with whom Circuit Judges T.G. NELSON and CLIFTON join, concurring in the result:
I concur in the result, that the conviction should be affirmed. If there were error, it would be, as the majority concludes, harmless. I dissent from the majority’s view that the psychotherapist-patient privilege applies even to a patient’s imminent, seriously intended, and properly disclosed threat to commit murder.
It is important to make clear what this case is not. The majority decision does not preserve confidentiality in therapeutic relationships. This is not a case about a patient who says to his psychotherapist “I have homicidal thoughts and feelings, and although I am not going to act on them, they disturb me and I need your help to get rid of them.” This case involves a patient who says “I am going to kill FBI agents,” has engaged in preparation and has the means to do so, and is understood by the psychotherapist to be voicing a serious intention to act imminently. The patient was understood by his psychotherapist to be past the point of seeking help that would prevent criminal action, so she felt it essential to warn his prospective victims and did so. His therapeutic confidentiality was gone.
No doubt many patients’ disclosures to psychotherapists would sound alarming were they repeated in court, and yet would not be “true threats.” But here, appellant does not put at issue whether he made a “true threat,”1 both in what he said to his psychotherapist, and what he said to the clinic operator. One might perhaps argue that statements made to a psychotherapist or her staff in the course of obtaining therapy cannot constitute a threat to federal officials under the statute of conviction,2 but that argument has not been *994made in this case, so we do not speak to it. Once it is assumed, as in this case, that a true threat to kill FBI agents made to a third party constitutes the crime, it follows that the psychotherapist observed the patient committing a crime in her office, just as she would have if she had seen the patient steal her receptionist’s purse on the way out. As a percipient witness to a felony, she ought to be required to testify to what she perceived.
The case is controlled by federal common law. Psychotherapists’ ethical obligations regarding patients’ rights to privacy are not the same thing as federal evidentiary privilege. Psychotherapists’ duties of confidentiality, and tort duties to potential victims of patients are also distinct from federal evidentiary privilege. These are matters of professional ethics and state law. Both have some relevance, because they educate us on what is generally expected of psychotherapists in therapeutic relationships with dangerous patients, but neither is controlling. As the Sixth Circuit has held, the connection between a state law duty to inform a prospective victim of a threat has only a “marginal connection, if any at all” to the federal evidentiary question of the scope of the psychotherapist patient testimonial privilege.3 The controlling Federal Rule of Evidence, Rule 501, provides that except as the Constitution, federal statutes, or federal rules may otherwise provide, and except for certain civil cases, “the privilege of a witness ... shall be governed by the principles of the common law as they may be interpreted by courts of the United States in the light of reason and experience.”4 Thus the extensive state authorities relied upon by the majority as a basis for decision, are, as a matter of law, prohibited to us as a controlling basis for decision. Of course we look to state authorities, both statutes and judicial decisions, for guidance insofar as they shine the “light of reason and experience” on the question,5 but we must be “governed,” Rule 501 says, by federal judicial decisions on the common law of privilege.
There is already a circuit split on whether there is a dangerous patient exception to the federal psychotherapist-patient privilege, as the majority recognizes.6 The Tenth Circuit in United States v. Glass7 held that a psychotherapist may testify about a patient’s threat if it was serious when made and disclosure was the only means of averting harm. The Sixth Circuit in United States v. Hayes8 held that there is no dangerous patient exception. The First Circuit in In re Grand Jury Proceedings (Gregory P. Violette)9 held that the crime or fraud exception to the attorney-client privilege applies to the psychotherapist-patient privilege.
*995The Supreme Court has spoken expressly to the issue in this case, saying that the privilege does not apply in cases such as the one before us. The only reason we have any room to opine to the contrary, as the majority does, is that the Court spoke in dicta. The holding in Jaffee v. Redmond10 is that there is indeed a psychotherapist privilege under Rule 501. The dictum critical to this case is a footnote, in which the Court says “we have no doubt that there are situations in which the privilege must give way, for example, if a serious threat of harm to the patient or to others can be averted only by means of a disclosure by the therapist.”11
The case before us is precisely the one described in the Jaffee footnote. The Supreme Court has said in the plainest English that in cases such as the one before us, “the privilege must give way.” We ordinarily treat Supreme Court dicta with “due deference”12 even though they are not binding. Because we are to interpret those decisions “in the light of reason and experience,” the Supreme Court’s dictum should speak even more persuasively than usual, since, dictum or not, what the Court says reflects its “reason and experience.”
The words “the privilege must give way” do not mean that “the right to out-of-court confidentiality must give way,” or that “the right to confidentiality is superseded by the duty of out-of-court disclosure to the prospective victim.” They mean what they say, that what must “give way” is the “privilege.” The “privilege” is the privilege not to testify in federal court. There is only one way to read the plain English of the Jaffee footnote, and that is that the privilege of a psychotherapist to refuse to testify in federal court, or her patient’s privilege to bar her testimony, does not exist “if a serious threat of harm to the patient or to others can be averted only by means of a disclosure by the therapist.”13 That is, when the serious threat occurred that could be averted only by disclosure, the privilege died.
The majority attempts to resuscitate the privilege not to testify. The privilege breathed its last when the “serious threat of harm” could be and was “averted only by means of a disclosure by the therapist.” 14 Artificial respiration of a privilege dead under Supreme Court language that speaks directly to it takes a miracle.15 I do not think the majority brings it off. It says “[w]e read that footnote as endorsing — albeit elliptically — a duty to disclose threats to the intended victim and the authorities16 The majority’s adverb, “elliptically” means “[mjarked by obscurity of style or expression.”17 “The privilege must give way” is not obscure. Any lawyer knows what the word “privilege” means. The context of the Jaffee footnote was the patient’s privilege to bar her psychotherapist’s testimony in federal court under Federal Rule of Evidence 501. The majority makes the surprising supposition that “the Jaffee footnote was intended to extend [state requirements of disclosure of threats] to psychotherapist-patient relationships to which federal law applies (such *996as treatment by federally employed psychologists at overseas army hospitals).”18 There is nothing in Jaffee to support that remarkable speculation.
The majority reads the words “the privilege must give way” to mean that the privilege does not give way, and what does give way is the psychotherapist’s duty to keep the patient’s communication secret from the prospective victim. But that is not what the Court said. And it makes no sense to say, as the majority apparently does, that the Court was speaking to the issue of whether a psychotherapist may disclose a serious threat to the prospective victim, rather than whether the psychotherapist may testify in court about it. Jaffee raised only the issue of testimonial privilege, not the issue of disclosure. The issue of testimonial privilege, to which the Court expressly spoke, is a federal common law issue under Federal Rule of Evidence 501. The issue of duty to disclose to the prospective victim is a state tort law issue, not a matter of federal law, and not within the scope of Jaffee. There is just no getting around the proposition that Jaf-fee said, and meant, that the psychotherapist-patient “privilege must give way,” referring to the privilege under Rule 501 to refuse to testify.19
But I concede that the Court’s remark does not bind us, because it is dictum (though it has the look of a footnote added to avert a risk that someone in the majority perceived were the opinion published without it). So let us turn to “reason and experience,” as Rule 501 commands. As the majority concedes, the states are divided with California recognizing a dangerous patient exception to the privilege and others not. The majority rejects any such exception on policy grounds, arguing that once the psychotherapist warns the prospective victim, there is little social value in obtaining the psychotherapist’s testimony and much value in excluding it to preserve a therapeutic relationship.
This argument is not compelling. In the case at bar, the psychotherapist told the patient, when he first voiced threats, that she would have a duty to disclose threats he made. Subsequently, she did disclose them to the FBI as she told him she would. This patient plainly did not need an assurance of confidentiality to speak honestly to his psychotherapist about what he felt and what he planned to do about his feelings. Nor did he have one. The psychotherapist had told him his threatening communications would be disclosed. In my view (as in the view of the dissenting judge in the Sixth Circuit ease in similar circumstances)20 this case could be simply and properly resolved on the ground that by communicating after the psychotherapist had told him she would not keep the communications secret, appellant waived the privilege. Communicating to a psychotherapist on express terms that the communication will be disclosed is an unprivileged medical communication, much like communication by a plaintiff to a defense physician under Federal Rule of Civil Procedure 35.
The majority opinion notes the distinction between disclosure to the prospective victim and testimony in court, but draws the wrong inference from it. The majority concludes that the privilege remains in force even though disclosure is made,21 but that does not make much sense. Once disclosure is made, the patient has lost the medical benefit of being able to speak to his psychotherapist in confidence that *997what he says will remain secret. His communications are disclosed, and he knows it. His hated enemy, whom he plans to kill, is now privy to his communication to his psychotherapist. Once the person the deranged individual hates so much that he plans to kill him knows his secrets, and the deranged individual knows that his psychotherapist refuses to keep his secrets from that person, there is not much therapeutic value in refusing later to tell this already-disclosed information to the judge and jury. After all, the deranged person does not hate them and his confidentiality is long gone. The majority is evidently concerned about deranged murderous individuals stopping valuable therapy because the psychotherapist reveals their confidences. But where that will occur, it will doubtless already have occurred where the psychotherapist betrayed their confidences to their worst enemies.
Psychiatry and psychotherapy are often of great value, and rely for their value on open and candid communication by the patient. I have no doubt that testimony by psychotherapists against their patients, or for that matter disclosure to prospective victims by psychotherapists, harms the therapeutic relationship. But the therapeutic relationship is not the only social value at stake. The prospective victims’ lives are at stake.
Sometimes a warning may suffice to protect the victim, sometimes not. FBI agents, the prospective victims in this case, carry guns and know how to use them, so perhaps they (but not their spouses and children) could protect themselves if they know who to look out for. But most people cannot. What, exactly, is one to do if a psychotherapist calls up and says “I have a deranged patient who plans to kill you, and he’s serious?” Call the police? They do not provide bodyguard services. Seek state civil commitment proceedings, as the majority opinion suggests? How shall the threatened individual assemble the money for lawyers and experts and persuade the involved bureaucracies and individuals to act fast enough to prevent realization of the threat? The fastest way to get someone locked up who threatens to kill a federal official in violation of the statute of conviction may well be a federal criminal proceeding in which the psychotherapist testifies about what the patient says.
Protecting federal officials from assassination is only part of the purpose of the law. The statute criminalizing threats against federal officials22 is not merely prophylactic, to prevent the harms threatened. It prohibits the threats themselves. Federal officials, high and low, are supposed to be able to do their jobs, not only without being killed, but also without facing death threats. The threats themselves inhibit the efficient functioning of government. Suggesting, as the majority opinion does, that threatened officials try to get the threateners civilly committed does nothing to alleviate the interference with government functioning caused by the threats.
Beyond these practical concerns, there is another concern, altogether ignored by the majority opinion. That is the concern with having the truth vindicated and justice done. The tradition of the common law, as Wigmore teaches, is that privileges are “distinctly exceptional, being so many derogations from a positive general rule” that “the public ... has a right to every man’s evidence.”23 The reason that the states found it necessary to provide by statute for a physician-patient privilege is that Lord Mansfield, in the Duchess of *998Kingston’s Trial (a House of Lords trial for bigamy) had rejected it as a matter of common law24 In the 1600’s “the obligations of honor among gentleman” was often sufficient ground for refusing to testify, but in the same “notorious Duchess of Kingston’s Case” in 1776, “the older point of view was definitely abandoned and the new one thoroughly promulgated.”25 The common law preference is for the truth coming out, rather than privilege.
The question for the jury in this case was whether Steven Gene Chase did or did not make true threats to kill FBI agents. As in any trial, little can be more important than that the verdict be true. That is why we call it a “verdict,” derived from Latin roots for “true speech.” Although in this case the verdict would plainly be true even were the trier of fact deprived of the psychiatrist’s testimony, in many cases that will not be so. Where the patient does not pose so serious a threat of harm that the psychiatrist realizes disclosure is needed to avert the threat, the risk of a false verdict may be worth taking for the social benefit of the therapy. That is the policy rationale for the holding in Jaffee.
But this is a case where the threat was understood by the psychiatrist to be so serious as to require disclosure, and she had disclosed what her patient told her. The confidentiality of the therapeutic relationship had already been breached, and the patient knew it. Where disclosure was necessary, the social interest in assuring that the judge and jury know the whole truth greatly exceeds the value of preserving any remaining shreds of the confidential therapeutic relationship. The jury ought, in such circumstances, to know the truth about what Chase said. The cat being already out of the bag, trial is no occasion for stuffing it back in.
. See Watts v. United States, 394 U.S. 705, 89 S.Ct. 1399, 22 L.Ed.2d 664 (1969).
. 18 U.S.C. § 115(a)(1)(B).
. United States v. Hayes, 227 F.3d 578, 583 (6th Cir.2000).
. Fed.R.Evid. 501 (emphasis added).
. Fed.R.Evid. 501. See also Jaffee v. Redmond, 518 U.S. 1, 12-13, 116 S.Ct. 1923, 135 L.Ed.2d 337 (1996) (policy decisions of states ''bear” on recognition of federal privileges).
. See In re Grand Jury Proceedings (Gregory P. Violette), 183 F.3d 71, 74 (1st Cir.1999) (finding crime-fraud exception to privilege) and United States v. Glass, 133 F.3d 1356, 1360 (10th Cir.1998) (therapist may testify if threat was serious and disclosure was the only means of averting harm); but see United States v. Hayes, 227 F.3d 578, 579 (6th Cir.2000) (no dangerous patient exception to psychotherapist-patient privilege).
. 133 F.3d 1356, 1360 (10th Cir.1998).
. United States v. Hayes, 227 F.3d 578 (6th Cir.2000).
. 183 F.3d 71 (1st Cir.1999).
. 518 U.S. 1, 116 S.Ct. 1923, 135 L.Ed.2d 337 (1996).
. Jaffee, 518 U.S. at 18, n. 19, 116 S.Ct. 1923 (emphasis added).
. United States v. Baird, 85 F.3d 450, 453 (9th Cir.1996).
. Jaffee, 518 U.S. at 18 n. 19, 116 S.Ct. 1923.
. Id.
. Cf. 2 Kings 4:31-37.
. Majority Op. at 984.
.’ The American Heritage Dictionary 446 (2d ed.1982).
. Majority Op. at 984, n. 2.
. Jaffee, 518 U.S. at 18 n. 19, 116 S.Ct. 1923.
. United States v. Hayes, 227 F.3d 578, 588 (6th Cir.2000) (Boggs, J., dissenting).
. Majority Op. at 979.
. 18 U.S.C. § 115.
. 8 John Henry Wigmore, Evidence in Trials at Common Law § 2192, at 70 (John T. McNaughton rev., 1961).
. Id. § 2380 at 818.
. Id. § 2286 at 531.