I dissent.
Admittedly at least two confessions were secured from defendant under circumstances violative of the rules set forth in Escobedo v. Illinois, 378 U.S. 478 [84 S.Ct. 1758, 12 L.Ed.2d 977], and People v. Dorado, 62 Cal.2d 338 [42 Cal.Rptr. 169, 398 P.2d 361]. It is well settled by both California and federal law that the erroneous admission of a confession is prejudicial per se and requires a reversal. This problem was fully, carefully and correctly considered in People v. Dorado, supra, which decision until now has been consistently followed.
In Dorado, supra, at pages 356-357, we stated:
“Finally, we cannot dispose of the introduction of the illegally obtained confession upon the ground that it constituted merely harmless error. Although under some circumstances the introduction into evidence of statements obtained from a defendant during police interrogation in violation of his right to counsel and his right to remain silent may constitute harmless error, we are convinced that the error is necessarily prejudicial when the statements are confessions. Indeed, Es*334cobedo itself follows Hamilton v. Alabama (1961) 368 U.S. 52 [82 S.Ct. 157, 7 L.Ed.2d 114], and White v. Maryland (1963) 373 U.S. 59 [83 S.Ct. 1050, 10 L.Ed.2d 193], eases which do ‘not rest ... on a showing of prejudice.’ (White v. Maryland, supra, at p. 60.)
“The use of an involuntary confession results in a denial of due process and requires reversal ‘ regardless of other evidence of guilt. ’ (People v. Matteson (1964) 61 Cal.2d 466, 469-470 [39 Cal.Rptr. 1, 393 P.2d 161] ; accord: People v. Brommel (1961) 56 Cal.2d 629, 634 [15 Cal.Rptr. 909, 364 P.2d 845]; People v. Trout (1960) 54 Cal.2d 576, 585 [6 Cal.Rptr. 759, 354 P.2d 231, 80 A.L.R.2d 1418] ; People v. Berve (1958) 51 Cal.2d 286, 290 [332 P.2d 97] ; Lynumn v. Illinois (1963) 372 U.S. 528, 537 [83 S.Ct. 917, 9 L.Ed.2d 922] ; Rogers v. Richmond (1961) 365 U.S. 534, 540-541 [81 S.Ct. 735, 5 L.Ed.2d 760]; Payne v. Arkansas (1958) 356 U.S. 560, 567-568 [78 S.Ct. 844, 2 L.Ed.2d 975].)
“The improper introduction of the confession which has been obtained in violation of the constitutional right to counsel transgresses the protection of due process no less than the illegal introduction of a confession which has been coerced. In either case courts cannot inquire into the prejudicial nature of the introduction of an illegally obtained confession for the reasons stated in People v. Parham (1963) 60 Cal.2d 378, 385 [33 Cal.Rptr. 497, 384 P.2d 1001] ; ‘Almost invariably ... a confession will constitute persuasive evidence of guilt, and it is therefore usually extremely difficult to determine what part it played in securing the conviction. [Citation omitted. ] These considerations justify treating involuntary confessions as a class by themselves and refusing to inquire whether in rare cases their admission in evidence had no bearing on the results. ’ Thus in the instant case the improper introduction of the confession obtained in violation of a constitutional right compels reversal. ’ ’
The United States Supreme Court cases cited above dealt with coerced confessions. As to them, the cases have established that a showing of prejudice is not required and that if improperly admitted a reversal is required regardless of the overwhelming evidence of guilt. In Dorado (supra, 62 Cal.2d 338, 356) as the above quote demonstrates, the coerced confession rule is applied to confessions admitted in violation of the constitutional right to counsel and the right to remain silent. Now what the majority have done is to create an exception to the rule of reversible per se that in fact emasculates that rule. They state that in multiple confession cases *335article VI, section 4%, of the California Constitution is applicable, and that the error is not reversible unless prejudicial. This is simply to adopt the views of the dissenting justices in the Dorado ease, supra. If we are going to reverse this basic holding of Dorado we should do so fairly and frankly, and after a full discussion of the problem. We should not do so by creating exceptions to a rule that has no exceptions.
If the problem of multiple confessions were an open one there might be some merit in the position taken by the dissenters in Dorado, supra, and by the majority in the instant ease. But the question is not an open one. The United States Supreme Court has spoken, and its determination as to the effect of a violation of a federal constitutional right is binding on us. The United States Supreme Court has not only held generally that there are no exceptions to the rule that erroneous admission of a confession is reversible per se, but on at least two occasions has held specifically that the reversible error per se is applicable in multiple confession eases. The majority concede that Stroble v. California, 343 U.S. 181 [72 S.Ct. 599, 96 L.Ed. 872], held that if there was one coerced confession introduced into evidence, it required reversal in spite of the fact there were several “good” confessions. The “multiple confession” exception created by the majority in the instant case was directly repudiated in Stroble. It held that this court had been in error in holding that the admission of a coerced confession did not result in prejudicial error. It would appear that the majority in the instant case are attempting to reinstate the very rule that was repudiated in Stroble. The majority seek to distinguish Stroble on the ground that what was said in that case was dicta, and therefore not binding on us. It may be that what was said in Stroble was dicta, but it was dicta based on a prior decision—Malinski v. New York, 324 U.S. 401 [65 S.Ct. 781, 89 L.Ed. 1029]—a case cited and relied on in Stroble. The factual situation in Malinski is very similar to that in the instant one.
In Malinski, a conviction of murder had been based upon testimony of the defendant’s multiple confessions. Two of the confessions were made to relatives and a third was made to a close friend. The latter confession contained a complete account of both the planning and the execution of the crime. In addition, the police had secured several confessions from the defendant and all but the first of these confessions were *336introduced into evidence. The defendant contended that the first confession obtained by the police, the one which had not been introduced into evidence, had been obtained by coercion. The United States Supreme Court agreed. The court then went on to hold that although the first confession had not been introduced into evidence, its existence had been alluded to by both the prosecution and the judge. Solely by virtue of the references to the confession the court concluded:
“It is thus apparent that the judgment before us rests in part on a confession obtained as a result of coercion. Accordingly a majority of the Court do not come to the question whether the subsequent confessions were free from the infirmities of the first one. ” (Id., 324 U.S. at p. 410.)
Thus, in Malinski, the court held that mere reference to a nonintroduced coerced confession by the prosecutor and trial court was sufficient to require a reversal of the cause no matter how many “good” confessions there were. The court in Malinski did not think it necessary to consider what part the confession played in the trial, and what was its evidentiary impact. It was sufficient for the United States Supreme Court that a coerced confession had been brought to the jury’s attention. As already noted, this holding was the basis of Stroble and that case must be viewed as stating a similar rule. Thus, by force of these eases we are compelled to reverse all judgments where a confession, obtained in violation of the defendant’s constitutional rights, is brought before the jury. This should be the end of this case.
But even if the majority were correct on this point, which they are not, they have fallen into another and very basic error. There were multiple confessions in this case. Some were clearly investigatory, and two were clearly accusatory and violative of Escobedo. The other confessions, at least as to several of them, occurred under circumstances that, dependent on the facts, they could be either accusatory or investigatory. But the court never passed on this issue. Because this case was tried before Escobedo neither the defense nor the prosecution was alerted to the importance of this issue. Thus the record is most incomplete. The majority, based on the incomplete record before us, determined as a matter of law that these confessions were secured under circumstances not violative of Escobedo. Thus, this court places itself in the place of the trial judge. I had always thought it to be the law that the defendant had a right, a constitutional right, to a decision by the fact finder of all factual issues in the case. *337But now we find that in situations that are debatable the majority of this court have held that this court can determine whether or not a confession has been secured in violation of the rules announced in Escobedo v. Illinois, supra. This is in direct violation of Jackson v. Denno, 378 U.S. 368 [84 S.Ct. 1774, 12 L.Ed.2d 908, 1 A.L.R.3d 1205], held to be applicable to the Escobedo-Dorado problem by People v. Schader, 62 Cal.2d 716, 727 [44 Cal.Rptr. 193, 401 P.2d 665].
As to those “confessions” overheard by and testified to by third persons we can assume, as a matter of law, that these were not secured in violation of the rule announced in Escobedo v. Illinois, supra, 378 U.S. 478, and People v. Dorado, supra, 62 Cal.2d 338. But as to the remaining confessions the problem is a very close one. The majority conclude that two of these, as a matter of law, were introduced improperly, and I agree, but then the majority rule that the other confessions, as a matter of law, were properly introduced. This conclusion is palpably unsound.
As already pointed out, the majority, in order to rationalize their conclusion, found it necessary to hold that some of the confessions were properly introduced while others should have been excluded. The important confessions were those that the majority found, as a matter of law, had been made during the investigatory phase of the proceedings. It was only because the majority were able to find these to be “good” confessions that they were able, by specious reasoning, to affirm the judgment in spite of the “bad” confessions. This is an abuse of the appellate function. The defendant has a right to have factual issues determined on a proper record and by a proper factfinding body and to have a proper review of the legal issues involved in or raised by those findings.
In Jackson v. Denno, supra, 378 U.S. 368, the United States Supreme Court commented as follows upon the respective roles of the appellate and trial courts in confession cases: “Where pure factual considerations are an important ingredient, which is true in the usual case, appellate review in this Court is, as a practical matter, an inadequate substitute for a full and reliable determination of the voluntariness issue in the trial court and the trial court’s determination, pro tanto, takes on an increasing finality. The procedures used in the trial court to arrive at its conclusions on the coercion issue progressively take on added significance as the actual measure of the protection afforded a defendant under *338the Due Process Clause of the Fourteenth Amendment against the use of involuntary confessions. These procedures must, therefore, be fully adequate to insure a reliable and clear-cut determination of the voluntariness of the confession, including the resolution of disputed facts upon which the voluntariness issue may depend.” (Id. at pp. 390-391.) That reasoning is applicable to the instant case. There is no legal difference between the fact of whether a confession was coerced and the fact of whether it was elicited by a process of interrogation. Here the trial court employed no procedures to insure that confessions that might run afoul of the standards of Escobedo and Dorado were eliminated. The trial court did not employ any such procedure because the case was tried before those decisions were rendered. Lacking trial court determinations on these issues, the majority have decided to make their own independent determinations of fact. Having assumed a factfinding function, the majority were then forced to rely upon a one-sided incomplete record, a record which was ill prepared to answer the questions involved. The majority were thereby forced to conclude on this issue, solely in accordance with the testimony adduced by the prosecution. Defendant’s counsel did not cross-examine on this issue, because he did not know it was involved. This was tantamount to directing a verdict, on this issue, against the defendant, and this, of course, cannot be done. (People v. Crowley, 101 Cal.App.2d 71, 75 [224 P.2d 748] ; People v. Conboy, 15 Cal. App. 97 [113 P. 703], Cf. United Brotherhood of Carpenters & Joiners v. United States, 330 U.S. 395, 408 [67 S.Ct. 775, 91 L.Ed. 673].)
Because this case, like several which have preceded it and some that will follow it, was tried prior to the Escobedo and Dorado decisions the records before this court were and are generally inadequate to answer the questions posed by the newly formulated rule. The defense could not have been expected to perfect a record on appeal upon a rule not yet formulated, and the manifest unfairness in so requiring has been previously recognized by this court when we held that no objection based upon the Escobedo-Dorado rule need be made to the introduction of confessions at the trials held before these decisions. (People v. Hillery, 62 Cal.2d 692, 711 [44 Cal.Rptr. 30, 401 P.2d 382].) This case indicates that it would be unfair to expect a defendant to perfect a record on an issue then unknown to him. Here the majority have abandoned the policy of Hillery. Here the majority have deprived the defendant of the opportunity to cross-examine the prose*339cution’s witnesses on this issue, to impeach these witnesses as to this issue or to adduce further evidence on this issue. Of course, no instructions were given on this issue.
The proposed holding that an appellate court cannot decide against the defendant on the Escobedo-Dorado issues in doubtful cases is not a departure from our previous practice. Whenever we have held that the accusatory stage was reached and that the defendant had not been advised of his rights and therefore reversed the judgment, the cause has been reversed and remanded to the trial court. Presumably these cases will be tried anew and presumably at this new trial the prosecution could introduce additional evidence to establish that the defendant’s Escobedo-Dorado rights were not violated. The defense will have an opportunity to meet and develop the issue. Those eases in which we have affirmed the judgment even though there was a violation of Escobedo by the introduction of statements, we admitted the error, but held it not prejudicial. That is, under our miscarriage of justice rule there was no need for a retrial. But in the instant case, the majority hold as a matter of law there was no error. This was necessary because, were the confessions made to the police held to be nonadmissible, a reversal would occur. This distinguishes this case from those already decided.
The transcript demonstrates the inadequacy of the evidence on this issue. Considering only the confessions made to the police officers, the record is hopelessly ambiguous. As to several of these confessions it cannot be ascertained with any certainty whether they were at the accusatory or investigatory stage. This is hardly surprising inasmuch as the testimony upon which the majority rely was only introduced for the purpose of demonstrating that the confessions were not coerced. No one knew or thought that the stage of the proceedings might be important.
Considering the confessions in the order in which they were made, the first full confession was given to Officer LeBlane. The majority approve of the introduction of this confession on two theories. First, they assert that the confession was not given in response to a process of interrogation and second, it was properly admitted under the rule announced in People v. Modesto, 62 Cal.2d 436, 446 [42 Cal.Rptr. 417, 398 P.2d 753].
Whether or not the confession was elicited by a process of interrogation requires an appraisal of all of the relevant circumstances. (People v. Stewart, 62 Cal.2d 571, 579 [43 Cal.*340Rptr. 201, 400 P.2d 97].) The majority opinion discloses that several questions had been previously asked of the defendant prior to his making the confession. Whether the confession was given in response to these questions and whether these questions amounted to a process of interrogation are queries that must be answered by the fact finder. Certainly it is not so clear that the accusatory stage had not been reached that we should be willing to hold that it would be unreasonable to believe that a fact finder could so conclude or that the defendant, if he knew the issue were involved, could not have introduced relevant evidence on the issue.
The majority are also incorrect in their discussion of the applicability of the Modesto rule in connection with the interrogation by Officer LeBlanc. The majority hold that there was a paramount interest in possibly saving the child’s life, but a reading of the transcript reveals no such intent on the part of the officer and no such emergency as was presented in Modesto. In that case the police had good reason to believe that the continued interrogation of the suspect might save the girl’s life. There is no such showing in the present case.1
The last confession found by the majority to be admissible was given during the ride back to the station. This is a most flagrant example of an attempt by the majority to make a finding of fact on an incomplete transcript. The only possible clue as to whether there was a “process” of interrogation (all of the other requirements for the application of the Escobedo-Dorado rule were clearly met), was the officer’s testimony that he had a “conversation” with the defendant on the way back to the station. Upon this the majority predicate their conclusion that there was no “process” but rather that the defendant’s statement was volunteered by him. Precisely what the word “conversation” may mean is difficult to ascertain, but it is not without the realm of possibility that it means something other than that the statement was volunteered.
*341The only complete confessions of first degree murder were made to, and related by, police officers. In all such situations as disclosed by the incomplete record, the fact finder might have ruled against the prosecution on the issue of the admissibility, had it been submitted to it. Certainly, if the parties had known the issue were involved, more evidence would and could have been introduced on it. It cannot be said that it is inconceivable that the taking of additional testimony would add nothing. On the contrary, from the state of the record before us it appears most desirable to have additional evidence to aid in the application of the Escobedo-Dorado standard to the confessions.
In this connection, we should not overlook the ruling of Townsend v. Sain, 372 U.S. 293 [83 S.Ct. 745, 9 L.Ed.2d 770]. In that case the United States Supreme Court was faced with the question of under what circumstances a federal district court upon a petition for writ of habeas corpus was required to hold a hearing and take additional evidence. The court concluded with the following test: “Where the facts are in dispute, the federal court in habeas corpus must hold an evidentiary hearing if the habeas applicant did not receive a full and fair evidentiary hearing in a state court, either at the time of the trial or in a collateral proceeding. In other words, a federal evidentiary hearing is required unless the state-court trier of fact has after a full hearing reliably found the relevant facts. ” (Id. at pp. 312-313.)
While it is true we are not compelled by the rule of Townsend to modify our state procedures to conform to the federal standard, “We recognize, however, the desirability of the maximum integration of such process in order to avoid the possibility of wasteful duplication.” (In re Shipp, 62 Cal.2d 547, 554 [43 Cal.Rptr. 3, 399 P.2d 571].)
There is still another error in the majority opinion that must be mentioned. Even were we to concede that the improperly admitted confessions were not prejudicial on the guilt trial (a concession I cannot make) their admission on the penalty trial would be demonstrably prejudicial error. The confessions admitted by the majority to have been secured in violation of Escobedo-Dorado were testified to, on the guilt trial, by the officers involved. On the penalty trial these confessions were proved by tape recordings, and the prosecuting attorney was permitted to argue not only the content of such confessions but that the tone of voice in which they had been delivered demonstrated a lack of re*342morse and an abandoned heart.2 To say that this error was not “substantial" within the meaning of People v. Hines, 61 Cal.2d 164, 169 [37 Cal.Rptr. 622, 390 P.2d 398], is to close one’s eyes to the realities of the situation. To permit the prosecutor to play the tapes in the presence of the jury and to permit him to argue that by the very tone of voice of the defendant he had demonstrated that defendant was a coldblooded killer without remorse and should, therefore, suffer the death penalty, could not help but sway the jury away from any possibility of life imprisonment. It must be remembered that the jury determines life or death in its discretion, and certainly one of the things that it can and should consider in making this critical determination is the character of the defendant and whether or not he is a coldblooded killer lacking any remorse. Particularly is this true under the federal standard of reversible error, adopted by the majority in the instant case. As stated by the majority the test is whether there is a “reasonable possibility that the errors complained of might have contributed to the conviction.” (Italics added.) That is, of course, the test adopted by the United States Supreme Court in Fahy v. Connecticut, 375 U.S. 85, 91 [84 S.Ct. 229, 11 L.Ed.2d 171], and is the proper test. Tested by that standard, or by the standard set forth in People v. Watson, 46 Cal.2d 818, 836 [299 P.2d 243], the errors here involved were incontrovertibly prejudicial, at least on the penalty trial. The fact that there is some other evidence to the effect that defendant acted without remorse is immaterial. That evidence, from the lips of third persons, is trifling when compared with the impact of hearing the defendant’s own voice demonstrating that he was without remorse. The impact of the tape recordings cannot be calculated.
I would reverse the judgment in its entirety.
Appellant’s petition for a rehearing was denied October 13, 1965. Peters, J., was of the opinion that the petition should be granted.
Perbaps the majority are holding that so long as the victim has not been found there is, as a matter of law, a paramount interest in attempting to save the victim’s life. There is no basis for such a broad reading of the Modesto holding. The paramount interest ought to be, in an appropriate ease, one of the factual determinations made as part of the foundation to the introduction of a confession. There appears to be no good reason to formulate a rule that, as a matter of law, any time the police have not yet found the victim’s body the police may interrogate the suspect, and certainly not in a case where they had been told the victim was dead.
Among other things the prosecutor argued to the jury that:
"I talked a few minutes ago about the lack of remorse. You don’t need to go any further than just the testimony or just the hearing of the tape recording that you heard this morning as to his attitude. Even were the story that he told on this witness stand true—but you, of course, found it isn’t—even assuming it true, that he just walked in there and discovered the child, no one but a cruel and callous man or a man incapable of belief could have acted and conducted himself as the tape shows he did.”