(dissenting). Defendant was convicted on the testimony of one witness: the complainant. The defense at trial was misidentification and the complainant’s hypnotic statement was critical to it. The descriptions of the assailant given by the complainant, both at the police station and in her trial testimony, did not completely fit defendant, and the description she gave while under hypnosis was plainly exculpatory. The trial court barred defense counsel from using the complainant’s hypnotic statement to impeach her, and the majority of our court now upholds that ruling on the ground that, under our prior decisions, hypnotically produced recollections are inherently unreliable. Because, in my view, defendant’s State and Federal constitutional rights to confront adverse witnesses were unduly restricted on the basis of an erroneous evidentiary ruling, I dissent.
I
In complainant’s statements to the police, as well as in the composite sketches drawn by a police artist and herself, the assailant was depicted as clean-shaven; in fact, at all relevant times, defendant wore a distinctive moustache. The assailant’s hair was portrayed as reaching his collar; but there was proof at trial that he didn’t wear his hair that long. The assailant was described as wearing a shirt bearing only the name "Exxon”; but the undisputed proof was that defendant’s workshirts all bore his name, "Mike H.”. And, most significantly, in her hypnotically produced statement, the complainant said that she did see a name on the assailant’s shirt, and that it was "Billie”.
That hypnotic statement would have underscored the disparities between the complainant’s pretrial descriptions of her assailant and defendant’s actual appearance; more than that, it would have added another particularly significant discrepancy — the assailant’s name. The trial court’s ruling, that the hypnotically produced description of the assailant could not be used, deprived defendant of a crucial prior inconsistent statement with which to impeach the complainant on cross-examination, and, thus, seriously curtailed his constitutional confrontation rights. While those rights are certainly not absolute (see, People v Cintron, 75 NY2d 249, 259; Delaware v Van *201Arsdall, 475 US 673, 679), the arguments made to justify the trial court’s ruling in this case do not withstand scrutiny.
The People contend that our decisions in People v Hughes (59 NY2d 523) and People v Tunstall (63 NY2d 1) declared hypnotic statements to be inherently unreliable and, thus, dictate the inadmissibility of any such statement, regardless of the purpose — whether by the People to prove their case-in-chief or, as here, by the defendant to impeach the prosecution witness who uttered it. This position, now adopted by the majority (see, majority opn, at 197), misconstrues those decisions, contravenes established rules governing the use of prior inconsistent statements, and, moreover, interferes with a basic trial right of a criminal defendant without sufficient justification.
II
In Hughes, we held that a rape victim’s hypnotically produced recollections — i.e., recollections occurring during the hypnotic procedure or coming to mind for the first time thereafter — could not be used by the prosecution as evidence to convict the defendant. The basis for our decision was not that such hypnotically produced recollections, themselves, were invariably false or unreliable. Rather, the basis was simply that there is, as yet, no sure way of determining whether such hypnotically produced recollections reflect the subject’s actual memory of events, as opposed to confabulations or fictitious recollections suggested during the procedure itself.
Our decision in Hughes rested solely on whether "hypnosis has gained general acceptance in the scientific community as a reliable means of restoring recollection” (59 NY2d, at 543); our answer was just as specific: "at the present time hypnosis has not achieved that status” (id.). While hypnotically produced memories can be accurate and reliable, we held that hypnotically produced recollections could not be permitted as a basis for convicting a defendant because science has yet to devise a way of separating true recollections from the false. And while we firmly rejected the use of such recollections against a defendant because of this uncertainty in their probative value (id., at 542-544), we never suggested that those recollections should be barred for all purposes — let alone for the impeachment of a previously hypnotized prosecution witness by the defendant, the very person sought to be protected by the guidelines adopted in Hughes (see, id., at 545-548).
*202Neither was any such prohibition suggested in Tunstall. Rather, in that case, we established the requirement for a pretrial hearing to determine whether the confidence of a prosecution witness in her prehypnotic recollections was artificially enhanced by the hypnotic process (63 NY2d, at 8-9).1 Significantly, our court made clear that the particular purpose for the safeguards in Tunstall, as for those in Hughes, is to ensure that "there has been no substantial impairment of the defendant’s ability to meaningfully cross-examine” a previously hypnotized witness (id., at 9). Surely, our holdings in those cases were never intended to restrict a defendant’s right to cross-examine such a witness. It is, at best, incongruous that those decisions should now be invoked in support of such a purpose.
But even assuming that Hughes and Tunstall could somehow be read as declaring that hypnotically produced recollections themselves are invariably unreliable, that would still not lead to the conclusion, drawn by the majority, that such recollections may not be used by a defendant for the sole purpose of impeaching the credibility of a prosecution witness who has made a prior inconsistent hypnotic statement. It is basic that the purpose of using a prior inconsistent statement on cross-examination is not to prove the facts contained therein (Richardson, Evidence § 501 [Prince 10th ed]). It is not offered for its truth; its reliability as an assertion of the truth is entirely beside the point (id.; 3A Wigmore, Evidence § 1018 [a] [Chadbourn rev]). Indeed, in most jurisdictions, including New York, it is the duty of the trial court to instruct the jury that a prior inconsistent statement has no independent testimonial value (see, e.g., People v Freeman, 9 NY2d 600; Matter of Roge v Valentine, 280 NY 268; see, CPL 60.35; 3A Wigmore, op. cit., § 1018 [b]).
Thus, a prior inconsistent statement is not used "assertively”. The fact finder is not asked to believe it, or even to *203choose between it and the contrary trial testimony (see, 3A Wigmore, op. cit., § 1018 [b]). Rather, the prior statement and the trial testimony are merely set against each other. The very fact that the witness necessarily erred in making one or the other — because both cannot be true — tends to undermine the witness’s credibility. In short, it simply makes no difference whether the prior inconsistent statement, when used as such, has any basis in truth. Hence, under settled rules, there is no basis for excluding a prosecution witness’s prior hypnotic statement, even if it is deemed unreliable as an assertion of fact, where it is offered solely for impeachment on the basis of its inconsistency.
The Supreme Court’s decision in Mincey v Arizona (437 US 385) in no way supports the opposite conclusion. There, the court barred the use of an involuntarily obtained confession to impeach the defendant, but not because of the confession’s unreliability (see, majority opn, at 198). The court never made a finding of unreliability; rather, it found only that defendant had not wished to answer police inquiries and that his "will was simply overborne.” (437 US, at 401-402.)
The court precluded the use of the confession, not because of any lack of probative value, but because of the core due process violation involved. Of course, as the Mincey court reminded us, where the "prophylactic” strictures of Miranda v Arizona (384 US 436; see, New York v Quarles, 467 US 649, 654) are violated, a defendant’s confession will still be admissible if its " 'trustworthiness * * * satisfies legal standards.’ ” (437 US, at 398, quoting Harris v New York, 401 US 222, 224.) But the court made clear that trustworthiness is irrelevant where the confession is truly involuntary. '[A]ny criminal trial use against a defendant of his involuntary statement,” the court emphasized, "is a denial of due process of law” (id. [emphasis in original]).
Hence, Mincey hardly stands for the proposition that an unreliable past statement may not be used to impeach a prosecution witness who uttered it. Instead, that decision underscores the proposition that the basic trial rights of an accused must be guarded with the utmost vigilance. The majority of this court now uses Mincey to support the totally antithetical proposition that a defendant’s right to cross-examine an adverse witness must be limited due to the possible unreliability of that witness’s prior statement.
As the Supreme Court has frequently observed, "the right of *204confrontation and cross-examination is an essential and fundamental requirement for the kind of fair trial which is this country’s constitutional goal” (Pointer v Texas, 380 US 400, 405; see also, Lee v Illinois, 476 US 530, 540). The ability of a criminal defendant to cross-examine adverse witnesses is "primarily a functional right that promotes reliability in criminal trials.” (Lee v Illinois, supra, at 540 [emphasis added].) Because of the importance of confrontation rights to our system of justice in providing an effective safeguard against false or questionable accusations (see, id.), both our court and the Supreme Court have insisted that they not be infringed unless compelled by some strong State interest (see, People v Cintron, supra, at 258-259; Delaware v Van Arsdall, supra, at 679). Such an interest is entirely absent here.
Even if, as the majority holds, hypnotically produced recollections are necessarily unreliable and prior inconsistent statements which are unreliable cannot generally be used for impeachment, still, that cannot justify the trial court’s ruling in this case. It hardly provides a compelling reason for denying a criminal defendant the opportunity to confront a prosecution witness with her hypnotic statement which not only conflicts with her trial testimony, but also exculpates the defendant. At the least, something more than an evidentiary rule of general application is necessary to limit the exercise of a basic trial right.
Hence, in Chambers v Mississippi (410 US 284), the Supreme Court invalidated the application of a state’s hearsay rule that restricted the defendant’s right to present witnesses on a matter critical to his defense. And more recently, in Rock v Arkansas (483 US 44), the court held that the state’s rule barring hypnotically refreshed testimony could not be used to infringe upon the defendant’s right to testify in his own behalf. A state’s evidentiary rule must " 'not be applied mechanistically to defeat the ends of justice,’ ” the court warned, "but must meet the fundamental standards of due process.” (Id., at 55, quoting Chambers v Mississippi, supra, at 302.) Those standards are violated, in my view, when the rules our court adopted in Hughes and Tunstall, to prevent a prosecutor’s use of potentially unreliable hypnotic recollections to obtain convictions, are applied against a defendant, to interfere with his right of cross-examination — here, defendant’s right to cross-examine the complaining witness about a prior inconsistent and exculpatory statement regarding the assailant’s identity, the sole issue at trial.
*205Ill
I would hold that the trial court erred in refusing to allow defendant to cross-examine complainant on the basis of her inconsistent hypnotic statement. The court’s refusal was particularly ironic here where it also ruled that the prosecution could use ’ complainant’s posthypnotic recollections against defendant, even though he was not permitted to test those recollections in a Hughes-Tunstall hearing.2
But the ironic effect of this ruling is magnified by a curious position taken in1 the majority opinion. Because defendant attempted to undercut complainant’s posthypnotic testimony by adducing expert evidence concerning the possible taint from the hypnosis, he is now held to have "conceded the unreliability of the complainant’s hypnotic statements” (majority opn, at 199) — and, presumably, also to have "conceded” that they could not be used by him for the purpose of cross-examination.3 It is unclear whether a defendant may use the complainant’s hypnotic statement for cross-examination only if he foregoes expert proof on the effect of hypnosis on the complainant’s direct testimony. The majority opinion does seem to advance this singular proposition, however, and uses it with telling effect on defendant here.
In short, the People were allowed to convict defendant on the basis of testimony which might have been tainted by hypnosis. Defendant, on the other hand, was deprived of his right to cross-examine the complainant by using a hypnotic statement which, for all that is known, might have been entirely accurate and trustworthy.
For these reasons, I believe it was error to restrict the exercise of defendant’s confrontation rights by denying him the use of complainant’s prior inconsistent hypnotic statement
*206on cross-examination; I would reverse the order of the Appellate Division, vacate defendant’s conviction and order a new trial.
Chief Judge Wachtler and Judges Simons, Kaye and Bellacosa concur with Judge Alexander; Judge Hancock, Jr., dissents and votes to reverse in a separate opinion in which Judge Titone concurs.
Order affirmed.
. In Hughes, we had also held that a subject’s prehypnotic recollections —i.e, what the subject recalled of the actual events before submitting to the hypnotic procedure — could be adversely affected by a subsequent hypnotic procedure (59 NY2d 523, 545-548). The subject’s confidence in the accuracy of her prehypnotic recollections could be enhanced, thereby making it more difficult to cross-examine her effectively. We held, nevertheless, that a subject should be permitted to testify against a defendant in a criminal trial concerning such prehypnotic recollections provided that the court has first determined in a hearing that the prehypnotic recollections were not enhanced or otherwise affected by the subsequent hypnosis (id., at 547-548).
. Because no hearing was held, there is no way of knowing whether "the victim’s prehypnotic recollections were not 'unduly influenced by the hypnotic sessions’ ” (People v Hughes [II] 72 NY2d 1035,1037). Also, and of equal importance, without a hearing the extent of the victim’s prehypnotic recollections — and the boundary of admissible testimony — was not established (id., at 1037). Thus, it cannot be determined whether defendant was convicted either on the basis of fictitious or confabulated recollections or prehypnotic recollections which were tainted by the subsequent hypnosis.
. It should be noted that the court in denying defendant’s motions with respect to the Hughes-Tunstall hearing and for a preclusion of the victim’s testimony, specifically provided as a condition of its denial that it would "certainly permit [defendant] to introduce evidence from [his] expert on hypnosis.” (Trial minutes, at 51.) It seems anomalous that defendant’s decision to avail himself of this permission should now be held against him.