Opinion

This defendant challenges the long-standing rule that prohibits voir dire from being conducted as a means to uncover bases for peremptory challenges. (People v. Edwards (1912) 163 Cal. 752, 753 [127 P. 58].) Like a moss-covered oak, the doctrine has seemed sturdy because of its venerable age, but we have only to examine its shallow roots and hollow substance to realize that it is precariously poised, ready to topple at the first challenging blow.

I

In the early morning hours of February 21, 1978, defendant Williams fatally shot a drinking companion, Travis King. Earlier in the evening, the two men had been arrested after a fight on a freeway ramp. Both men appeared intoxicated. After King was released from jail, he went to defendant’s house to retrieve his car. King told defendant’s son he wanted to talk to defendant; he followed the son to the bedroom in which defendant had been sleeping with his one-year-old grandson. When defendant told his son that he was asleep and did not desire to talk to anyone, King pounded on the door, shouted obscenities, and demanded that defendant come out to fight. The door opened, a struggle ensued, and King was shot.

Defendant was charged with murder. The central issue at trial was whether in shooting King he had acted reasonably in defense of himself, his grandson, and his home. The jury acquitted him of murder but found him guilty of voluntary manslaughter. Defendant was placed on three years’ probation on condition that he serve five months in county jail.

At the beginning of voir dire the court inquired of the veniremen as a group whether they would follow the court’s instructions on the law regardless of their personal opinions about what the law is or should be. The prospective jurors responded en masse that they could do so. The court also asked whether they had any quarrel with the propositions that the People must prove defendant’s guilt beyond a reasonable doubt *398and that defendant is presumed innocent. They responded that they did not.

Thereafter, counsel were permitted to question the veniremen individually (Pen. Code, § 1078). Defense counsel attempted to ask one prospective juror whether, if she were instructed to apply a “reasonable man” standard of conduct, she could conceive of a “hypothetical, reasonable and prudent man.” The prosecution’s objection was sustained. The court also sustained an objection to counsel’s request that a juror give him “a brief idea of your feeling about the right of a person to defend himself in his own home.”

After the first day of voir dire, counsel sought permission to ask the veniremen whether they would willingly follow an instruction to the effect that a person has a right to resist an aggressor by using necessary force and has no duty to retreat. The court denied the request. Counsel was permitted, however, to ask members of the panel whether they would follow self-defense instructions even if they disagreed with the law. One venireman candidly replied that he could not “truthfully answer” and did not “know for sure” whether he could follow instructions with which he disagreed. Another, understandably finding it difficult to answer hypothetically without first being informed of the probable content of the instructions, asked for an example. Despite these expressions of uncertainty, the court stood by its original ruling.

II

The sole issue on appeal is whether defense counsel should have been allowed to ask the excluded questions, either because they may have led to challenges for cause and were therefore properly within the scope of the existing voir dire standard, or because they could have assisted counsel in the intelligent exercise of peremptory challenges and should therefore have been allowed despite California precedent to the contrary.

We hold that the existing standard is unnecessarily restrictive and arbitrarily applied, and adopt the rule prevailing in most other jurisdictions that counsel may make reasonable inquiries to assist in the intelligent exercise of peremptory challenges. We further hold that a question about a prospective juror’s willingness to apply a specific doctrine of law if so instructed should be permitted if the doctrine is likely to be applied at trial; this, as will be discussed infra, is subject to reasonable limitations and the discretion of the trial judge.

*399The existing rule in California comes to us from a 1912 case, People v. Edwards, supra, 163 Cal. 752, 753 (see also People v. Plyler (1899) 126 Cal. 379, 381 [58 P. 904]; People v. Britton (1897) 118 Cal. 409, 412 [50 P. 664]; People v. Hamilton (1882) 62 Cal. 377, 382; People v. Trask (1907) 7 Cal.App. 103, 105 [93 P. 891]), in which the court purported to justify its holding with the following observation: “there is an increasing tendency to prolong the proceedings inordinately by allowing counsel on either side to indulge in tedious examinations of jurors apparently with no definite purpose or object in view, but with the hope of eliciting something indicating the advisability of a peremptory challenge, and . .. the supposed privilege of doing this has been greatly abused.” Therefore, in an attempt to streamline the voir dire process, the court foreclosed counsel from asking questions immaterial to challenges for cause.

In so doing, the court created a standard arbitrary and difficult to apply, erratic in achievement of its desired end, and insensitive to the constitutional mandate that the defendant be tried before a fair and impartial jury.

Although the Edwards rule was no doubt intended to substantially restrict the scope of voir dire, and in application has often had that effect, if it were applied with strict logic it would impose no significant limitation on “tedious examination.” A venireman may be excused for cause if it is shown that he harbors actual bias “in reference to the case, or to either of the parties, which will prevent him from acting with entire impartiality” (italics added; Pen. Code, § 1073), or if he possesses one of several characteristics from which bias is implied as a matter of law.1 Because counsel can know if a question will expose bias only by awaiting the answer, and because the judge must rule on the propriety of the question before the answer is given, the only limitation on the admissibility of a particular question under the Edwards rule is the judge’s ability or willingness to conceive of a possible response that would reveal legally cognizable bias. Furthermore, Edwards imposed no limit on the form of the questioning; as long as the questions were material to a challenge for cause, they could be as endlessly repetitive or confusing as counsel cared to make them, subject only to the limits of the court’s patience.

*400Confronted with an unwieldy standard, courts necessarily developed their own limitations. The results have been understandably erratic.

Some decisions have been based on an ad hoc balancing test, comparing the likelihood that a particular question would ultimately lead to a challenge for cause with the probability that it would result in a peremptory challenge. If the latter was the more predictable result, then the question was excluded even though strict adherence to the Edwards standard would allow it. Thus in People v. Estorga (1928) 206 Cal. 81 [273 P. 575], a rape case, the trial court refused to allow questions concerning the veniremen’s marital status, family, or close association with any victim of rape. In affirming, this court held that the questions were “not so much for the purpose of laying the basis of challenges for cause as to pave the way for possible peremptory challenges .. .. ” (Id. at p. 87.)

But because of the broad discretion left in the trial courts to apply Edwards literally and therefore liberally, the Legislature perceived as unabated the practice of allowing virtually unrestrained voir dire so long as the remote possibility existed that bias would be uncovered thereby. (People v. Crowe (1973) 8 Cal.3d 815, 822 [106 Cal.Rptr. 369, 506 P.2d 193].) It therefore responded in 1927 by directing trial judges to allow only “reasonable examination of prospective jurors by counsel .. . .” (Pen. Code, § 1078.) It is clear from its face, however, that the statute recognizes the need and safeguards the opportunity to *401make reasonably thorough inquiry to secure a “fair and impartial jury.” (Id.; see also People v. Estorga, supra, 206 Cal. at p. 84.) Nonetheless, many courts apparently viewed the statute as further justification for restricting counsel’s opportunity to make specific inquiries to uncover latent prejudice.

As a result, both before and after the enactment of the statute, most courts limited voir dire by asking broad questions at the outset about the veniremen’s impartiality. Under Edwards, they could then foreclose as unlikely to reveal bias other questions related to particular areas in which bias was in fact likely to be discovered. (See, e.g., People v. Casserio (1936) 16 Cal.App.2d 223, 227 [60 P.2d 505]; People v. Edenburg (1928) 88 Cal.App. 558, 561-563 [263 P. 857]; People v. Hill (1913) 20 Cal.App. 407, 410-411 [129 P. 475]. Voir dire thereby was often reduced to a “stark little exercise consuming minutes rather than hours and often eliciting no verbal responses at all.” (Babcock, Voir Dire: Preserving “Its Wonderful Power” (1975) 27 Stan.L.Rev. 545, 549.) For example, in Casserio, supra, 16 Cal.App.2d at page 227, the following procedure was approved: “The court asked the panel collectively, if any of them knew of any reason why they could not act fairly and impartially without bias or prejudice in the case. The court then stated that the prosecution must prove the offense charged beyond a reasonable doubt and to a moral certainty, and asked the jury whether or not they were in sympathy with such a rule. The court also said: ‘You all understand I assume, a defendant in every criminal case is presumed to be innocent until his guilt is established beyond a reasonable doubt and to a moral certainty. If there is anyone who does not understand that, please raise your hand.’” Counsel was not permitted to question the jurors further about bias or their willingness to adhere to the enunciated standards.

Viewed in retrospect, these opinions appear to reflect the belief that actual bias means “consciously held bias. ... Thus a juror who swore he would judge the case solely on the evidence presented at trial was taken at his word, no matter what his actual predisposition might have been.” (Note, Voir Dire: Establishing Minimum Standards to Facilitate the Exercise of Peremptory Challenges (1975) 27 Stan.L.Rev. 1493, 1495; see also Note, Selection of Jurors by Voir Dire Examination and Challenge (1949) 58 Yale L. J. 638, 638.) But such a restrictive and wooden approach to voir dire is insupportable in light of the unassailable truth that direct and general inquiries about juror bias cannot be expected to uncover all forms of partiality.

*402Our courts have become increasingly aware that bias often deceives its host by distorting his view not only of the world around him, but also of himself.2 Hence although we must presume that a potential juror is responding in good faith when he asserts broadly that he can judge the case impartially (People v. Preston (1973) 9 Cal.3d 308, 313 [107 Cal.Rptr. 300, 508 P.2d 300]), further interrogation may reveal bias of which he is unaware or which, because of his impaired objectivity, he unreasonably believes he can overcome.3 And although his protestations of impartiality may immunize him from a challenge for cause (see Pen. Code, § 1076), they should not foreclose further reasonable questioning that might expose bias on which prudent counsel would base a peremptory challenge. For instance, although a juror has asserted his willingness to presume defendant’s innocence, “careful counsel would exercise a peremptory challenge if a juror replied that he could accept this proposition of law on an intellectual level but that it troubled him viscerally because folk wisdom teaches that where there is smoke there must be fire.” (United States v. Blount (6th Cir. 1973) 479 F.2d 650, 651.

This court observed in People v. Crowe, supra, 8 Cal.3d at p. 831, fn. 31, that “Although each of the jurors asserted that he knew of no reason why he could not serve as an impartial juror, such assertions do not eliminate the necessity for additional reasonable inquiry to uncover con*403cealed or subtle bias.... [II] ... a biased juror may be unwilling to confess that bias openly, and . . . questions to which there is a ‘right’ and a ‘wrong’ answer may be less likely to reveal such bias than more open-ended questions.”4 To foreclose such an open-ended question simply because it does not directly inquire about prejudice sufficient to sustain a challenge for cause is unnecessarily and unrealistically to limit counsel’s opportunity to secure an impartial jury.

Moreover, little psychological insight is needed to realize that the setting in which voir dire is conducted creates additional pressures for the venireman to answer questions as he believes the judge would have him answer, or in conformity with the answers of the preceding panelists.5 In Irvin v. Dowd (1961) 366 U.S. 717, 728 [6 L.Ed.2d 751, 759, 81 S.Ct. 1639], the Supreme Court observed, “No doubt each juror was sincere when he said that he would be fair and impartial ... but the psychological impact requiring such a declaration before one’s fellows is often its father.”

Despite increased recognition of the need to search for subtle manifestations of bias, and of the limits of questions bluntly and broadly directed at overt prejudice to aid in that search, our courts have generally continued to cling dogmatically to the artificial limitation of the Edwards rule. (See, e.g., People v. Rigney (1961) 55 Cal.2d 236, 244 *404[10 Cal.Rptr. 625, 359 P.2d 23, 98 A.L.R.2d 186]; Rousseau v. West Coast House Movers (1967) 256 Cal.App.2d 878, 882 [64 Cal.Rptr. 655].) But it can no longer be argued seriously that voir dire is properly restricted to a search for legally cognizable bias only.

It has always been understood that public support for a system of justice depends not only on the capacity of that system to render justice in fact, but also on its ability to project to the parties and public an unimpeachable image of fairness.6 The peremptory challenge has long acted as a fixative for preserving that image. Its purpose has always been to insure that criminal trials are conducted before jurors who not only proclaim their impartiality, but whose ability to be evenhanded is not seriously questioned by the parties. In the words of Blackstone, “As everyone must be sensible .. . how necessary it is that a prisoner (when put to defend his life) should have a good opinion of the jury, the want of which might totally disconcert him, the law wills not that he should be tried by any one man against whom he has conceived a prejudice, even without being able to assign a reason for such dislike.” (4 Commentaries *353.)

Although it is still ordinarily true that no reason must be given for the exercise of peremptory challenges (but see People v. Wheeler (1978) 22 Cal.3d 258, 281-282 [148 Cal.Rptr. 890, 583 P.2d 748]), it is also true they are of little value to the accused or the state if counsel is forced to utilize them unaided by relevant information. Such intelligence may or may not be uncovered by questions directed at challenges for cause.7

In fact, largely because they have been deprived of a reasonable opportunity to gather helpful information, counsel have often resorted of necessity to means of jury selection that are reprehensible and in some cases of dubious constitutionality. Thus in People v. Wheeler, supra, 22 *405Cal.3d 258, 281, we held that jurors are not to be peremptorily challenged on the basis of group bias alone. But unless counsel is given a significant opportunity to probe under the surface to determine the potential juror’s individual attitudes, he may be relegated to a Catch-22 alternative of making his decision on the superficial basis we held impermissible in Wheeler, or making it on no basis at all.

Faced with that choice, affluent litigants have in recent years resorted to costly pretrial surveys of the population from which the jury is drawn. From these surveys, a profile is constructed of the most or least desirable jurors based solely on general characteristics such as race, sex, age, religion, and level of education.8 Not only does this practice tend to perpetuate the problem we addressed in Wheeler, it also creates the untoward image of a system in which the opportunity to rationally exercise peremptory challenges is governed by the size of the litigant’s bank account or defense fund. (See Miller, Justice Ain't Cheap—A Defense Counsel's View of the Joan Little Case (1978) 7 U. San Fernando Val.L.Rev. 59, 63-66; Note, supra, 49 So.Cal.L.Rev. 597.)

Because the peremptory challenge is a critical safeguard of the right to a fair trial before an impartial jury (see Swain v. Alabama (1965) 380 U.S. 202, 219-221 [13 L.Ed.2d 759, 771-773, 85 S.Ct. 824]; Pointer v. United States (1894) 151 U.S. 396, 408 [38 L.Ed. 208, 213-214, 14 S.Ct. 410]), questions directed at its intelligent exercise manifestly fall within the bounds of the “reasonable inquiry” to which counsel are entitled. (Pen. Code, § 1078.)

The federal courts agree, acknowledging that “if [the constitutional right to an impartial jury] is not to be an empty one, the defendants must, upon request, be permitted sufficient inquiry into the background and attitudes of the jurors to enable them to exercise intelligently their peremptory challenges.” (United States v. Dellinger (7th Cir. 1972) 472 F.2d 340, 368; see also Nebraska Press Assn. v. Stuart (1976) 427 U.S. 539, 602 [49 L.Ed.2d 683, 722-723, 96 S.Ct. 2791] (conc. opn. of Brennan, J.); Swain v. Alabama, supra, 380 U.S. 202, 218-219 [13 L.Ed.2d 759, 771]; United States v. Robinson (D.C.Cir. 1973) 475 F.2d 376, *406381; United States v. Lewin (7th Cir. 1972) 467 F.2d 1132, 1138.)9 As the United States Supreme Court recently observed, “lack of adequate voir dire impairs the defendant’s right to exercise peremptory challenges . ...” (Rosales-Lopez v. United States (1981) 451 U.S. 182, [68 L.Ed.2d 22, 28, 101 S.Ct. 1629].) The high court in Rosales-Lopez reaffirmed Swain v. Alabama, supra, 380 U.S. at pp. 218-219 [13 L.Ed.2d at p. 759], by recognizing the connection between voir dire and the exercise of peremptory challenges: “The voir dire in American trials tends to be extensive and probing, operating as a predicate for the exercise of peremptories . . . . ”

The same rule is applied in most state courts. (47 Am.Jur.2d, Jury, § 201, p. 791; see, e.g., Gossett v. Commonwealth (Ky. 1968) 426 S.W.2d 485, 487; Emanus v. State (Tex.Crim.App. 1975) 526 S.W.2d 806, 808; Lamb v. State (1978) 241 Ga. 10 [243 S.E.2d 59, 61]; People v. Harrell (1976) 398 Mich. 384 [247 N.W.2d 829, 830]; State v. Anderson (1972) 30 Ohio St.2d 66 [59 Ohio Ops.2d 85, 282 N.E.2d 568, 571].) The last-cited case makes the following trenchant observation: “‘The right to examine prospective jurors on their voir dire is granted to litigants in order to enable them to select a jury composed of men and women qualified and competent to judge and determine, without bias, prejudice or partiality, facts in issue. For the proper exercise of this right, the Legislature has deemed it wise to give to litigants the right of peremptory challenge and challenge for cause. This former right is to be exercised at their discretion and free from any limitation or restriction. Any rule of law which denies a litigant reasonable latitude in the examination of prospective jurors as to their qualifications, in order to *407enable him to exercise such peremptory challenges judiciously and intelligently, deprives him of a substantial right.’”

Even in our own state, the compelling wisdom of the majority rule, together with the difficulty encountered in applying Edwards, have led some courts to suggest that the latter decision is not worthy of continued application. In People v. Terry (1964) 61 Cal.2d 137, 141 [37 Cal.Rptr. 605, 390 P.2d 381], for instance, Justice Tobriner observed that “Neither the defendant nor the prosecution should suffer an improper restriction upon a reasonable voir dire examination of prospective jurors or a frustration of an intelligent exercise of peremptory challenges or challenges for cause.” And two Courts of Appeal defiantly upheld the propriety of certain questions they found relevant only to the intelligent exercise of peremptory challenges. (People v. Boorman (1956) 142 Cal.App.2d 85, 89-91 [297 P.2d 741]; Walker v. Greenberger (1944) 63 Cal.App.2d 457, 463-464.) Instead of rebuking those courts, we acknowledged in People v. Crowe, supra, 8 Cal.3d at pp. 821-822, fn. 2, that they created a legitimate conflict in our law.

Moreover, currently our courts often allow a broad range of questions even when purporting to apply Edwards, thus implicitly recognizing the legitimacy of voir dire directed at the intelligent exercise of peremptory challenges.10 The time has come to make that recognition explicit in our law.11 We therefore hold that counsel should be allowed to ask questions reasonably designed to assist in the intelligent exercise of peremptory challenges whether or not such questions are also likely to uncover grounds sufficient to sustain a challenge for cause.

*408In so doing, however, we leave intact the considerable discretion of the trial court to contain voir dire within reasonable limits. (See, e.g., Ham v. South Carolina (1973) 409 U.S. 524, 528 [35 L.Ed.2d 46, 51, 93 S.Ct. 848]; People v. Riordan (1926) 79 Cal.App. 488, 496 [250 P. 190].) Under this standard, trial courts need not and should not permit the inordinately extensive and unfocused questioning that prompted the adoption of the Edwards rule. Nonetheless, “expedition should not be pursued at the cost of the quality of justice.” (United States v. Blount, supra, 479 F.2d 650, 652.) Although it is impossible to dictate a priori the proper balance to be struck by the trial courts in each case, we commend to them the guidance provided by the District of Columbia Circuit, i.e., that counsel should at least be allowed to inquire into “matters concerning which either the local community or the population at large is commonly known to harbor strong feelings that may stop short of presumptive bias in law yet significantly skew deliberations in fact.” (United States v. Robinson, supra, 475 F.2d 376, 381.) In addition, if a particular juror has given the court some reason to suspect he harbors such feelings, even though the general population does not, further questioning would be appropriate.

We reaffirm that it is not “a function of the examination of prospective jurors to educate the jury panel to the particular facts of the case, to compel the jurors to commit themselves to vote a particular way, to prejudice the jury for or against a particular party, to argue the case, to indoctrinate the jury, or to instruct the jury in matters of law.” (Rousseau v. West Coast House Movers, supra, 256 Cal.App.2d 878, 882.)12 Therefore, a question may be excluded if it appears to be intended solely to accomplish such improper purpose. In addition, the court need allow only reasonable questions—although it cannot exclude questions proper in scope, it is free to require that they be phrased in neutral, nonargumentative form. On the other hand, a question fairly phrased and legitimately directed at obtaining knowledge for the intelligent exercise of peremptory challenges may not be excluded merely because of its additional tendency to indoctrinate or educate the jury. The Legislature has recently reaffirmed its commitment to a voir dire process in which attorneys are allowed reasonable participation. (Pen. *409Code, § 1078, as amended by Stats. 1974, ch. 960, § 1, p. 2000.) In such a setting, the potential for anticipatory argument, even under the existing rule, “is an unavoidable consequence of the voir dire jury examination.” (United States v. Blount, supra, 479 F.2d 650, 652.)13

Ill

Under the Edwards rule, the cases are in conflict as to the admissibility of questions such as those defense counsel here attempted to ask. While certain cases hold or suggest that counsel may inquire whether jurors are willing to apply particular doctrines of law (People v. Love (1960) 53 Cal.2d 843, 852, fn. 1 [3 Cal.Rptr. 665, 350 P.2d 705]; People v. Tuthill (1947) 31 Cal.2d 92, 98 [187 P.2d 16]; People v. Creeks (1915) 170 Cal. 368, 371-372 [149 P. 821]; Kramm v. Stockton Electric R. R. Co. (1913) 22 Cal.App. 737, 746-747 [136 P. 523]), others hold that only questions about the juror’s willingness to apply general doctrines, applicable in all cases, are permitted. (People v. Bennett (1926) 79 Cal.App. 76, 90-91 [249 P. 20].) Still others declare that even questions concerning general propositions of law are inapposite once the juror’s willingness to apply the law as instructed is ascertained. (People v. Soltero (1978) 81 Cal.App.3d 423, 428 [146 Cal.Rptr. 457]; People v. Orchard (1971) 17 Cal.App.3d 568, 576 [95 Cal.Rptr. 66]; People v. Edenburg (1928) 88 Cal.App. 558, 562 [263 P. 857]; People v. Harrington (1955) 138 Cal.App.2d Supp. 902, 905 [291 P.2d 584].)

Instead of undertaking to resolve the conflict apparently caused in part by the failure of Edwards to provide pragmatic guidance, we prefer to remove that cause and decide the issue under the standard enunciated above.

The Attorney General contends that even under the foregoing standard, the questions proposed in this case were properly excluded because of the independent rule that jurors are not to be tested on their *410knowledge of the law. (People v. Love, supra, 53 Cal.2d at p. 852.) It appears from the face of the questions however that their answers require no preexisting knowledge of legal doctrines, but only an expression of the venireman’s attitude toward those doctrines, the relevant content of which is contained in the questions.

The Attorney General further contends that questions testing the veniremen’s willingness to apply particular legal doctrines are unnecessary after the court has ascertained their willingness to apply the law, whatever its content, as instructed. He claims that allowing such questions would constitute an admission that jurors falsify in order to be empanelled, and would therefore be contrary to the presumption of veracity to which their answers are entitled. (See People v. Preston, supra, 9 Cal.3d 308, 313.) But in addition to the problem of subtle or unconscious bias that makes a general proclamation of fairmindedness untrustworthy, here the ordinary venireman has a very limited fund of knowledge from which to evaluate the broadly phrased question. His answer may be true to the extent that he is willing generally to act as the judge instructs him. But it is untenable to conclude that the veniremen’s general declaration of willingness to obey the judge is tantamount to an oath that he would not hesitate to apply any conceivable instruction, no matter how repugnant to him.14 Hence the answer is merely a predictable promise that cannot be expected to reveal some substantial overtly held bias against particular doctrines. We therefore hold that in general a reasonable question about the potential juror’s willingness to apply a particular doctrine of law should be permitted when from the nature of the case the judge is satisfied that the doctrine is lijcely to be relevant at trial. Reversal will be required, however, only if the doctrine is actually relevant, and the excluded question is found substantially likely to expose strong attitudes antithetical to defendant’s cause.

Of the questions to which the prosecution objected at voir dire in this case, all were arguably relevant and could properly have been admitted, but only one was erroneously and prejudicially excluded. The judge acted within his discretion in excluding inquiry concerning the prospective juror’s ability to conceive of a “hypothetical, reasonable and *411prudent man.” The question bears little potential for exposing prejudice or resistance to application of the law. Nor did the judge abuse his discretion in excluding the question as to a particular venireman’s opinion of the right of a person to defend himself in his own home. The panelist to whom this question was put had expressed no reservations about application of the law of self-defense in response to prior questions. And of all the aspects of self-defense law, the principle allowing a person to defend himself in his own home is perhaps the least likely to meet with serious opposition. Counsel therefore had little if any reason to suspect that the venireman would quarrel with this generally accepted tenet.

On the other hand, there is a real possibility the average juror might disagree with the controversial rule that a person may use force in self-defense even though an avenue of retreat is open. As most attorneys and judges are aware, the law regarding the duty to retreat varies substantially throughout the country, reflecting a wide range of attitudes on the subject. (40 Am.Jur.2d, Homicide, § 162, p. 449.) Moreover, even if the judge was not initially inclined to allow the question regarding the existing doctrine in California, it should have become evident in the course of voir dire that the query might reveal important reservations about that rule. As noted above, two prospective jurors confessed doubt about their willingness to apply self-defense principles with which they disagreed. Refusing counsel’s request to further explore these panelists’ attitudes erroneously foreclosed questioning that could potentially expose either bias or strong feelings short of presumptive bias. And although neither of these panelists ultimately served on the jury, counsel should have been allowed to put the same interrogatory to those who did.

The court was apparently aware when it denied counsel’s motion that the outcome of the case was likely to turn on whether the jury found that defendant killed in justifiable self-defense. A juror who firmly believed that all passive means of self-defense should be utilized before employing deadly force would find it difficult to judge defendant’s conduct objectively under our law. A statement of general willingness to apply unspecified self-defense principles is not a sufficient guarantee of impartial application of a rule so important to the case and so likely to encounter opposition. As noted above, “Careful counsel would exercise a peremptory challenge if a juror replied that he could accept [a] proposition of law on an intellectual level but that it troubled him viscerally ....” (United States v. Blount, supra, 479 F.2d 650 at p. 651.)

*412Because the inquiry foreclosed by the court bore a substantial likelihood of uncovering jury bias, the court abused its discretion by refusing to allow the proposed question. Although in many contexts a procedure depriving defendant of the right to secure an impartial jury necessarily dictates reversal (see, e.g., People v. Wheeler, supra, 22 Cal.3d 258, 283; People v. Carmichael, supra, 198 Cal. 534, 547), that standard should not apply if the potential for bias relates only to a particular doctrine of law. When antipathy to a legal rule is the issue, its potential effect is limited by the significance of the rule to the case’s outcome. If in light of the evidence no reasonable, impartial juror would apply the rule, or if the judge determines the rule to be irrelevant as a matter of law, or if the verdict is consistent with an application of the rule favorable to defendant, any bias would obviously be harmless. Here, however, the jury was instructed to excuse defendant’s failure to retreat if it found certain facts, for which there was substantial evidentiary support. Moreover, the jury’s finding of guilt was inconsistent with application to defendant of the retreat doctrine, Therefore, the error in excluding questions regarding that doctrine was prejudicial.15

The judgment is reversed.

Tobriner, J., Newman, J., Reynoso, J.,* and Grodin, J.,* concurred.

Voir dire should not be limited solely to questions involving challenges for cause. The former limitation on voir dire was án unwarranted restriction on an accused’s right to a jury composed of impartial individuals. Further, it was unworkable in practice. Trial courts throughout the state differed widely in their application of the rule, some using it as a justification for sharply curtailing voir dire, others reading it broadly to allow extensive questioning of potential jurors.

However, if this court is to scrap the old rule, we should not replace it with a new standard which is more difficult to apply than its predecessor. That is exactly what the majority have done. They state that *413voir dire must be allowed “at least” into “‘matters concerning which either the local community or the population at large is commonly known to harbor strong feelings that may stop short of presumptive bias in law yet significantly skew deliberations in fact.’” (Majority opn., ante, at p. 408, quoting United States v. Robinson (D.C.Cir. 1973) 475 F.2d 376, 381.) Exactly what does that mean?

I submit that this new guideline will be no easier to apply in a uniform manner in practice than the old. Trial courts will continue to differ widely in their evaluation of what issues are sufficiently controversial to justify voir dire questioning.

The majority’s own examples are an excellent illustration of the problem. Under their formulation, a question about a prospective juror’s attitude toward the right to self-defense in the home is excluded, because this right is unlikely “to meet with serious opposition,” and is a “generally accepted tenet” with which few are likely to quarrel. (Majority opn., ante, at p. 411.)

On the other hand, the right of a person to use force in self-defense despite the existence of an avenue of retreat is found to be a “controversial rule,” and thus one about which the defendant must be allowed to question the jury panel. (Majority opn., ante, at p. 411.)

The distinction drawn between these two lines of questioning is ephemeral at best. The right to use deadly force against an intruder, even in the home, is certainly controversial in our society. Some people would probably find it even more controversial than the no-retreat rule. The majority’s ability to draw a neat line between these two questions is more a comment on the justices’ personal views of the two issues than a reflection of an objective difference in community attitudes.

I fear this new standard will lead to unpredictable results because judges will not agree on what issues are sufficiently controversial or emotional to meet the requirements. To implement this rule, a trial judge would have to be able to predict accurately the existence of possible bias on a given issue as it cuts across all social strata in a community. I doubt that the judiciary is capable of making such decisions and this court should not ask it to do so.

Further, there is no reason to exclude questioning about an issue such as self-defense in the home, even though it might arguably be consid*414ered somewhat less controversial than another issue. Appellant’s right to defend himself in his home was central to his defense. Each juror’s attitude toward this issue was crucial to a fair adjudication of appellant’s guilt.

Voir dire should not be limited to “commonly known” prejudices. The chief purpose of voir dire is to uncover any biases harbored by the particular group of people who will serve on a jury. Why should a person be subjected to the risk that one of his jurors harbors doubts about a crucial legal principle merely because most community members might not harbor such doubts?

A limitation on voir dire is necessary to prevent endless voir dire questioning. However, the standard set down to achieve such a limitation should not rest on a trial judge’s ability to evaluate community attitudes toward the legal and factual issues that each case presents. Rather, inquiry should be permitted into all significant legal and factual aspects of a case.1 The trial court’s discretion to control voir dire would be better served under a standard similar to Evidence Code section 352.2 In controlling voir dire in this manner, trial judges would be making decisions they are well-trained to make, weighing the relevance of questions against their potential to create confusion or waste time.

Although this standard would leave a good deal of discretion with the trial courts, it will direct attention to the real issues that should be raised in voir dire questioning. Attorneys must be allowed great latitude in voir dire in order to ensure a fair and impartial jury. The courts may legitimately limit voir dire to prevent the needless waste of time or confusion. However, a trial judge’s gut level feelings about the possibility of bias in the community are not a proper standard for determining whether or not a particular question is proper voir dire.

I have no quarrel with the efforts of my colleagues of the majority as they seek to articulate a broad rule which would permit counsel “to ask questions reasonably designed to assist in the intelligent exercise of peremptory challenges ...” (ante, *415p. 407), and to preserve considerable discretion of “the trial court to contain voir dire within reasonable limits.” (Ibid.) Similarly, I agree with my colleague of the concurrence who points to the lack of specificity of the proposed new standard and the broad generality of the analysis and test advanced in extended fashion by the majority. I take no issue with my colleague’s efforts in this direction because it is my belief and observation that this is exactly what trial courts in the area of voir dire examination have been doing for years.

I respectfully dissent, however, from the majority’s conclusion that, assuming the trial court erred in unduly restricting the voir dire questioning, such error was so prejudicial as to warrant a reversal of the judgment.

The majority attempts to justify its reversal of the judgment by relying upon cases wherein the defendant has been deprived of an impartial jury. (E.g., People v. Wheeler (1978) 22 Cal.3d 258, 283 [148 Cal.Rptr. 890, 583 P.2d 748] [exclusion of black citizens from jury venire]; People v. Carmichael (1926) 198 Cal. 534, 547 [246 P. 62] [restriction on voir dire questioning to establish bias].) In the present case, any infringement upon defendant’s rights is much less severe—restricting exploratory voir dire questioning which might disclose some basis for a peremptory challenge.

Article VI, section 13, of the California Constitution provides in pertinent part that “No judgment shall be set aside, ... in any cause, .. . for any error as to any matter of procedure, unless, after an examination of the entire cause, including the evidence, the court shall be of the opinion that the error complained of has resulted in a miscarriage of justice.” (Italics added.)

Twenty-five years ago we considered the foregoing constitutional mandate in People v. Watson (1956) 46 Cal.2d 818, 836 [299 P.2d 243], and held that “a ‘miscarriage of justice’ should be declared only when the court, ‘after an examination of the entire cause, including the evidence,’ is of the ‘opinion’ that it is reasonably probable that a result more favorable to the appealing party would have been reached in the absence of the error.” (Italics added.) The Watson standard has been applied in hundreds of subsequent decisions and is acknowledged as representing the present test for measuring the harmlessness of trial errors under the state constitutional standard. (People v. Jackson (1980) 28 Cal.3d 264, 307 [168 Cal.Rptr. 603, 618 P.2d 149]; see Wit-*416kin, Cal. Criminal Procedure (1963) Reversible Error, § 741, at p. 712 et seq.)

In the present case, as the majority acknowledges, “At the beginning of voir dire the court inquired of the veniremen as a group whether they would follow the court’s instructions on the law regardless of their personal opinions about what the law is or should be. The prospective jurors responded en masse that they could do so.” (Ante, p. 397 italics added.) In addition, defense counsel was permitted to ask panel members whether they would follow self-defense instructions even if they disagreed with the law. (Ante, p. 398.) The only restriction placed on voir dire examination in this regard concerned questions aimed at determining the prospective jurors’ specific views regarding the use of force in self-defense where an avenue of retreat was available.

Under the foregoing circumstances, how can we hold that a “miscarriage of justice” (art. VI, § 13) occurred in this case, or that “it is reasonably probable” (Watson, supra, 46 Cal.2d 818) that a more favorable verdict would have been rendered but for the voir dire restrictions? In my view the majority errs in so readily assuming the probability of prejudice in this case.

Our continued and studied neglect of the constitutional mandate of article VI, section 13, has not gone unnoticed. (MacLeod, The California Constitution and the California Supreme Court in Conflict Over the Harmless Error Rule (1981) 32 Hastings L.J. 687.) This thoughtful analysis correctly observes that “as interpreted in Watson, article VI, section 13 allows an appellate court to reverse a verdict only if it first finds that the result probably would have been different had the error not occurred; ... The court in Watson emphasized that reasonable probability, not mere possibility, of a different result is required to overturn a trial court verdict.” (Pp. 689-690, italics in original, fns. omitted.) The principal thesis of the Hastings commentator is that we have “effectively abrogated the state constitutional rule on harmless error” (p. 687) by, among other things, use of an “ipse dixit approach, in which the court, without support from its own analysis of the record, attempts to satisfy the Watson test by transmuting its own determination that the error possibly was prejudicial into a holding that it probably was prejudicial; ...” (P. 691, italics in original, fns. omitted.) In my view, there is much merit in the article’s conclusion that our misapplication of the constitutional mandate results in reversals of *417judgments “despite the harmlessness of the error, shaking public confidence in the system of justice.” (P. 705.)

Because my reading of the record in this case does not persuade me that it is reasonably probable that the trial court’s restriction upon defense counsel’s voir dire examination affected the jury’s verdict or resulted in a miscarriage of justice, I would affirm the judgment of conviction.

Respondent’s petition for a rehearing was denied July 3, 1981. Richardson, J., was of the opinion that the petition should be granted.