This is an appeal from a Manitowoc county circuit court judgment convicting Ronald Fencl of first-degree murder in violation of sec. 940.01, Stats., and an order denying a new trial. This appeal was certified to this court by the court of appeals, pursuant to sec. 809.61. We affirm the judgment of conviction and order of the circuit court.

Debra Sukowaty disappeared on September 24, 1977. On October 1, 1977, the police received a purse containing Sukowaty’s identification and several other items which were found in a plastic bag in a nearby river. Among the items contained in the bag was a parking ticket, traceable to Ronald Fencl’s car. Detective Geigel of the Two Rivers Police Department visited Fencl that same day to inquire about Sukowaty. Fencl stated that *226he did not know Sukowaty or anything about the items found in the river.

At approximately 4 p.m. on October 2, 1977, Geigel again visited Fencl. At this meeting Fencl told Geigel that he wanted to talk to his attorney and that he would get back to him. Half an hour later Fencl went to the police station. He told Geigel that he had found the items in his car and threw them into the river in order to avoid any trouble with the police. Geigel had to cut their conversation short because he received a call informing him that a body had been found in a nearby gravel pit. Fencl agreed to meet with him later that evening. In the meantime, the body was identified as Sukowaty. The police then impounded Fencl’s car.

At 7 p.m. that same day, Fencl returned to the police station with his attorney, Steven Alpert. Fencl said nothing. His attorney spoke to Geigel only to ask why Fencl’s car had been impounded. Two Manitowoc Police Department detectives talked to Fencl and gave him his Miranda rights. Fencl was allowed to leave while the investigation continued. On November 4, 1977, a criminal warrant was issued charging Fencl with first-degree murder. He was arrested the next day. Alpert represented Fencl until just after the preliminary hearing. At that time new counsel was substituted because it appeared that Alpert might be called as a witness against Fencl.

During the trial the state made several references to Fencl’s pre- and post -Miranda silence. In his opening statement the district attorney referred to the 4 p.m. meeting on October 2, 1977, between Detective Geigel and Fencl. He said that Fencl did not want to answer too many questions and that Fencl wanted to talk to his attorney. Detective Geigel also testified about this statement. Geigel testified three times about his 7 p.m., October 2, 1977, meeting with Fencl and Alpert. Each *227time Geigel indicated that Fencl said nothing. In his closing argument the district attorney once again referred to the 4 p.m. meeting of October 2, 1977, between Geigel and Fencl when he stated:

“He [Geigel] said as long as you’re not mixed up in the disappearance of Debbie Sukowaty we’re not interested in prosecuting. As long as your [sic] not interested. As long as you’re not involved in Debbie Sukowaty’s disappearance, that’s alright [sic]. We’re not interested in prosecuting you. He made that quite clear. At that point Fencl said he wanted to talk to his lawyer, so Geigel left.”

The jury found Fencl guilty of first-degree murder, and the court sentenced him to life imprisonment. Fencl moved for a new trial on September 4, 1979. During a hearing on this motion, it was revealed that Fencl’s first attorney, Alpert, had engaged in some questionable practices in connection with his representation of Fencl. Nevertheless, the court denied Fencl’s motion for a new trial by order entered October 27, 1980. Fencl’s appeal of the judgment and the order was certified by the court of appeals and accepted by this court pursuant to sec. 809.61, Stats.

There are two issues presented on this appeal: (1) Did Attorney Alpert’s questionable conduct deny Fencl his constitutional right to effective counsel? (2) Did the district attorney unconstitutionally imply Fencl’s guilt by referring at trial to Fencl’s prearrest silence?

(1) Assistance of Counsel

Fencl contends that Alpert’s conduct deprived him of his constitutional right to effective assistance of counsel. It is well established that a defendant facing a felony charge has a constitutional right to effective assistance of counsel. McMann v. Richardson, 397 U.S. *228759, 771 (1970). Fencl, however, was neither charged with a felony nor arrested at the time Alpert committed several of the alleged errors. Therefore, Fencl’s assistance of counsel argument raises the question of whether a person has a constitutional right to effective counsel during the investigative stage, prior to the commencement of criminal proceedings. We do not reach this issue because we hold that Alpert’s representation was sufficiently effective.

This court set forth the standard for assessing counsel’s effectiveness in State v. Harper, 57 Wis. 2d 543, 557, 205 N.W.2d 1 (1973).

“Effective representation is not to be equated, as some accused believe, with a not-guilty verdict. But the representation must be equal to that which the ordinarily prudent lawyer, skilled and versed in criminal law, would give to clients who had privately retained his services.” A defendant “is not entitled to the ideal, perfect defense or the best defense but only to one which under all the facts gives him reasonably effective representation.”1

Fencl claims that Alpert made several errors which competent counsel would have avoided. The fact that Alpert should have acted differently, however, does not establish that his assistance was ineffective. State v. Rock, 92 Wis. 2d 554, 560, 285 N.W.2d 739 (1979). This court has long disapproved of this “hindsight-is-better-than-foresight” approach. Weatherall v. State, 73 *229Wis. 2d 22, 26, 242 N.W.2d 220 (1976); Lee v. State, 65 Wis. 2d 648, 657, 223 N.W.2d 455 (1974); Ameen v. State, 51 Wis. 2d 175, 186, 186 N.W.2d 206 (1971). In evaluating the effectiveness of Alpert’s assistance to Fencl, we must first determine whether there was a basis in reason for Alpert’s actions. State v. Rock, supra at 560; Weatherall v. State, supra at 28. If we find that Alpert’s conduct was unreasonable and contrary to the actions of an ordinarily prudent lawyer, we must then determine whether such action was prejudicial to Fencl.2 See Weatherall v. State, supra at 32. We conclude that, although Alpert’s conduct was undesirable, it was not so unreasonable as to render his assistance ineffective.

Fencl’s first contention involves some advice given by Alpert. In late September Fencl told Alpert that he had found items belonging to Sukowaty in his car and asked what he should do with them. Alpert suggested three options: turning the items over to the police, turning them in to the post office lost and found, and disposing of them. Fencl then informed Alpert that he had already thrown the items into the river. Fencl contends that an ordinarily prudent attorney would have advised him to retrieve the items and turn them over to the police. The likelihood that Fencl could have found the items was remote. Therefore, Alpert’s failure to advise Fencl to retrieve them was not unreasonable.

Fencl’s second contention is that Alpert, in order to divert attention from Fencl, passed on to the district attorney unsubstantiated stories involving other possible suspects which he had encouraged Fencl to gather. While this action seems improper at first blush, it was not *230without basis in reason. According to the record, Alpert thought that Fencl had been a police informant in the past. Thus, it was reasonable for Alpert to believe that Fencl might be able to obtain useful information about Sukowaty’s murder. Therefore, because Alpert believed Fencl was innocent, he had a reasonable basis for advising Fencl to attempt to clear himself from further investigation by assisting the police in the search for the real murderer.

Fencl’s third contention is that Alpert unwisely advised him to take the Fifth Amendment at the John Doe hearing. Fencl maintains that his silence at the hearing cast suspicion upon him. Here, Fencl is clearly using hindsight to judge Alpert’s conduct. Alpert’s decision to advise Fencl to take the Fifth Amendment at his John Doe hearing is similar to a decision regarding trial tactics. We have long held that “[i]n deciding upon trial tactics, ‘ “ ‘. . . [I]t is the considered judgment of trial counsel that makes the selection among available defenses, not the retroactive conclusion of postconviction counsel.’ ” ’ ” Weatherall v. State, 73 Wis. 2d at 26 [quoting Kain v. State, 48 Wis. 2d 212, 222, 179 N.W.2d 777 (1970)]. We conclude that Alpert’s strategy had a sufficient basis in reason. It is not unusual for defense counsel to advise a client to exercise his or her Fifth Amendment rights during criminal proceedings. Indeed, this tactic should not be dismissed without careful consideration, for in many instances silence is the best defense.

Fencl’s fourth contention is that Alpert informed the police that Fencl had lied to him in the past and might be withholding information about the case. Because we cannot find a reasonable basis for this action, we must determine whether Alpert’s indiscretion prejudiced Fencl. We conclude that it did not. Fencl has failed to persuade us that Alpert’s remarks affected his trial or the *231police investigation. Fencl had substantial contact with the police prior to this incident. According to Geigel, Fencl acknowledged that he had been in trouble with the police before. Moreover, Alpert testified that Fencl had been an informant with the Two Rivers Police Department. Thus, the police had an independent basis for judging Fencl’s credibility. In any event, by the time Alpert made the remark, Fencl’s veracity had already come into question because of his inconsistent statements relating to his knowledge of the victim. Therefore, it is highly unlikely that Alpert’s comment had any prejudicial impact.

Fencl’s fifth contention is that Alpert’s attempts to sell information to the authorities and his plans to write a book about Fencl’s case created a conflict of interest. Fencl and Alpert collaborated in the preparation of the information and the plan to sell it to the authorities. Notwithstanding the fact that this was a joint effort, Fencl maintains that Alpert’s pecuniary interest in these ventures precluded his undivided loyalty and effective assistance. Where a conflict of interest is alleged, a different analysis applies. Rather than evaluate the reasonableness of Alpert’s conduct, we must determine whether Fencl has shown by clear and convincing evidence that an actual conflict of interest existed.3 The mere possibility or suspicion that a conflict could arise under hypothetical circumstances is not sufficient. State v. Kaye, 106 Wis. 2d 1, 8, 315 N.W.2d 337 (1982); State v. Medrano, 84 Wis. 2d 11, 28, 267 N.W.2d 586 (1978). When an attorney’s misconduct involves a conflict of interest, a defendant need not demonstrate preju*232dice in order to obtain relief. State v. Kaye, supra at 9.4

Although Alpert’s conduct was clearly undesirable, we hold that Fencl has failed to establish by clear and convincing evidence that an actual conflict of interest existed. According to the record, neither of these ventures was consummated. The authorities declined to purchase any information from Alpert, so he gave it to them without charge. Moreover, there is no evidence in the record that Alpert entered into any contract relating to a book. Alpert testified that before Fencl’s preliminary hearing the book had been no more than the “germ of an idea,” taking up only about half an hour of his time. Therefore, the evidence fails to establish an actual conflict of interest.

We hold that Alpert’s representation was sufficiently effective to satisfy Fencl’s right to effective assistance of counsel.

(2) References to Prearrest Silence

Fencl argues that the prosecution violated his due process right to a fair trial and his Fifth Amendment right against self-incrimination by referring at trial to his prearrest silence.5 The references in question can *233be divided into two categories: those involving pre-Miranda silence and those involving post -Miranda silence. Detective Geigel made the post-Miranda references while testifying about his 7 p.m., October 2, meeting with Fencl. Fencl was given the Miranda warning at that meeting.6 Detective Geigel testified on three separate occasions that during the 7 p.m. meeting Fencl said nothing. The first questioned reference came on the district attorney’s direct examination of Geigel:

“Q. And did Mr. Fencl — did he come back at that time, at 7? A. Yes. He came back with his Attorney Steve Alpert.

“Q. And at that time did Mr. Fencl make any statements to you in regards to the missing girl? A. No sir.

On redirect Geigel again testified about the 7 p.m. meeting:

“Q. What about the third meeting? A. The third meeting he said nothing.”

Finally, in response to questions from the court, Geigel testified a third time about the 7 p.m. meeting:

“The Court: — You had another meeting with him?

“The Witness: We had prearranged a meeting for 7 p.m., but at this time he came in with his attorney and he didn’t say a word; his attorney did all the talking.”

We hold that the above-cited references constitute constitutional error. It is fundamentally unfair and a denial of due process to assure a defendant through the Miranda warning that he has the right to remain silent *234and then penalize him for exercising that right. In Doyle v. Ohio, 426 U.S. 610, 618-19 (1976), the United States Supreme Court stated:

“Moreover, while it is true that the Miranda warnings contain no express assurance that silence will carry no penalty, such assurance is implicit to any person who receives the warnings. In such circumstances, it would be fundamentally unfair and a deprivation of due process to allow the arrested person’s silence to be used to impeach an explanation subsequently offered at trial. Mr. Justice White, concurring in the judgment in United States v. Hale, supra, at 182-183, put it very well:

“‘[W]hen a person under arrest is informed, as Miranda requires, that he may remain silent, that anything he says may be used against him, and that he may have an attorney if he wishes, it seems to me that it does not comport with due process to permit the prosecution during the trial to call attention to his silence at the time of arrest and to insist that because he did not speak about the facts of the case at that time, as he was told he need not do, an unfavorable inference might be drawn as to the truth of his trial testimony. . . . Surely Hale was not informed here that his silence, as well as his words, could be used against him at trial. Indeed, anyone would reasonably conclude from Miranda warnings that this would not be the case.’ ” (Footnote omitted.)

It must be noted that, unlike the instant case, Doyle dealt with postarrest silence. This distinction is, however, immaterial. The Doyle decision was based upon the fact that governmental action (i.e., giving the Miranda warning) encouraged or induced silence by assuring the defendant that such silence is protected. Receipt of the Miranda warning is the important factor in the Doyle analysis, not whether the defendant has been arrested. Therefore, the Doyle rationale protects post-MmwwZa silence whether occurring before or after arrest.

The state contends that even under Doyle the references were not constitutional error because they were not used against Fencl. We disagree. We conclude that *235the prosecution’s questions were intended to suggest “a tacit admission of guilt on the part of the defendant.” Reichhoff v. State, 76 Wis. 2d 375, 378, 251 N.W.2d 470 (1977). The prosecutor directly asked Geigel whether Fencl made a statement regarding Sukowaty at the 7 p.m. meeting. Geigel responded, no. On redirect, the prosecutor again elicited testimony from Geigel that Fencl said nothing at the 7 p.m. meeting. These references to Fencl’s silence did riot help prove the facts the state contends it was trying to prove. Since the probative value of this testimony was very low, we conclude that it was introduced to tacitly imply Fencl’s guilt. Therefore, these references to Fencl’s silence were used against him and, accordingly, constitute constitutional error.

References to Fencl’s prearrest, pre-Miranda silence were made by the prosecutor during his opening and closing statements and by Detective Geigel during his testimony. The prosecutor asked Geigel how Fencl responded to his inquiry about Sukowaty during their 4 p.m. meeting on October 2, 1977. Detective Geigel’s answer was:

“A. He was very friendly. He said I want to talk to my lawyer and I’ll get back to you later.”

Relating to the same interview, the district attorney stated in his opening statement:

“And this time Ron Fencl didn’t want to answer too many questions. He said I want to talk to my lawyer first and then maybe I’ll talk to you.”

During his closing statement the district attorney argued:

“He said as long as you’re not mixed up in the disappearance of Debbie Sukowaty we’re not interested in prosecuting. As long as your [sic] not interested. As long as you’re not involved in Debbie Sukowaty’s disappearance, that’s alright [sic]. We’re not interested in prosecuting you. He made that quite clear. At that point *236Fencl said he wanted to talk to his lawyer, so Geigel left.”

Because the preceding references are to Fencl’s pre-Miranda silence, a strict interpretation of Doyle renders the due process rationale inapplicable. Jenkins v. Anderson, 447 U.S. 231, 240 (1980). The question then is whether such references are violative of Fencl’s Fifth Amendment privilege against self-incrimination.7 We hold that the protections of the Fifth Amendment do extend to pre-Miranda, prearrest silence.

“Our cases have consistently held it improper to comment upon a defendant’s choice to remain silent at or before trial.” State v. Wedgeworth, 100 Wis. 2d 514, 526, 302 N.W.2d 810 (1981).8 The protection from reference to silence arises from the Fifth Amendment guarantee against self-incrimination. Rudolph v. State, 78 Wis. 2d 435, 442, 254 N.W.2d 471 (1977) (Per Curiam) . We have construed the Fifth Amendment to protect a defendant’s silence during the early stages of a criminal proceeding. In Reichhoff v. State, 76 Wis. 2d at 380, we held that reference to a defendant’s silence at the time of arrest was constitutional error. We stated:

“The difference between prosecutorial use of the defendant’s silence at trial and prosecutorial use of the defendant’s silence at time of arrest is miniscule. We believe that in both circumstances reference to the defendant’s silence does ‘no more than turn on the red light *237of potential prejudice involving the defendant’s fifth amendment rights.’ United States v. Arnold, 425 F.2d 204, 206 (10th Cir 1970).”

In Rudolph we held that the prosecution may not affirmatively use in its case in chief the fact that the defendant chose to remain silent in the face of accusations during a custodial interrogation. 78 Wis. 2d at 441-42. We have not, however, directly decided whether the protections of the Fifth Amendment also extend to prearrest, pre-Miranda silence.

The privilege against self-incrimination found in the Fifth Amendment provides that no person “shall be compelled in any criminal case to be a witness against himself.”9 The state contends that, unless silence is compelled by arrest or a custodial interrogation, it is not protected by the Fifth Amendment. We disagree. The Fifth Amendment protects a person from compelled self-incrimination at all times, not just upon arrest or during a custodial interrogation. Any time an individual is questioned by the police, that individual is compelled to do one of two things — either speak or remain silent. If both a person’s prearrest speech and silence may be used against that person, as the state suggests, that person has no choice that will prevent self-incrimination. This is a veritable “Catch-22.” Thus the state’s theory places an impermissible burden on the exercise of Fifth Amendment rights. We hold that a person is entitled to the protection of the Fifth Amendment even prior to arrest or a custodial interrogation.10 Therefore, Fencl’s silence at *238his first meeting on October 2 with Detective Geigel was protected by the Fifth Amendment.

The state next argues that, even assuming Fifth Amendment protection, there was no constitutional error because Fencl’s silence was not used against him at trial. We conclude, however, that the state’s references to Fencl’s silence were subtle, albeit relatively ineffective, attempts to imply Fencl’s guilt. There is no other plausible reason for these references. Therefore, we hold that the state’s reference to Fencl’s prearrest, pre-Miranda silence was constitutional error.

Because we find constitutional error, we must next determine whether that error was harmless beyond a reasonable dc^bt. A constitutional error is harmless beyond a reasonable doubt if there is no reasonable possibility that the error might have contributed to the conviction. Chapman v. California, 386 U.S. 18, 23-24 (1967). We have considered the following factors in determining whether a constitutional error was harmless beyond a reasonable doubt: (1) the frequency of the error; (2) the nature of the state’s evidence against the defendant; and (3) the nature of the defense. Rudolph v. State, 78 Wis. 2d at 443. In light of the foregoing factors, we hold that the constitutional error in this case was harmless beyond a reasonable doubt.

The unconstitutional references to Fencl’s silence cannot be viewed in a vacuum but, rather, must be examined within the entire context of the trial. The trial in this case ran from June 5 through June 9, 1978. Throughout this five-day period, twenty-nine witnesses testified and numerous exhibits were introduced. While there were six references to Fencl’s silence, these references amounted to only brief, relatively isolated comments. The state did not make a concentrated, overt effort to imply Fencl’s *239guilt through references to his silence.11 Moreover, trial “counsel’s failure to raise the errors is indicative of counsel’s view of the seriousness of the error.” Odell v. State, 90 Wis. 2d 149, 155, 279 N.W.2d 706 (1979) (Per Curiam). Evidently he viewed the errors to be insufficiently egregious to warrant an objection or motion for mistrial.

Fencl argues that the state’s case was based on weak, circumstantial evidence. A finding of guilt, however, may be based upon circumstantial evidence alone. Rudolph v. State, 78 Wis. 2d at 444; see also Bautista v. State, 53 Wis. 2d 218, 223, 191 N.W.2d 725 (1971); Bethards v. State, 45 Wis. 2d 606, 612, 173 N.W.2d 634 (1970). We are convinced that there was sufficient evidence to support Fencl’s conviction. In any event, “[e]ven assuming the evidence in this case was less than overwhelming, that is but one factor to be considered in determining whether the constitutional error complained of was harmless.” Rudolph v. State, 78 Wis. 2d at 444. We are also convinced that the references to Fencl’s silence did not have a sufficient impact upon the jury to adversely affect its ability to fairly evaluate his defenses.

We hold that the state’s references to Fencl’s silence were too brief and unobtrusive to be prejudicial. Therefore, the constitutional error was harmless beyond a reasonable doubt.

*240Fencl’s next contention is that it was prejudicial error to admit testimony concerning the three investigative suggestions which he and Alpert delivered to the authorities. According to Fencl this testimony was irrelevant and prejudicial within secs. 904.01 and 904.03, Stats. This contention is without merit. The trial court properly exercised its discretion in determining that the evidence was admissible because its probative value outweighed its prejudicial impact.

Fencl’s last argument, that the trial court committed reversible error by giving Wis. J I — Criminal 1100, is equally without merit. This court recently approved that instruction in Muller v. State, 94 Wis. 2d 450, 289 N.W. 2d 570 (1980).

Because we find the errors in this case to be harmless, we hold that a new trial is not warranted.

By the Court. — Judgment and order affirmed.

(concurring). Although I am dissatisfied with the majority’s statement of the test of lawyer deficiency, I concur in the mandate of the court. Even in the absence of a standard that can serve as a guideline for future cases, I see no egregious deficiency in representation here that had the likelihood or the reality of prejudicing the defendant. Perhaps, to a degree, the majority’s view of the question of ineffectiveness of counsel is similar to Justice Stewart’s statement in respect to obscenity, when he said that, although he was not sure he could define it, “I know it when I see it.” Jacobellis v. Ohio, 378 U.S. 184, 197 (1964).

I share Justice Abrahamson’s concern that this court attempt a reasoned formulation of a standard. Whether a conclusively definitive standard can in fact be achieved *241is perhaps doubtful, but we cannot be satisfied with the lack of a formulation that appears to satisfy the majority.

My problem in this case, however, is not solely with the majority’s formulation or nonformulation of a standard. Rather, I am concerned with the methodology utilized in arriving at the ultimate conclusion that counsel is not effective and, particularly, with the lack of careful explication of how the determination that prejudice has occurred is to be applied in arriving at the final result.

The majority implies that there are degrees of effectiveness and that, if counsel is not fully effective, then the question of prejudice is raised. Representation of a criminal defendant is either effective or ineffective. If it is ineffective, reversal is required. There cannot be a determination that ineffective counsel was harmless, because the determination of ineffectiveness impliedly incorporates the finding of an adverse impact resulting from counsel’s deficiency. The deficiency or failure should be evaluated, initially at least, according to the prudent-lawyer standard of State v. Harper, 57 Wis. 2d 543, 205 N.W.2d 1 (1973). If the failure could have had no adverse effect on the defendant, the representation would not have been any more effective had that failure not occurred.

Because the concept of ineffectiveness involves the idea of adverse consequences, it is redundant to consider the question of prejudice after finding that counsel has been found ineffective. It is logically inconsistent to find that ineffective counsel did not harm the defendant.

Unreasonable or even egregiously deficient representation is not necessarily ineffective. Hence, the test used by the majority is incomplete, because it does not apply the test of the likelihood of adverse consequences and does not clearly require a determination of prejudice before labeling conduct as ineffective. The prerequisite *242to finding that counsel is ineffective is (1) inappropriate conduct of an egregious nature, (2) that is likely to result in an adverse result, and (3) in respect to which the state has failed to assume the burden of showing that the conduct was harmless. Only if all of these factors are considered and appropriate conclusions reached can counsel be found to be ineffective.

I recognize that the analysis of this problem presents a morass of philosophical and semantic problems,1 but if, as I believe it must be conceded, a defendant is constitutionally entitled to effective counsel, we should not make a finding of ineffectiveness or unreasonably deficient representation and then explain it away as being nonprejudicial. Rather, the appropriate methodology is to determine whether egregiously deficient conduct of counsel seriously impaired the defense or whether the likelihood that it did was great. Then counsel may be found ineffective and, subject to the state’s rebuttal, a per se rule of reversal is required.

Accordingly, the question of prejudice, applied at the appropriate step in the analysis, is important and necessary. But it is logically incongruent to conclude that ineffective counsel, once found, is not per se prejudicial.

Also, I believe it is clear that, where the defendant demonstrates that the record reveals an egregious deficiency in representation that is likely to adversely affect the result, it is the burden of the state — the party which has reaped the benefit from the error — to show that it is not prejudicial.2 Once it is determined that there is a revealed deficiency in representation likely to cause harm, the burden rests upon the government to prove *243absence of harm and to prove it beyond a reasonable doubt. Chapman v. California, 386 U.S. 18 (1967). Hence, it is the duty of the defendant to show lack of competent representation likely to cause harm; and it is the duty of the state to show, if it wishes to avoid the automatic reversal which ensues upon a finding of ineffectiveness, that no harm occurred.

When a finding of ineffective counsel is made, the inquiry is over. No question of prejudice remains to be decided. That question has, or should have, been addressed in determining whether counsel was effective.

Perhaps the majority does, in fact, follow the methodology I suggest, but I believe that this methodology should be set out so that it is clear that prejudice is an essential prerequisite to a conclusion there has not been effective counsel.

I concur in the result.

(concurring). While I have misgivings and doubts about the majority’s evaluation of the claims of ineffective assistance of counsel in this case and of the prejudicial effect of the state’s references to the defendant’s silence,1 I concur in the result *244the majority reaches and write separately to express my reservations about the majority’s analysis of the ineffective assistance of counsel claim.

The majority opinion is the latest in a series of opinions of this court that depart from the standard we established in State v. Harper, 57 Wis. 2d 543, 557, 205 N.W.2d 1 (1973), to test ineffective assistance of counsel claims. In determining whether the defendant was denied his state or federal constitutional right to effec*245tive assistance of counsel, the majority states that it tests the alleged ineffectiveness of counsel’s representation by using the Harper standard, but phrases the standard by which it actually measures the allegations in terms of whether there was a “basis in reason” for counsel’s actions, whether the “conduct was unreasonable,” and whether the conduct was “contrary to the actions of an ordinarily prudent lawyer.” It is unclear whether the majority has rejected the Harper test, whether the majority states one standard in several ways or whether it states several standards in the alternative or conjunctive. The majority opinion thus further confuses an already complex area of the law.

The use of several standards to evaluate effective assistance of counsel claims is not unique to this opinion or this court.2 In State v. Harper, 57 Wis. 2d 543, 557, 205 N.W.2d 1, 9 (1973), this court abandoned the “sham and mockery of justice” standard and adopted a more stringent standard for reviewing ineffective assistance of counsel claims. The Harper standard states that the defendant’s representation must be equal to that which the ordinarily prudent lawyer, skilled and versed in criminal law, would give to clients.

Some of our opinions since Harper, like this majority opinion, have purported to use the Harper standard (the ordinarily prudent lawyer standard) but in reality have used a “basis in reason” or “any rational basis” standard. The first opinion in which this court moved away from Harper to this new standard was Weatherall v. State, 73 Wis. 2d 22, 242 N.W.2d 220 (1976), cert. den., 429 U.S. 923 (1976), where the court stated the Harper *246standard, and then casually added the following language:

“We are not required to predict the outcome of the case if it had been tried as the postconviction counsel now states he would then have tried it. The sole question before us is whether there was a basis in reason or any rational basis for the trial counsel recommending to his client that the defense of entrapment not be attempted.” 73 Wis. 2d at 28 (emphasis added).

Later cases picked up the Weatherall “basis in reason,” “any rational basis” language, expanded upon it, and applied it. See State v. Rock, 92 Wis. 2d 554, 285 N.W.2d 739 (1979); State v. Nye, 100 Wis. 2d 398, 302 N.W.2d 83 (Ct. App. 1981), summarily aff’d 105 Wis. 2d 63, 312 N.W.2d 826 (1981).

The Harper standard and the Weatherall “basis in reason” standard are not necessarily the same. The cases seem to establish two standards which require distinct modes of analysis and levels of proof.

The Harper standard appears to call for proof and determination of a legal professional’s alleged ineffectiveness in the same manner as proof and determination of any professional’s ineffectiveness. Ineffectiveness may not necessarily be the same as malpractice, but this court in Harper recognized that the standards for professional legal services are analogous to the standards for professional medical and dental services. One would therefore expect to read in the records the testimony of expert witnesses as to the degree of care and skill required of an ordinarily prudent lawyer. I have not read such testimony. Perhaps the court, being skilled in the practice of law, may without testimony determine both the appropriate standard of care and whether the lawyer’s conduct conformed to the standard. The court is its own expert and some would say is taking judicial notice of the appropriate standard of care and the deviation therefrom. *247Is such judicial notice advisable or permissible under ch. 902, Stats. 1979-80?

The ordinarily prudent lawyer standard appears to call for three levels of proof and analysis: first, the parties must present proof of, and the court must make a determination of, what the attorney did or did not do and what was the basis of the conduct giving rise to the claim of ineffectiveness; second, the parties must present proof of, and the court must make a determination of, what an ordinarily prudent lawyer would do in like circumstances ; and third, the court must determine whether the lawyer’s challenged representation met that standard.

The “basis in reason” standard (or the “any rational basis” standard, which I assume is the same standard) appears to call for two levels of proof and analysis: first, the parties must present proof of, and the court must make a determination of, what the attorney did or did not do and what was the basis of the conduct; second, the court must determine whether the lawyer’s conduct has a basis in reason or any rational basis. The lawyer whose representation is being challenged may have an after-the-fact explanation for his or her conduct or the court itself may hypothesize an explanation for the conduct. This standard, unlike the Harper standard, does not appear to call for an objective evaluation of an attorney’s conduct in light of professional standards, but for the court’s subjective evaluation of an attorney’s conduct.3 Using the “basis in reason” standard, then, *248may very well come full circle back to the “sham and mockery of justice” standard we rejected in Harper. 57 Wis. 2d at 551-52. The “sham and mockery of justice” standard, like the “basis in reason” standard, calls for a subjective evaluation of the record rather than an objective evaluation of the challenged conduct in light of professional standards.

Perhaps one reason the court has drifted away from the Harper test back toward the “sham and mockery” test is because the court has been unwilling to give meaning to the Harper standard. While the Harper standard presents a measurable objective standard to evaluate effective assistance of counsel claims, its generality has made it, as Judge Bazelon has said, an “empty vessel into which content must be poured.” Judge Bazelon further observed that “the courts have failed to address these questions [that is, what is customary or reasonable for a lawyer to do] with the result that the new test has made little change.” Bazelon, The Realities of Gideon and Argersinger, 64 Geo. L.J. 811, 820 (1976).

In view of the erosion of the Harper standard and increased litigation raising claims of ineffective counsel, I suggest that we reexamine the very difficult question of how we, as judges and as lawyers, should evaluate effectiveness of counsel claims. The assumption underlying my discussion is, as we recognized in Harper, 57 Wis. 2d at 557, n. 8, that lawyers need guidelines for their conduct and courts need guidelines to evaluate effective representation. Guidelines can educate attorneys, inform the parties about the evidence needed to prove ineffectiveness of counsel, and provide the courts with a framework to assess the attorney’s conduct.

This court has recognized the need for such guidelines, *249for filling “the empty vessel,” and we adopted in Harper several sections of the ABA Standards for the Defense Function as partial guidelines for deciding future cases. 57 Wis. 2d at 557, n. 8. See also State v. Simmons, 57 Wis. 2d 285, 298-99, 203 N.W.2d 887 (1973).

In the case at bar the majority might have considered as guidelines for the conduct of an ordinarily prudent lawyer sec. 3.4 of the ABA’s Standards for the Defense Function dealing with publication contracts4 and sec. 3.7 dealing with attorneys’ advice on a client’s anticipated unlawful conduct.5 We did not adopt these sections in Harper but they might be helpful in assessing the attorney’s representation in this case.

The ABA standards provide a good starting place, but they do not give complete guidance to the legal community to evaluate ineffective counsel claims. The ABA standards are not specific enough, and they do not address many problems that attorneys face every day, as the case at bar illustrates. Therefore the court must develop additional guidelines for the ordinarily prudent lawyer.

*250The Code of Professional Responsibility provides additional guidelines. See Barnes v. Jones, 665 F.2d 427 (2d Cir. 1981), cert. granted- U.S. -, 102 S. Ct 2902 (1982). In the case at bar, several provisions of the Code might be helpful as guidelines for the conduct of an ordinarily prudent lawyer.6

*251Unfortunately the majority opinion never refers to the ABA standards or the Code of Professional Responsibility or any other guidelines in evaluating the attorney’s conduct. Instead the majority’s evaluation is in subjective and conclusory terms.

Furthermore, the majority opinion pays no heed to the determination of the circuit court. Although the majority opinion is silent on the standard an appellate court should apply in reviewing the trial court’s holding that there was or was not a violation of the defendant’s constitutional right to effective counsel, it appears that the majority has not deferred to the circuit court. In contrast, in State v. Rock, 92 Wis. 2d 554, 563, 564, 285 N.W.2d 739 (1979), Justice Callow, writing for the court, concluded that the standard for review was whether the circuit court’s determination was against the great weight and clear preponderance of the evidence. In Rock the court said that “the trial court’s finding and conclusion that [the attorney’s] conduct . . . did not constitute ineffective counsel is not against the great weight and clear preponderance of the evidence” and that “the defendant has not demonstrated that the great weight and clear preponderance of the evidence is contrary to the trial court’s finding.” Is the majority now abandoning this standard of review ?

Determining what an ordinarily prudent lawyer would do and what this lawyer did does not necessarily answer the question of whether the conviction should be reversed. The majority concludes that “if we find that [the attorney’s] conduct was unreasonable and contrary to the actions of an ordinarily prudent lawyer, we must then determine whether such action was prejudicial to Fenel.” Supra, p. 229. See Slappy v. Morris, 649 F.2d 718 (9th Cir. 1981), cert. granted, - U.S. -, 102 S. Ct. 1748, 2005 (1982).

*252Separating the concepts of “ineffectiveness” and “prejudicial” is a difficult and perhaps futile task.7 I share the concerns expressed by Justice Heffernan in his concurring opinion. Professor Tague explains the relationship between the two concepts as follows:

“The defendant cannot meet his burden of establishing ineffective representation simply by showing that his attorney violated one of the ABA guidelines.143 In defining his burden, courts can choose between two approaches. They could make it relatively easy for the defendant to establish a constitutional violation, but refuse to reverse the conviction if the attorney’s failures were harmless error. Or, they could require the defendant to make a significant showing of prejudice caused by his attorney’s failures,144 and reverse automatically, without considering harmless error, if the defendant met that burden. DeCoster II chose the first approach. In Beasley v. United States, 145 the Sixth Circuit chose the second approach in a decision where that circuit also rejected the ‘farce’ test.

“In choosing between DeCoster IPs and Beasley's, approaches, courts must resolve two issues concerning the purpose of post-conviction review. First, should courts try to protect only the individual defendant against an unfair verdict, or should they try to prod all attorneys to provide better representation? Second, is a trial fair if the verdict appears reliable, or is it fair only if the defendant had a full opportunity to present his own case and to dispute the government’s case?

“DeCoster II focuses on the attorney’s conduct, while Beasley concentrates on the reliability of the verdict. . . .”149

*253Courts as well as commentators have struggled to establish a method for deciding cases involving claims of ineffective counsel. The DeCoster case, United States v. DeCoster, 624 F.2d 196 (D.C. Cir. 1979), is required reading, because the DeCoster opinions (DeCoster I, II, and III) set forth several divergent views on the difficult issues in ineffective assistance of counsel claims, such as standards, burden of proof, prejudice, harmless error, and when is reversal warranted.

*254Judge Leventhal’s plurality opinion in DeCoster III sets forth a standard for evaluating effectiveness of counsel, adopts an interest of justice test for reversal, and puts the burden on the defendant to show prejudice. Judge Leventhal summarizes his analysis of the issues as follows:

“In the absence of a governmental impediment to effective assistance of counsel, the court cannot lightly vacate a conviction on the basis of its own appraisal of the performance of defense counsel. The door is open, but only for cases of grievous deficiency and where the court has serious misgivings that justice has not been done.

“So "far as the present case is concerned, ultimately dispositive of the appeal are the strength of the government’s case and the failure of appellant to demonstrate a likelihood of effect on the outcome.”

“Our analytic structure permits reversal in the interest of justice, but without inappropriate rigidity. The claimed deficiency must fall measurably below accepted standards. To be ‘below average’ is not enough, for that is self-evidently the case half the time. The standard of shortfall is necessarily subjective, but it cannot be established merely by showing that counsel’s acts or omissions deviated from a checklist of standards.” 624 F.2d at 214-15.

In contrast, Judge Robinson, in a concurring opinion in DeCoster III, puts the burden on the government to show no prejudice:

“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 [the majority’s view], 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 *255right to effective assistance of counsel of a great deal of the value it was intended to have.” 624 F.2d at 262.

Judge Bazelon in a dissenting opinion in DeCoster III suggests a three-step inquiry:

1. Did counsel violate one of the articulated duties ?

2. Was the violation substantial ?

3. Has the government established that no prejudice resulted ?

Judge Skelly Wright, writing for himself and Judges Bazelon and Robinson, states in a separate opinion in DeCoster III that although the three differ on some points they agree that there are two fundamental disposi-tive principles in ineffective counsel cases:

“(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).” 624 F.2d at 300.

In summary, I am very uneasy that the majority bows in the direction of the Harper standard but uses the “basis in reason” test and simply concludes that there was a rational basis for the lawyer’s conduct.

We must begin to address claims of ineffective representation — claims of denial of state and federal constitutional guarantees — in less simplistic and superficial terms than we have in the past or than the majority does in this case. I know that it is not easy to formulate the issues here and that the answers are even more difficult to find. Even those who have given a great deal of thought to the constitutional right to effective assistance *256of counsel have not been successful in proposing a generally accepted and readily workable model to use in cases raising claims of ineffective assistance of counsel. The model will, of necessity, evolve on a case-by-case basis. The majority opinion in this case does not, however, assist in the development of the law.