with whom Judges WARDLAW, PAEZ and BEA join, concurring:
Judge Gould spends much effort proving something that no one disputes: That a criminal defendant who represents himself may use standby counsel to take certain actions on his behalf. Judge Gould’s worry that we are “hostile to the ability of a pro se defendant to delegate trial tasks to standby counsel,” Gould Concurrence at 752, is entirely unfounded. Cases where a pro se defendant voluntarily delegates tasks that he could perform himself remain unaffected by today’s opinion, which turns on a very different question: Whether petitioner was allowed to appear at the bench conference to speak for himself — as he was entitled to do under Faretta v. California, 422 U.S. 806, 95 S.Ct. 2525, 45 L.Ed.2d 562 (1975), and McKaskle v. Wiggins, 465 U.S. 168, 104 S.Ct. 944, 79 L.Ed.2d 122 (1984)— or whether he was forced to communicate only through standby counsel?
On this record, this is hardly a trivial issue. Consideration of a jury note in the midst of deliberations is a critical stage of the proceedings where both sides are entitled to express their views. See McKaskle, 465 U.S. at 178, 104 S.Ct. 944. When the jury sent its note, petitioner was not in court but was being held in custody elsewhere. Frantz’s trial judge thus had a responsibility to have Frantz brought before him for the bench conference. Frantz, after all, was counsel for the defense. The prosecutor was invited to appear and did so; defense counsel — in this case Frantz — had an equal right to be there. But Frantz could not get there on his own: A prisoner in lockup has no choice about whether to come into court. If the judge orders him to appear, he is slapped into manacles and brought in, like it or not; if the judge does not order him to appear, the prisoner remains in his cell, no matter how badly he may want to go before the judge. Frantz’s absence suggests that the trial judge probably denied petitioner his right to self-representation by failing to have him brought from his cell to participate in the bench conference.
It is conceivable that the judge here sent Frantz a message inviting him to come if he wished, and Frantz chose to remain in his cell and speak through Lamb instead. Judges seldom send prisoners engraved invitations with stamped RSVP cards, but maybe this judge was uncommonly polite and considerate. If Frantz was given the option of attending and chose not to, then there is no Faretta or McKaskle issue because Frantz would have given up his right to appear pro se at the bench conference. This is at least a theoretical possibility, and the district court on remand can easily resolve it by taking evidence on the narrow question of whether Frantz’s absence was the result of his voluntary decision to remain in his cell and communicate through his standby counsel, or was forced on him by the judge’s wrongful failure to order Frantz brought before him.
And this is why the question that the concurrence finds dispositive — whether Frantz authorized Lamb to speak for him at the conference, Gould Concurrence at 751-52, 752 — is actually beside the point. It doesn’t matter whether Frantz authorized Lamb to appear for him, if Frantz had no choice in the matter. If Frantz was not allowed to attend the conference and speak for himself, what else could he do except speak through Lamb? Frantz’s decision to do so, rather than refuse to provide any defense input whatsoever, can hardly be *747construed as waiving his right under Far-etta and McKaskle to appear and speak for himself. If Frantz was not allowed to appear, then he was denied his rights under Faretta, and the fact that he authorized Lamb to speak for him cannot cure that error^ — even if Lamb accurately conveyed Frantz’s instructions. See Bittaker v. Enomoto, 587 F.2d 400, 402-03 (9th Cir.1978) (Faretta error is not subject to harmless error doctrine).
Which brings me to the concurring opinion’s curious silence on the issue that the majority finds dispositive, namely whether Frantz voluntarily gave up his right to appear at the bench conference. Since the majority sees the issue as pivotal, one would think the concurrence would at least mention it, if only to dispute its significance. But the concurrence says nothing about it, leaving a complete mystery as to what the author of the concurrence, and the judges joining him, think about this issue.
The concurrence seems to see some virtue in this, suggesting that it is somehow premature to address such issues before the record is further developed. Gould Concurrence at 752-53. I find this very strange, because the concurrence is not the least bit coy about expressing views on any number of legal issues that turn on facts not yet found: “If Frantz told standby counsel how to handle the bench conference concerning the jury’s question about reading the 911 tape, then there is simply no Faretta violation whatsoever because a self-representing defendant may properly use standby counsel to assist in defense. Similarly, if Frantz directed standby counsel concerning the conference, then there is no McKaskle issue because McKaskle merely sets guidelines for when standby counsel may act without the assent of the self-representing defendant.” Gould Concurrence at 746-53 (emphasis added). And again: “In my view, so long as Frantz adequately instructed Lamb concerning the tape, there is no violation.” Id. at 752 (emphasis added). And yet again: “There is no McKaskle error if Lamb had received Frantz’s authority to speak at the conference.” Id. (emphasis added). And still yet again: “If Lamb was authorized by Frantz to attend the conference and state his position on the 911 tape, then there is no issue under Faretta or McKaskle.” Id. at 753 (emphasis added).
Of course, there’s nothing wrong with this; appellate opinions usually discuss the facts and the law applicable to the case. The quaint notion that we may not properly discuss the law until the district court finds the facts is contradicted not only by common sense and common experience but by the concurrence’s own exuberant willingness to express legal conclusions based on hypothetical facts. It also raises some difficult questions, such as: How is the district court to know what facts it needs to find, and how are the parties to know what evidence to present, unless we explain to them the applicable legal theory?
If we are dissatisfied with the district court’s fact-finding the first time around, we must tell the judge and the parties what facts matter and, thus, what they are to do on remand. To this end, we must adopt the legal framework pursuant to which further proceedings are to be conducted. Or, to come at it from the other direction: If we ask the district court to find facts without first determining that those facts are of legal significance, aren’t we asking the district judge and the parties to engage in a futile exercise? Should we be free, as the concurrence seems to believe, to say to the district court and the parties after remand, “sorry to have bothered you, but all those facts you found at our direction turn out to be of no legal significance after all”?
*748No responsible appellate court would say anything like that. If we wish the district court and the parties to conduct another trial, that’s our prerogative, but we must tell them what facts matter and why. And to do that, we must develop a legal theory that differentiates the legally significant facts from the trivial ones. There is nothing premature or theoretical about this; it’s what appellate courts do all the time. If our concurring colleagues disagree with the majority’s key legal ruling, that it is dispositive whether Frantz was allowed to attend the bench conference in the flesh, they should say so forthrightly and explain why. This magisterial gesture of refusing to express a view as to selected legal issues, unless and until the parties and the district court find facts that may turn out to be irrelevant, is not judicial restraint; it is disregard for the time and effort of the other players in the system.
Contrary to the suggestion in the concurrence, Judge Berzon’s opinion expresses no hypothetical views, nor is its discussion in the least premature. Rather, the opinion correctly applies the law with a commendable economy of analysis. I am pleased to join it.