clear that evaluation was to be made not by the experts but by the jury. On the basis of past experience we feared that any concrete definitions we might offer the jury for guidance would promptly find their way into the standard repertoire of psychiatric testimony, capable again of yes-or-no answers, thereby returning the ultimate issues to the keeping of the psychiatrists.
For inspiration we turned to a model long familiar to the law, the method of assessing fault in negligence cases. We articulated no precise definition of responsibility, as the law articulates no precise definition of negligence. Instead in each case we allow the jury to hear all relevant information and ask it to decide whether by prevailing community standards the defendant was at fault. Thus the jury not only makes the factual determination that the defendant behaved in a certain manner, or that his mental condition was of a certain character, but also fixes the legal norm against which the mental condition and its relationship to the behavior must be measured. The role of the expert is to tell the jury anything he can about the relationship between the defendant’s behavior and his state of mind. The jury decides in the light of that information whether the defendant can justly be held responsible for the harmful consequences of his act, or whether, on the contrary, the harm should be attributed to chance, for example, or to mental disorder.
The principle of Durham was to impose responsibility only if an act was the product of a free choice on the part of the defendant, and not of a mental disease or defect. In adopting the term “product” we borrowed again from the law of torts. In negligence cases the law is concerned with the relationship between the defendant’s conduct and the resultant injury. Even when it is possible to establish some sort of causation, the requirement of “proximate cause” permits the jury to decide that the negligence was too slight or the causal connection too remote to have legal significance. Durham uses the term product not to limit liability but to limit its avoidance. Nevertheless, the function of productivity is similar to the function of proximate cause. In criminal cases the law is concerned with the relationship between the defendant’s mental condition and his unlawful conduct. Even though it is usually possible to establish some sort of causal relationship between almost any mental illness and the unlawful conduct, the requirement of productivity permits the jury to decide that the illness was too slight or the causal connection too remote to have legal significance.
Critics of Durham said the product requirement was illusory, because no psychiatrist would be able to deny the possibility of a causal connection between the illness and the act.16 Consequently, it *1017was argued, responsibility would turn solely on the question of mental disease, a question clearly within the scope of psychiatric expertise. Thus Durham would increase the power of the expert to decide legal and moral questions, rather than cut it down. We intended, however, that the inquiry into productivity would define the moral determination inherent in any determination of responsibility and commit it to the judgment of the jury rather than that of the experts. We expected that if a mental illness was present, and the experts could not exclude causation entirely, the issue would ordinarily go to the jury as a question of degree.
III. THE NEED TO ABANDON THE DURHAM-McDONALD TEST
Ever since this court announced its new test of responsibility in 1954, we have been struggling with the problem of distinguishing between the uniquely psychiatric elements of the determination of responsibility, and the legal and moral elements of that determination. We have repeatedly urged psychiatrists to avoid using the conclusory labels of either psychiatry or law.17 Testimony in terms of the legal conclusion that an act was or was not the product of mental disease invites the jury to abdicate its function and acquiesce in the conclusion of the experts. Testimony in terms of psychiatric labels obscures the fact that a defendant’s responsibility does not turn on whether or not the experts have given his condition a name and the status of disease.18
Since Durham we have been engaged in a continuing process of refining and explicating the rule of that ease. Carter v. United States defined the term “product” in broad terms designed to restrict conclusory expert testimony and expand the basic factual information available to the jury.19 Later McDonald v. United States sought to do the same for the term “mental disease or defect”20 by discouraging the use of psychiatric labels which often served to hide the fact that *1018the experts were providing virtually no information about the defendant's underlying condition.21 The point in both cases was to invite all the information that modern knowledge could provide, to guide the jury’s consideration of the moral, legal, and medical elements in the issue of responsibility. But most psychiatrists declined the invitation and continued to decide themselves when an illness should relieve a defendant of responsibility. Therefore in Washington v. United States we reluctantly took the step of prohibiting all psychiatric testimony in terms of the issue of productivity, on the ground that such testimony was particularly likely to usurp the jury’s function of resolving the ultimate question of guilt.22 We said that the existence of disease was a medical question *1019which psychiatrists could properly answer, but the question of productivity was the ultimate question for the jury, involving a mixture of medical information and moral judgment.
The reason for the Washington rule was to reserve exclusively for the jury one part of the determination of criminal responsibility. By prohibiting testimony in terms of the ultimate question of productivity we sought to focus on the need for testimony in depth concerning the nature, extent, and manifestations of the defendant’s disability. The purpose was to give the jury an adequate basis for deciding whether the disability was such that it would be unjust to condemn the defendant for his conduct.
In practice, however, under Durham and its progeny psychiatrists have continued to make moral and legal judgments beyond the proper scope of their professional expertise.23 Even after Washington, the District Court continues to commit defendants to the public mental hospital for examination under a standard order that asks for a conclusion on productivity.24 And the doctors who perform most of the examinations have shown little reluctance to answer that an illness was present but the act was not its product. The testimony takes a form that indicates that the psychiatrists are in fact making a moral judgment, that they are finding an illness which in their view is not sufficiently serious or sufficiently related to the act to warrant acquittal. And that, of course, is precisely the judgment that we have entrusted not to the experts but to the jury.
Moreover, the productivity requirement tends to focus the attention of expert witnesses and the jury on extraneous and inappropriate issues, and to divert them from the core of the question of responsibility. Durham suggested that the government could establish criminal responsibility either by proving freedom from illness or by proving that the illness did not cause the act. And one way to prove that the illness did not cause the act is to prove that the defendant would have done it anyway. Carter even more explicitly than Durham invited the government to establish responsibility by proving that the defendant would have committed the act even if he had not been ill. Carter stated that productivity amounted to causation of the “but for” variety: an act is the product of mental disease if “the accused would not have committed the act he did commit if he had not been diseased as he was.”25 This approach invited experts and juries to speculate about the defendant’s character, and convict him on the ground that he would have been “bad” if he had not been sick.26
The abuses of the productivity inquiry are strikingly illustrated by the record *1020in this case. Since the expert witnesses agreed that appellant suffered from a substantial disorder, his conviction would seem to depend on the jury’s resolution of the question of productivity. The government’s two expert witnesses both found mental illness without productivity. The testimony of these two witnesses is open to at least two interpretations. It may be that they regarded appellant’s illness as highly specific in its operation: that its only effect on appellant’s behavior was to produce an occasional reflex-like explosive reaction following instantly on the heels of the triggering event rather than an hour or two later; that the illness could have no relation to behavior of the type that resulted in appellant’s prosecution. Putting aside the conclusory parts of their testimony, the balance of their testimony so understood could support a jury finding of criminal responsibility.
But it is not clear whether the conflict among the experts related to the scope of the illness or to its legal significance. In other testimony the government witnesses seem to reject such a tightly compartmentalized view of appellant’s mental and emotional processes.27 There is reason to suspect that their conclusion was based not on a professional judgment about the scope of the illness but rather on the view that the illness was irrelevant because appellant would have committed the crime in any event. Their testimony suggests that they regarded appellant’s act as a normal response for someone in his circumstances.28
Clearly, firing a shotgun through a closed door is not a normal response for everyone who is hurt in a fight, though it may well be for some people. The criminal law assumes that there is a spectrum of “normality,” and that some “normal” people commit crimes while others do not. We cannot allow either the experts or the jury to speculate about where on that spectrum the defendant would belong if he were not mentally ill. That sort of speculation is especially pernicious because it is likely to discriminate systematically against inner-city slum residents like appellant, since violent unlawful behavior is more common in the slums than in middle class neighborhoods. To regard behavior as the product of illness in the suburbs but “normal” in the slums is to establish an odious double standard of morality and responsibility.
The insanity defense is based on the premise that it is unjust to convict a man for behavior he could not control. There is a high incidence of mental illness in inner-city slum areas,29 and we are bound to give it the same significance in dealing with their residents as we do in dealing with other people.30 If ap~ *1021pellant’s behavior controls were substantially impaired by mental illness, he should not be held responsible on the ground that it is “normal” for those in his environment to behave that way,31 or even because the examining psychiatrist believed that “under the same circumstances I would want to get even with somebody who broke my jaw.” 32
IV. THE COURT’S ARTICULATED REASONS FOR REPLACING DURHAM-McDONALD WITH ALI -DONALD
In my view, the ALI test of criminal responsibility, with or without the McDonald side bar, cannot remedy the problems in the administration of the defense which have come to light as a result of our efforts to implement the Durham-McDonald rule. While I assume my brethren do not share my pessimistic appraisal of the new test, they make no real effort to justify this change. Pages 981-985 of the Court’s opinion are devoted to the articulation of two reasons for adopting the ALI test: (A) the “interest of uniformity of judicial approach and vocabulary, with room for variations and adjustments,” majority opinion at 984-985; and (B) the “need to depart from [the] ‘product’ formulation and undue dominance by experts,” majority opinion at 981-983.
A. THE INTEREST OF UNIFORMITY
At issue here is no garden variety “uniformity.” Whatever the benefit of having the Circuits in substantial alignment on important questions of law, that is clearly not the benefit which the Court is seeking here. The other Circuits that have adopted the ALI test have taken a variety of substantially different approaches to the interpretation of its language.33 And today this Court adopts a variation on the ALI theme which differs, in design at least, from the approach of every other court. But the Court makes clear that uniformity in substance is not our goal, but only uniformity in vocabulary. If all of us speak the same language, the Court reasons, judicial communication will be enhanced.
Even accepting the argument at face value, it contributes very little to the resolution of the issue. For the argument does not even purport to demonstrate that the ALI test is inherently preferable to Durham-McDonald or any other test. I do not mean to disparage the value of judicial communication, but it is surely a concern of markedly less importance than the substantive merits of the test.
In any ease, it is far from clear that our adoption of the new test will, in fact, enhance communication. The Court assumes that the lessons we have learned from Durham-McDonald have been “lost *1022in translation” to the other Circuits, and that their lessons have been similarly-lost to us. The problem apparently arises because of “blockage due to jargon.” Majority opinion at 984. It is not clear to me how one would prove or disprove that assertion. But as a matter of logic —-and surely as- a reason for adopting the ALI test — the assertion is entirely unpersuasive. In fact, if I were to devise a paradigm scheme for blocking communication, I would arrange for courts to hide major differences behind uniform language, so that the differences would be overlooked by all but the most astute observers. That is precisely the result we have achieved by adopting the all-things-to-all-people language of the ALI test. By contrast, the singular phrasing of our prior rule encouraged its analysis by courts and commentators, and forced us to compare our approach with the approach of other courts. I do not see how we can reasonably expect our adoption of the ALI test to enhance our communication with other courts.
B. THE NEED TO DEPART FROM THE PRODUCT FORMULATION
The questions initially raised on this appeal pertained to our decision in Washington v. United States, 129 U.S.App.D.C. 29, 390 F.2d 444 (1967), barring conelusory expert testimony on the issue of productivity. Appellant insisted at the outset that Washington’s prohibition had been disregarded at trial, and he asked us to demonstrate our opposition to expert domination of the process by enforcing the Washington rule. This Court, sua sponte, altered the focus of the inquiry by calling into question our test of responsibility, and by asking the parties and amici curiae to canvass the arguments for and against a change in the Durham-McDonald rule.
In examining the ALI test now adopted by this Court, it is important to keep in mind the origins of this case and the problem which the adoption of a new test is designed to solve. The great bulk of the Court’s opinion is devoted to an explication of the ALI test as adopted in this jurisdiction. Since the clarification of ambiguous language now may minimize litigation later, that is, of course, an important undertaking. But the critical question before us is whether or not the adoption of the ALI test is likely to make any significant inroad on the problem of expert domination. The answer to that question depends in large part upon an understanding of the reasons why Durham’s productivity requirement became a convenient handle for expert domination.
Durham did not invent the question of causality. Every responsibility test demands (or assumes) some link between the defendant’s act and his impairment; Durham merely gave explicit recognition to the issue. Thus, the critical question is not whether the act must be related to the impairment (“mental disease,” “defect of reason,” or whatever), but rather how directly, if at all, the jury’s attention should be focused on the question.
It is still not clear precisely what the concept of causality means in this context, or whether it is an intelligible concept at all. Yet Durham forced the concept out from its hiding place behind the murky language of other responsibility tests and made the productivity question the ultimate issue for the jury. In our effort to clarify the question we held that an act was to be considered the “product” of a mental disease only if it would not have been committed but for the disease — the disease must have “made the effective or decisive difference between doing and not doing the act.” Carter v. United States, 102 U.S.App.D.C. 227, 236, 252 F.2d 608, 617 (1957). That definition gave the false impression that we understood the concept of causality fully and could draw meaningful and distinct lines between sufficient and insufficient cause — between acts that were caused by mental illness and acts that were not.
Notwithstanding the expectations to the contrary of the courts and commenta*1023tors who examined the Durham, rule,34 the productivity requirement became a formidable obstacle to the presentation of a successful responsibility defense. Even as recently as 1970 one federal court suggested that the “product portion of the test seems superfluous,” reasoning that once a disability had been established, it would ordinarily be impossible to prove that it had no relationship to the unlawful act. Wade v. United States, 426 F.2d 64, 69 (9th Cir. 1970). Nevertheless, psychiatrists in this jurisdiction have often concluded — and convinced juries-— that a mentally ill defendant should be convicted because his act was not the product of his illness. See United States v. Eichberg, 142 U.S.App.D.C. 110, 113, 439 F.2d 620, 628 (1971) (Bazelon, C. J., concurring).
This development may have allayed the fears of some who expected Durham to produce a vast increase in the number of insanity acquittals. After all, it was only the productivity requirement that stood between the liberalized concept of mental illness and acquittal; insistence on a rigid, and often impossible, showing of causal connection was an obvious means of reining in the defense. But the primary drawback of the productivity requirement was not that it reduced the number of insanity acquittals, for it is extremely unlikely, in my opinion, that juries would have acquitted many more defendants if the product formulation had never been devised. The real difficulty was that the superficial simplicity of the productivity question made it seem susceptible of an unambiguous medical or scientific answer. As a consequence, jurors too often relied on the conclusions of the experts, failing to see that the “gravity of an impairment and its relevance to the acts charged are both questions of degree, which can only be resolved with reference to the community’s sense of when it is just to hold a man responsible for his act.” United States v. Eichberg, 142 U.S.App.D.C. 110, 113, 439 F.2d 620, 623 (1971) (concurring opinion).
As I understand the Court’s opinion, the rationale for the switch from Durham-McDonald to ALI -McDonald can be summarized as follows: The primary flaw of our prior test was its emphasis on productivity, which permitted “undue dominance” by the expert witnesses who testified on the issue of responsibility. Majority opinion at 981. The remedy is not to depart from the product requirement (which would hardly be possible in any case since the requirement is an integral part of every responsibility test), but to depart from the product formulation. The ALI test retains the “core requirement” of productivity, in the sense that there must be a “meaningful relationship between the mental illness and the incident charged.” Id. at 983. But the question of causality does not occupy a position of prominence under ALI comparable to the position that the product requirement occupied under Durham. By eliminating the term “product” we can eliminate the vocabulary which was “conducive to a testimonial mystique permitting expert dominance and encroachment on the jury’s function.” Id. The foregoing reasoning of my brethren represents the primary articulated justification for adoption of the ALI test, and the validity of the analysis must, therefore, be considered with great care.
1. The Court’s reasoning suggests that our primary goal is to deemphasize the question of productivity or causality. Yet there is strong reason to suspect that adopting the ALI test will not bring us closer to that goal. The difficulty of applying the ALI productivity requirement —and hence the amount of attention which the requirement will attract — is likely to vary with the nature of the defendant’s impairment. If the defendant cannot “distinguish ‘good and evil,’ i. e., if he ‘doth not know what he is doing, *1024no more than * * * a wild beast,’ ” 35 he may well lack the capacity to appreciate the wrongfulness of any act or to conform any act to the requirements of law. In that case, a jury is likely to conclude that the defendant’s impairment “caused” his act, irrespective of the act he allegedly committed. If the defendant’s impairment is not so severe as to render him a “wild beast,” the question of productivity is still unlikely to present great difficulty so long as the impairment is deep enough and pervasive enough to compel the conclusion that most of his acts are substantially affected by the impairment. A finding of non-productivity will not often be made where the defendant is suffering from a psychosis or other severe disorder, because the defendant’s act will, in all probability, bear a strong and obvious relationship to the impairment.36 And even where the question is close, juries may often resolve their doubts in favor of a finding of causality in order to insure that the defendant is committed to a hospital rather than a penitentiary.
The real difficulty with the causality requirement arises when the defendant’s impairment is a neurotic condition or personality disorder. It appears that these conditions are often encapsulated, in the sense that they may have a significant impact on some aspects of the defendant’s behavior, while leaving his personality substantially intact.37 In these *1025eases disputes will often arise concerning the relationship of the act to the impairment. And snce the impairment may be much less severe than a psychosis, the defense will often not be aided by a presumption that hospitalization is the appropriate disposition. In fact, the jury may be convinced that the non-psychotic defendant deserves criminal punishment even though the experts consider him mentally ill. And they may use the productivity requirement as a handle for the rejection of the responsibility defense.
In short, the most efficient means of eliminating the productivity problem (but not the productivity question) is to limit the definition of exculpatory mental illness to those conditions which are so severe that in most cases a finding of productivity can readily be made. It can reasonably be argued that the intent of the ALI draftsmen was to make the responsibility defense available only to defendants suffering from psychoses or other severe disabilities.38 Under that interpretation, which is apparently accepted by at least some other federal jurisdictions,39 the productivity issue should rarely present great difficulty. But that interpretation is plainly not the one adopted by the Court in today’s opinion. As I read the Court’s opinion, the retention of the McDonald definition of mental illness reaffirms our longstanding view that, in the words of Mr. Dempsey’s amicus brief, “the defense is not restricted to persons suffering from the gravest types of mental disorders. While the jury must find that the defendant’s ‘mental or emotional processes’ have been ‘substantially affected’ and his ‘behavior controls’ ‘substantially impaired,’ the jury is not bound by whether those consequences flow from what the psychiatrists label a ‘psychosis,’ ‘pyschoneurosis,’ a ‘sociopathic personality,’ an ‘emotionally unstable personality,’ or whatever.” 40 If we are indeed to retain the expansive definition of mental illness implicit in Durham and formalized in McDonald, then the productivity question will remain a source of controversy and debate.
Unlike Durham, which focused on the relationship between the defendant’s mental illness (impairment) and his act, the ALI test focuses on the relationship between the defendant’s mental illness and his impairment. In the words of the test, the impairment must exist “as a result” of mental illness. But productivity in the Durham sense — the relationship between the impairment and the act — is not abolished; it is concealed in two questions which are implicit in the test; Could the defendant appreciate the wrongfulness of the particular act he committed? Could he have conformed that particular act to the requirements of law? So long as the defendant’s impairment is a psychosis or other severe disability and is roughly consonant with his act, the answers to those questions should be sufficiently obvious that the questions will not even seem to arise. But where the defense is predicated on a less severe form of impairment — as it apparently can still be in this jurisdiction — those questions will assume the burden that has been carried up to now by Durham’s explicit requirement of productivity.
*1026The operation of the causality requirement implicit in the ALI test can be illustrated by considering how Brawner would have been tried under the new test. The expert witnesses would presumably characterize his condition as an explosive personality disorder, manifested in an inability to deal with provocation. The act which Brawner committed— shooting through a closed door in retaliation for a blow to his jaw a short while before — -is surely consistent with his condition. It could thus be said that in at least some situations Brawner apparently lacked substantial capacity to conform this kind of behavior to the requirements of law. But I have little doubt that the government would seek to introduce expert testimony, as it did under Durham, that Brawner committed this act not because of his personality disorder, but rather because he wanted “to get even with somebody who broke [his] jaw.” See page 1014 supra. The issue raised by this line of testimony need not be called a productivity or causality question. But whatever it is called, it is functionally identical to the productivity question that routinely arose under Durham.
The Court undoubtedly recognizes that retention of McDonald’s open-ended definition of mental illness will require an inquiry into causality in a large number of cases, in marked contrast to the opinions of the other federal courts that have adopted the ALI test,41 the Court’s opinion places great emphasis on the causality question. Superficially, the Court’s references are directed only at the first stage of the causality question under the ALI test — the relationship between the illness and the impairment rather than’ the relationship between the impairment and the act. But the question raised by that first stage is' so trivial and the Court’s references to causality are so numerous that it is hard to avoid the implication that the references are primarily aimed at the second stage of the productivity question. Those references carry an implicit assurance that acquittal under the ALI test will be no less difficult for a defendant without a pervasive disability than it has always been under Durham.
The critical question, therefore, is how the productivity issue will be presented to the jury. As I pointed out above, the Durham formulation gave the false impression that the question required a medical or scientific answer. The ALI language could fare better, since it does not invite the expert witnesses to offer a flat and seemingly scientific answer that the impairment did or did not “cause” the act. But while there is some promise in the language of the ALI test, I fear that the Court’s construction of that language may dissipate much of that promise. The ALI test provides that “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 * * * ” (Emphasis supplied.) The Court maintains that the causality requirement lurks in the term “as a result,” suggesting that the “mental disease of a kleptomaniac does not entail as a ‘result’ a lack of capacity to conform to the law prohibiting rape.” Majority opinion at 991. The term “as a result” does, of course, contain a requirement of causality. But it refers only to the first stage of the requirement under the ALI test, indicating that the impairment must be caused by the mental disease. But the crucial question of causality — the link between the impairment and the act — is not reflected in the term “as a result.” It inheres in the concepts of “appreciating wrongfulness” and “conforming conduct.” Thus, if kleptomania is an abnormal condition of the mind, then for *1027purposes of the ALI test a kleptomaniac “lacks substantial capacity as a result of mental disease” regardless of the act he allegedly committed. But if he is charged with rape, his responsibility defense would presumably fail because, even though he may lack capacity to appreciate the wrongfulness of theft or to conform his conduct to the requirements of the law prohibiting theft, he may in fact have substantial capacity to appreciate the wrongfulness of rape and to conform his conduct to the requirements of the law prohibiting rape.
My concern with the source of the ALI productivity requirement is not intended as an exercise in the splitting of hairs. By making the term “as a result” carry not only the unimportant first stage of the causality question, but also the critical second stage, the Court repeats precisely the mistake it correctly identifies in Durham: the articulation of a catchphrase that facilitates conclusory expert testimony and that obscures the moral and legal overtones of the productivity question. Where a psychiatrist would formerly have testified that the act was not the “product” of the disease, he can now assert that the disease of the defendant does not entail as a “result” the kind of impairment that could have produced the act in question. Under my view of the ALI language, a psychiatrist attempting to present a conclusory no-productivity argument would have to lead the jury through the murky waters of “appreciating wrongfulness” and “conforming conduct,” and in all likelihood the jury would be lost almost from the outset. If the causality requirement cannot readily be expressed as an uncomplicated yes-no question, there is a good chance that juries would begin to recognize that the requirement subsumes the moral and legal questions which lie at the heart of the responsibility defense.
2. Our opinion in Washington recognized that the productivity requirement can lead to domination by the expert witnesses not so much because they testify about the issue, but because they testify about the issue in conclusory terms. For that reason, we barred conclusory testimony on this issue, and urged the experts to disclose the factual data from which the jury could draw reasonable inferences about the defendant’s condition. Inexplicably, the Court now concludes that Washington is “superseded— on this point — by our change today of the ultimate rule,” majority opinion at 1003. Yet, as the Court repeatedly makes clear, the change of the ultimate rule leaves standing the causality requirement. The net effect of today’s decision is, therefore, to require the experts to drop the term “product” in favor of the term “result,” and to permit them once again to tell the jury in eonclusory terms that the act was not caused by the defendant’s impairment.42 To be sure, a mystique has developed around the term “product,” and the elimination of that term should undercut the mystique. But I see no reason to assume that the term “result” is immune to the identical development, especially in view of the Court’s unexplained determination that experts should once again be permitted to testify in conclusory terms on the issue of causality.
3. If our primary goal is, in fact, to achieve a reduction in expert domination of the process, the gratuitous overruling of one aspect of Washington v. United States43 is not the only — and perhaps not the most important — step backward. The Court identifies the productivity requirement as the chief villain that per*1028mits the experts to encroach on the jury’s function. But there is another aspect of the test which is at least as susceptible to expert domination. Like Durham, the ALI test demands a “mental disease” as a condition of non-responsibility. And the Court today holds that the definition of “mental disease” announced in McDonald will be applicable to the ALI test. Nevertheless, Brawner’s discussion of the term suggests at least a partial erosion of the McDonald view that “mental disease” is a legal concept, and that, “neither the court nor the jury is bound by ad hoc definitions or conclusions as to what experts state is a disease or defect.” 114 U.S.App.D.C. 120, 124, 312 F.2d 847, 851 (1962).
The Court today asserts that it has rejected “suggestions to adopt a rule that disentangles the insanity defense from a medical model,” and adds that a successful responsibility defense must be predicated on the existence of an “ascertainable condition characterized by ‘a broad consensus that free will does not exist.’ ” Majority opinion at 995. I fear that counsel, the experts, and the trial courts will view that requirement as a delegation of sweeping new authority to the medical experts.
Of course, the Court does point out that a defendant can make a broad presentation to the jury, offering all of the evidence, even if not strictly medical, which is pertinent to an abnormal condition of the mind. But that broad presentation is already guaranteed by the traditional rules of evidence. The real impact of the Court’s decision is to establish a barrier which will prevent some defendants from taking any evidence at all to the jury on the issue of responsibility. The power to open and close that barrier is effectively delegated to the psychiatric experts.
We can only speculate on the impact of this requirement, but it seems likely to produce very substantial distortions of the process. First, it focuses attention on an entirely irrelevant issue. If a defendant is prepared to present evidence that his mental or emotional processes and behavior controls were in fact impaired, it is not clear why anything should turn on the experts’ view of his condition in the abstract.
Second, the requirement obliges the defendant to make a vastly greater showing to have the issue of reponsibility submitted to the jury than to have any other issue submitted. We held many years ago that “sanity is an ‘essential’ issue which, if actually litigated — that is, if ‘some proof is adduced’ tending to support the defense — must be submitted to the jury under the guidance of instructions.” Tatum v. United States, 88 U.S.App.D.C. 386, 389, 190 F.2d 612, 615 (1951). Conceding “that any attempt to formulate a quantitative measure of the amount of evidence necessary to raise an issue can produce no more than an illusory definiteness,” we pointed out that “so long as there was some evidence relevant to the issue * * * the credibility and force of such evidence must be for the jury, and cannot be matter of law for the decision of the court.” 88 U.S.App.D.C. at 390, 190 F.2d at 616, quoting from Kinard v. United States, 68 App.D.C. 250, 253-254, 96 F.2d 522, 525-526 (1938). As I read the Court’s opinion, a defendant who can introduce “some evidence” that his capacity to control his behavior was in fact impaired cannot take the responsibility issue to the jury unless he can also offer, should the question be put in issue, “convincing evidence” that he is suffering from a medically-recognized condition characterized by a broad consensus that free will does not exist.44
Still, the greatest difficulty is not that the requirement shifts attention onto an extraneous issue or that it imposes an unwarranted obstacle to the presenta*1029tion of an affirmative defense. Those difficulties could be tolerated if the requirement of a “broad consensus that free will does not exist” reflected the Court’s effort to achieve some important purpose of the responsibility defense. At no point in its opinion does the Court explain why the boundary of a legal concept — criminal responsibility — should be marked by medical concepts, especially when the validity of the “medical model” is seriously questioned by some eminent psychiatrists.45 Nor does the Court explain what it means by “convincing evidence” of the existence of a “broad con-census.” If five psychiatrists are prepared to assert that a particular condition does tend to impair free will, how many psychiatrists must be willing to testify that it does not have such an effect before we can preclude a responsibility defense on the ground that there is no “broad consensus” that the defendant’s condition tends to impair free will ? How many psychiatrists must be convinced that a particular condition is “medical” in nature before a defendant will be permitted, within the confines of the “medical model,” to predicate a responsibility defense on such a condition ?
The Court similarly fails to explain how medical experts can be expected to provide information about the impairment of free will, when free will would seem to be a philosophical and not a medical concept. If psychiatrists will be required to frame their testimony in terms of this non-medical concept, then the Court will have resurrected M’Naghten with one ironic twist. Under M’Naghten, medical experts effectively answered moral and legal questions, and cloaked the answers in medical terminology. The Court now seems to ask experts to make moral and legal determinations about the nature of an exculpatory condition, and invites them to state their conclusions in non-medical terms.
It is possible, however, that the Court’s reference to free will is not intended to carry moral or philosophical implications, but is nothing more than a short-hand for the component of the ALI test which refers to substantial capacity to conform conduct to the requirements of law.46 If *1030so, it is unclear why the Court omits reference to the second component of the ALI test: namely, the capacity to appreciate the wrongfulness of conduct. Is that omission premised on a determination that the cognitive element is irrelevant to responsibility? Or does it mean, perhaps, that the element of cognition is subsumed within the concept of behavior control? See United States v. Currens, 290 F.2d 751, 774 (3d Cir. 1961). These questions, and others which are no less extraneous to the question genuinely in issue, will have to be answered in the course of applying this new requirement.
Of course, the fact that the requirement is illogical, unwieldy, and an invitation to expert domination does not necessarily mean that it should not be adopted. I suggested in a recent opinion that adoption of an explicit medical model may be the only available means of fending off a number of difficult questions concerning our handling of a dangerous defendant who has been found not guilty for lack of responsibility, but who cannot be committed to a medical institution for medical care. In that same opinion I outlined several alternative approaches and attempted to point out the advantages and disadvantages of each. See United States v. Alexander & Murdock, 152 U.S.App.D.C.-at---, 471 F.2d 923 at 960-965 (April 21, 1972). But the Court does not disclose the reasoning that underlies its adoption of the medical model. Nor does it provide any indication of the purpose of this limitation on the legal concept of responsibility. The disadvantages of clinging to a medical model are shouldered without acknowledgment or explanation. What does emerge clearly from the Court’s opinion is that we have now turned over to the experts a substantial part of the inquiry, without making clear why expert domination in this context — as opposed to the context of productivity — is unobjectionable.
V. THE ADVANTAGES OF A RULE THAT INSTRUCTS THE JURY TO ACQUIT THE DEFENDANT IF HE CANNOT JUSTLY BE HELD RESPONSIBLE
The effort to preserve the jury’s function from encroachments by the experts must begin with a clear understanding of what that function is. In determining the responsibility issue, a jury has two important tasks:
In the first place it measures the extent to which the defendant’s mental and emotional processes and behavior controls were impaired at the time of the unlawful act. The answer to that question is elusive, but no more so than many other facts that a jury must find beyond a reasonable doubt in a criminal trial. * * * The second function is to evaluate that impairment in light of community standards of blameworthiness, to determine whether the defendant’s impairment makes it unjust to hold him responsible. The jury’s unique qualification for making that determination justifies our unusual deference to the jury’s resolution of the issue of responsibility.47
Nothing in the Court’s opinion today suggests a departure from our long-standing view that the second of these two functions — the evaluation of the defendant’s impairment in light o,f community standards of blameworthiness — is the very essence of the jury’s role. The Court points out, for example, that
[i]t is the sense of justice propounded by those charged with making and declaring the law — legislatures and courts — that lays down the rule that persons without substantial capacity to know or control the act shall be excused. The jury is concerned, with applying the community understanding of this broad rule to particular lay and medical facts. Where the matter is unclear it naturally will call on its *1031own sense of justice to help it determine the matter. [Emphasis added.]
Majority opinion at 988. And again,
The doctrine of criminal responsibility is such that there can be no doubt “of the complicated nature of the decision to be made — intertwining moral, legal, and medical judgments,” * * * * [J]ury decisions have been accorded unusual deference even when they have found responsibility in the face of a powerful record, with medical evidence uncontradicted, pointing toward exculpation. The “moral” elements of the decision are not defined exclusively by religious considerations but by the totality of underlying conceptions of ethics and justice shared by the community, as expressed by its jury surrogate. [Emphasis added; citations omitted.]
Majority opinion at 982.48 See also majority opinion at 990.
Against this background it is clear that Durham focused the jury’s attention on the wrong question — on the relationship between the act and the impairment rather than on the blameworthiness of the defendant’s action measured by prevailing community standards. If the ALI test is indeed an improvement, it is not because it focuses attention on the right question, but only because it makes the wrong question so obscure that jurors may abandon the effort to answer it literally.
Instead of asking the jury whether the act was caused by the impairment, our new test asks the jury to wrestle with such unfamiliar, if not incomprehensible, concepts as the capacity to appreciate the wrongfulness of one’s action, and the capacity to conform one’s conduct to the requirements of law. The best hope for our new test is that jurors will regularly conclude that no one — including the experts — can provide a meaningful answer to the questions posed by the ALI test. And in their search for some semblance of an intelligible standard, they may be forced to consider whether it would be just to hold the defendant responsible for his action. By that indirect approach our new test may lead juries to disregard (or at least depreciate) the eonclusory testimony of the experts, and to make the “intertwining moral, legal, and medical judgments” on which the resolution of the responsibility question properly depends. The Court’s own opinion hints at this approach, maintaining that “[t]here is wisdom in the view that a jury generally understands well enough that an instruction composed in flexible terms gives it sufficient latitude so that, without disregarding the instruction, it can provide that application of the instruction which harmonizes with its sense of justice. The ALI rule generally communicates that meaning.” Majority opinion at 988-989.
The Court’s approach may very well succeed and encourage jurors to look behind the testimony and recommendations of the experts. But, as I have tried to demonstrate above, there is also a significant possibility that our new test will leave the power of the experts intact — or even make possible an enlargement of their influence. In my opinion, an instruction that tells the jurors candidly what their function is, is the instruction most likely to encourage the jurors to resist encroachments on that function. In itself, that might not be sufficient justification for adopting such a test if it were clear that its adoption would entail substantial costs as a necessary by-product. But I am unaware of any costs that compel us to adopt instead the ALI test, which offers so much less promise of dealing with the problems *1032that initially brought this ease to our attention.
Our instruction to the jury should provide that a defendant is not responsible if at the time of his unlawful conduct his mental or emotional processes or behavior controls were impaired to such an extent that he cannot justly be held responsible for his act. This test would ask the psychiatrist a single question: what is the nature of the impairment of the defendant's mental and emotional processes and behavior controls? It would leave for the jury the question whether that impairment is sufficient to relieve the defendant of responsibility for the particular act charged.49
The purpose of this proposed instruction is to focus the jury’s attention on the legal and moral aspects of criminal responsibility, and to make clear why the determination of responsibility is entrusted to the jury and not the expert witnesses. That, plainly, is not to say that the jury should be east adrift to acquit or convict the defendant according to caprice. The jury would not be instructed to find a defendant responsible if that seems just, and to find him not responsible if that seems just. On the contrary, the instruction would incorporate the very requirements — -impairment of mental or emotional processes and behavior controls — that McDonald established as prerequisites of the responsibility defense.
The proposed instruction has the additional advantage of avoiding any explicit reference to “mental disease” or “abnormal condition of the mind.” As used in our prior tests, these terms were never intended to exclude disabilities that originate in diseases of the body,50 but simply reflect the fact that the defense of non-responsibility has traditionally been associated with mental illness, or in the language of an earlier day, “insanity.” Washington v. United States, 129 U.S.App.D.C. at 37 n. 23, 390 F.2d at 452 n. 23. Moreover,
our experience has made it clear that the terms we use — “mental disease or defect” and “abnormal condition of the mind” — carry a distinct flavor o,f pathology. And they deflect attention from the crucial functional question —did the defendant lack the ability to make any meaningful choice of action —to an artificial and misleading excursion into the thicket of psychiatric diagnosis and nomenclature.
United States v. Alexander & Murdock, 152 U.S.App.D.C. - at---, 471 F.2d 923 at 960-961 (April 21, 1972), (dissenting opinion).
I would adopt an instruction based on the language of McDonald, which seems to me more comprehensible than the language of the ALI test. The capacity to appreciate the wrongfulness of conduct and the capacity to conform conduct to the requirements of the law are, I fear, concepts with little meaning to experts or to jurors. But ,for the present purpose, the critical aspect of the proposed jury instruction is not the use of the McDonald terminology or the omission of any reference to an “abnormal condition of the mind.” If the Court is convinced that the terminology of the ALI test would illuminate the jury’s inquiry, or that the terms “mental disease” or “abnormal condition of the mind” should, for whatever reason, be retained, it is still possible to draft an instruction that clearly describes the jury’s role in *1033deciding when the defendant’s incapacity is sufficient to warrant exculpation. In fact, a minority of the ALI draftsmen (along with Professor Wechsler, the reporter of the Model Penal Code) proposed a test providing that a person
is not responsible for criminal conduct if at the time of such conduct as a result of mental disease or defect his capacity either to appreciate the criminality of his conduct or to conform his conduct to the requirements of law is so substantially impaired that he cannot justly be held responsible.
By contrast, the majority ALI test, now adopted by this Court, provides that 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 criminality of his conduct or to conform his conduct to the requirements of law.
The difference between the two approaches does not pertain to the type of showing a defendant must make. Both require the defendant to demonstrate a particular form of incapacity. The approaches diverge in establishing a standard to determine when the incapacity is sufficient to exculpate the defendant. Under the ALI majority view, the jury must acquit if the defendant’s capacity is substantially impaired. The ALI/ minority would require acquittal where the defendant’s capacity is so substantially impaired that he cannot justly be held responsible.
The ALI ultimately rejected the minority approach because “[s]ome members of the Council deemed it unwise to present questions of justice to the jury, preferring a submission that in form, at least, confines the inquiry to fact.” 51 The Court apparently shares this view, and rejects an instruction “overtly cast in terms of ‘justice’ ” on the grounds that such an instruction “cannot feasibly be restricted to the ambit of what may properly be taken into account but will splash with unconfinable and malign consequences.” Majority opinion at 987. That argument seems to present two separate justifications for pretending that the inquiry is confined to fact.
First, the argument apparently reflects a concern that adoption of the “justice” approach would permit the introduction at trial of extraneous information. But under the approach urged by a minority of the ALI Council, a defendant must still demonstrate that proffered evidence is relevant to an impairment of capacity. The test does not provide him with a license to introduce evidence merely for the purpose of engendering sympathy for him in the jury. Adoption of the “justice” approach would still leave standing all of the traditional obstacles to the introduction of irrelevant evidence.
The Court’s second ground of objection is apparently that an instruction cast in terms of justice would permit the jury to convict or acquit without regard to legal standard. The Court points out, for example, that
[i]t is one thing * * * to tolerate and even welcome the jury’s sense .of equity as a force that affects its application of instructions which state the legal rules that crystallize the requirements of justice as determined by the lawmakers of the community. It is quite another to set the jury at large, without such crystallization, to evolye its own legal rules and standards of justice.
Majority opinion at 989. I take it that in the Court’s view the majority version of the ALI test offers the jury “legal rules that crystallize the requirements of justice as determined by the lawmakers of the community,” and that the minority version sets the jury adrift without such crystallized rules. What, then, are these crystallized rules? I pointed out above that while the minority version asks the jury to measure the impairment in terms of its own sense of justice, the majority *1034version requires acquittal if the incapacity is substantial, and requires conviction if the incapacity is insubstantial. Can we seriously maintain that the majority ALI instruction is preferable because its determination that the impairment must be “substantial” reflects a crystallization of the requirements of justice by the lawmakers of the community? Naturally, we would all prefer a rule that could, as a matter of law, draw a bright line between responsible and non-responsible defendants. But the ALI test adopted by this Court is plainly not such a rule. It offers the jury no real help in making the “intertwining moral, legal, and medical judgments” that all of us expect. In fact, because it describes the question as one of fact it may lull the jury into the mistaken assumption that the question of responsibility can best be resolved by experts, leaving the jury at the mercy o,f the witness who asserts most persuasively that, in his expert judgment, the defendant’s capacity was or was not substantially impaired.
It is not at all clear that the approach I have suggested — whether based on the terminology of the ALI test or McDonald —would finally bar encroachments on the jury’s function. Nevertheless, this approach — unlike the majority ALI test adopted by the Court — comes directly to grips with the problem of expert domination in a manner that is at least responsive to our experience under Durham. The majority ALI test merely reshuffles and obfuscates the Durham components; it does nothing to sort out for the jury the difference between its function and the function of the expert witnesses. Our instruction should make clear that in order to convict a defendant the jury must first determine, on the basis of expert opinion and the factual background disclosed by the experts, the extent to which the defendant’s mental and emotional processes and behavior controls were impaired, and then find, on the basis of community moral standards, that the degree of impairment was sufficiently slight that the defendant can fairly be blamed and held responsible for his act like any other person.52
To expand the scope of the inquiry in this way would not invite a significant increase in the number of acquittals. It would, however, encourage greater commitment to the effort to understand how each criminal defendant came to act as he did. Even if juries were consistently to set the standard of responsibility so low that virtually every defendant would meet it, they would still have to confront the causes of criminal conduct in a way that might teach us all something about human behavior. And they would be giving defendants the kind of careful, individual study that should precede any decision as consequential as the imposition of moral condemnation on another human being.
VI. PRACTICAL PROBLEMS OF THE DEFENSE AND THE DISPOSITION OF THIS CASE
In a distressing number of recent cases this Court has been asked to consider questions unrelated to the substantive test of responsibility, but which have, as a practical matter, far greater impact on the operation of the defense than the language of the rule. The Court’s *1035decision to abandon Durham-McDonald in favor of ALI-McDonald does nothing to obsolete these questions or the Court’s responses to them. If our paramount goal is an improvement of the process of adjudication of the responsibility issue, our attention should be focused on these questions rather than on the ultimate definition of the test. Obviously, these questions cannot all be resolved by one opinion. But the Court’s approach to the disposition of this case offers some indication of the manner in which these questions will be handled in the future.
1. The one consistent note in the Court’s analysis of our experience under Durham is the objection to domination by the experts accomplished through the productivity requirement. We attempted to deal with that problem in Washington v. United States by barring conelusory, expert testimony on the issue of productivity. Virtually all of the expert witnesses at Brawner’s trial agreed that he was suffering from an abnormal condition of the mind. The issue in dispute was productivity — the ultimate issue for the jury. And the transcript is riddled with conclusory, expert testimony on that issue. It is hard to imagine a case which could make a stronger appeal for enforcement of the Washington rule.
After hearing one of his expert witnesses state that Brawner had a “personality disorder connected with epilepsy,” the prosecutor asked the witness:
Did you also come to any opinion concerning whether or not the crimes in this case were causally related to the mental illness which you diagnosed?
After defense counsel’s objection to the question was overruled, the prosecutor asked again:
What was your conclusion as to whether or not there was a causal relationship between the two matters?
The witness replied:
It was my conclusion that there was no causal relationship between his mental disorder and the alleged offense.
Transcript at 464. To be sure, this testimony was not phrased in terms of “product,” but the jury could hardly avoid the message that causality was the cutting edge of the responsibility test and that at least some of the experts were convinced that causality did not exist in this case. Nevertheless, the Court refuses to overturn the conviction despite this patent violation of the letter and the spirit of the Washington rule.
I suggested above that the abandonment of the term “product” may have some beneficial effect in reducing the mystique that surrounds the causality question in this jurisdiction. But I also noted that the Court has made available a new handle for conclusory testimony on the issue of causality — “result”-—and at the same time it has lifted the ban on conclusory testimony on this issue. The transcript of Brawner’s trial offers a glimpse of what we can expect from responsibility trials under the ALI test. The Court’s unwillingness to reverse Brawner’s conviction on this ground makes clear that this Court and the trial courts no longer have any weapons to combat the problem of conclusory testimony and the resulting domination by experts.53
*10362. Since 1895 the federal courts have taken the position that if the defendant introduces “some evidence” of insanity, the issue will be submitted to the jury and the government will bear the burden of proving responsibility beyond a reasonable doubt. Davis v. United States, 160 U.S. 469, 484, 16 S.Ct. 353, 40 L.Ed. 499 (1895). Yet as the responsibility defense has developed under our case law, it has become increasingly clear that the defendant carries an overwhelming practical burden which is not acknowledged in the traditional rule. As a practical matter, the defendant often has very great difficulty obtaining adequate expert assistance to gather the information necessary for the presentation of a significant defense. If he can obtain such information, his defense will often prove vulnerable to attack unrelated to the real merit of his responsibility claim. And even if the attack is very weak the defendant will rarely be entitled to a directed verdict. See United States v. Eichberg, 142 U.S.App.D.C. 110,112-113, 439 F.2d 620, 622-623 (1971).
With limited access to expert psychiatric assistance, indigent defendants normally rely on the government to provide an adequate psychiatric examination at the hospital to which the defendant is committed for observation. In a large number of eases the government’s experts are called to testify on behalf of the defense, and their testimony has often proved inadequate. In one recent case, for example, the trial court concluded that the testimony of a government expert testifying for the defense was completely unacceptable under the principles of Washington v. United States, and he struck the testimony as inadmissible. Yet the trial court refused to grant the defendant’s motion for a mistrial or a new mental examination by experts capable of explaining their findings to a court. And this Court affirmed that ruling. United States v. Alexander & Murdock, 152 U.S.App.D.C. ---at---, 471 F.2d 923 at 952-957 (April 21, 1972) (Bazelon, C. J., dissenting). See also United States v. Leazer, 148 U.S.App.D.C. 356 at 362, 460 F.2d 864 at 870 (Jan. 19, 1972) (Bazelon, C. J., concurring). If an indigent defendant relies' on the government for assistance in preparing his case and if there is no remedy when the government’s assistance is legally inadequate, it will be little consolation to the defendant that the government still carries the burden of persuasion on that issue.
The practical burden on the defendant is greatly enhanced by the ease with which defense testimony can often be torn to pieces on cross-examination. Where a psychiatrist testifying for the government asserts that the defendant did not suffer from any abnormal condition which could impair his mental processes or behavior controls, defense counsel must have considerable expertise in psychiatry to pick out the weak points in the analysis. Yet “very few attorneys, if any, possess the requisite expertise, and we have no automatic procedure for enabling them to consult with psychiatric experts in the preparation and conduct of the defense.” United States v. Leazer, 148 U.S.App.D.C. 356 at 363, 460 F.2d 864 at , 871 (Jan. 19, 1972), (Bazelon, C. J., concurring). Even where the defendant has obvious symptoms of mental disorder, defense counsel is frequently helpless to rebut the suggestion by government psychiatrists that the defendant is malingering. If he produces testimony from a private psychiatrist that the defendant is not a malingerer, he is almost sure to find that the government and its expert witnesses will disparage that testimony on the grounds that it was based on an insufficient period of observation. See, e. g., United States v. Bennett, 148 U.S.App.D.C. 364 at 366-367, n. 4, 460 F.2d 872 at 874-875, n. 4 (Jan. 19, 1972), *1037United States v. Schappel, 144 U.S.App.D.C. 240, 445 F.2d 716 (1971); Rollerson v. United States, 119 U.S.App.D.C. 400, 343 F.2d 269 (1964).54
There are other grounds on which the testimony of defense psychiatrists is extremely vulnerable. A psychiatrist or psychologist who testifies that the defendant suffered from some mental illness exposes himself to what the Court appropriately terms “know-nothing appeals to ignorance.” Majority opinion at 1004. For example, “by requiring the witness to describe in isolation the most minute ‘symptoms’ on which the diagnosis rests — the defendant’s answer to a particular question or his reaction to a particular ink-blot — the prosecution may succeed in making these symptoms seem trivial or commonplace.” United States v. Leazer, 148 U.S.App.D.C. 356 at 363, 460 F.2d 864 at 871 (Jan. 19, 1972), (Bazelon, C. J., concurring). At Brawn-er’s trial, the prosecutor ridiculed the testimony of a defense psychologist in his summation to the jury:
Ladies and gentlemen, then we came to that ink blot, and the doctor said, well, the usual thing about that was those anatomical things, and how many of them were there. Well, let’s see, and he counts, and there are four. How many responses? Fourteen of them. Fourteen responses and four of them turn out to be anatomical things — • hearts or whatever it happened to be. Is there something unusual about that? Is a man crazy when he sees a heart or something else four times, four different anatomical things or maybe the same things in those little drawings, these little ink blots ? After all, they are just blots of ink. Is a man crazy when he sees them?
Transcript of closing arguments at 36-37. We have seen almost identical efforts to ridicule defense experts in other cases. See United States v. Alexander & Murdock, 152 U.S.App.D.C.-at---, 471 F.2d 923 at 955 (April 1972), (Bazelon, C. J., dissenting); United States v. Leazer, 148 U.S.App.D.C. 356 at 363-364, 460 F.2d 864 at 871-872 (Jan. 19, 1972) (Bazelon, C. J., concurring); United States v. McNeil, 140 U.S.App.D.C. 228, 231-235, 434 F.2d 502, 505-509 (1970) (Bazelon, C. J., concurring). The difficulty of presenting credible expert testimony is a major part of the burden on the defendant.
The defendant might be able to cope with these obstacles to the successful use of the defense if we were willing to set aside jury verdicts unsupported by the evidence. In fact, we have been extremely reluctant to overturn a jury verdict even in the face of substantial evidence that the defendant’s act was the product of a condition which impaired his mental or emotional processes and behavior controls. See, e. g., United States v. Eichberg, 142 U.S.App.D.C. 110, 439 F.2d 620 (1971). If the burden of proof does rest on the government, then acquittal should be required not only when non-responsibility is proved, but also when there is a reasonable doubt about responsibility.
At Brawner’s trial, both the prosecution and the defense offered evidence that the defendant was suffering from an abnormal condition of the mind which *1038could impair behavior controls. While the testimony on productivity was expressed largely in conelusory terms, the record does contain a substantial amount of evidence which could support the view that the act was very closely tied to the impairment. In my view, there are two theories which can explain our failure to reverse the conviction on the grounds that a reasonable man must have had a reasonable doubt about the defendant’s criminal responsibility. First, our deference to the jury’s resolution of this issue may be attributable to its special role in evaluating the defendant’s impairment in light of community concepts of blameworthiness, to determine whether that impairment makes it unjust to hold him responsible. See United States v. Eichberg, 142 U.S.App.D.C. 110, 114-115, 439 F.2d 620, 624-625 (1971) (Bazelon, C. J., concurring). But it becomes increasingly difficult to rely on that explanation in the face of this Court’s refusal to make the special function of the jury explicit in the jury instruction. And reliance on the jury’s special function seems dangerously misplaced in a case, such as this one, where the testimony on the only issue in dispute was phrased in such conelusory terms that expert domination is almost inevitable. If we will not take meaningful action to curtail domination by the experts, then we should not rely, in upholding the jury’s verdict, on the jury’s supposed ability to make a kind of judgment that it almost surely did not make.
A second possible explanation for our refusal to set aside the verdict is that we have relaxed the standard of proof in responsibility cases. In fact, Congress enacted a statute in 1970 which purports to shift onto the defendant the burden of establishing insanity by a preponderance of the evidence. 24 D.C.Code § 301 (j). Under that standard one could reasonably conclude that the verdict should not be set aside. But the constitutional validity of the statute is open to very serious question. United States v. Trantham, 145 U.S.App.D.C. 113, 120, 448 F.2d 1036, 1043 (1971) (statement in support of rehearing en banc); United States v. Eichberg, 142 U.S.App.D.C. 110, 114, 439 F.2d 620, 624 (1971) (concurring opinion). See In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368 (1970).
The Court declines to consider the constitutionality of the statute and instead provides the district court with alternative instructions on the burden of proof. In my opinion, we should resolve at this time the question of the statute’s constitutionality. If the statutory change is invalid and the government must prove beyond a reasonable doubt that the defendant was responsible for his conduct, we can no longer pretend not to notice that defendants are being overwhelmed by an invisible burden of proof. And if the statute’s attempt to shift the burden of persuasion onto the defendant is constitutional, then we must still take steps to facilitate the production at trial of meaningful information by both the government and the defense.
3. I applaud the Court's decision to overturn Fisher v. United States, 80 U.S.App.D.C. 96, 149 F.2d 28 (1945), aff’d, 328 U.S. 463, 66 S.Ct. 1318, 90 L.Ed. 1382 (1946) and Stewart v. United States, 107 U.S.App.D.C. 159, 275 F.2d 617 (1960), rev’d on other grounds, 366 U.S. 1, 81 S.Ct. 941, 6 L.Ed.2d 84 (1961), and to make clear that a defendant can introduce psychiatric and other expert testimony to negative specific intent. I suggested in two recent cases that Fisher and Stewart did not preclude our adoption of this doctrine, inaptly termed “diminished responsibility,” and that it was therefore unnecessary to overrule those cases. See United States v. Bryant, 153 U.S.App.D.C.-, 471 F.2d 1040 (April 21, 1972) (dissenting opinion); United States v. Alexander & Murdock, 152 U.S.App.D.C. -, 471 F.2d 923 (April 21, 1972) (dissenting opinion). Two panels of this Court rejected my view and concluded that the doctrine could not be accepted without an en banc decision of the Court. The Court now sits en banc and concludes that expert testimony is relevant to the determina*1039tion of specific intent where the defendant is charged with murder in the first degree.
The Court points out, however, that it does not decide whether the doctrine is applicable to cases of second-degree murder, where the prosecution must prove that the defendant acted with a state of mind called “malice.” In Murdock, where the defendant was charged with second-degree murder, I discussed the argument against applying the doctrine so as to reduce the offense of second-degree murder to manslaughter. The argument rests on the premise
that malice refers not to a state of mind, but to an objective set of circumstances ; it can be negated by evidence of circumstances that would provoke a reasonable man to act in the heat of passion, but not by evidence of actual subjective provocation and passion. * * * In a recent series of cases, however, we reviewed with some care the concept of malice, and concluded, inter alia, that it is not entirely an objective matter, but has subjective elements as well.
152 U.S.App.D.C. at-, 471 F.2d at 950 (footnotes omitted). But even though the Court apparently concedes that in some cases malice is established on a subjective standard, it concludes that the “matter * * * requires further analysis and reflection,” and “[t]he problem is [therefore] remitted to future consideration.” Majority opinion at 1002 n. 75.
While I am convinced that the question can be resolved without delay, I would have no objection to the Court’s cautious approach if the question had no application to the case before us. But it should be clear that the question is directly relevant to the disposition of this case. Although originally charged with first-degree murder, Brawner was acquitted on that count by the trial court before the case was submitted to the jury. He was convicted of murder in the second-degree. The Court thus resolves the question of diminished responsibility up to the point where it becomes relevant to this case, and it remits to future consideration the only aspect of the issue which could have any bearing on the outcome of the case before us. That bizarre result is justified with the comment that “future consideration * * * will be aided by the availability of a specific factual context.” Majority opinion at 1002, n. 75. The Court’s refusal to consider the question in the case before us, where a “specific factual context” plainly exists, seems to me entirely inconsistent with the fair and efficient administration of justice.
VII. CONCLUSION y
This Court’s search for a new set of words to define the elusive concept of responsibility has a distinctly archaic quality. The arguments for and against the Durham wording, the wording of the majority and minority versions of the ALI test, and the wording of McDonald, were clearly articulated many years ago. What should by now be clear is that the problems of the responsibility defense cannot be resolved by adopting for the standard or the jury instruction any new formulation of words, The practical operation of the defense is primarily controlled by other factors, including the quality of counsel, the attitude of the trial judge, the ability of the expert witnesses, and the adequacy of the-pretrial mental examination. If the adoption of the ALI test produces some improvement in the quality of adjudication of the responsibility issue, that, of course, is all to the good. But we cannot allow our search for the perfect choice of words to deflect our attention from the far more important practical questions. For it is on those questions that the rationality and fairness of the responsibility defense will ultimately turn.
. Our far-ranging experience with the responsibility defense has led me in recent years to urge fundamental changes in the defense. See United States v. Alexander & Murdock, 152 U.S.App.D.C.-, 471 F.2d 923 (April 21, 1972) (separate opinion) ; United States v. Leazer, 148 U.S.App.D.C. 356, 460 F.2d 864 (Jan. 19, 1972) (concurring opinion) ; United States v. Trantham, 145 U.S.App.D.C. 113, 448 F.2d 1036 (1971) (statement in support of rehearing en banc) ; United States v. Eichberg, 142 U.S.App.D.C. 110, 439 F.2d 620 (1971) (concurring opinion).
. Prior to our decision in Durham, the test of criminal responsibility in this jurisdiction was the rule established in M’Naghten’s Case, 8 Eng.Rep. 718 (1843), joined with the so-called irresistible impulse test. See Smith v. United States, 59 App.D.C. 144, 36 F.2d 548 (1929).
. McDonald defined mental disease in legal terms as “any abnormal condition of the mind which substantially affects mental or emotional processes and substantially impairs behavior controls.” 114 U.S.App. at 124, 312 F.2d at 851.
. To be sure, the Court’s decision does have the important intention of abolishing the unnecessary and misleading emphasis on productivity that has characterized tiie adjudication of the responsibility issue in this jurisdiction. But see pages 1022-1027, infra.
. Cf. United States v. Carter, 141 U.S.App.D.C. 46, 56, 436 F.2d 200, 210 (1970) (concurring opinion) : “It may well be that we simply lack the resources — to say nothing of the understanding — that would be required if those who stole to feed their addiction were removed from the criminal process on the ground that they are not responsible for their actions. But if this is so, we should recognize the fact, and not rationalize our treatment of narcotics addicts on the false premise that their crimes are the result of a wrongful exercise of free will.”
. D.C.Code § 24-301(a) (Supp. V 1972) ; Winn v. United States, 106 U.S.App.D.C. 133, 270 F.2d 326 (1959).
. This is of course the legal test of responsibility set forth in Durham v. United States, 94 U.S.App.D.C. 228, 240-241, 214 F.2d 862, 874-875 (1954), as modified by McDonald v. United States, 114 U.S.App.D.C. 120, 124, 312 F.2d 847, 851 (1962) (en banc).
. “Psychologic brain syndrome associated with convulsive disorder,” “personality disorder associated with epilepsy,” “explosive personality, or epileptic personality disorder,” “explosive personality with epileptoid personality disorder.”
. Dr. Hamman reported exaggerated emotional responses, and a tendency to go into long lasting rages. Dr. Stammeyer, the clinical psychologist, found an explosive response to threats and a tendency to perseverate — to fix on an idea and remain preoccupied with it after it was no longer appropriate. Dr. Weickhardt and Dr. Platkin stated that in their view appellant’s disorder would result in impulsive, purposeless, and uncontrolled acts in immediate and exaggerated response to a situation.
. Which is not surprising in view of his “dull normal” I.Q. of 82, his sixth grade education, his rejection by the Armed Forces for failure to pass the aptitude test, and his release from various jobs because of asthmatic attacks.
. 94 U.S.App.D.C. 228, 240-241, 214 F.2d 862, 874-875 (1954).
. M’Naghten’s Case, 8 Eng.Rep. 718, 722 (1843) ; Smith v. United States, 59 App.D.C. 144, 146, 36 F.2d 548, 550 (1929) (adding irresistible impulse test to M’Naghten). For critical discussion of the various tests of responsibility, see A. Goldstein, The Insanity Defense (1967).
. See Durham v. United States, 94 U.S.App.D.C. at 236-240, 214 F.2d at 870-874 and sources cited. Professor Gold-stein has argued persuasively that the traditional rules can be given an expansive reading that meets these objections. “Knowledge” of right and wrong can be read to include emotional appreciation as well as cognition. And the control tests can be read to reach a wide range of impaired behavior controls, in addition to the well-known “irresistible impulse.” Goldstein argues that only their own narrow vision prevents defense lawyers and psychiatrists from introducing contemporary psychiatric insights under the traditional tests of responsibility. He suggests that psychiatric testimony has seldom in practice been limited by narrow judicial application of the rules. A. Goldstein, The Insanity Defense 45-79 (1967). In this jurisdiction, however, the traditional rules were regularly applied to prevent psychiatrists from explaining their insights in broad terms. See, e. g., Durham v. United States, 94 U.S.App.D.C. at 234, 214 F.2d at 868. Furthermore, any testimony that passed the initial hurdle of admissibility then had to pass through the eye of a needle in the form of the jury instruction. While I sympathize with Goldstein’s proposal for expanding the old tests, it seems to me more effective to adopt a new one, unencumbered with restrictive interpretations. I have no illusions about the power of a new verbal formula to affect the behavior of lawyers and experts, courts and jurors with respect to the troublesome issue of responsibility. Gold-stein, supra, at 94-95. But our obligation to supervise the administration of the test carries with it the obligation to offer such guidance as we can in the form of analysis and restatement.
. See, e. g. M. Guttmacher & H. Weihofen, Psychiatry and the Law 406-07 (1952), and sources cited. A group of psychiatrists urging a change in the test of responsibility was unusually candid about what they saw as an obligation to temper psychiatric conclusions with moral judgments. “The pivotal assumption [of M’Naghten] is that in some degree a disorder of the cognitive faculty (knowledge) is the only basis for the determination of responsibility. This confines the psychiatrist to an exceedingly short tether and is usually his undoing. There is much more to be said about mental disease, motivations and of behavior, and the psychiatrist can do little else but cut corners on the question of ‘knowledge’.” Committee on Forensic Psychiatry, Group for the Advancement of Psychiatry, Criminal Responsibility and Psychiatric Expert Testimony 17 (Preliminary Report). The last sentence was deleted from the final version, GAP Report No. 25 (1954).
. See R. Leifer, In the Name of Mental Health 196-98, (1969); T. Szasz, Law, Liberty, and Psychiatry 136-37 (1963).
. See, e. g., Wechsler, The Criteria of Criminal Responsibility, 22 U.Chi.L.Rev. 367 (1955) ; Frigillana v. United States, 113 U.S.App.D.C. 328, 331, 307 F.2d 665, 668 (1962) ; State v. Lucas, 30 N.J. 37, 70-71, 152 A.2d 50, 67 (1959).
. E. g., Washington v. United States, 127 U.S.App.D.C. 29, 390 F.2d 444 (1967) ; McDonald v. United States, 114 U.S.App.D.C. 120, 312 F.2d 847 (1962) (en banc) ; Campbell v. United States, 113 U.S.App.D.C. 260, 307 F.2d 597 (1962) ; Briscoe v. United States, 101 U.S.App.D.C. 318, 248 F.2d 640 (1957) (statement of Bazelon, J. in relation to petition for leave to appeal in forma pauperis) ; Stewart v. United States, 101 U.S.App.D.C. 51, 247 F.2d 42 (1957) ; Stewart v. United States, 94 U.S.App.D.C. 293, 214 F.2d 879 (1954).
. As a result of the sudden decision of St. Elizabeth’s Hospital in 1957 to treat “soeiopathic personality disturbance” as a mental illness, the court has been very sensitive to the danger of allowing psychiatric labels to determine legal results. See Blocker v. United States, 110 U.S.App.D.C. 41, 48-50, 288 F.2d 853, 860-862 (1961) (Burger, J., concurring) ; Blocker v. United States, 107 U.S.App.D.C. 63, 274 F.2d 572 (1959). Compare United States v. Collins, 139 U.S.App.D.C. 392, 400-401 n. 3, 433 F.2d 550, 558-559 n. 3 (1970) (concurring and dissenting opinion) (changing psychiatric views of narcotic addiction) ; Salzman v. United States, 131 U.S.App.D.C. 393, 407-408 n. 43, 405 F.2d 358, 372-373 n. 43 (1968) (Wright, J., concurring) (alcoholism).
. 102 U.S.App.D.C. 227, 252 F.2d 608 (1957). The trial court had instructed the jury that the act must be the direct consequence or natural result of the illness. This court disapproved that instruction and formulated a test that purported to include virtually any mode of effective causation. The illness might have been the source of the defendant’s urge to do the act, or it might have exaggerated an otherwise ordinary response to an external threat; it might have impaired his ability to control impulses normally repressed, or it might liave impaired his appreciation of the necessity for such control. So long as “the disease made the effective or decisive difference between doing and not doing the act,” then the act was the product of the disease for the purpose of Durham.
. 114 U.S.App.D.C. 120, 124, 312 F.2d 847, 851 (1962) (en bane). In order to - prevent expert opinion from controlling the question of responsibility, we set forth a legal definition of mental illness against which expert testimony could be measured.
. The use of conclusory psychiatric labels without description of the underlying data carries an assurance of certainty while systematically hiding from the jury the considerations that influenced the expert’s opinion. Unfortunately, it is not uncommon for extraneous and undisclosed factors to be taken into account. For example, the witness may have moulded his testimony so as to make possible the legal disposition which he considered most beneficial to the defendant or the public. Or, if the witness is a psychiatrist at a government hospital which receives persons found not guilty by reason of insanity, his diagnosis may be designed to insure that a person he considered untreatable or a troublemaker would not be returned to the custody of his hospital. And even if psychiatrists do not make a deliberate effort to frame their diagnoses in terms of a preferred legal disposition, their conclusions may still be colored by other factors not directly related to the subject’s condition. See, e. g., Brown, The Doctor’s Reasons for Referral, in The Prevention of Hospitalization 131 (M. Greenblatt, R. Moore, R. Albert & M. Solomon eds. 1963) ; Katz, Cole & Lowery, Studies of the Diagnostic Process: The Influence of Symptom Perception, Past Experience, and Ethnic Background on Diagnostic Decisions, 125 Am.J.Psych. 937 (1969).
Moreover, the use of conclusory labels often makes it impossible for the jury to appreciate other potentially critical shortcomings of the expert opinion. The expert witness may have been unable to prepare adequately for trial, perhaps because he lacked the time or the facilities to carry out an adequate examination of the accused. Or he may have lacked the ability or the training to evaluate an unusual condition. Perhaps the state of psychiatric knowledge generally would not permit an unequivocal diagnosis of the defendant’s condition by even the most outstanding expert. So long as these deficiencies hide behind the experts’ jargon and conclusions, the jury cannot rationally deal with even the medical component of the responsibility question. The solution is not, of course, to bar expert testimony except where that testimony is utterly unimpeachable. In the words of the Washington instruction to expert witnesses in responsibility cases,
We recognize that an opinion may be merely a balance of probabilities and that we cannot demand absolute certainty. Thus you may testify to opinions that are within the zone of reasonable medical certainty. The crucial point is that the jury should know how your opinion may be affected by limitations of time or facilities in the examination of this defendant or by limitations in present psychiatric knowledge. The underlying facts you have obtained may be so scanty or the state of professional knowledge so unsure that you cannot fairly venture any opinion. If so, you should not hesitate to say so. And, again, if you do give an opinion, you should explain what you did to obtain the underlying facts, what these facts are, how they led to the opinion, and what, if any, are the uncertainties in the opinion.
129 U.S.App.D.C. 29, 43, 390 F.2d 444, 458 (1967). The Court’s retention today of the bulk of the Washington instruction, see page 1027 n. 43 infra, including the portion quoted above, reaffirms our determination to advise the jury fully about what we know — and what we do not know — about the defendant.
. 129 U.S.App.D.C. 29, 40-41, 390 F.2d 444, 455-456 (1967). The reluctance of the court to enforce a rigid limitation on expert testimony has been apparent from the start. Although the Washington opinion expressly prohibits testimony “directly in terms of ‘product,’ or even ‘result’ or ‘cause’ ” the sample instruction to expert witnesses appended to the opinion does not prohibit such testimony; it merely advises the expert that “it will not be necessary for you to express an opinion on whether the alleged crime was a ‘product’ of a mental disease or defect and *1019you will not be asked to do so.” Id. at 42, 390 F.2d at 457. See also the concurring opinion of Judge Fahy. Id. at 45-46, 390 F.2d at 460-461.
. There is, of course, great pressure on the experts to decide the moral questions tacitly in the guise of making a scientific determination, and thereby relieve society of the need to make some difficult decisions. We can easily understand why the experts yield to that pressure, but we cannot approve the result or stop trying to force the moral questions out of the scientific domain and into the public arena.
. See pages 1013-1014 supra.
. 102 U.S.App.D.C. at 236, 252 F.2d at 617.
. Garter acknowledged the problem at issue here, but dismissed it as a “logician’s nicety,” stating that in the ordinary case we made the “tacit assumption that if the disease had not existed the person would have been a law-abiding citizen.” Id. Taken literally, that assumption amounts to an assumption of productivity and hence of nonresponsibility. Perhaps it would have been more accurate to say that in the ordinary case it is not anticipated that the inquiry into productivity will focus on the likelihood that the defendant would have committed the act without his illness. As a prediction that statement has proved false, and the problem should not be ignored.
. Dr. Weickhardt testified that people with appellant’s illness certainly do not have as good control over their behavior as other people, that in some cases they become irritable and angry faster than normal people, and that under stress they) may react impulsively. Dr. Platkin testified that appellant’s illness involved emotional instability, “a low fuse level of tolerance,” and a general pattern of getting involved in fights and reacting way out of proportion to a situation.
. See page 1014 supra.
. A. Hollingshead & F. Redlich, Social Class and Mental Illness (1958).
. Compare In re Betty Jean Williams, No. 27-220-3 (D.C.Juv.Ct., Oct. 20, 1959) (denying motion for mental examination) :
Her precocious sexual experiences are certainly pathetic but neither in themselves nor in conjunction with the associated mental symptoms are they indicative of a mental disturbance sufficient to prevent her understanding of the proceedings or assisting in her defense. Such experiences are far from being uncommon among children in her socioeconomic situation with the result that the traumatic effect may be expected to be far less than it would be in the case of a child raised by parents and relatives with different habits.
Memorandum opinion at 2.
The assumption that poor people are less seriously affected by traumatic events and by mental illness provides a convenient rationale for society’s refusal to provide the resources that would be necessary to deal honestly with their problems. The *1021underlying explanation for the Williams decision is not that Miss Williams failed to present a convincing claim, of mental disorder, but rather that her claim was no more compelling than that of many other children, more numerous than the court could possibly help. “In view of respondent’s personal history it is scarcely surprising that she feels ‘tense and unhappy and in need of psychiatric help.’ But so do a vast number of the children coming before this court.” Id.
. Compare United States v. Carter, 141 U.S.App.D.C. 46, 55 n. 14, 436 F.2d 200, 209 n. 14 (1970) (concurring opinion of Bazelon, C. J.) (rejecting psychiatrist’s speculation that, had appellant not suffered from anxiety, he might nevertheless have become addicted to drugs, because he lived in an environment where drug addiction was common).
. See page 1014 supra.
. The variety of approaches which hide behind ALI’s language are carefully delineated in the excellent brief filed by William H. Dempsey, the amicus appointed by this Court, at 980-986. Mr. Dempsey’s brief focuses on the construction of the ALI test by federal appellate courts. It would thus be a most useful line of inquiry to determine how the differences on the appellate level are reflected in the transcripts of cases tried in these other jurisdictions. It may well be that the variation in approaches revealed by Mr. Dempsey’s study is only the tip of the iceberg.
. United States v. Eichberg, 142 U.S.App.D.C. 110, 118 n. 40, 439 F.2d 620, 628 n. 40 (1971) (Bazelon, C. J., concurring). See page 1016 and n. 16 supra.
. Durham v. United States, 94 U.S.App.D.C. 228, 235, 214 F.2d 802, 869 (1954), quoting Glueck, Mental Disorder and the Criminal Law 138-39 (1925), and Rex v. Arnold, 16 How.St.Tr. 695, 764 (1724).
. See, e. g., Royal Comm’n on Capital Punishment 1949-53, Report § 280 at 99:
Where a person suffering from a mental abnormality commits a crime, there must always be some likelihood that the abnormality has played some part in the causation of the crime; and, generally speaking, the graver the abnormality and the more serious the crime, the more probable it must be that there is a causal connection between them.
J. Page, Psychopathology: The Science of Understanding Deviance 30-32 (1971) :
The label of “psychosis” is essentially restricted to the most severe behavior disorders that occur in adults and children. * * * A significant distinctive feature of psychotic behavior is that it is relatively independent of voluntary control or external reality. * * * A second distinctive feature consists of varying degrees of personality distintegration with consequent significant impairment in personal and social functioning. As p a rule, the behavior of psychotic individuals is so defective and disorganized that they require care or supervision. * * *
The psychotic person finds it difficult or impossible to differentiate between fantasies and actual experiences. Wishes tend to be confused with facts; imagined dangers, slights, and misdeeds are accepted as real, and real ones are grossly exaggerated or misinterpreted. Whereas normal people attempt to adapt their behavior to the expectations and demands of the physical and social environment, the psychotic’s reactions are more or less exclusively dominated by inner dictates. * * *
The personality disintegration and reality distortion characteristic of psychoses are strikingly apparent in delusions and hallucinations. * * *
Other types of mental patients, and normal persons ns well, may also hold onto false beliefs and experience hallucinations. What distinguishes psychotic behavior is not the presence or absence of delusions and hallucinations per se, but rather the extent to which they pervade, dominate, and distort the person’s perceptions, feelings, decisions, and actions.
(Emphasis supplied.)
. See, e. g., Page, supra note 36, at 33-35. Of course, even some psychoses may be substantially encapsulated, in which case a productivity problem could obviously arise. Page points out that “[i]n contrast to other psychotic reactions that are accompanied by a general disorganization of personality and gross impairment in general functioning, paranoia and paranoid states consist mainly of a capsulated persecutory or grandiose delusional system in an otherwise relatively intact personality. The patient’s functioning is unaffected in areas outside of his delusional system.” Id. at 32. Thus, it is possible to imagine a perplexing question of productivity in a case involving, for example, a businessman who *1025had cheated on his income tax for many years, and who eventually developed a full-blown paranoid delusional system. If he continued to falsify his return after the onset of his illness, the question of productivity would presumably present great difficulty.
. See amicus brief of William H. Dempsey at 18-19, citing, inter alia, an explanatory comment by the Reporter of the Model Penal Code, Professor Wechsler: “[I]t was thought that the criterion should ask if there was * * * a deprivation of ‘substantial capacity’ to know or to control, meaning thereby the reduction of the capacity to the vagrant and trivial dimensions characteristic of the most severe afflictions of the mind.’’ Wechsler, Codification of Criminal Law in the United States: The Model Penal Code, 68 Colum.L.Rev. 1425, 1443 (1968) (emphasis added).
. See amicus brief of William H. Dempsey at 19-31.
. Id. at 16 (footnotes omitted).
. Of the other Circuit opinions cited by the Court at page 979 supra, only one even acknowledges the existence of a causality question under the ALI test. See United States v. Freeman, 357 F.2d 606, 623 (2d Cir. 1966) :
Relieved of their burden of divining precise causal relationships, the judge or jury can concentrate upon the ultimate decisions which are properly theirs, fully informed as to the facts.
[Emphasis added.]
. Washington, of course, made clear that psychiatrists “should not speak directly in terms of ‘product,’ or even ‘result’ or ‘cause.’ ” 129 U.S.App.D.C. 29, 41, 390 F.2d 444, 456 (1967) (emphasis added).
. The Court does retain, however, the portion of the Washington instruction unrelated to the issue of productivity. I applaud that decision because of the significant salutary effect that the instruction has had on the adjudication of the responsibility issue in this jurisdiction. See page 1018 n. 21 supra.
. Compare Heard v. United States, 121 U.S.App.D.C. 37, 40, 348 F.2d 43, 46 (1965) : “[T]here was no evidence that appellant’s capacity to control his Jehavior was impaired. * * * [T]he McDonald standard for submission of the criminal responsibility issue was not met * * * ” (emphasis supplied).
. See United States v. Eichberg, 142 U.S.App.D.C. 110, 116 & n. 31, 439 E.2d 620, 626 & n. 31 (1971) (Bazelon, C. J., concurring), citing
J. Elkes, Word Fallout: or, on the Hazards of Explanation, in The Psychopathology of Adolescence 118 (1970) (presidential address, Am. Psychopathological Ass’n) ; R. Leifer, In the Name of Mental Health, 196-98 (1969) ; K. Menninger, Toward A Unitary Concept of Mental Illness, in A Psychiatrist’s World 516 (1959) ; K. Menninger, Changing Concepts of Disease, in A Psychiatrist’s World 670 (1959) ; M. Roth, Seeking Common Ground in Contemporary Psychiatry, 62 Proceedings of the Royal Soc’y of Medicine 765 (1969) (presidential address, section of psychiatry) ; M. Susser, Community Psychiatry 10-20 (1968) ; T. Szasz, The Myth of Mental Illness (1961).
The medical model of mental illness has been questioned ever more extensively by behavioral scientists outside psychiatry. See, e. g., G. Albee, The Uncertain Future of Clinical Psychology, 25 American Psychologist 1071 (1970) (presidential address, Am. Psychological Ass’n) ; E. Wolf, Learning Theory and Psychoanalysis, 39 British Journal of Medical Psychology 525 (1969) (paper and critical evaluations) ; The Mental Patient: Studies in the Sociology of Deviance (S. Spitzer & N. Denzin ed. 1988).
Cf. Blocker v. United States, 110 U.S.App.D.C. 41, 48, 288 F.2d 853, 860 (1961) (Burger, J., concurring) (“no rule of law can possibly be sound or workable which is dependent upon the terms of another discipline whose members are in profound disagreement about what those terms mean”) ; Campbell v. United States, 113 U.S.App.D.C. 260, 266, 307 F.2d 597, 603 (1962) (Burger, J., dissenting) .
. The circularity of our new test becomes apparent when the three facets — ALI, McDonald, and the “broad consensus”— are read together and in proper sequence. Henceforth, a person is not criminally responsible if, as a result of mental disease or defect, which is (a) an abnormal condition of the mind which substantially affects mental or emotional processes and substantially impairs behavior controls, and (b) an ascertainable condition characterized by a broad consensus that free *1030will does not exist, he lacks substantial capacity either to appreciate the wrongfulness of his conduct or to conform bis conduct to the requirements of law.
. United States v. Eichberg, 142 U.S.App.D.C. 110, 114-115, 439 F.2d 620, 624-625 (1971) (Bazelon, C. J., concurring) .
. The Court continues with a quotation from Williams v. Florida, 399 U.S. 78, 100, 90 S.Ct. 1893, 1906, 26 L.Ed.2d 446 (1970), pointing out that the essential feature of a jury “lies in the interposition between the accused and his accuser of the commonsense judgment of a group of laymen, and in the community’s participation and shared responsibility that results from that group’s determination of guilt or innocence.”
. Cf. Holloway v. United States, 80 U.S.App.D.C. 3, 4, 148 F.2d 665, 666 (1945) : “The application of these tests [McNaghten and irresistible impulse], however they are phrased, to a borderline case can be nothing more than a moral judgment that it is just or unjust to blame the defendant for what he did.”
. Thus, it is distressing to find in the case at bar that the prosecutor and the court below seemed concerned with establishing that appellant’s epileptoid disorder may have been “physiological as over against mental.” The determination of criminal responsibility cannot turn on the outcome of a debate about whether epilepsy is a mental illness or a physical one.
. Model Penal Code § 4.01, Comment at 159 (Tent.Draft No. 4, 1955) (emphasis added). See ALI Proceedings 206-20 (May 21, 1955) (unpublished).
. In United States v. Alexander & Murdock, 152 U.S.App.D.C.-at---, 171 F.2d 923 at 960-965 (April 21, 1972), (separate opinion), I pointed out that changes in the reach of the responsibility defense could have important ramifications for the doctrine of civil commitment. If we diminish the class of persons who can be found criminally responsible, wo may produce a concomitant expansion in the class of persons who can be subjected to involuntary civil commitment. Adoption of a jury instruction like the minority ALI test would presumably not give rise to such an expansion since the test does not expand the category of persons who can be exculpated by a responsibility defense. It merely gives explicit recognition to the jury’s function in resolving a question of degree. That same function is implicit in every test of criminal responsibility.
. The Court’s refusal to reverse the conviction rests in part on the doctrine of “curative admissibility.” There may be cases in which a party’s introduction of irrelevant or otherwise inadmissible testimony confers on his adversary the right to introduce in rebuttal further evidence that would otherwise be inadmissible. But such a rule is discretionary and cannot be invoked when it would subvert a fundamental substantive policy like .that of Washington, to preserve for the jury its critical role in assessing criminal responsibility. See United States v. Winston, 145 U.S.App.D.C. 67, 447 F.2d 1236 (1971) ; United States v. Thompson, 150 U.S.App.D.C. 403, 465 F.2d 583 (May 8, 1972). See generally 1 J. Wigmore, Evidence § 15 (3d ed. 1940, Supp.1964). Defense counsel’s inquiry into productivity here was undoubtedly inspired by the certain knowledge that the government would ground its case on evidence of nonproductivity. Before any expert testimony was presented, the trial court correctly stated the Washington rule, but *1036failed to apply it during the ensuing examination of both prosecution and defense witnesses. We cannot say that the effect of the prosecution’s impermissible testimony was neutralized by that of the defense. The proper approach was not to admit both but to exclude both.
. One recent study of jury behavior in responsibility cases suggests that jurors have a systematic tendency to view a defendant as sane when the expert testimony is in conflict. See Klein & Temerlin, On Expert Testimony in Sanity Cases, 149 J. Nervous & Mental Disease 435 (1969). Summarizing the study, which was based on the behavior of 96 mock juries, the authors conclude that jurors were influenced by expert testimony when the testimony was not conflicting. In cases of conflicting expert testimony, or when there was no expert witness, there was a tendency for the jury to vote sane even though the defendant was clearly psychotic by the usual clinical criteria.
Id. at 438. Earlier in the same article, the authors described the “tendency” as “overwhelming.” Id. at 437. See also R. Simon, The Jury & the Defense of Insanity (1967).