concurring:
I agree with the majority that the defendant was denied his constitutional right to counsel at a critical stage of the proceedings. I also agree that this error entitles the defendant to a new trial. I write separately, however, because I do not agree that harmless error analysis is appropriate in reviewing this error. Thus, I specially concur in parts III and IV of the majority opinion.
In Arizona v. Fulminante, 499 U.S. 279, 306-14, 111 S.Ct. 1246, 1263-66, 113 L.Ed.2d 302 (1991), the Supreme Court held that the erroneous admission of a coerced confession could be reviewed under a harmless error analysis. The Court reasoned that the admission of an involuntary confession is a “ ‘trial error,’ similar in both degree and kind to the erroneous admission of other types of evidence.” Id. at 310, 111 S.Ct. at 1265. Thus, an appellate court may “simply review[ ] the remainder of the evidence against the defendant to determine whether the admission of the confession was harmless beyond a reasonable doubt.” Id. The Court cited a number of eases where harmless error analysis was applied to errors of constitutional magnitude and found that
[t]he common thread connecting these cases is that each involved ‘trial error’— error which occurred during the presentation of the case to the jury, and which may therefore be quantitatively assessed in the context of other evidence presented in order to determine whether its admission was harmless beyond a reasonable doubt.
Id. at 308-09, 111 S.Ct. at 1264. The Court distinguished these “trial errors” from “structural defects,” which are not subject to harmless error analysis. Structural defects, it held, are “defects in the constitution of the trial mechanism;” they are errors “affecting the framework within which the trial proceeds, rather than simply an error in the trial process itself.” Id. at 310, 111 S.Ct. at 1265; see also Wayne R. LaFave & Jerold H. Israel, Criminal Procedure § 27.6, at 1170 (2d ed. 1992) (“[T]he characteristic of ‘structural’ defects most frequently mentioned in Supreme Court opinions is their ‘inherently indeterminate’ impact upon the outcome of the trial.”).
Unlike Fulminante and the cases cited therein for support (the “Fulminante cases,”) the error in the instant case is not a trial error susceptible to quantitative analysis. In the Fulminante cases, the reviewing courts were provided with the opportunity to examine the record to determine the impact of the error upon the jury in light of other evidence or other instructions also considered by the jury. Here, there is no information in the record to allow an objective analysis of the effect of the district court’s ex parte communication with the jury. It is possible, as noted by the majority, see maj. op. at 825, that the judge’s ex parte communication to the jury pressured the jury to reach a verdict by the end of the first afternoon of deliberations.1 Two jurors in particular had significant reason to wish to resolve the case by the end of the day: Juror Hurst was to be married on January 1, the day following the date on which the judge suggested the jury should reconvene and indicated that she hoped the court would not reconvene on that day; and Juror Nichting was leaving town and was not scheduled to return until the date which the judge suggested they reconvene.2 By not recalling counsel to *829be present while he spoke to the jury about potential dates to reconvene, the trial judge prohibited counsel from protecting the defendant’s right to be tried by a fair and impartial jury. Had counsel been present at the time, he could have taken a number of precautions to ensure that the jury was not coerced into rendering a verdict.3 Because counsel was not present, and these precautions were not taken, it is impossible to quantify the degree of prejudice here.
The majority implies that there is a connection between the duration of the asserted error and whether it is to be classified as a trial error or a structural error:
In this case, the defendant was deprived of his right to counsel during only one stage of the proceedings — a scheduling conference with jurors during their deliberations. This does not amount to a “total deprivation” of his right to counsel, nor can such error be classified as a “structural defect affecting the framework in which the trial proceed[ed].” Apart from the short, impromptu scheduling conference, counsel for the defendant was present for the entire proceeding. Unlike Gideon, where the defendant was erroneously denied any representation of counsel throughout the proceedings, the error in this case is more appropriately viewed as a “trial error,” the impact of which can be quantitatively assessed on appellate review. Therefore, we agree with the court of appeals that the harmless error doctrine should be applied in this case.
Maj. op. at 826-27 (citation omitted).4 I disagree that the duration of the error is determinative of whether it is a structural error affecting the framework of the entire trial. The events which unfold at any one of the critical stages of a trial may effect an entire proceeding. Whether a defendant is denied the right to counsel at one critical stage or at all critical stages is irrelevant; in both cases the dispositive fact remains whether or not the impact of the error can be quantified.5
In summary, I agree with the majority that the defendant was denied his constitu*830tional right to counsel at a critical stage of the proceedings. It is my belief, however, that the state of the record, viewed in light of the nature of the violation, prohibits a quantitative assessment of the impact of this deprivation. I would therefore hold that the error must be considered a structural defect and is not susceptible to harmless error analysis.
I am authorized to say that Justice KIRSHBAUM joins in this special concurrence.. The record is silent as to how quickly the jurors reached a 'verdict after the judge spoke to them about the possibility of reconvening on New Year’s Eve. Some guidance is provided by the judge’s comment that the jurors had deliberated for "about three hours” prior to the ex parte communication and the judge's indication that the jurors could not stay past 5:30 that afternoon to deliberate.
. Even prior to trial, the judge was informed that several of the jurors were concerned about the length of the trial because it was scheduled for the week before Christmas. Also prior to trial, the judge was informed by Juror Hurst that she *829was getting married on January 1st and would be on her honeymoon after that date.
.Among the options asserted by the petitioner as available to counsel had he been present to preserve the defendant's right to a fair trial are the following:
First, counsel could have requested a mistrial on the grounds that the jury was unable to reach a unanimous decision. Second, counsel could have requested a mistrial on the grounds that there was no acceptable time to reconvene within the next few weeks. Third, counsel could have objected to the court's comments and lack of proper instructions which had the tendency to coerce the jury into reaching a verdict. Fourth, counsel could have requested that the jury be given the option of returning on Saturday, December 22, 1990, rather than on New Year's Eve. Fifth, counsel could have objected when the court told the jury that it could continue to deliberate that evening, but could not stay past 5:30. Sixth counsel could have objected to the court’s request that the jury reconvene on December 31, 1990, which was completely unacceptable to the jury.... Seventh, counsel could have asked the court to inquire whether or not the jurors were dissatisfied with the December 31st date. Eighth, counsel would have observed the jurors reactions to the court's suggestions that they reconvene on December 31st, and could have put such reaction on the record. Ninth, counsel could have requested the court give the jury a supplemental instruction directing them not to surrender their honest convictions as to the weight or effect of the evidence because the opinion of their fellow jurors or for the mere purpose of returning a verdict.
. Under the majority’s “total deprivation” rationale, I am left to wonder whether harmless error analysis would be utilized even if a defense counsel was absent during voir dire, opening argument, and direct and cross-examination, but was present for closing argument.
. I would not hold that in every instance where a defendant has been denied the right to counsel at a critical stage of the proceedings, harmless error analysis is inappropriate. In some instances, the error may be measured quantitatively by the evidence which was admitted or not admitted due to counsel's absence. See, e.g., Satterwhite v. Texas, 486 U.S. 249, 257, 108 S.Ct. 1792, 1798, 100 L.Ed.2d 284 (1988) (finding harmless error where "the effect of the Sixth Amendment [right to counsel] violation is limited to the admission into evidence of Dr. Grigson’s testimony.”); Moore v. Illinois, 434 U.S. 220, 232, 98 S.Ct. 458, 466, 54 L.Ed.2d 424 (1977) (remand for determination of whether admission of pretrial identification evidence obtained in violation of defendant’s right to counsel was harmless error). Additionally, like the majority, see maj. op. at 825, I would not hold that every communication between a judge and jury constitutes a critical stage of the trial.