# United States v. Willie Decoster, Jr., (Decoster Iii)

> Court of Appeals for the D.C. Circuit · July 10, 1979 · 624 F.2d 196

URL: https://www.frixlaw.com/law-library/cases/9466830

## Case

- **Full name:** UNITED STATES of America v. Willie DECOSTER, Jr., (Decoster III), Appellant
- **Court:** Court of Appeals for the D.C. Circuit
- **Decided:** July 10, 1979
- **Citations:** 624 F.2d 196; 199 U.S. App. D.C. 359
- **Precedential status:** Published
- **Opinion:** Concurrence by Robinson
- **Judges:** III, Robinson, Bazelon, Wright, Ba-Zelon, McGowan, Tamm, Leven-Thal, MacKinnon, Robb, Wilkey, Leventhal
- **Cited by:** 264 later opinions in the Frix Law Library

## Citator (automated)

- **Yellow flag:** Declined to follow by Washington v. Strickland, 693 F.2d 1243 (1982).
- Negative treatments: 1
- Distinguished by: 0
- Full citator and citing cases: https://www.frixlaw.com/law-library/cases/9466830

## How later opinions describe it (automated extraction)

- stating that "the problem of inadequate representation of the indigent cannot be solved by the courts alone" and recommending steps to be taken by the bar and the legislatures

## Opinion text

SPOTTSWOOD W. ROBINSON, III, Circuit Judge,
concurring in the result:
I agree with the majority of my brethren that the conviction in this case should be affirmed. I am unable, however, to subscribe to one link in the chain of reasoning they forge in reaching this result. While I applaud the collective action of the full court in now assuring to those accused of crime a level of legal assistance commensurate with the demands of the Sixth Amendment, I deplore the court’s allocation to the accused of a burden of demonstrating that he was jeopardized by established dereliction of duty on the part of his counsel. Judges Leventhal and MacKinnon, and my colleagues concurring with them, make prejudice to the defense — likely harm in the one instance and actual injury in the other — an indispensable prerequisite to any finding of ineffective assistance, and assign the onus of proof on that score to the defendant. 1 In so holding they stray rather widely, I believe, from established principles of pertinent jurisprudence.
These considerations summon this opinion, and dictate the general course it will take. I first summarize in brief fashion the position of the court. 2 Next I set forth my understanding of the test imposed by the Sixth Amendment for measuring the sufficiency of the service rendered by counsel for an accused. 3 I then elucidate my stand on burden of proof of prejudice in ineffective-assistance cases. 4 Lastly, I explain why I conclude that counsel’s performance in this case was constitutionally ineffective but nevertheless was harmless error. 5
I. THE COURT’S POSITION
The critical issue we are convened to resolve is the standard appropriately to be utilized in evaluating claims that defense counsel’s performance was constitutionally infirm. Early cases in this circuit shunned the Sixth Amendment as a source of entitlement to effective aid by a member of the bar. 6 Relying instead upon the Fifth, those decisions measured counsel’s adequacy by the impact of any deficiency on the fairness of the trial; 7 resultantly, there was concern only if execution of the defense function was so abominable that it rendered the trial “a farce and a mockery of justice.” 8 As late as Bruce v. United States 9 in 1967, this court maintained that while “[t]hese words [were] not to be taken literally,” they nevertheless were “a vivid description of the principle that the accused has a heavy burden in showing requisite *246 unfairness.” 10 And though it is now said that Bruce implicitly took a Sixth Amendment approach to the problem, 11 the Bruce court acknowledged no more than that in “rare and extraordinary” instances “an accused may obtain relief . . . if he shows both that there has been gross incompetence of counsel and that this has in effect blotted out the essence of a substantial defense either in the District Court or on appeal.” 12 That was then the court’s concept of the accused’s constitutional due, at least where the question was presented by collateral attack upon a conviction. 13
As Judge Leventhal’s survey of judicially-enunciated formulae for gauging ineffective-assistance claims discloses, every test thus far developed in this and other circuits, however expressed in words, has imposed an initial burden on the defendant to establish that his counsel’s performance at trial was abnormally deficient. 14 Every opinion announced today espouses a standard incorporating that thesis centrally as a hurdle that the defendant must first clear. Judge Leventhal, for a plurality of the court, insists upon proof of “serious incompetency, inefficiency or inattention of counsel — behavior of counsel falling measurably below that which might be expected from an ordinary fallible lawyer.” 15 Judges Bazelon and MacKinnon would require the defendant to demonstrate, not merely a violation of particular duties, but a “substantial” violation. 16 I myself believe that the defendant must point to some substantial deviation from a norm of reasonable competence. 17 Despite the terminological differences in the heft of the showing to be made, each formulation emphasizes that counsel’s breach must be serious, and that the defendant bears the onus of making it out.
Perhaps more importantly, the court is agreed that the “gross incompetence” standard of Bruce is dead, and that in my view is how it should be. For nearly four decades the guaranty of competent representation in federal criminal proceedings has had Sixth 18 as well as Fifth Amendment 19 underpinnings, a verity long calling for a thorough reexamination of this circuit’s criteria for proving and assessing asserted viola *247 tions. So, while a defendant must show some substantial dereliction in his counsel’s performance in order to qualify for relief, the level of the representation required is now considerably higher than it once was. On the other hand, a majority of my colleagues retain in some form the other aspect of Bruce — the defendant’s burden of proof vis-a-vis resulting prejudice — as a constitutional element in the adjudication of ineffective-assistance contests. A plurality of the court sets the required showing as likely harm; Judge MacKinnon says it should be actual harm. 20 I am unable to perceive any sound justification for saddling the defendant with the additional obligation of establishing that a demonstrated transgression of his right to effective aid by a lawyer either probably or actually influenced the outcome of his case.
II. THE CONSTITUTIONAL TEST
The Sixth Amendment solemnly proclaims that “[i]n all criminal prosecutions, the accused shall enjoy the right . to have the Assistance of Counsel for his defence.” 21 So plain and potent an injunction obviously demands a great deal more than mere appointment and physical presence of a legal representative. Indeed, it has long been recognized that one accused of crime must be afforded, even apart from the Sixth Amendment mandate, a lawyer’s help in substance as well as form. In Powell v. Alabama 22 in 1932, the Supreme Court held that, simply as a matter of due process, it was not enough that counsel for a defendant in a state capital prosecution is “assign[ed] at such a time or under such circumstances as to preclude the giving of effective aid in the preparation and trial of the case.” 23 And during the nearly half-century since Powell was decided, it has become increasingly clear that the Sixth Amendment entitlement to “assistance of counsel,” no less than its due process counterpart, is fully the right to effective assistance of counsel. 24
The Supreme Court has never undertaken to delineate the content of the “effective aid” of which Powell spoke, or to comprehensively define the standard of counsel-aid constitutionally demanded. That might be taken as good reason for believing that the expression is to have its natural and ordinary meaning; in any event, its connotation can hardly be mistaken. Counsel, to be sure, is not required to win the case, but “effective aid” certainly contemplates that counsel will endeavor to pursue a course reasonably calculated to achieve for the accused the most advantageous resolution of the case possible under the circumstances. Equally certain it is that an accused is due more than a lawyer whose performance barely escapes the label “grossly incompe *248 tent” — the standard espoused in Bruce. 25 In my view, “effective” assistance is a call for reasonably competent assistance, 26 for anything less robs “effective” of far too much of its evident meaning.
That the Supreme Court intends “effective aid” to signify reasonable competence is evident from a number of its decisions. In McMann v. Richardson 27 the Court declared that “defendants . . . are entitled to the effective assistance of competent counsel,” 28 and that counsel’s advice must be “within the range of competence demanded of attorneys in criminal cases.” 29 Moreover, the Court has utilized the term “effective” to zealously guard from outside interference activities of counsel designed to maintain the representation at a wholesome level. 30 Defense counsel must be free to decide whether and when to put his client on the witness stand. 31 He must be permitted to present a closing summation. 32 He must be allowed to confer with his client during an overnight recess. 33 It would be incongruous for the Court to protect so scrupulously the right to effective assistance when it is endangered by outside tampering but contemporaneously to ignore all but gross incompetence when substantially deficient service is traced directly to shortfalls in counsel’s ability or effort.
The farce-and-mockery test of effective assistance, to which the gross-incompetence test is first cousin, is definitely on the wane. 34 A majority of the federal circuits now have adopted some version of reasonable competence. 35 With that I agree, and I *249 cannot improve on the DeCoster I formulation: “a defendant is entitled to the reasonably competent assistance of an attorney acting as his diligent conscientious advocate.” 36 All of the standard-of-performance rhetoric in today’s opinions seems to come down ultimately to essentially this yardstick for measurement of defense counsel’s performance.
I perceive no policy consideration sufficiently forceful to persuade me that utilization of this standard bodes ill for either the attorney-client relationship or the adversary system. Reasonable competence is the concept traditionally and universally employed as the measure of the lawyer’s civil liability, without apparent untoward effect. 37 I cannot see how it could take on a destructive propensity merely because the object is reversal of a conviction rather than an assessment of damages. 38 And even assuming that resort to the familiar doctrine of reasonableness upon judicial evaluation of defense counsel’s performance may lead to a modicum of trial-court involvement in defense activities, some policing of counsel’s rendition is both necessary and appropriate. 39 As the Supreme Court has admonished,
if the right to counsel guaranteed by the Constitution is to serve its purpose, defendants cannot be left to the mercies of incompetent counsel . . . . [JJudges should strive to maintain proper standards of performance by attorneys who are representing defendants in criminal cases in their courts. 40
Nor do I have any quarrel with the plurality’s position that the criterion for measuring effectiveness of the assistance must preserve counsel’s freedom “to make quick judgment,” 41 and that a “shortfall by defense counsel that is perceptible but is modest . . is no basis for judicial interposition.” 42 A reasonableness standard is eminently consistent with these concerns, for it is breached only when counsel’s conduct deviates so substantially from an acceptable norm as to merit the label “unreasonable.” Plainly, neither “quick judg *250 ments” nor “modest shortfalls,” merely by reason of their character, are sufficiently off the mark to beckon the hand of the courts. 43
Lastly, there is the query on the caliber of counsel’s performance in the instant case, and for myself it is enough to give a short answer. Whatever the full range of his constitutional duty to his client, 44 it is clear that counsel was obligated to conduct a suitable investigation into the facts of the case and to plot the defensive strategy accordingly. 45 For reasons subsequently appearing, I am satisfied that he did not properly discharge that responsibility, 46 and on this at least seven members of the court concur. 47 Where I part company with the majority is the point at which we come to consider whether it was incumbent upon the client to discharge a burden of demonstrating more.
III. THE BURDEN OF PROOF ON PREJUDICE
The burden of proving unconstitutionality is upon him who asserts it. 48 “That burden,” we have said, “extends to production of the facts essential to a determination respecting the constitutional claim.” 49 Resultantly, the defendant who would charge ineffective representation “must,” we have added, “set forth evidence upon which the elements of a constitutionally *251 deficient performance might properly be found.” 50 We remain divided, however, on the question whether the demonstration incumbent upon the defendant includes a showing that his counsel’s allegedly subpar rendition actually or potentially affected the outcome of the case.
A majority of the court considers detriment to the defendant’s interests an indispensable ingredient of his constitutional claim, and thus a factor for him to prove. For them, the defendant must establish adversity resulting from the deficient conduct, else there is no infringement of the constitutional right. For Judge Leventhal and subscribers to his opinion, the crucial item is likely prejudice, 51 for Judge MacKinnon and those who join him, it is actual prejudice. 52 Judge MacKinnon goes so far as to insist that overwhelming evidence of guilt relieves defense counsel of the duty to conduct any more than minimal investigation on his client’s behalf. 53
I cannot agree with either of these formulations. Careful review of the caselaw convinces me that the burden-of-proof allocation they direct is an unwarranted distortion of the role that prejudice ordinarily plays in constitutional determinations and an impermissible expansion of the limited function of harmless error. In my view, the defendant establishes a constitutional violation when he makes out a substantial breach of duty by his counsel, and it is then up to the Government to demonstrate lack of ensuing prejudice if it can.
A. The General Role of Prejudice in Constitutional Adjudications
In Chapman v. California, 54 Supreme Court was asked to decide whether a violation of the Federal Constitution can ever be considered innocuous. The Court’s response, in essence, was that the answer depends upon the nature of the constitutional entitlement at stake. The Court declared that “there are some constitutional rights so basic to a fair trial that their infraction can never be treated as harmless error.” 55 The Court refused, however, “to hold that all federal constitutional errors, regardless of the facts and circumstances, must always be deemed harmful.” 56 Rather, said the Court, “there may be some constitutional errors which in the setting of a particular case are so unimportant and insignificant that they may, consistent with the Federal Constitution, be deemed harmless, not requiring the automatic reversal of the conviction.” 57 Thus an absence of prejudice may or may not be a valid judicial concern when a federal constitutional transgression is under investigation.
Chapman also addressed the allocation of burden of proof in those instances where harmlessness is entitled to some role. “Certainly error, constitutional error, in illegally admitting highly prejudicial evidence or comments,” the Court cited as an example, “casts on someone other than the person prejudiced by it a burden to show that it was harmless”; 58 [i]t is for that reason,” the Court noted, “that the original common-law harmless-error rule put the burden on the beneficiary of the error either to prove that there was no injury or to suffer a reversal of his erroneously obtained judgment.” 59 And not only is the burden thus to be assigned but, when the asserted error is of constitutional dimension, it is a burden *252 of peculiar weight. The rule appropriate, the Court declared, “requir[es] the beneficiary of a constitutional error to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.” 60 So it was that, for cases wherein harmlessness is a factor at all, 61 the bottom line was drawn: “[Bjefore a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt.” 62
Chapman did not, of course, speak to the precise question dividing us today — whether the defendant must prove prejudice as an element of his constitutional ineffective-assistance claim. But indubitably implicit in Chapman is the central theme that for a great many constitutional violations — and for perhaps the decided majority — the defendant need not demonstrate harm, either actual or potential, in order to obtain relief. Rather, it may be permissible, but only in some instances, for the Government to attempt to show lack of prejudice, and even then the proof must establish it beyond a reasonable doubt.
Chapman remains the seminal precedent today, 63 and obviously it demands two vital inquiries in the case at bar. Is prejudice a legitimate consideration in the assessment of a charge of ineffective assistance of counsel? If so, upon whom rests the burden of proof? Stating the second question somewhat differently, is a showing of threatened or consummated harm from a proven breach of counsel’s duties an essential element of the defendant’s claim, or is a demonstration of actual harmlessness a matter for the Government to undertake?
Many constitutional errors in criminal trials invoke a per se rule. The constitutional violation triggers spontaneous reversal of an ensuing conviction without any exploration into its real or probable effect upon the trial. Just when that will be the case is a question answerable only upon careful analysis of the nature of the right invaded and its capacity to withstand the inherent fallibility of an investigation into prejudice. From a host of diverse considerations that may deserve attention in the analysis, several come immediately to the fore.
One is the constitutional, statutory or judicial recognition the right has been accorded, 64 as well as the purpose the right sub-serves. 65 Another is the degree of prejudicial propensity of a trespass upon the right. 66 Still another is the feasibility of an *253 effort to measure the impact of the constitutional violation upon the outcome of the trial. 67 Not the least may be an uncompromising policy of deterring repetition of the same unconstitutional conduct in the future. 68 A modest sampling of Supreme Court decisions will illustrate the interplay of these and other factors.
Conviction by a judge having personally a direct and substantial interest in convicting necessitates reversal “[n]o matter what the evidence was against” the accused because “he had the right to have an impartial judge” 69 stemming from long-standing judicial realization that a biased tribunal violates fundamental due process. 70 Conviction by a jury selected through use of discriminatory techniques demands the same result because “[i]t is in the nature of [that evil] that proof of actual harm, or lack of harm, is virtually impossible to adduce,” and “there is no way to determine what jury would have been selected under a constitutionally valid selection system, or how that jury would have decided the case.” 71 Inflammatory publicity massively and pervasively surrounding a trial vitiates it without any special showing of consequent harm when “the totality of circumstances” indicates inherent prejudice. 72 Televising courtroom proceedings in a criminal case has been held violative of due process even absent proof of injury because it is innately harmful, 73 its adverse effects are too subtle to prove 74 and the practice has met widespread condemnation. 75
Similarly, admission into evidence of a coerced confession requires reversal as the sanction responsive to offensive police activity, irrespective of evidence of guilt dooming any argument that the admission was actually prejudicial. 76 Committing the *254 jury to continuous custody by deputy sheriffs who also were the principal prosecution witnesses denies due process “even if it could be assumed that the deputies never did discuss the case directly with any members of the jury,” for “it would be blinking reality not to recognize the extreme prejudice inherent in [the] continual association . .” 77 And incorporation of an unconstitutional presumption into the court’s instructions to the jury invalidates the verdict even though it is amply sustained by evidence apart from the presumption; the reason is that “[i]n view of the place of importance that trial by jury has in our Bill of Rights, it is not to be supposed that Congress intended to substitute the belief of appellate judges in the guilt of an accused, however justifiably engendered by the dead record, for ascertainment of guilt by a jury under appropriate judicial guidance, however cumbersome that process may be.” 78 And we ourselves have held that denial of the accused’s fundamental statutory, quasi-constitutional right to appear pro se is not redeemed by “the subsequent conclusion that [his] practical position [was not] disadvantaged,” 79 for the right “is designed to safeguard the dignity and autonomy of those whose circumstances or activities have thrust them involuntarily into the criminal process.” 80
As Chapman made clear, however, not every mistake of constitutional magnitude in a criminal trial leads inexorably to reversal. The majority opinion was careful to point out “that there may be some constitutional errors which in the setting of a particular case are so unimportant and insignificant that they may ... be deemed harmless . . . ,” 81 The concurring opinion similarly noted that “constitutional rights are not fungible goods,” and that “[t]he differing values which they represent and protect may make a harmless-error rule appropriate for one type of constitutional error and not for another.” 82 Indeed, the particular violation dealt with in Chapman 83 was held to invoke, not the per se rule of automatic reversal, but the special federal harmless-error rule fashioned in that case for infringements of those constitutional rights that might tolerate it. 84 It bears repeating, however, that when harmlessness is permitted any sway at all, a much higher-than-normal standard for af-firmance obtains: “[B]efore a federal constitutional error can be held harmless, the *255 court must be able to declare a belief that it was harmless beyond a reasonable doubt.” 85
Since Chapman’s day, decisions holding constitutional errors harmless are legion. 86 Usually, they have involved transgressions in the presence of the court, 87 instances in which assessment of the harm can be made with relative safety. 88 As the Supreme Court recently remarked,
[ i ]n the normal case where a harmless error rule is applied, the error occurs at trial and its scope is readily identifiable. Accordingly, the reviewing court can undertake with some confidence its relatively narrow task of assessing the likelihood that the error materially affected the deliberations of the jury. 89
It is evident, however, that there are out-of-court contexts in which it is entirely feasible to ascertain whether the accused almost assuredly would have been convicted even absent the cited error. 90 One such occasion rather clearly is when “the case against [the defendant is] so overwhelming that [the court can] conclude that [the] violation . . . was harmless beyond a reasonable doubt . . ..” 91
B. The Role of Prejudice in Right-to-Counsel Cases
Supreme Court decisions on the function of prejudice in right-to-counsel eases reflect essentially the same considerations pertinent in other areas of constitutional error. 92 *256 Two factors combine to frequently render Chapman’s special harmless-error rule inappropriate. First, entitlement to assistance of counsel is explicitly conferred by the Sixth Amendment, 93 with effectiveness of the assistance as its soul. 94 Second, harm to the accused’s interests is inherent in many denials of the right 95 and the scope of consequent injury all too often is not readily identifiable or probable. 96
Thus, where the defendant had no counsel at all at a critical stage of his trial, automatic reversal of his conviction is usually in order. 97 That result follows also when counsel became available too late for adequate preparation, 98 or when he labored under a conflict of interest precluding completely loyal and effective service to his client. 99 In these instances, harm to the accused is well nigh inescapable, 100 and intelligent assessment of its range is elusive because the effects of the violation might well permeate the entire trial. 101
In some circumstances, however, not even a total lack of counsel necessarily commands these conclusions. Prejudice may not approach the plane of inevitability, or an acceptable judicial appraisal thereof may not be out of reach. 102 Examples of resort *257 to the harmless-error rule are to be found when the sole consequence of the violation is the admission of evidence tainted by the absence of counsel at the time of its acquisition. An identification of an unrepresented suspect at a pretrial lineup is legally inadmissible, 103 as is an in-court identification attributable only to the lineup, 104 but the use of such an identification does not necessitate reversal if the Government proves its harmlessness beyond a reasonable doubt. 105 The admission into evidence of a voluntary confession taken from an uncounselled arrestee without constitutionally-required warnings, 106 or out of the presence of counsel after his appointment, 107 may similarly be cured, 108 in both instances, if the error is discrete and its adverse effects measurable. Want of counsel at some other pretrial proceedings — preliminary hearing, 109 arraignment, 110 entry of a not-guilty plea 111 — or even during short periods of the trial itself 112 may be found to be harmless. In each situation, by reason of the nature of the proceeding or the brevity of counsel’s absence, the range of possible negative consequences is limited 113 and possibly amenable to evaluation.'
Impairment of the right to effective assistance of counsel shares the characteristics of other right-to-counsel violations treatable as harmless error. Prejudice may not be invariably a concomitant of counsel’s delinquency, or the presence and extent of injury may be susceptible to an acceptable *258 degree of accurate measurement. 114 It is worth noting in this connection that in meeting his burden of showing inadequate representation, it is not enough that the defendant merely protest that his counsel was incompetent; he must stake out the shortfalls of which he complains. 115 This sort of particularization, in turn, supplies the specific points of reference that facilitate the inquiry on prejudice. In Chambers v. Maroney, 116 for example, upon a charge that counsel was inadequate because he was appointed too close to trial to permit preparation, the Supreme Court examined the deficiencies cited and found that in each instance the blunder, if any, was noninjurious. 117
I conclude, then, that the right to effective assistance of counsel is amenable to the harmless-error rule. This is not to say, however, that prejudice to the accused, either threatened or consummated, is a sine qua non of the constitutional claim. The defendant must demonstrate a substantial failure by counsel to discharge an important duty, but when he does so I think the claim is sustained although counsel’s mistake may be inconsequential — and thus may not incur reversal — because as matters turn out it doubtless did not influence the end result.
C. Allocation of the Burden of Proof on Prejudice
Proof of actual or potential harm is not normally an element of the showing prerequisite to establishing a violation of a right specifically enumerated in the Constitution. 118 To be sure, prejudice is a factor relevant though not indispensable to a determination respecting observance of the Sixth Amendment right to speedy trial; 119 certainly, too, “in most cases involving claims of due process deprivations . a showing of identifiable prejudice to the accused” is required. 120 But prejudice is *259 presumed for many due process denials, 121 and so too it generally is for trespasses on the Sixth Amendment right to counsel. 122 Sometimes, as has been seen, that presumption is conclusive in the sense that any effort to demonstrate an absence of injury-in-fact is totally foreclosed. 123 And even when the presumption is not fully preclu-sive, it permits no more than an attempted showing by the accused’s adversary that the constitutional transgression was harmless beyond a reasonable doubt. 124
In no uncertain terms, a positive guaranty of assistance of counsel is enshrined in the Sixth Amendment. 125 It is, I reiterate, unmistakably a pledge of the effective assistance of counsel. 126 I perceive no reason why that right, like the vast majority of others that the Constitution makes explicit, should not be fully honored upon the usual presumption, as distinguished from proof of prejudice from its denial.
The right we deal with was first articulated by the Supreme Court nearly a half-century ago. 127 It has enjoyed full stature in the Court ever since. 128 It has been proclaimed with regularity in every federal circuit. 129 Though sometimes attributed to the exigencies of due process, its Sixth Amendment origin has long been recognized; 130 though originally construed narrowly, 131 it has in recent years received in *260 creased judicial attention and protection. 132 The implication is that once it is realized that the right to effective assistance of counsel is grounded on the express command of the Sixth Amendment as well as encompassed in the generality of the due process concept, no justification for requiring the defendant to prove prejudice is apparent.
Ineffective assistance of counsel has a built-in potential for harm to the client. The right to effective assistance thus shares with most other constitutional guaranties a characteristic which normally obviates any need for proof of prejudice, and sometimes even forecloses consideration of arguments that in the particular situation a denial of the right might have been wholly innocuous. 133 Indeed, ineffective assistance is not far removed from total lack of assistance, which frequently calls for automatic reversal. 134 And while the harmless-error rule is in vogue when there was counsel — though inadequate counsel 135 — the Supreme Court has yet to levy on the defendant a burden of showing that demonstrated incompetence threatened or wrought damage to his cause. 136 Nothing in the nature of this fundamental constitutional right suggests to me that we should make such an imposition today. 137
In sum, I cannot accept the theory that proof of actual or potential harm to the accused is an element of an ineffective-as- *261 sistanee claim. I think the claim is established by a suitable showing that counsel defaulted on an obligation owed the accused, 138 and that any asserted lack of injury therefrom is to be treated here just as it normally is in any other instance of curable constitutional error. 139 This means, of course, that the burden rests upon the Government to prove absence of harm to the accused, and to prove it beyond a reasonable doubt. 140
The Leventhal-MacKinnon approach, I submit, confuses two independent questions commonly arising in ineffective-assistance litigation. One is whether defense counsel measured up to the constitutional standard of reasonably competent representation. This entails scrutiny of the quality of the service rendered. The other is whether proven deficiency in counsel’s performance clearly lacked adverse impact. The task here is simply to determine whether the error could have contributed in any material way to the result reached in the case. Harm is the focal point of the second; it has no bearing whatever on the first. And harmlessness is a doctrine serving only to avoid needless retrials where, owing to the innocuousness of the violation, the outcome would likely be the same. 141
The critical distinction between the defendant’s burden to show a constitutional transgression and the Government’s burden to demonstrate lack of ensuing prejudice becomes apparent when we look back three years to our Pinkney decision. 142 Following a conviction on two drug charges and before sentencing thereon, the Government filed an “allocution memorandum” containing information purporting to link Pinkney with narcotics trafficking and advocating the maximum penalty. After imposition of less severe though stiff sentences, Pinkney moved the District Court to reconsider them, insisting that his counsel had not discussed the Government’s memorandum with him and reminding that counsel had not disputed its contents at sentencing. The motion was denied and on appeal we declined to upset that ruling. 143 We acknowledged the constitutional implications of the asserted breach of duty 144 but deemed it unimportant because “[t]he record . . . [did] not support the contention that counsel’s alleged derelictions frustrated [Pinkney’s] opportunity to present his side of the controversy.” 145 We pointed to Pinkney’s obligation at this juncture to “set forth evidence upon which the elements of a constitutionally deficient performance might properly be found,” 146 and we found that Pinkney’s motion did not survive this requirement for two reasons. In the first place, he did not verify the deficiency complained of. 147 Beyond that, after learning of the central allegation of the allocution memorandum, he did not utilize open opportunities to convey to the sentencing judge anything he might have wished to say. 148
These omissions are very different from a failure to carry the burden on an issue of prejudice from an established violation. 149 *262 Although, we did not reach the question of prejudice in Pinkney, 150 we took pains to explain the distinction:
Our conclusion . . in no way impinges upon the rule . . . that once a substantial violation of counsel’s duties is shown, the Government’s burden is to demonstrate lack of prejudice therefrom. . In the case before us, we deal only with a procedural prerequisite to a hearing on appellant’s assertion that the representation afforded at sentencing fell below the constitutional norm. The essence of appellant’s contention is that sentencing counsel deprived him of the opportunity to combat allegations of the Government’s allocution memorandum by failing to inform him of the memorandum. . . . Only if the evidentiary elements of that claim had appeared in appellant’s motion would he have been entitled to a hearing, and only if evidence offered at a hearing tended to establish the elements would the Government have been summoned to disestablish prejudice. But if, on the other hand, appellant had met these preconditions, the Government would then have encountered the burden of proving that counsel’s dereliction did not harm appellant — for example, because the allocution memorandum actually had no effective role in the sentencing process. 151
This is the major point of deviation between the position of a majority of the court’s members and mine. In my view, the claimant before us needed only to show that his counsel fell substantially short of the standard of reasonable competence; in theirs, threatened or consummated injury therefrom is an additional required part of the showing. I believe the majority err in their approach and denude the constitutional right to effective assistance of counsel of a great deal of the value it was intended to have.
IV. THE PRESENT CASE
Turning now to Decoster’s arguments that the assistance furnished by his trial counsel was constitutionally ineffective, I find inescapable the conclusion that counsel failed miserably in responding to his obligation to conduct a reasonably competent investigation into the facts of the case. Prior to trial, as Judge Bazelon studiously recounts, 152 counsel made no real effort to tap known or likely sources of information, which included codefendants as well as prosecution witnesses. The duty to investigate is vital, and its violation is obviously fraught with danger to the interests of the client. 153 Here the investigative responsibility was almost wholly unmet, and I cannot view the dereliction as less than appalling.
I am equally convinced, however, that the record firmly establishes the violation as harmless. Two police officers witnessed the robbery in progress. One chased De-coster from the spot the short distance to the hotel lobby wherein he was apprehended, never losing sight of him for so much as a moment. Within minutes the robbery victim, in the presence of both officers, identified Decoster as one of the culprits, as the officers themselves were later to do. Decoster’s own testimony aside, no basis for impeaching these witnesses on these vital points surfaces on the record either of the preliminary hearing or the trial. 154 In sharp contrast, Decoster’s alibi, initially feeble, met disaster after he called *263 to the stand one of his codefendants only to hear him testify that he saw Decoster in a fight with the victim at the scene of the crime. This is not to say that every prose-cutorial presentation in which the evidence is so lopsided may fairly be characterized as overwhelming, 155 for that appearance may be attributable, at least in part, to counsel’s deficiencies. But considering here the Government’s direct and positive proof, the number of Government witnesses, 156 the consistency of their testimony, and the improbability of misinterpretation of the criminal activity or misidentification of Decoster as a participant, I see no reason for supposing that pretrial interviews with these witnesses would have turned up anything but ominous news for the defense.
Counsel’s omission to hunt for alibi witnesses was similarly unhurtful. Aside from the clerk on duty in the hotel lobby when Decoster was taken into custody, there is little to indicate that the prospect of locating any such witnesses was better than highly remote. More importantly, I cannot hypothesize any appreciable probability that the jury’s verdict would have differed had counsel found someone able to testify that Decoster and the victim drank together at a bar shortly before the robbery occurred, or someone who had seen Decoster enter the hotel lobby shortly after the robbery transpired. 157 No witness in one or the other place could possibly have established Decoster’s absence from the scene of the offense at the instant it was perpetrated. It is clear, of course, that at that point in time Decoster was somewhere between the bar and the lobby, but the question is whether that somewhere was the site of the crime. The most that can be said is that such a witness might have tended weakly and circumstantially to corroborate Decoster’s claim that he was not there. But it would blink reality to seriously suggest that, in the face of the Government’s powerful eyewitness case, so little would have carried the day. 158
I think, then, that in the area of pretrial investigation counsel’s performance was *264 substantially deficient and therefore constitutionally ineffective. But I believe, too, that in the circumstances counsel’s inadequacies were harmless beyond a reasonable doubt. 159 On that ground, I concur in af-firmance of the conviction.
BAZELON, Circuit Judge, with whom J. SKELLY WRIGHT, Chief Judge, joins, dissenting:
Willie Decoster was denied the effective assistance of counsel guaranteed by the Sixth Amendment because he could not afford to hire a competent and conscientious attorney. His plight is an indictment of our system of criminal justice, which promises “Equal Justice Under Law,” but delivers only “Justice for Those Who Can Afford It.” Though purporting to address the problem of ineffective assistance, the majority’s decision ignores the sordid reality that the kind of slovenly, indifferent representation provided Willie Decoster is uniquely the fate allotted to the poor. Underlying the majority’s antiseptic verbal formulations is a disturbing tolerance for a criminal justice system that consistently provides less protection and less dignity for the indigent. I cannot accept a system that conditions a defendant’s right to a fair trial on his ability to pay for it. Like Justice Black, I believe that “[tjhere can be no equal justice where the kind of trial a man gets depends on the amount of money he has.” 1 The Constitution forbids it. Morality condemns it. I dissent.
I.
The evolution of the right to the Assistance of Counsel reflects a growing awareness of the barriers faced by the indigent defendant seeking a fair trial, and of the challenge these obstacles pose to our ideal of justice without regard to wealth. By any reckoning, the barriers are formidable. The “street crime” that clogs our courts is bred by poverty and discrimination. It is committed by the dispossessed, the disadvantaged and the alienated of our society— those who most need the advice of a trained advocate. In the words of Justice Sutherland:
Even the intelligent and educated layman has small and sometimes no skill in the science of law. ... He requires the guiding hand of counsel at every step in the proceedings against him. If that be true of men of intelligence, how much more true is it of the ignorant and illiterate, or those of feeble intellect. 2
And the cruel irony, of course, is that the indigent are the very people who are least able to obtain competent representation. For the most part, “you get what you pay for” in legal representation. 3
*265 Only recently have we even recognized that the lack of effective counsel inevitably deprives the poor of the right to a fair trial. For a great many years, the shameful truth was that only the rich could obtain counsel, since only the rich could afford to pay counsel. One hundred and forty-one years after the adoption of the Bill of Rights, the Supreme Court first held, in Powell v. Ala bama 4 that due process requires the appointment of counsel for an indigent defendant in a capital case. Not until Gideon v. Wainwright, 5 more than thirty years later, did the Court acknowledge that the “noble ideal” of equal and fair justice could not be realized “if the poor man charged with crime has to face his accusers without a lawyer to assist him.” 6 Only then did the Court extend the right to counsel to all state felony prosecutions. And not until 1972, in Argersinger v. Hamlin, 7 did the Court affirm the right to counsel in all criminal prosecutions resulting in the deprivation of the accused’s liberty.
The Supreme Court’s effort to-eliminate second-class justice for the poor has not been confined to providing counsel for the indigent. 8 But the right to counsel is most essential in assuring fair and equal justice, 9 *266 for without the conscientious and knowledgeable advice of a trained legal advocate, an accused can secure none of the safeguards of the criminal process intended to protect all defendants. “The [right to] counsel is often a requisite to the very existence of a fair trial.” 10 To the extent that the indigent defendant receives inadequate representation, markedly inferior to that available to a defendant who can afford “competent and conscientious counsel,” a dual system of justice endures.
Inevitably there will be disparities in the quality of representation; some lawyers are simply more able or more conscientious than others. What offends the Constitution, however, is not merely that there are variations in the quality of representation, but that the burden of less effective advocacy falls almost exclusively on a single subclass of society — the poor. In constructing standards for assessing the ineffective assistance of counsel, we must therefore consider not only what measures are necessary to assure a fair trial in the case of any particular defendant. We also must structure our approach to eliminate the gross disparities of representation that make a mockery of our commitment to equal justice. We must institutionalize and enforce standards of attorney competence designed to assure adequate representation for a 11 defendants.
Because my colleagues in the majority divorce their analysis from the economic and social reality underlying the problem of ineffective assistance of counsel, their decision leaves indigent defendants nothing more than an empty promise in place of the Sixth Amendment’s commitment to adequate representation for all defendants, rich and poor. At best, the majority’s approach might help to rectify a few cases of blatant injustice. But their standards do nothing to help raise the quality of representation provided the poor to a level anywhere approaching that of the more affluent. On the contrary, my colleagues condone callous, back-of-the-hand representation by dismissing the basic duties of competent lawyering as “aspirational.” The majority thus provides no incentive or structure to improve the caliber of defense advocacy. By focusing exclusively on the consequences of counsel’s dereliction, their approach encourages an attorney who believes that his client is guilty to “cut corners,” with little risk that he will be held accountable for the inadequacies of his representation. The majority opinions may say “we don’t commend this,” or “we don’t approve of that,” but their bottom line is “Affirmed.’’
In its holding, the majority turns its back on the evolution in this circuit of the standard for evaluating claims of ineffective assistance. 11 In the earliest cases, we approached the problem solely from a due process-fundamental fairness viewpoint, requiring a defendant seeking relief to show that the proceedings were a “farce and a mockery of justice.” 12 In Bruce v. United *267 States, 13 we reconsidered that standard. We explained that the “farce and mockery” requirement was not to be taken literally, but was meant only to demonstrate that in order to obtain relief the accused bears a heavy burden of showing that “there has been gross incompetence of counsel and that this has in effect blotted out the essence of a substantial defense . . . .” 14 Shortly after Bruce, however, we explicitly recognized that the requirement of effective assistance of counsel derives not only from the Due Process Clause, but from the Sixth Amendment itself. 15 Consequently, in our original opinion in this case, United States v. DeCoster (DeCoster I), 16 this court adopted á standard for direct appeals consistent with the Sixth Amendment’s “more stringent requirements”: 17 “a defendant is entitled to the reasonably competent assistance of an attorney acting as his diligent conscientious advocate.” 18
DeCoster I represented a major advance in this court’s recognition of the realities of ineffective assistance. In that case, this court shifted the focus of judicial inquiry away from the prejudice to the defendant in any particular case and toward the task of articulating basic duties counsel owes his client. This approach, for the first time, gave content to what previously had been empty verbal formulations. Even more importantly, it recognized that the very lack of effective trial counsel might preclude a defendant from later establishing prejudice. Thus the court concluded that the only way to assure that every defendant receives a fair trial is to promulgate and enforce standards of adequate representation that apply across-the-board. Underlying DeCoster I, therefore, was a commitment to the basic principle that every defendant — rich or poor, innocent or guilty — is entitled to the reasonably competent assistance of an attorney acting as his diligent conscientious advocate.
II.
Appellant Willie Decoster and two code-fendants, Douglas Eley and Earl Taylor, were arrested for the robbery of Roger Crump on the evening of May 29, 1970. Two police officers on plainclothes patrol observed three men accosting Crump in the parking lot of the Golden Gate Bar. When the officers jumped from their car, the robbers fled and were pursued by the police. Officer Box and the victim found Decoster in the lobby of a nearby hotel, the D.C. Annex, where he was immediately arrested and identified by Crump. 19
At his trial on November 15, 1971, appellant testified that on the evening of the crime he had been drinking with Crump at the Golden Gate Bar. 20 Decoster claimed that he left Crump at the bar and walked directly to his hotel, where, while standing by the desk waiting to obtain his room key, he was arrested. One of appellant’s alleged *268 accomplices, Douglas Eley, was also called as a defense witness but his testimony contradicted Decoster’s in several respects. Most importantly, on direct examination he claimed to have seen appellant and Crump fighting in the parking lot outside the bar at the time of the alleged robbery. Following the trial, appellant was convicted of armed robbery and sentenced to 2-8 years. 21
On appeal, this court was troubled by a number of actions taken by Decoster’s court-appointed counsel which, taken together, suggested that Decoster may not have received the effective assistance of counsel. The record showed that on Nov. 4, 1970, several months after appellant’s arrest, the trial judge received a letter from Decoster in which he requested new counsel because his attorney was not providing adequate representation. 22 Decoster charged that although he had been accepted for pretrial custody by the Black Man’s Development Center on October 12, defense counsel had not kept his promise to file for bond review. 23 Appellant also requested a copy of the transcript of his preliminary hearing, which he had been unable to obtain through his attorney. 24
Although Decoster’s counsel finally filed the requested bond review motion on November 9, 25 he not only failed to mention that third-party custody had been arranged, but he also filed the motion in the wrong court. 26 On November 18, the district court advised counsel of his error and continued the motion to await review by the proper court, as required by law. 27 Again, however, counsel delayed filing; not until De *269 cember 9, did he file a motion for bond review in the proper court. 28
We also noted that events at the beginning of trial raised serious questions about the adequacy of counsel’s pretrial preparation and communication with his client. As the trial was about to start, and after counsel had asserted that he was prepared to proceed, appellant himself stepped forward and asked if the court would subpoena his two codefendants, 29 explaining that he “didn’t have a chance” to discuss the matter with his lawyer. Defense counsel then told the court that he had considered the possibility of issuing subpoenas, “except for the fact that we have no address for the other defendants.” 30 The prosecutor immediately volunteered that codefendant Eley was in jail with Decoster; 31 an address for Taylor was subsequently provided from the court records. 32 The court thereupon ordered defense counsel to “take care of the situation.” 33
Moments later, after defense counsel again announced that he was ready for trial, the prosecutor informed the court that the Government had not received any response to its alibi-notice demand. Defense Counsel replied that although he might rely on an alibi defense, no response was needed because the Government had not given the twenty days’ notice required by the local rules. The trial judge ordered the defense to provide the names of alibi witnesses anyway, whereupon defense counsel relented and stated, “We will proceed without the alibi witnesses.” 34
Defense counsel then informed the court that his client wished to waive jury trial. When asked if he was aware that the trial *270 judge already had heard evidence concerning Decoster’s case while presiding over the trial of his codefendants, counsel responded that he was not. 35 After attempting unsuccessfully to find another judge who could hear the case at such a late date, the trial judge ruled that he could not hear the case himself but would instead preside over a jury trial. Appellant’s case thereupon proceeded to trial before a jury.
In the midst of all this confusion, Decoster again complained to the court about his attorney’s efforts on his behalf.
THE DEFENDANT: Your Honor, I feel that this case should be continued because this is, I can’t get proper representation that I should be getting and too I think I should have an accurate statement of what happened here when the other two defendants was in court. 36
Defense counsel then requested to withdraw from the case “because apparently I have caused some dissatisfaction to the defendant. . .” 37 The district judge, however, did not inquire into the basis of the defendant’s complaints. Instead, after receiving counsel’s assurances that he had prepared the case and was ready to go to trial, the court denied the request for a continuance and refused to appoint new counsel. 38
*271 In view of the foregoing, in our original opinion 39 we remanded for supplementary hearings on the adequacy of trial counsel’s representation and granted leave for appellate counsel to move for a new trial. At the hearings on remand, 40 the district court elicited further information about trial counsel’s preparation and his explanations for his actions. Counsel admitted that he had not interviewed the robbery victim or either of the police officers. 41 He also admitted that he had made no attempt to contact or interview the hotel desk clerk or, for that matter, anyone else at either the D.C. Annex hotel or the Golden Gate bar. 42
As for the codefendants, counsel conceded that he had not interviewed Taylor, 43 but claimed that he had talked with Eley in the cellblock behind the courtroom on the *272 second day of the trial. 44 Counsel also admitted that he never obtained a transcript of the preliminary hearing, but stated that since he had conducted most of the cross-examination at that hearing, he saw no need for the transcript. 45 Moreover, counsel testified that the U.S. Attorney’s Office usually makes a copy of the transcript available during discovery. Although he did not specifically remember Decoster’s case 46 counsel said he assumed that the government’s copy had been available and that he had read it. 47
In attempting to defend his actions, counsel testified that he had not interviewed any witnesses because, until shortly before trial, appellant had never mentioned any possible alibi witnesses. Counsel explained that, to the best of his recollection, Decoster had continuously maintained that he had joined Crump for a drink in the bar, had left him there, and had just returned to his hotel when he was arrested 48 Then, on the eve of trial, 49 counsel received a letter in which appellant changed his story, alleged *273 that he had fought with Crump but did not rob him, and asserted that his codefendants would support this version of the offense. 50 Explaining why he had not interviewed De-coster’s codefendants even after receiving this letter, counsel stated that “it was my feeling at that time that any testimony that might be given by either of these defendants might be contradictory to what I had already heard from the Defendant.” 51 Counsel claimed that he did interview Eley after appellant insisted at trial that his codefendants be subpoenaed, and that Eley told him that Decoster was not at the scene of the crime. 52 Believing that Eley would say this in court, counsel decided to put him on the stand, but Eley instead testified that he saw Decoster and Crump fighting outside the bar. 53
Counsel also was asked at the remand hearing to explain the reasons underlying certain “tactical decisions” he had made. He could not recall why the motion for bond review was filed in the wrong court, or why he failed to mention appellant’s acceptance by the Black Man’s Development Center in the original motion. With respect to the waiver of jury trial, counsel said that although he opposed the idea, he had requested a bench trial at his client’s insistence. 54 Finally, counsel stated that he gave no opening statement because he had felt it to be unnecessary, and not because he had no defense theory at the time the trial started. However, counsel could not recall why he had concluded that an opening statement was unnecessary. 55
In its findings on remand, the district court isolated seven 56 particular acts or *274 omissions by defense counsel that were alleged to have deprived appellant of the effective assistance of counsel. 57 With respect to three of the allegations — counsel’s waiver of opening statement, his attempt to waive jury trial, and his failure to see that Decoster was given credit for time served as ordered by the sentencing judge 58 — the district court found no ineffective assistance. 59 Two other claims — the delay in moving for bond review and the failure to obtain a transcript of the preliminary hearing — were rejected because the court found that the appellant had not been prejudiced by the violations. 60 On the final two allegations — counsel’s failure to interview witnesses and his premature announcement that he was ready for trial — the district court’s conclusions can be interpreted as holding either that there was no constitu *275 tional violation or simply that no prejudice to appellant was shown. 61
III.
The analysis of this case should be guided by the principles established in DeCoster J. 62 We there held that upon showing a substantial violation of any of counsel’s specified duties, a defendant establishes that he has been denied effective representation and the burden shifts to the government to demonstrate that the violation did not prejudice the defendant. Thus, DeCoster I prescribed a three-step inquiry for determining whether a claim of ineffective assistance of counsel warrants reversing a conviction:
1) Did counsel violate one of the articulated duties?
2) Was the violation “substantial” ?
3) Has the government established that no prejudice resulted?
The heart of this approach lies in defining ineffective assistance in terms of the quality of counsel’s performance, rather than looking to the effect of counsel’s actions on the outcome of the case. If the Sixth Amendment is to serve a central role in eliminating second-class justice for the poor, then it must proscribe second-class performances by counsel, whatever the consequences in a particular case. Moreover, by focusing on the quality of representation and providing incentives in all cases for counsel to meet or exceed minimum standards, this approach reduces the likelihood that any particular defendant will be prejudiced by counsel’s shortcomings. In this way, courts can safeguard the defendant’s rights to a constitutionally adequate trial without engaging in the inherently difficult task of speculating about the precise effect of each error or omission by an attorney. Although the question of prejudice remains part of the court’s inquiry, it is distinct from the determination of whether the defendant has received effective assistance. Rather, prejudice is considered only in order to spare defendants, prosecutors and the courts alike a truly futile repetition of the pretrial and trial process.
A. Violation of Articulated Duties
In DeCoster I, this court attempted to give substantive content to the Sixth Amendment’s mandate by setting forth minimum requirements of competent per *276 formance. 63 The obligations were described as “duties owed by counsel to client,” 64 and thus were not offered as merely “aspirational” guidelines to which attorneys should strive. Indeed, the duties announced in De-Coster I represent the rudiments of competent lawyering guaranteed by the Sixth Amendment to every defendant in a criminal proceeding. 65
The duties set forth in DeCoster I were derived from the American Bar Association’s Standards for the Defense Function. 66 These ABA Standards summarize the consensus of the practicing Bar on the crucial elements of defense advocacy in our adversary system. Even though these standards were not intended by their drafters to serve “as criteria for judicial evaluation of effectiveness[,]” 67 this court noted that “they are certainly relevant guideposts in this largely uncharted area.” 68 Naturally, given the complexities of each case and the constant call for professional discretion, it would be a misguided endeavor to engrave in stone any rules for attorney performance. 69 Nonetheless, preserving flexibility *277 is not incompatible with establishing minimum components of effective assistance, and the ABA Standards give helpful guidance in pursuing both aims. 70
In DeCoster I this court was sensitive to these concerns and so did not attempt to prescribe categorical standards of attorney performance. Instead, we took pains to note that the articulated duties were “meant as a starting point for the court to develop, on a case by case basis, clearer guidelines for courts and for lawyers as to the meaning of effective assistance.” 71 We recognized, however, that there were certain tasks, such as the ones we enumerated in our decision, that can never be ignored: conferring with the client without delay and as often as necessary; fully discussing potential strategies and tactical choices; advising the client of his rights and taking all actions necessary to preserve them; and conducting appropriate factual and legal investigations 72 I submit that no one can dispute that a reasonably competent lawyer, absent good cause, 73 would or should do less. Counsel should proceed in the representation of his client under the guidance of these minimal duties, departing only when the particular needs of his client compel a different course of action.
Prominent among the duties of defense counsel is the obligation to “conduct appropriate investigations, both factual and legal, to determine what matter of defense can be developed.” 74 As the Commentary of the ABA Standards stresses, “[I]nvestigation and preparation are the keys to effective representation . . . . It is impossible to overemphasize the importance of appropriate investigation to the effective and fair administration of criminal justice.” 75
Investigation is crucial for several reasons. First, the proper functioning of our adversary system demands that both sides prepare and organize their case in advance *278 of trial. There can be no justice where one party to the battle has made no effort to arm itself with the pertinent facts and law. 76 Second, in a very practical sense, cases are won on the facts. Proper investigation is critical not only in turning up leads and witnesses favorable to the defense, but in allowing counsel to take full advantage of trial tactics such as cross-examination and impeachment of adverse witnesses. And of course, adequate legal investigation is necessary to ensure that all available defenses are raised and that the government is put to its proof. 77 “[I]t is axiomatic among trial lawyers and judges that cases are not won in the courtroom but by the long hours of laborious investigation and careful preparation and study of legal points which precede the trial.” 78
Moreover, the necessity for exhaustive investigation is not limited to its value in preparation for trial. As a leading manual for defense lawyers emphasizes:
“The facts are counsel’s most important asset not only in arguing before a jury but in every other function counsel performs: seeking advantageous terms of bail, urging the prosecutor to drop or reduce charges, negotiating with him about a plea, urging a favorable sentence recommendation on a probation officer or sentencing disposition on a judge.” 79
At a minimum, the duty to investigate requires counsel (or his investigator) 80 to *279 contact persons whom he has or should have reason to believe were witnesses to the events in question, to seek witnesses in places where he has or should have reason to believe the events occurred, and to conduct these interviews and investigations as promptly as possible, before memories fade or witnesses disappear. 81
In the present case, Decoster’s attorney did none of these things. Although the failure to interview a particular witness, by itself, may not rise to the level of inadequate assistance, defense counsel’s investigation and preparation for this case was so perfunctory that it clearly violated his duties to his client 82 The prosecution called three witnesses at trial — Roger Crump and Officers Box and Ehler. Despite the cardinal rule that proper investigation begins with interviews of those witnesses whom the government intends to call, 83 particularly the arresting and investigating officers, 84 defense counsel made no attempt to interview any of these witnesses at any time prior to trial. Nor did he request or obtain a transcript of the preliminary hearing where these witnesses testified. 85 Defense counsel did not even contact and interview Decoster’s codefendants, Eley and Taylor, before trial. Nor did he seek or talk to any witnesses at the hotel or bar. In fact, defense counsel made absolutely no effort to discover, contact, or interview a single witness prior to trial. 86 Apparently, he was willing to go to trial without having made any real effort to determine what could be elicited by way of defense or to evaluate the strengths and weaknesses of his client’s case. 87
*280 Moreover, defense counsel’s violations of the duties owed to his client were not limited to an egregious failure to investigate. There are several indications that counsel did not “confer with his client ... as often as necessary to elicit matters of defense, or to ascertain that potential defenses are unavailable.” 88 Surely, many of the problems that developed at and just prior to trial could have been eliminated had counsel more fully prepared himself and discussed the case with his client. 89 In addition, coun *281 sel was derelict in his duty to “promptly advise his client of his rights and take all actions necessary to preserve them.” 90 For example, 50 days elapsed from the time appellant was accepted for third-party custody until his attorney filed a proper bond review motion. 91 Finally, counsel’s representation of his client at the sentencing hearing was anything but diligent and conscientious. Despite the critical need for effective advocacy at what “may well be the most important part of the entire proceeding,” 92 counsel’s total contribution at appellant’s hearing consisted of the following “allocution”:
If the Court please, Counsel is aware that Your Honor has a fully comprehensive a [sic] detailed probation report, and Counsel is aware of the report and would submit based on said report. 93
In sum, counsel violated each of the duties enunciated in DeCoster I as the prerequisites of a reasonably competent performance. Appellant’s court-appointed attorney provided the kind of shoddy representation that none of us would tolerate for ourselves — a slovenly, slipshod job, almost totally lacking in preparation, characterized by repeated failures to protect his client’s rights and an obvious indifference to his client’s fate.
B. “Substantial” Violations
Contrary to the intimations of the majority, we do not contend that the slightest departure from a checklist of counsel’s duties establishes ineffectiveness and requires reversal. 94 Since counsel’s decisions must be adapted to the complexities of a given case, the proper performance of an attorney’s obligations necessarily entails considerable discretion. Moreover, the human animal is too fallible and the task of defense counsel too complex to expect that every action taken by an attorney will prove correct on hindsight. We have repeatedly cautioned that “[t]his court does not sit to second guess strategic and tactical choices made by [defense] counsel.” 95 The Sixth Amendment demands that counsel’s conduct be conscientious, reasonable, and informed by adequate investigation and preparation; it does not demand that counsel’s performance be flawless.
Thus, like the majority, we recognize that counsel’s conduct must be evaluated in the *282 context of a particular case and that not every deviation from a perfect, or even average performance makes out a claim of ineffective assistance. Instead, counsel’s violations must be substantial to offend the Sixth Amendment right to effective assistance of counsel. 96 The duties articulated in DeCoster I, like the ABA Standards and the obligations prescribed by the Fourth Circuit in Coles v. Peyton, 97 describe the minimum components of a competent performance and provide the court with an objective basis for assessing the adequacy of representation. A demonstration that counsel has violated one of these duties compels further inquiry into counsel’s conduct to determine whether, in this specific case, counsel’s departure from the prescribed standards was either “excusable” or “justifiable.” The first of these inquiries recognizes that even the most diligent and conscientious attorney may occasionally falter in fulfilling his responsibility; one minor error in an otherwise commendable performance does not automatically render the representation inadequate. 98 The second inquiry is necessary because the “reasonably competent” attorney must tailor his actions to fit the unique circumstances presented by a given case; some particular situations may justify or even mandate a course of action that transgresses the general list of duties, a list that of necessity was designed to govern defense counsel’s conduct in the typical criminal case.
In this case, the frequency and pervasiveness of defense counsel’s omissions and failures certainly belie any notion that these actions were isolated and excusable events. The violation in this case was not simply that counsel failed to interview certain named witnesses. The record reveals that counsel conducted almost no investigation whatsoever in the 17 months preceding trial. Consequently, he began trial unaware of what the prosecution witnesses would say and as a result was unable to refute their stories, was ignorant of the possible defenses and witnesses he might present, and was even unsure of his own client’s version of the events.
Nor do any special circumstances justify counsel’s breach of his obligations. In some cases prudential judgments or tactical considerations may be involved in counsel’s decision about whom to interview. 99 In the present case, however, there simply is no *283 possible justification for counsel’s near-total lack of investigation and preparation.
Defense counsel’s failure to investigate cannot be justified on the basis that he felt he was familiar enough with the facts of this case to judge for himself that his client was guilty. To begin with, the assertion that defense counsel had sufficient knowledge of what an investigation would reveal is manifestly unsupportable. From all that appears affirmatively in the record, defense counsel’s entire knowledge of the events in question derived solely from two sources: conversations he may have had with his client and his representation of the defendant at the preliminary hearing. As to the former, the record reveals only that appellant and his attorney appeared in court together on six occasions; 100 counsel presented no evidence on remand on the extent of his communications with his client. As to the latter, the preliminary hearing occurred 17 months before trial, lasted all of 20 minutes and consisted entirely of the testimony of Officer Ehler, who was not even the arresting officer. Thus, defense counsel’s total knowledge of the case in fact consisted entirely of two conflicting versions of the events — one from a police officer and the other from his own client. 101
Perhaps counsel concluded from this limited information that his client had no alibi defense and was guilty, and that therefore counsel was excused from conducting any investigation. 102 But the suggestion that a client whose lawyer believes him to be guilty deserves less pretrial investigation is simply wrong. An attorney’s duty to investigate is not relieved by his own perception of his client’s guilt or innocence. 103 I can think of nothing more destructive of the *284 adversary system than to excuse inadequate investigation on the grounds that defense counsel — the accused’s only ally in the entire proceedings — disbelieved his client and therefore thought that further inquiry would prove fruitless. 104 The Constitution entitles a criminal defendant to a trial in court by a jury of his peers — not to a trial by his court-appointed defense counsel. 105
In this case, however, court-appointed counsel failed to interview even the prosecution witnesses, ostensibly because he was already “aware” of the main points of their likely testimony. By virtue of his attendance at the preliminary hearing perhaps counsel obtained a good indication of Officer Ehler’s likely testimony at trial. 106 But there can be no justification for counsel’s *285 failure to interview either complainant Crump or Officer Box, the critical witnesses to the circumstances surrounding the arrest of appellant. 107
As for the codefendants Taylor and Eley, the district court concluded that, “[a]s a result of the information the defendant had given him and his knowledge of the prior ‘guilty’ pleas of the co-defendants, [counsel] considered that their testimony might be contradictory to that of the defendant.” 108 My colleagues apparently believe that this finding justifies counsel’s failure to interview the codefendants. But counsel’s speculations and beliefs are no substitutes for facts. No matter how experienced an attorney is, no matter how astute his predictions of the content of the witnesses’ testimony might be, it is inexcusable not to interview such key potential witnesses on the ground that counsel feels that their testimony might contradict what the defendant has told him. Counsel’s reasoned and informed judgment that a witness’ testimony would not aid the defense may justify not calling that witness to testify, but it cannot excuse the failure even to interview him. 109
It is true that appellant complicated defense counsel’s task when, sometime before trial, he sent counsel a letter suggesting a new version of what happened on the day of his arrest. 110 Again, however, this has no *286 bearing on counsel’s failure to investigate. 111 By his own account, counsel did not learn of the self-defense theory until the day before or the day of the trial. Appellant’s conflicting stories, therefore, obviously cannot excuse counsel’s inaction during the previous seventeen months. If anything, appellant’s differing accounts should have emphasized the need for an independent investigation to determine which, if either, version was accurate and could be presented as a defense. 112 Yet, even after receiving appellant’s letter, counsel was ready to go to trial without having attempted to contact the codefend-ants to learn their version of the events on the night of the robbery.
In the end, the majority’s conclusion that appellant was not denied the effective assistance of counsel rests on their perception that the record contains overwhelming evidence of appellant’s guilt. 113 [Ultimately, “there was a total failure of appellant to show that it was likely that counsel’s deficiencies had any effect on the outcome of [the] trial.” 114 The logic of their position seems to be as follows: If the accused was probably guilty, then nothing helpful could have been found even through a properly conducted investigation. Thus, any violation of that duty — no matter how egregious — was inconsequential and hence excusable. 115
*287 Even on its own terms, such reasoning is faulty. It assumes that the value of investigation is measured only by information it yields that will exonerate the defendant. Yet, even if an investigation produces not a scintilla of evidence favorable to the defense — an unlikely hypothesis — appellant still will benefit from a full investigation. One of the essential responsibilities of the defense attorney is to conduct an independent examination of the law and facts so that he can offer his professional evaluation of the strength of the defendant’s case. 116 If this full investigation reveals that a plea of guilty is in the defendant’s best interests, then the attorney should so advise his client and explore the possibility of initiating plea discussions with the prosecutor. 117 It is no secret that in the majority of criminal prosecutions the accused is in fact guilty, notwithstanding any initial protestations of innocence. It is also no secret that the vast majority of criminal prosecutions culminating in conviction are settled through plea bargaining. 118 Indeed, the Supreme Court has recognized that plea bargaining will remain “an essential component of the administration of justice” in this country until the courts’ resources are greatly expanded. 119 In many cases, therefore, perhaps the most valuable function that defense counsel can perform is to advise the defendant candidly that a thorough investigation — conducted by his own representative and seeking any glimmer of exonerating evidence — has turned up empty. Only then can the defendant truly evaluate his position and make an informed decision whether to plead guilty or whether to continue to assert his innocence at trial. 120
More importantly, the majority’s position confuses the defendant’s burden of showing that counsel’s violation was “substantial” with the government’s burden of proving that the violation was not “prejudicial.” The former entails a forward-looking inquiry into whether defense counsel acted in the manner of a diligent and competent attorney; it asks whether, at the time the events occurred, defense counsel’s violations *288 of the duties owed to his client were justifiable. In contrast, the inquiry into “prejudice” requires an after-the-fact determination of whether a violation that was admittedly “substantial,” nevertheless did not produce adverse consequences for the defendant.
All that the accused must show to establish a Sixth Amendment violation is that counsel’s acts or omissions were substantial enough to have deprived him of the effective assistance of counsel in his defense. He need not prove that counsel’s violations were ultimately harmful in affecting the outcome of his trial. 121 Quite simply, the inquiry into the adequacy of counsel is distinct from the inquiry into guilt or innocence. The Constitution entitles every defendant to counsel who is “an active advocate in behalf of his client.” 122 Where such advocacy is absent, the accused has been denied effective assistance, regardless of his guilt or innocence. The majority opinions nevertheless force the appellant to shoulder the burden of proving that counsel’s acts or omissions actually or likely affected the outcome of the trial. 123 To thus condition *289 the right to effective assistance of counsel on the defendant’s ability to demonstrate his innocence is to assure that only the constitutional rights of the innocent will be vindicated. Our system of criminal justice does not rest on such a foundation.
Recent Supreme Court decisions affirm that a distinct showing of prejudice is unnecessary to establish a Sixth Amendment violation. In Geders v. United States, 124 for example, the defendant had been ordered not to consult with his attorney during the overnight recess between his direct and cross-examination. The Court of Appeals affirmed the conviction on the grounds that the defendant failed to claim any prejudice from his inability to confer with counsel. 125 The Supreme Court, however, found that the petitioner had been deprived of his Sixth Amendment right and made no inquiry into whether he had been prejudiced in any way by the trial court’s order. That the defendant had been deprived of the assistance of counsel established the constitutional violation; no showing of prejudice was necessary. 126
More recently, in Holloway v. Arkan sas, 127 the Court found a violation of the Sixth Amendment in the trial court’s failure to appoint separate counsel in the face of the defense attorney’s assertions that conflicting interests might prevent him from providing effective assistance for each of three codefendants. In Holloway, as in Geders, the petitioners’ appeal below had been rejected on the ground that the record demonstrated no actual conflict of interest or prejudice to the defendants. Again, however, the Supreme Court did not inquire into whether the defendants were prejudiced or even into whether trial counsel’s claim of possible conflicting interests was valid. Because the trial court failed, in the face of counsel’s objections, either to appoint separate counsel or to ascertain whether the risk of conflict of interests was too remote to require such appointment, the defendants were deprived of their right to the effective assistance of counsel. 128 Upon finding a constitutional violation, the Court then proceeded, in a separate part of its *290 opinion, to determine whether reversal of the petitioners’ convictions was required. 129
C. Was the Substantial Violation Prejudicial?
Having determined in this case that counsel’s violation of his duty to his client was substantial, 130 and that appellant consequently was denied the effective assistance of counsel, we now must consider whether this violation of the Sixth Amendment mandates reversing appellant’s conviction. 131 Our inquiry is governed by Chap *291 man v. California, 132 in which the Supreme Court concluded that “there may be some constitutional errors which in the setting of a particular case are so unimportant and insignificant that they may ... be deemed harmless . . .” 133 For these errors, a return to earlier stages in the criminal process would merely be an exercise in futility because the second proceedings would be certain to reach the same result as the first.
Under Chapman the burden in each case rests squarely on the government to prove beyond a reasonable doubt that an error was harmless before the defendant’s conviction can be allowed to stand. 134 To place the burden on the defendant would require him to establish the likelihood of his innocence. The presumption of innocence 135 that cloaks the accused cannot be stripped by a conviction obtained in something less than a constitutionally adequate trial.
To satisfy its burden of establishing lack of prejudice, it is not enough for the government simply to point to the evidence of guilt adduced at trial, no matter how overwhelming such evidence may be. 136 In the first place, “proof of prejudice may well be absent from the record precisely because counsel has been ineffective.” 137 When, as in this case, ineffectiveness is founded upon gross omissions of counsel rather than spe *292 cific errors, counsel’s violations so permeate the trial that they necessarily cast doubt on the entire adjudicative process. 138 Even where the consequences of counsel’s omissions are less pervasive, it will generally be impossible to know precisely how the proceedings were affected, 139 and the resulting prejudice will be “incapable of any sort of measurement.” 140 As the Supreme Court has emphasized, “ ‘The right to have the assistance of counsel is too fundamental and absolute to allow courts to indulge in *293 nice calculations as to the amount of prejudice resulting from its denial.’ ” 141
Moreover, “prejudice” to the defendant may take many forms. The likelihood of acquittal at trial is not the only touchstone against which the consequence of counsel’s failures is to be measured. The duties of an attorney extend to many areas not necessarily affecting the outcome of trial. As the present case highlights, inadequate investigation and preparation may prejudice the defendant not only at trial but before trial — in counsel’s inability to offer informed, competent advice on whether to plead guilty and whether to demand a jury trial — as well as after trial — in providing ineffective representation at sentencing. 142
These principles, in fact, might suggest that a per se rule is appropriate in all cases in which counsel’s representation fails to meet the standards of the Sixth Amendment. 143 It may be that the prejudice to the defendant from the denial of effective assistance of counsel is so great, and the likelihood that the government can prove lack of prejudice so small, that reversal should be required whenever a substantial violation of counsel’s duties is shown. Indeed, the Supreme Court has frequently emphasized that “the assistance of counsel is among those ‘constitutional rights so basic to a fair trial that their infraction can never be treated as harmless error.’ ” 144 And it may be that only a rule requiring automatic reversal can provide the deterrent effect necessary to insure that all defendants — innocent or guilty — receive the effective assistance of counsel. 145
Nevertheless, there may be cases — however few — in which the reviewing court can isolate specific deficiencies in counsel’s performance and can accurately gauge the consequences of counsel’s acts or omissions. 146 *294 For example, when a defendant alleges ineffectiveness because his counsel has failed to object to the introduction of arguably inadmissible evidence, the consequences of counsel’s violation may readily be measured. There, the government may be able to prove harmlessness either by showing that a suppression motion would not have succeeded (and the evidence would have been admitted anyway), 147 or by proving that there was no “reasonable possibility that the evidence complained of might have contributed to the conviction.” 148 Similarly, where counsel violates his duties by failing to interview a particular witness, the government may be able to carry its burden by proving, through proffer of the witness’ testimony, that the witness had nothing relevant to offer even if he had been interviewed. As these examples demonstrate, in appropriate circumstances, reversing the defendant’s conviction may not be required because rectifying counsel’s errors could not possibly benefit the defendant. 149
On the record before us in the present case, I would conclude that the government has failed to discharge its burden of proving that no adverse consequences resulted from counsel’s gross violations of his duties to his client. Several important questions on the matter of prejudice remain unanswered, and in the absence of any evidence on these critical issues, I am unable to find that counsel’s violations were “so unimportant and insignificant” 150 that reversing appellant’s conviction would be a futile exercise. No inquiry was made for example, on the relationship between counsel’s failure to investigate and Decoster’s decision to go to trial rather than to seek and possibly accept a plea bargain comparable to that of his codefendants. 151 Nor was there exploration *295 of whether counsel’s failure to offer any allocution at the sentencing hearing had any bearing on the trial judge’s decision to sentence Decoster to a prison term of 2-8 years while his codefendants received only probation. 152
In DeCoster I, the court expressly stated that reversal would be required, “unless the government, ‘on which is cast the burden of proof once a violation of these precepts is shown, can establish lack of prejudice thereby.’ ” 153 Thus, any doubts about the harmlessness of counsel’s violations in this case would ordinarily be resolved against the government, and the case would be reversed and remanded for a new trial. Yet, despite our prior remand, it is not clear that the government was ever required to satisfy its burden of proving that the denial of Decoster’s right to the effective assistance of counsel was harmless beyond a reasonable doubt. In ruling on several of appellant’s contentions the district court found that counsel had not violated any duty owed to his client. As to these issues, therefore, the question of the prejudicial effect of counsel’s conduct was never reached. On several other allegations, the district judge appears to have required appellant to demonstrate that he was prejudiced by counsel’s actions in order to establish a constitutional violation. Thus, the government was not put to its proof in establishing harmlessness. Moreover, the initial remand hearing was conducted without benefit of this opinion’s elucidation of the principles set forth in DeCoster I governing questions of ineffective assistance. I therefore would remand the case to allow the government an opportunity to satisfy its burden of proving harmlessness.
IV
Of course, even reversing Deeoster’s conviction would not remedy the pervasive problem of ineffective representation of the indigent. 154 The disparity between representation of the poor and of the well-to-do reflects the larger inequality of riches in our affluent society. The imbalance in the quality of legal assistance provided for the indigent and the wealthy is only one of a host of inequities in our society — inequality of educational opportunity, of jobs, of housing, of health care.
It is not the province of the judiciary to remedy all these inequities. We have neither the means nor the competence to redress all of society’s imbalances. We however do have the duty, entrusted to us by the Bill of Rights, to assure that no individual is deprived of liberty by our courts of *296 law without a constitutionally adequate trial. We violate this duty when we place our imprimatur of “Equal Justice Under Law” on the incompetent performance of court-appointed counsel in cases like the one before us. 155
An appellate court’s role is limited. We can promulgate standards that specify the minimum requirements of the constitutionally mandated competent performance. We can closely scrutinize the records of those cases in which effectiveness is at issue, carefully monitoring trial counsel’s performance to ensure that the attorney’s obligation have been fulfilled. 156 When substantial violations are uncovered, we can enforce the Sixth Amendment’s guarantee by vacating the defendant’s conviction and remanding for a new trial in which the effective assistance of counsel will be provided.
The real battle for equal justice, however, must be waged in the trenches of the trial courts. Although reversing criminal convictions can have a significant deterrent effect, an appellate court necessarily depends upon the trial courts to implement the standards it announces. No amount of rhetoric from appellate courts can assure indigent defendants effective representation unless trial judges — and ultimately defense counsel themselves — fulfill their responsibilities. The Supreme Court, too, has recognized the duty of the trial court to fulfill the Sixth Amendment’s promise:
[I]f the right to counsel guaranteed by the Constitution is to serve its purpose, defendants cannot be left to the mercies of incompetent counsel, and judges should strive to maintain proper standards of performance by attorneys who are representing defendants in criminal cases in their court. 157
Because, as this case demonstrates, ineffective representation is often rooted in inadequate preparation, a first step that a trial judge can take is to refuse to allow a trial to begin until he is assured that defense counsel has conducted the necessary factual and legal investigation. The simple question, “Is defense ready?” may be insufficient to provide that assurance. Instead, we should consider formalizing the procedure by which the trial judge is informed about the extent of counsel’s preparation. 158 Before the trial begins — or before a guilty *297 plea is accepted 159 — defense counsel could submit an investigative checklist certifying that he has conducted a complete investigation and reviewing the steps he has taken in pretrial preparation, including what records were obtained, which witnesses were interviewed; when the defendant was consulted, and what motions were filed. 160 Although a worksheet alone cannot assure that adequate preparation is undertaken, it may reveal gross violations of counsel’s obligations; at a minimum, it should heighten defense counsel’s sensitivity to the need for adequate investigation and should provide a record of counsel’s asserted actions for appeal.
The trial judge’s obligation does not end, however, with a determination that counsel is prepared for trial. Whenever during the course of the trial it appears that defense counsel is not properly fulfilling his obligations, the judge must take appropriate action to prevent the deprivation of the defendant’s constitutional rights. “It is the judge, not counsel, who has the ultimate responsibility for the conduct of a fair and lawful trial.” 161
My colleagues fear that judicial “inquiry and standards . . [may] tear the fabric of [our] adversary system.” 162 But for so very many indigent defendants, the adversary system is already in shreds. Indeed, until judges are willing to take the steps necessary to guarantee the indigent defendant “the reasonably competent assistance of an attorney acting as his diligent conscientious advocate,” 163 we will have an adversary system in name only. The adversary system can “provide salutary protection for the rights of the accused” 164 only if *298 both sides are equally prepared for the courtroom confrontation. 165
Some of my colleagues are also concerned that a wide-ranging inquiry into the conduct of defense counsel would transform the role of the trial judge. To emphasize the supposed hazards of such a result, the majority refers to the warning of Judge Prettyman in Mitchell v. United States: 166
If the trial judge were required, after a trial has been concluded, to judge the validity of the trial by appraising defense decisions, he would also be under an obligation to protect those rights of an accused as the trial progressed. [Emphasis added]
Yet this is the very role that the Constitution has assigned the trial judge. His is the ultimate responsibility for ensuring that the accused receives a fair trial, with all the attendant safeguards of the Bill of Rights. It is no answer to say that defense counsel will fulfill the function of protecting the accused’s interest; the very essence of the defendant’s complaint is that he has been denied effective assistance of counsel. 167 The trial judge simply cannot “stand idly by while the fundamental rights of a criminal defendant are forfeited through the inaction of ill-prepared counsel . . .” 168
*299 However vigilant the judge, the problem of inadequate representation of the indigent cannot be solved by the courts alone, The bench, bar and public must jointly renew our commitment to equal justice. 169 The bar certainly must increase its efforts to monitor the performances of its members and to take appropriate disciplinary action against those attorneys who fail to fulfill their obligations to their clients. 170 Addi *300 tional funding is needed to increase the number of public defender positions and to provide those organizations with better support services. We must increase the compensation of court-appointed counsel to attract high-quality legal talent and to ensure that those who represent the indigent on a regular basis do not have to sacrifice all economic security to perform this vital role. We must reduce the caseloads of both public defenders and court-appointed counsel to manageable levels. And we must establish procedures to insulate appointed counsel from the pressure to curry favor with the judges who appoint them and fix their compensation.
That the ultimate solution does not lie exclusively within the province of the courts does not justify our ignoring the situation nor our accepting it as immutable. The people have bestowed upon the courts a trust: to ensure that the awesome power of the State is not invoked against anyone charged with a crime unless that individual has been afforded all the rights guaranteed by the Constitution. We fail that trust if we sit by silently while countless indigent defendants continue to be deprived of liberty without the effective assistance of counsel.
J. SKELLY WRIGHT, Chief Judge, joined by BAZELON and SPOTTSWOOD W. ROBINSON, III, Circuit Judges:
I write only to note that while there is no majority opinion of the court, Judges Bazel-on and Robinson and I agree on the two fundamental principles dispositive of this case: (1) The constitutional standard of effective assistance of counsel in a criminal case is the reasonably competent assistance of an attorney acting as the defendant’s diligent, conscientious advocate; and (2) where that standard is shown by the defendant not to have been satisfied, the defendant has been denied his constitutional right to counsel and his conviction must be reversed unless the Government proves beyond a reasonable doubt that the ineffective assistance of counsel was harmless. Chapman v. California, 386 U.S. 18 , 87 S.Ct. 824 , 17 L.Ed.2d 705 (1967).
APPENDIX
Opinion After Remand (Criminal 2002-71).
Before BAZELON, Chief Judge, and WRIGHT and MacKINNON, Circuit Judges.
Opinion for the Court filed by BAZEL-ON, Chief Judge.
Dissenting Opinion filed by MacKIN-NON, Circuit Judge.
BAZELON, Chief Judge:
Appellant was convicted of armed robbery and was sentenced to 2-8 years. 1 On *301 appeal, this court su a sponte noticed several indications that appellant’s sixth amendment right to the effective assistance of counsel had been violated. United States v. DeCoster, 159 U.S.App.D.C. 326, 328-330 , 487 F.2d 1197, 1199-1201 (1973). Unwilling to speculate about whether these indications would reveal a failure to provide effective assistance, we remanded the record for supplementation.
Appellant moved in the district court for a new trial. After holding evidentiary hearings and oral argument, the district judge denied the motion. The record has been returned to us, and we find that appellant’s conviction must be reversed because his Sixth Amendment rights were infringed.
I
The facts surrounding the offense are set forth in our first opinion and can be briefly summarized. The victim, Roger Crump, testified that on May 27, 1970, his wallet containing $110 was taken by three men, one of whom held a knife. He further testified that while this was occurring, two persons, who he later learned were plainclothes officers Box and Ehler, jumped out of a car and chased the three men. Because his eyesight was impaired as a result of an automobile accident occurring months after the alleged offense, Crump was unable to identify the appellant at trial. He stated, however, that immediately after the alleged robbery he had identified all three persons who were arrested.
Officers Box and Ehler did identify appellant as one of the persons they had seen robbing the victim. Officer Box also testified that he had chased the appellant into a nearby hotel, the D.C. Annex, where appellant was arrested while standing at the hotel desk. At the hotel, Box stated, Crump had identified appellant. Officer Ehler testified that he had chased, arrested, and searched Earl Taylor and had found a straight razor in his pockets. The wallet and money were never found.
In his own defense, appellant testified that on the afternoon of the crime he had been drinking with Crump at the Golden Gate bar, near the scene of the alleged crime. (Crump admitted having been in the bar but could not recall whether he had seen appellant there.) Appellant claimed that after leaving the bar he had walked to the hotel where he was staying, and was arrested while obtaining his key from the desk clerk. He denied having been with Earl Taylor, and denied even knowing the third alleged robber, Douglas Eley, at the time of the offense. The only other defense witness was Eley; he agreed that appellant and Crump had been together at the bar but stated that he had subsequently seen them fighting outside the bar.
The facts surrounding trial counsel’s efforts require a more detailed statement. The appellant was arraigned in the Court of General Sessions on May 30,1970, and bond was set at $5,000. At that time the lawyer who was to represent appellant at trial was appointed. A preliminary hearing was held on June 8 at which trial counsel represented appellant and did most of the questioning. On the basis of Officer Ehler’s testimony, the three defendants were held for the Grand Jury.
After being indicted, the three defendants were arraigned in United States District Court, where appellant’s counsel’s appointment was reaffirmed. On November 4, appellant wrote to the district judge claiming he was guilty only of “assault by self defence” [sic] and requesting a new lawyer. The only specific charge leveled against counsel was that appellant had been accepted for pretrial custody by the Black Man’s Development Center on October 12, and no bond review motion had been filed. On November 9 counsel filed such a motion in district court, but did not mention the acceptance by Black Man’s. One week later, appellant filed a pro se motion for bond review (which also did not mention Black Man’s). On November 18, 1970, the district judge, as required by law, 2 continued the *302 motions to await review by the Court of General Sessions, which had originally set the bail. Counsel did not make the motion for review in General Sessions until December 8, and it was denied December 12.
On January 12, 1971, the case was called for trial in district court, but a continuance was granted after the prosecutor indicated Crump was hospitalized following his automobile accident. Two days later, the district judge granted the bond review motion and released appellant to Black Man’s. On January 21, 1971, appellant absconded, and shortly thereafter a bench warrant issued. On June 17, trial of the two codefendants commenced, but in the middle of trial they pleaded guilty.
Appellant was rearrested on September 2, 1971, on unrelated charges for which he was ultimately convicted in Superior Court. Trial was set in this ease for November 15, 1971. On the day of the trial, appellant asked the court to subpoena the two code-fendants, explaining that he “didn’t have a chance” to talk to his lawyer about this. Counsel indicated he had considered the possibility of subpoenas, but did not have the codefendants’ addresses. The prosecutor reported that Eley was in jail (where he had been for six weeks); 3 the court read from the court file the address Earl Taylor had listed at the time of his release on personal recognizance eleven months earlier. That address proved out of date.
After counsel announced he was ready for trial, the prosecutor informed the court that the Government had served an alibi-notice demand and received no response. Defense counsel argued that while he might rely on an alibi, no response was necessary because the Government had not given 20 days notice as required by the local rules. The court decided that although the Government had been dilatory, the names of alibi witnesses should be provided nonetheless. Counsel then announced that he would “proceed without the alibi witnesses.” 4
Counsel next informed the court that his client wished to be tried without jury. When asked if he had considered the fact that the trial judge already had listened to some of the evidence in the codefendants’ trial, counsel stated that he thought their pleas had been entered before any evidence was heard. At this point the defendant requested a continuance because he felt, “I can’t get proper representation.” Counsel then requested to withdraw because of his client’s dissatisfaction, but his motion was denied. Defendant was convicted on November 16 and sentenced March 3, 1972. Our opinion issued October 4, 1973.
This much was clear to us on the original appeal and aroused our concern. 5 At the hearings on remand, held February 6, 11, and 13, 1974, additional information was elicited concerning counsel’s preparation and his explanation for his actions. Counsel admitted that he had not interviewed the victim, at least one and perhaps both of *303 the police officers, 6 anyone at the Golden Gate bar or D.C. Annex Hotel, 7 or the code-fendant Taylor. Counsel claimed that he had interviewed codefendant Eley on the morning of the second day of trial, and that Eley had maintained that appellant was not present during the robbery. 8 Counsel also admitted he had not obtained a transcript of the preliminary hearing, but stated that he had held several conferences with the prosecutors, and surmised, based on their usual practice, that they had made the transcript available to him and that he had read it. 9 He also guessed, based on his usual practice, that he had obtained the 251 Form from the police department.
By way of explanation counsel testified that not until shortly before trial had appellant ever mentioned any witnesses, and then just the two codefendants. Counsel further stated that at about the same time he had received a letter from appellant in which appellant admitted he had fought with Crump but denied having robbed him. 10 The letter indicated that the code-fendants would support this claim. Counsel testified that this letter — which was consistent with Decoster’s earlier letter to the district judge and with Eley’s trial testimony — was the first time appellant had indicated to counsel that appellant had seen Crump after leaving the bar. Until that time, counsel testified, appellant had maintained, as he testified at trial, that he had gone directly to the hotel. 11 Based on the letter, counsel concluded that the testimony of the codefendants “might be devastating” (presumably to the alibi defense). He nevertheless agreed at trial to subpoena the codefendants, and called Eley after interviewing him and establishing that Eley would support appellant’s alibi testimony. 12
In denying the new trial request, the district court identified seven acts or omissions by counsel which appellant alleged had deprived him of his right to effective assistance of counsel. 13 With respect to two allegations — the delay in moving for bond review and the failure to obtain a transcript of the preliminary hearing — the court found that the defendant had not been prejudiced by the violations. With respect *304 to three other allegations — counsel’s attempt to waive jury trial, his waiver of an opening statement, and his failure to see that Lorton Reformatory gave appellant credit for time served as ordered by the sentencing judge — the court found no ineffective assistance. And with respect to the final two allegations — the failure to interview witnesses and counsel’s premature announcement that he was ready — the court made findings of fact and several conclusions of law, reprinted in pertinent part in the margin. 14 These conclusions can be read as either holding that there was no constitutional violation and in any event no prejudice, or simply holding that no prejudice was shown.
II
A.
The benchmarks for adjudicating this case are set forth in our original opinion.
In DeCoster I we unanimously held that, at least when counsel’s performance is challenged on a direct appeal, appellants need not show that “the proceedings were a farce and a mockery of justice” 15 or that “gross incompetence of counsel has in effect blotted out the essence of a substantial defense.” 16 Rather, following a number of other circuits, 17 we adopted a more stringent standard: “a defendant is entitled to the reasonably competent assistance of an attorney acting as his diligent conscientious advocate.” 159 U.S.App.D.C. at 331 , 487 F.2d at 1202 . Moreover, recognizing that “ ‘reasonably competent assistance’ is only a shorthand label, and not subject to ready application,” we articulated several duties owed by counsel to a client:
In General — Counsel should be guided by the American Bar Association Standards for the Defense Function. .
*305 Specifically — (1) Counsel should confer with his client without delay and as often as necessary to elicit matters of defense, or to ascertain that potential defenses are unavailable. Counsel should discuss fully potential strategies and tactical choices with his client. (2) Counsel should promptly advise his client of his rights and take all actions necessary to preserve them. . (3) Counsel must conduct appropriate investigations, both factual and legal, to determine what matters of defense can be developed. . . [I]n most cases a defense attorney, or his agent, should interview not only his own witnesses but also those that the government intends to call, when they are accessible. The investigation should always include efforts to secure information in the possession of the prosecution and law enforcement authorities. And, of course, the duty to investigate also requires adequate legal research. (Footnotes omitted.)
Id. at 332-33 , 487 F.2d at 1203-04 . And we stated that these duties are only “a starting point for the court to develop, on a case by case basis, clearer guidelines for courts and for lawyers as to the meaning of effective assistance.” Id. at 332 n.23, 487 F.2d at 1203 n.23.
The requirements set forth in DeCoster I are indisputably the minimal components of “reasonably competent assistance.” Even so, not every violation of one of the duties warrants reversing a conviction for ineffective assistance. Rather, DeCoster I contemplates a three step inquiry: did counsel violate one of his articulated duties; was the violation “substantial”; and was the substantial violation “prejudicial.” Id. at 333 , 487 F.2d at 1204 .
If all defense attorneys had the dedication, skill and experience of a Clarence Darrow, or if all clients had the sophistication and resources of an Andrew Carnegie, then perhaps the concern embodied in DeCoster I might be unnecessary. This is not to say that such lawyers always will render, or such clients always receive effective assistance of counsel; the task of criminal representation is too difficult and the human animal too fallible. But in a world of Dar-rows and Carnegies, perhaps it would be tolerable for judges to assume a more passive role.
We do not live in that kind of world, however. In the real world of criminal justice, the vast majority of defendants lack the means to afford effective representation and/or the sophistication to vindicate their right to it. The governing principle is clear: “There can be no equal justice where the kind of trial a man gets depends on the amount of money he has.” Griffin v. Illinois, 351 U.S. 12, 19 , 76 S.Ct. 585, 591 , 100 L.Ed. 891 (1956).
B.
“[Ijnvestigation and preparation,” as the Commentary to the ABA Standards for the Defense Function recognize, “are the keys to effective representation. . . . ”
[I]t is axiomatic among trial lawyers and judges that cases are not won in the courtroom but by the long hours of laborious investigation and careful preparation. . . . The adversary process assumes and its proper functioning demands that both sides have prepared and organized their case in advance of trial. 18
Moreover, as the Standards themselves state, the “duty to investigate exists regardless of the accused’s admissions or statements to the lawyer of facts constituting guilt.” 19
The duty to investigate is not necessarily fulfilled simply by interviewing those persons whom a client names as defense witnesses; 20 it demands that counsel “make an *306 independent examination of the facts, circumstances, pleadings and laws involved. ,” 21 Minimally, this requires counsel (or his investigator) to contact persons whom he has or should have reason to believe were witnesses to the events in question; to seek witnesses in places in which he has or should have reason to believe the events occurred; and to conduct these interviews and investigations as promptly after his appointment as is possible, before memories fade or witnesses disappear. 22
In this case, according to his own admissions — and the district court’s factual findings — neither trial counsel nor an investigator did any of these things. He did not interview codefendant Taylor, and delayed interviewing Eley until the second day of trial. 23 He did not interview the complainant or the arresting officers, and he failed to search for witnesses at the hotel or the bar. From all that appears in the record, counsel advised his client on whether to go to trial, and then conducted the trial, without making any real effort to determine what could be elicited by way of defense.
The dissent argues at length that counsel knew all along that the codefendants would say the alibi was false, and that appellant had participated in the crime to which they pleaded guilty. There is nothing in counsel’s testimony at the hearing on remand, however, to support this conclusion. Instead of evidence, the dissent relies on logic, reasoning that counsel’s “decision not to contest the finding of probable cause [at the preliminary hearing] necessarily involved knowledge by defense counsel for Decoster and Taylor (the counsel whose conduct is here in question) that could only have been obtained by prior discussion of the offense with these men and by consultation with Eley or his counsel.” Dissent at - of 199 U.S.App.D.C., at 319 of 624 F.2d. The circularity of this deduction is transparent: by assuming precisely what is at issue here — namely, that counsel rendered reasonably effective assistance — the dissent is able to spin a web of facts, which, if supported by the record, would at least present a more difficult question. As matters stand, there is no basis for assuming that counsel had discussions with the code- *307 fendants that he failed to mention at the hearing on remand.
Although counsel conducted no interview, it is possible that special circumstances justify this omission, and that therefore the duty to investigate was not breached. To be sure, there is less need or room for tactical decisions in deciding who not to interview than, for example, in deciding who not to call. But tactical and prudential judgments still may be involved, 24 and this court does not sit to “second guess” informed judgments of this sort unless they are manifestly unreasonable. 25 In this case, however, we find the explanations proffered by counsel or hypothesized by the government lack plausibility. 26
1. Codefendants Taylor and Eley. Three arguments are offered by the Government to support the failure to or delay in interviewing the codefendants:
(a) It is argued that the fact that appellant gave his counsel conflicting accounts of the events — the alibi and the claim of a fight — somehow excuses the lack of prompt interviews. While the existence of these conflicts might be relevant were counsel’s failure to call witnesses at issue, the conflicts can hardly justify the failure to interview. The defendant did not offer the self-defense claim to his counsel until a day or two before trial, long after the interviews should have been conducted. Moreover, even if defendant had contradicted himself earlier, the importance of careful investigation would have been heightened, rather than lessened, since counsel would have needed to determine which defense could or should have been presented.
(b) The Government argues that locating Taylor would have been a “formidable task” after appellant’s flight. But before the flight, from September 1970 to January 1971 when trial was scheduled to begin, Taylor was available in the D.C. Jail. Moreover, one week after appellant was rearrested, and three days before a trial date was set, Taylor was sentenced in district court to probation. It strains credulity to believe that Taylor could not have been found had appellant’s counsel contacted Taylor’s on-and-off employer who had written a letter on Taylor’s behalf prior to sentencing; Taylor’s probation officer, with whom the record reveals Taylor was in regular contact at the time of appellant’s trial; or perhaps even Taylor’s lawyer.
(c) The district court found that not until the day of trial had appellant suggested to his counsel that the codefendants might be helpful to the defense. But counsel knew that the codefendants were alleged to have participated in a robbery with his client, and knew that his client claimed not to have been there. Surely counsel should have realized that the co-defendants were at least potential witnesses in support of the alibi, and they should have been interviewed.
*308 2. The Government witnesses. The district court found that counsel was justified in not interviewing Officer Ehler because counsel had examined him at the preliminary hearing. We agree. But with respect to Officer Box, the district court noted only that his testimony was “generally consistent with that of Officer Ehler,” and with respect to the victim, Crump, that “[tjhere was nothing incredible about [hisj testimony.” Whatever relevance the substance of their trial testimony may have to the question of the effect of the failure to interview, it can hardly provide a tactical justification for not conducting pretrial interviews. 27 The government contends that after appellant’s flight, interviewing Crump was “impracticable,” since he was living in Georgia. But again, the Government ignores the failure to interview for the several months between the time of the offense and the accident, even though the very occurrence of the accident demonstrates the importance of prompt interviews.
3. The desk clerk at the hotel. The district court found no reason to seek out the desk clerk because “[tjhere was no dispute as to when the defendant entered the D.C. Annex or when and where he was arrested.” However, the trial record reveals that appellant claimed he had walked from the bar to the hotel and into the lobby, while Office Box testified that he had chased appellant. Surely the desk clerk should have been contacted to ascertain whether he saw the appellant enter the hotel or remembered anything relevant about appellant’s demeanor while at the desk.
4. Other witnesses. No other potential witnesses were identified by name or job position. Nevertheless, the record reveals two potentially fruitful places for investigation that were not tapped: the hotel lobby and the bar. For the same reasons that the hotel clerk’s recollections might have been useful, i. e., to resolve the dispute as to how appellant reached the hotel, guests or residents who had been in the lobby at the time should have been interviewed. At the very least, the clerk should have been asked for the names of persons he remembered having seen in the lobby. Similarly, appellant testified that persons unknown to him had been in the bar at the same time he was there. Such witnesses could have been helpful if they could have corroborated appellant’s claim that he and Crump had been drinking together, or, perhaps, if they had overheard conversation, or seen Crump and appellant leave. Counsel at least could have questioned employees of the bar to see if they had useful information or could supply the names of customers who were at the bar at the time.
In sum, we hold that counsel’s failure to interview Taylor, Crump, Officer Box, or the desk clerk; his delay in interviewing Eley; and his failure to seek out witnesses from the hotel or the bar were not supported by tactical considerations, informed or otherwise, and violated the duty to conduct a factual investigation. 28 Of course, counsel was “under no duty to assist in the fabrication of a defense,” as the district court wisely noted. But counsel was under a duty to investigate whether there was a non-fabricated defense that could be presented. The dissent may well be correct that there were no such defenses available in this case, although it may be significant that the two codefendants pled guilty only to robbery and not armed robbery. But *309 even if the dissent is correct, investigation is still necessary not only so th

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/cases/9466830. Public record. Not legal advice.