People v. Alls

83 N.Y.2d 94 (1993) 629 N.E.2d 1018 608 N.Y.S.2d 139

The People of the State of New York, Respondent,
v.
Robert Alls, Appellant.

Court of Appeals of the State of New York.

Argued November 19, 1992. Reargued September 8, 1993. Decided December 21, 1993.

Francis J. Constantine, Livonia, for appellant.

Thomas Moran, District Attorney of Livingston County, Rochester (Melvin Bressler of counsel), for respondent.

Judges TITONE, HANCOCK, JR., and SMITH concur with Judge LEVINE; Chief Judge KAYE dissents in part and votes to reverse, grant defendant's motion to suppress and order a new trial, in a separate opinion; Judge SIMONS dissents and votes to affirm in another opinion in which Judge BELLACOSA concurs.

LEVINE, J.

Defendant, an inmate at the Groveland Correctional Facility in Livingston County (hereinafter the "facility"), was involved in an incident leading to his indictment for sodomy in the first degree and assault in the second degree committed against another inmate. After the alleged victim complained to facility staff concerning the incident, Correction Sergeant Michael Rhodes was directed by his watch commander to question defendant regarding "a fight or an assault between the two inmates" (emphasis supplied). Rhodes testified that he went to the building where defendant was housed and, "after finding out that the basement area was empty, I took him, Alls, down to the basement to interview him and I asked him about the incident last night". Defendant made certain admissions regarding assaulting the complainant.

Concededly, Rhodes questioned defendant without first administering Miranda warnings (Miranda v Arizona, 384 U.S. 436). County Court held, however, that the questioning of defendant was not custodial interrogation so as to require advising him of his Miranda rights. The court concluded that Miranda warnings are only required, inter alia: "(1) [w]here an individual is questioned in `custodial settings that have inherently coercive pressures which tend to undermine the individual's will to resist and compel him to speak'". County Court then found that the specific circumstances of defendant's interrogation by Rhodes were not in an "`inherently coercive atmosphere'", because (1) he was not told that he was not free to leave or to break off the interview; and (2) he was only removed to an "office-like setting in the basement which could afford the parties privacy".

Accordingly, defendant's motion to suppress his admissions to Rhodes was denied, and his statement was introduced against him at the trial that followed. The jury acquitted defendant of the sodomy charge, but convicted him of the assault count of the indictment. The Appellate Division affirmed, indicating its agreement with County Court that defendant was not in custody when he gave his statement to Rhodes (170 AD2d 996). Moreover, the Court held, there was overwhelming proof of guilt rendering harmless any error in admitting defendant's statement. A Judge of this Court granted leave to appeal. We now modify and remit for a de novo suppression hearing.

Relying on Mathis v United States (391 U.S. 1), defendant's primary contention on appeal is that, since an inmate in a correctional facility is clearly not permitted to leave the confines of the facility, any questioning of the prisoner is per se custodial interrogation, requiring Miranda warnings. We disagree.

First, reliance on Mathis must be tempered by the more recent decision of the United States Supreme Court on prison inmate interrogation, Illinois v Perkins (496 U.S. 292). Perkins was incarcerated on charges unrelated to the unsolved murder the police suspected him of committing. An undercover officer posing as another inmate was placed in his cell and, through subtle questioning, elicited incriminating statements from Perkins regarding the murder, of course, without prefacing his questions by Miranda warnings. Rejecting application of the syllogism suggested by defendant's reading of Mathis here, the Court in Perkins held that Miranda warnings were not required, despite the undisputed facts that Perkins was in custody and inculpated himself in response to questioning by a government agent.

Moreover, there is dictum in Illinois v Perkins (supra) that, in our view, precludes reading it as representing only a narrow, undercover agent exception to the general rule that questioning an inmate is per se custodial interrogation for purposes of Miranda v Arizona. The Court stated that "[t]he bare fact of custody may not in every instance require a warning even when the suspect is aware that he is speaking to an official, but we do not have occasion to explore that issue here" (Illinois v Perkins, supra, at 299). The Court explained that, although the prisoner there was in detention and, thus, "in custody in a technical sense" (id., at 297), "[c]onversations between suspects and undercover agents do not implicate the concerns underlying Miranda" (id., at 296 [emphasis supplied]). Rejecting the argument that failure to follow a per se custody rule in prison settings would undercut the need for a "bright-line rule" for applying Miranda, the Court stated that "[t]he interests protected by Miranda are not implicated in these cases, and the warnings are not required to safeguard the constitutional rights of inmates who make voluntary statements to undercover agents" (id., at 300 [emphasis supplied]).

Thus, in Illinois v Perkins, the Supreme Court held that incarcerated persons are not entitled to Miranda warnings before being questioned by undercover agents, because in such "cases" the Fifth Amendment interests protected by the Miranda warnings are not implicated. As previously noted, the Court further suggested the existence of other kinds of cases involving prison inmates where the interests protected by Miranda are also not implicated.

The majority's dictum in Illinois v Perkins, that there may be additional instances where the questioning of a prison inmate is not deemed custodial interrogation, is supported by other Supreme Court decisions holding that, despite the conceded existence of some form of detention of the person questioned during a confrontation with law enforcement authorities, the interrogation was not considered per se custodial for purposes of Miranda. In Berkemer v McCarty (468 U.S. 420), Justice Marshall writing for the majority candidly acknowledged "that a traffic stop significantly curtails the `freedom of action' of the driver and the passengers, if any, of the detained vehicle" and "few motorists would feel free either to disobey a directive to pull over or to leave the scene of a traffic stop without being told they might do so" (id., at 436). Nevertheless, despite meeting the standard criteria for determining whether a suspect is in custody for Miranda purposes (see, Miranda v Arizona, 384 US, at 444, supra; People v Rodney P., 21 N.Y.2d 1, 8-9), the Court in Berkemer held that police roadside questioning of a motorist during a traffic stop does not by itself constitute custodial interrogation. Additionally, the Miranda opinion itself recognized that not all temporary detentions of witnesses for on-the-scene investigative questioning trigger Miranda warnings (Miranda v Arizona, supra, at 477-478). Likewise, questioning of a suspect during a "Terry stop" (Terry v Ohio, 392 U.S. 1) does not per se constitute custodial interrogation (Berkemer v McCarty, supra, at 439-440; United States v Brignoni-Ponce, 422 U.S. 873, 880-881).

The common elements of the foregoing situations where questioning during some measure of detention by law enforcement authorities is not considered custodial interrogation requiring Miranda warnings are twofold. First, they typically do not involve the kind of inherently coercive atmosphere with which Miranda was most concerned. Thus, in Berkemer v McCarty, Justice Marshall pointed out that the detention in a traffic stop situation "is presumptively temporary and brief" (supra, at 437), and that "the typical traffic stop is public" and "the detained motorist typically is confronted by only one or at most two policemen" (id., at 438 [emphasis supplied]). And, in sustaining Miranda-less questioning during investigative border patrol stops, the Supreme Court relied on the facts that such detention "`usually consumes no more than a minute' * * * `[a]ll that is required of the vehicle's occupants is a response to a brief question or two and possibly the production of a document evidencing a right to be in the United States'" (United States v Brignoni-Ponce, supra, at 880 [emphasis supplied]). Also, it was pointed out in Berkemer v McCarty (supra) that questioning in a "Terry stop" (Terry v Ohio, supra) is not deemed custodial interrogation because "[t]ypically, this means that the officer may ask the detainee a moderate number of questions to determine his identity and to try to obtain information confirming or dispelling the officer's suspicions * * * And, unless the detainee's answers provide the officer with probable cause to arrest him, he must then be released" (Berkemer v McCarty, supra, at 439-440 [emphasis supplied]).

The second common element in the foregoing cases is that, in each situation, requiring Miranda warnings would seriously undermine other important social values. As Justice Marshall explained in Berkemer v McCarty (supra), a per se rule applying Miranda to all traffic stops "would substantially impede the enforcement of the Nation's traffic laws — by compelling the police either to take the time to warn all detained motorists of their constitutional rights or to forgo use of self-incriminating statements made by those motorists — while doing little to protect citizens' Fifth Amendment rights" (468 U.S. 420, 441, supra; see also, United States v Brignoni-Ponce, supra, at 879).

In the context of a correctional facility environment, despite the undeniable fact of an inmate's general confinement in the sense of not being free to leave the facility, we can envisage confrontations between correctional officers and prisoners analogous to the relatively brief, generally public, or otherwise on-the-scene investigatory detentions in nonprison settings found not custodial for Miranda purposes. On the other hand, to impose a requirement that correctional authorities must administer Miranda warnings before even the most casual or spontaneous interchanges with inmates integral to prison life would have an enormous impact upon the ability of correctional authorities to maintain prison order and discipline, in cases where the Fifth Amendment rights of inmates are not seriously implicated (Berkemer v McCarty, supra, at 441; Illinois v Perkins, 496 U.S. 292, 299, supra). For these reasons, we decline to adopt a per se rule that any questioning of an inmate in a correctional facility is custodial interrogation under Miranda v Arizona (supra).

When, however, the circumstances of the detention and interrogation of a prison inmate are no longer analogous to those kinds of detentions found not custodial in nonprison settings, but instead entail added constraint that would lead a prison inmate reasonably to believe that there has been a restriction on that person's freedom over and above that of ordinary confinement in a correctional facility, Miranda warnings are necessary. This standard is not unlike that enunciated in Cervantes v Walker (589 F.2d 424, 428 [9th Cir]).

Although County Court concluded that defendant was not subjected to custodial interrogation or restraints over and above that of general incarceration, its determination was based upon an erroneous legal standard. First, County Court articulated an erroneous standard for determining whether an interrogation is custodial so as to require Miranda warnings. It stated that "[s]uch warnings are required: (1) [w]here an individual is questioned in `custodial settings that have inherently coercive pressures which tend to undermine the individual's will to resist and compel him to speak'" (emphasis supplied). The court then applied that erroneous standard to defendant's case: "[t]he Court thus does not find any [emphasis in the original] indicia that the interview between Rhodes and the defendant were [sic] inherently coercive" (emphasis supplied). Attempts such as this to limit the requirement of giving Miranda warnings exclusively to custodial settings which had other coercive aspects have been repeatedly rejected, because Miranda presumes that custody by law enforcement authorities is inherently coercive, even in a nonthreatening atmosphere (see, Orozco v Texas, 394 U.S. 324, 326-327; see also, New York v Quarles, 467 U.S. 649, 654; Dunaway v New York, 442 U.S. 200, 213-214, 214, n 16).

The foregoing erroneous legal standard also tainted County Court's determination that, because defendant was not handcuffed, shackled or subjected to protracted questioning or questioning in a massive, law enforcement-dominated atmosphere during Rhodes' interrogation, there were no "indicia that the interview between Rhodes and the defendant * * * entailed a measure of restriction upon the defendant's movement `over and above' that generally imposed by reason of his incarceration" (emphasis supplied). In thus requiring additional, specific coercive official conduct as a sine qua non for application of the additional restraints test here, County Court clearly was proceeding under its erroneous legal supposition that only questioning in a coercive atmosphere of detention constitutes custodial interrogation for Miranda purposes (see, supra, at 97). Moreover, County Court was only able to limit its focus of inquiry exclusively to the ambience of the basement "interview" by reason of its entirely speculative factual assumption, for which there is no evidentiary support in the record, that Rhodes' "taking" of defendant to the basement was somehow mutually agreed upon to obtain privacy.

Through the foregoing errors, the suppression court effectively converted the added imposition of restraint test into an added imposition of coercive pressures test. The added restraint test was designed as a fair, but limited, accommodation of the Miranda rules to a correctional facility environment, because applying the Miranda definition of custody literally would have required warnings any time a correction officer asked an inmate a question (Cervantes v Walker, 589 F.2d 424, 427, supra). Such a per se rule would have afforded an inmate "greater protection * * * than * * * his nonimprisoned counterpart" (id., at 427). County Court's application of the test, to require not merely further constraint upon the interrogated inmate, but additional coercive restraint, disregarded the authoritarian realities of a correctional facility and afforded defendant significantly less protection of the Miranda rules than his civilian counterpart.

The foregoing legal errors require remittal for a new determination of the facts under the proper legal standard (see, People v Davis, 75 N.Y.2d 517, 524; People v Morales, 65 N.Y.2d 997, 998). Because the suppression hearing was conducted under County Court's erroneous conception of the added restraint test, the record is, in our view, insufficient to determine whether defendant was actually subjected to custodial interrogation under the criteria we have enunciated herein; therefore, it is appropriate to direct that a new suppression hearing be held, in which the parties will have a full opportunity to develop the facts (see, People v Dodt, 61 N.Y.2d 408, 417-418; People v Havelka, 45 N.Y.2d 636, 642-643).

On such a de novo hearing, the circumstances under which Correction Sergeant Rhodes "took" defendant to the basement to question him should be fully explored, particularly in the light of prison regulations requiring an inmate to comply with the directions of a correction officer relating to movement within the facility or face severe disciplinary sanctions (7 NYCRR 270.2 [B] [10] [iii]). If defendant's movement to the basement area of his facility housing unit was at Rhodes' direction, defendant was under some compulsion to comply, in the absence of proof of other circumstances which could have led defendant reasonably to believe he was free to decline to follow the correction officer's directions, such as Rhodes' offering him a choice as to where to be interviewed. The People could only have dispelled the inference of defendant's compulsion to accompany Rhodes by evidence that defendant "was actually offered a choice" (People v Dodt, supra, at 417). The taking of a suspect under apparent compulsion by a law enforcement official to a place of isolation has repeatedly been held to constitute custodial restraint, even when such asportation is to a place less coercive than a police headquarters (see, Florida v Royer, 460 U.S. 491 [isolation of a suspected drug courier in a small room at an airline terminal]; United States v Baron, 860 F.2d 911 [9th Cir], cert denied 490 U.S. 1040 [suspect taken from a parking lot and isolated in another suspect's apartment]). Indeed, the court in United States v Baron identified as "the crucial factor" in the case "the coerciveness created by isolating a suspect in a private space controlled by the police" (id., at 916).^[*]

Moreover, the foregoing circumstances, if established, would not bear any resemblance or analogy to the relatively brief, generally public, or otherwise on-the-scene investigatory detentions in nonprison settings found not custodial for Miranda purposes.

On the issue of the admissibility of defendant's inculpatory statement, it remains for us to address Judge Simons' test for determining custodial interrogation in a prison context. The test would limit Miranda's application for prison inmate interrogations to those instances where the circumstances of the case show a "`danger'" of coercion from "`the interaction of custody and official interrogation'" (dissenting opn, Simons, J., at 121). The test clearly requires a fact-intensive, individualized inquiry as to whether the particular prisoner interrogation was conducted in a coercive manner, or whether the questioning took place in an inherently coercive atmosphere — similar to the test applied by County Court in holding that no custodial interrogation took place here. Moreover, it seems self-evident that the test suggested by Judge Simons represents at least a partial undermining of the function and objective of the Miranda rules as, in a sense, a codification of proper interrogation procedures for ready application by law enforcement agencies and courts and to withstand evasion by the police as well as by trial courts in exercising their fact-finding powers. The same kind of fact-intensive inquiry, to determine whether the particular circumstances of a custodial setting were or were not potentially coercive, was suggested not long after Miranda in Orozco v Texas (394 U.S. 324, 328-331 [White, J., dissenting], supra). Indeed, Justice White's description of the noncoercive atmosphere of the postarrest interrogation in Orozco, that the questioning was brief and in familiar surroundings and that, therefore, "[t]he same danger of coercion [as apprehended in Miranda] is simply not present in such circumstances" (id., at 331), is virtually a replication of Judge Simons' view of defendant's interrogation here. Recognizing the inconsistency of that approach with a prophylactic function of Miranda, the majority in Orozco rejected the attempt to so limit Miranda's application. And a similar more flexible approach was rejected in Dunaway v New York (442 U.S. 200, 213-214, 214, n 16, supra). We have concluded that such an approach is also inappropriate here, despite the concededly difficult challenge posed here to fairly apply Miranda to inmate interrogations in a prison setting.

Finally, in our view, the Appellate Division's alternative harmless error rationale for affirming defendant's conviction is insupportable. In this case, the jury's verdict convicting defendant of assault but acquitting him of sodomy indicates that the jury did not fully credit the complaining witness's testimony. Thus, it is readily inferable that the jury was influenced by defendant's admissions solely to having committed an assault. The constitutional error committed here is only harmless if "there is no reasonable possibility that the error might have contributed to defendant's conviction" (People v Crimmins, 36 N.Y.2d 230, 237). Here, there clearly is a reasonable possibility that the defendant's statement influenced the verdict.

Upon remittal for a de novo suppression hearing, if County Court determines that defendant is entitled to suppression, then a new trial must be ordered. If County Court determines that suppression is not warranted, the judgment of conviction should be amended to reflect that result.

Order modified, and the case remitted to Livingston County Court for further proceedings in accordance with this opinion and, as so modified, affirmed.

Chief Judge KAYE (dissenting in part).

This appeal concerns criminal prosecution of an inmate who was brought to the prison basement one day after an assault and questioned without the benefit of Miranda warnings. Mathis v United States (391 U.S. 1) — Supreme Court precedent on point — controls here and mandates suppression. The refusal by both the majority and dissent to acknowledge Mathis as binding in these circumstances is troubling, to say the least.

I hasten to add, however, that the fact that this case has generated three opinions may obscure the Court's unanimity on certain essential points. We all agree that the Miranda right is not lost by reason of incarceration. We all agree that Miranda should not apply per se to every custodial interrogation, and we all agree that Miranda must be harmonized with the unique attributes of prison life. What separates us is how to determine whether a particular prison interrogation must be preceded by Miranda warnings.

The majority rejects the view that prisoners are, by virtue of incarceration, in custody for Miranda purposes. They would require additional restraints on the prisoner before warnings are necessary, and conclude that a new hearing is required to determine whether such restraints were present here. The dissent believes that courts must determine, on a case-by-case basis, whether the interaction of custody and interrogation presents a danger that the suspect would feel coerced, and concludes that there was no such danger here.

In my view, prisoners are in custody for Miranda purposes without the need to consider other variables. The reconciliation of Miranda and the needs of prison administration should be accomplished not by modifying the established definition of custody, but by recognizing a prison exception to Miranda in appropriate cases.

I.

The evidence adduced at a Huntley hearing revealed that on March 21, 1988, Correction Sergeant Michael Rhodes was assigned to interview defendant — an inmate at Groveland Correctional Facility — in connection with an altercation that occurred the previous night. Rhodes was told by his superior that there had been a fight or assault between two inmates, but was not given details.

After confirming that the basement, containing kitchen and recreation facilities, was empty, Rhodes retrieved defendant from H block and (according to Rhodes) "took" him down to the basement for questioning. Rhodes asked defendant if he had been involved in some sort of fight. Defendant responded that, after a fellow inmate, Prentice, jumped onto him, he grabbed the man's neck, pushed him against the wall and hit him several times in the chest. At one point Rhodes asked defendant whether he knew that he was admitting to an assault, and defendant replied that he did. Rhodes conceded, however, that he did not advise defendant of the Miranda warnings at any time, explaining that they are not used in the Department of Correctional Services and were not part of his training.

Indicted for assault and sodomy, defendant moved to suppress his statement. County Court denied the motion, finding that, under the circumstances, defendant was not in custody for purposes of Miranda v Arizona (384 U.S. 436) because he was not interrogated in an inherently coercive setting, nor were there restrictions on his movement "`over and above' that generally imposed by reason of his incarceration."

Defendant proceeded to trial, where the prosecution used his statement on its direct case. Prentice, the only eyewitness to the incident testifying for the People, claimed that defendant assaulted and sodomized him. The jury, however, apparently disbelieved Prentice and, consistent with defendant's statement, convicted of assault but acquitted of sodomy. Defendant was sentenced to an additional 3½ to 7 years of incarceration.

On appeal, the Appellate Division affirmed, agreeing with County Court that Miranda warnings were not required, and that in any event, use of the statement was harmless in light of the overwhelming evidence of guilt. I would reverse, grant the motion to suppress and order a new trial.

II.

The entire Court agrees that prisoners, like everyone else, enjoy a constitutional privilege against compelled self-incrimination (Minnesota v Murphy, 465 U.S. 420, 426). To safeguard this constitutional right, Miranda v Arizona (384 U.S. 436, supra) held that a "custodial interrogation" must be preceded by advice that the suspect has a right to remain silent and that any statements could be used in court (384 US, at 444). The Supreme Court stressed that the coercive aspects of modern interrogation techniques are psychological rather than physical (384 US, at 448), and that in-custody interrogation "contains inherently compelling pressures" that undermine an individual's will (384 US, at 467).

Just two years after Miranda was decided, the Supreme Court considered whether prisoners are in custody for purposes of Miranda, and made clear that they are. In Mathis v United States (391 U.S. 1, supra), an individual serving a sentence for a State crime was questioned, without Miranda warnings, by IRS agents performing a "routine tax investigation" (391 US, at 4). Over the protest of Justice White — who argued that "[a]lthough petitioner was confined, he was at the time of interrogation in familiar surroundings" (391 US, at 7, supra [White, J., dissenting]) — the Court suppressed the statements, adhering to the "well-considered conclusions of Miranda with reference to warnings to be given to a person held in custody." (391 US, at 4; emphasis added; see also, id., at 4; emphasis added ["We reject the contention that tax investigations are immune from the Miranda requirements for warnings to be given a person in custody"].)

The following year, the Court (citing Mathis) expressly rejected Justice White's "familiar surroundings" argument and suppressed defendant's statements made in his own home, on his bed (Orozco v Texas, 394 U.S. 324, 326). Thereafter, the Supreme Court itself has applied Miranda to the prison setting without the need to find "additional restrictions" on the prisoner's liberty — the test embraced by the majority today (majority opn, at 100) (see, Estelle v Smith, 451 U.S. 454, 467 ["Respondent was in custody at the Dallas County Jail when the examination was ordered and when it was conducted"; suppressing psychiatric evaluation because defendant not given Miranda warnings]; compare, Berkemer v McCarty, 468 U.S. 420, 434 ["no question" that defendant was in custody when statements were taken at county jail, at least from the moment he was formally arrested and placed in police car]; see also, Bradley v State, 559 A2d 1234, 1244-1245 [Del]; Commonwealth v Chacko, 500 Pa 571, 459 A2d 311, 314-315; Young v State, 234 So 2d 341 [Fla]).

Similarly, the established test for determining custody in other contexts is whether, from the perspective of the reasonable person (Berkemer v McCarty, 468 US, at 442), "there is a `formal arrest or restraint on freedom of movement' of the degree associated with a formal arrest" (California v Beheler, 463 U.S. 1121, 1125, quoting Oregon v Mathiason, 429 U.S. 492, 495; see also, New York v Quarles, 467 U.S. 649, 655). Unquestionably prisoners satisfy this test (see, e.g., United States v Conley, 779 F.2d 970, 973 [4th Cir], cert denied 479 U.S. 830).

Accordingly, applying Supreme Court precedent to the undisputed facts of this case, it is plain that defendant was in custody, by virtue of incarceration, when he was questioned by Rhodes.^[*]

III.

County Court's conclusion that the interrogation was noncustodial stemmed from application of the added-imposition-of-restraint test, first announced in Cervantes v Walker (589 F.2d 424), a 2 to 1 decision of the Court of Appeals for the Ninth Circuit. The majority in that case reasoned that if the established definition of custody were applied, it could "totally disrupt prison administration" (id., at 427; see also, United States v Conley, 779 F.2d 970, 973, cert denied 479 U.S. 830, supra). Thus, the court required a restriction on the prisoner's freedom of movement "over and above that in his normal prisoner setting" (id., at 428). In this case, County Court did not find such restrictions and accordingly determined that defendant was not entitled to be advised of his rights.

Cervantes and similar cases have been criticized (see, Cervantes v Walker, 589 F2d, at 429, supra [Anderson, J., dissenting]; Bradley v Ohio, 497 U.S. 1011, 1012, n [Marshall, J., dissenting from denial of certiorari]; United States v Morales, 834 F.2d 35, 39 [2d Cir] [Oakes, J., concurring]; United States v Cadmus, 614 F Supp 367, 370 [SD NY]; Comment, Putting the Fifth Amendment Behind Bars, 55 Brooklyn L Rev 455, 475-476 [1989]). Indeed, under that test, reminiscent of the pre-Miranda "voluntariness" standard, bizarre holdings have resulted. For example, United States v Conley (779 F2d, at 973-974, supra) held that a prisoner who "wore handcuffs and, at some points, full restraints" was not in custody — because inmates were commonly transported in that fashion. As Judge Oakes has observed, courts applying the added-imposition test "invariably conclude that the challenged interrogations are noncustodial, suggesting that the factors may merely disguise a denial of prisoners' Fifth Amendment rights." (United States v Morales, 834 F2d, at 39, supra [Oakes, J., concurring]; see also, Comment, op. cit., 55 Brooklyn L Rev, at 475 ["The `added imposition' requirement bears many logical absurdities and inconsistencies with the purposes of Miranda"].)

Miranda was premised on the idea that custodial interrogations contain inherently coercive pressures that may undermine a suspect's will to resist (Miranda v Arizona, 384 US, at 467). A requirement of added restraint on an inmate is inconsistent with that principle because the prison environment is already enormously coercive, physically and psychologically. In the majority's view, if the questioning took place in the basement at defendant's option, he would not be entitled to Miranda warnings (majority opn, at 102). I disagree. Even in those circumstances, the coercion inherent in the situation would warrant Miranda warnings. Accordingly, I believe the majority errs today in adopting added-restraint as the test of custody.

IV.

The added-restraint test was inspired by recognition that a forthright application of the established definitions of custody and interrogation could seriously disrupt prison administration (see, e.g., United States v Conley, 779 F2d, at 973, supra; Cervantes v Walker, 589 F2d, at 427). I agree that "there must be mutual accommodation between institutional needs and objectives and the provisions of the Constitution that are of general application." (Wolff v McDonnell, 418 U.S. 539, 556.) However, the Cervantes solution to the problem — ratchetingup the definition of custody, which then applies to all prison interrogations, regardless of circumstances — is obsolete in light of a subsequent Supreme Court decision, New York v Quarles (467 U.S. 649, supra).

In Quarles, a police officer restraining a reportedly armed suspect in a store asked him where the gun was, and he nodded in the direction of empty cartons and responded, "the gun is over there." The Supreme Court reversed our suppression order and developed a "public safety" exception to Miranda, leaving intact the established definitions of custody and interrogation. In language that could conceivably apply to prisons, the Court concluded that "the need for answers to questions in a situation posing a threat to the public safety outweighs the need for the prophylactic rule protecting the Fifth Amendment's privilege against self-incrimination." (467 US, at 657.)

Had Quarles been decided before Cervantes, the Ninth Circuit could have established a prison exception to Miranda that was "circumscribed by the exigency which justifies it" (Quarles, 467 US, at 658). Instead, inasmuch as Miranda was then absolute, the Cervantes court modified the definition of custody to balance the needs of prison administrators with the rights of inmates. I suggest that, in light of Quarles, a more satisfactory approach would be to simply recognize a "prison exception" to Miranda.

Such an exception should be tailored to situations where application of Miranda in prison would be unworkable. It might apply, for example, to situations presenting threats to safety or order. This case, however — involving questioning of defendant in the basement one day after the incident — does not provide an appropriate vehicle for exploring the contours of the exception. Indeed, to do so here would violate the norms of case-by-case adjudication, deciding no more than the case presents. As we observed in People v English (73 N.Y.2d 20, 24): "While we are urged to demarcate categories when Miranda warnings are and are not required, the facts of this case do not call upon us to decide whether the nature of the parole officer-parolee relationship is such that even routine, noncustodial questioning must be preceded by Miranda warnings if it is concerned with possible criminal activity." That is no less true in the present case.

Focus on the needs of prison administration, rather than an artificial inquiry into whether the inmate is in "custody," has the twin benefits of faithfulness to Miranda and flexibility to correctional officers. When there is no prison-related need to avoid Miranda, as in the present case, an interrogation should be preceded by advice that the suspect may decline to answer — the setting is already coercive enough to trigger Miranda, without additional restraints or coercion. By the same token, in certain exigent circumstances it may be necessary to place added restraints on a prisoner and receive immediate answers. In these cases, fair application of the added-restraint test should result in suppression. Under my proposed analysis the statement would be admissible.

V.

My disagreement with the added-restraint test precludes me from joining the majority, and I find Judge Simons' approach — which in practical application appears indistinguishable from added-restraint — similarly problematic.

The dissent would affirm under Illinois v Perkins (496 U.S. 292), which it reads as announcing a new Miranda test requiring a finding that there was a danger of coercion resulting from the "interaction" of custody and interrogation (dissenting opn, at 121). I cannot agree that Perkins has changed the law of Miranda in the manner suggested.

In Perkins, the issue was whether Miranda warnings are required when an undercover officer posing as an inmate speaks to a fellow prisoner. That the case was decided on the narrow ground that the suspect simply did not know that he was talking to a government official is plain from the Court's holding in the opening paragraph — "Miranda warnings are not required when the suspect is unaware that he is speaking to a law enforcement officer and gives a voluntary statement" (496 US, at 294) — as well as the logic throughout the opinion.

The dissent, however, argues that Perkins should be read more broadly as establishing a new "interaction" test requiring courts to examine all the circumstances to determine whether there was a possibility or danger of coercion (see, dissenting opn, at 122). But the whole point of Miranda was to eliminate the old "voluntariness" test, whereby courts determined case-by-case whether defendant's will was overborne by coercion (see, e.g., Haynes v Washington, 373 U.S. 503, 514; Payne v Arkansas, 356 U.S. 560, 561-562), and to replace it with prophylactic rules.

While maintaining that the prophylaxis of Miranda is preserved, because under its test the inquiry would focus on "danger" or "possibility" of coercion rather than actual coercion (dissenting opn, at 122), it is evident that the dissent advocates a return to pre-Miranda case-by-case factual inquiries, the very antithesis of prophylactic rules. Moreover, the dissent fails to articulate which factors are relevant in determining whether the forbidden "interaction" is present, and does not adequately distinguish between its test and added-restraint. Because the dissent is content to rely on the facts as found by County Court under the added-restraint test — instead of remitting for a new hearing to explore the interaction of custody and interrogation under its own newly announced test — the conclusion is unavoidable that the two tests are essentially similar.

While concluding with language praising Miranda as a vital source of protection for all our citizens (dissenting opn, at 122), the dissent's application of its own test reveals just how little protection would remain for interrogated prisoners. Here, under legal compulsion (7 NYCRR 270.2 [B] [10] [iii]), defendant was taken to an empty prison basement and interrogated about a prison crime that occurred the previous day. Defendant of course had to answer truthfully or risk discipline (7 NYCRR 270.2 [B] [8] [iii]). Yet the dissent fails to see a possibility or danger of coercion in these circumstances.

I think the majority misreads Perkins too, by claiming that it in some way "tempered" Mathis' analysis of custody (majority opn, at 97). Perkins expressly distinguished Mathis solely on the ground that the "suspect in Mathis was aware that the agent was a Government official, investigating the possibility of non-compliance with the tax laws" (496 US, at 299). In fact, the dictum upon which both the majority and the dissent so heavily rely — "[t]he bare fact of custody may not in every instance require a warning even when the suspect is aware that he is speaking to an official, but we do not have occasion to explore that issue here" (496 US, at 299) — does not suggest that the test of custody should be different in prison, but that warnings may not always be required. That is consistent with my view that prisoners are in custody but that there should, in appropriate circumstances, be a prison exception to Miranda.

VI.

I agree with the Court that the Appellate Division's alternative harmless error rationale is unsupportable. In this case, the jury did not fully credit the complaining witness's testimony, convicting defendant of assault but acquitting of sodomy — a verdict consistent with defendant's statement. Constitutional error is harmless only if "there is no reasonable possibility that the error might have contributed to defendant's conviction" (People v Crimmins, 36 N.Y.2d 230, 237). That standard is clearly not met here, inasmuch as there is a reasonable possibility that defendant's statement was used by the jury in reaching its verdict.

For the foregoing reasons, I would grant the motion to suppress without the need for a new hearing.

SIMONS, J. (dissenting).

The majority has rejected the idea that prison inmates are per se in custody for purposes of Miranda. I agree. Rejection of the per se rule accurately reflects law to be found in recent Supreme Court decisions and is consistent with the underlying purposes of Miranda. Moreover, that view recognizes the obvious, that the traditional two-prong custody and interrogation test of Miranda is unworkable in the context of prison life.

I disagree with the majority, however, for three reasons. First, I find no basis for its decision to remit this case for further fact finding. The courts below used the same legal test that the majority adopts today. Those courts concluded that defendant was not subject to added restraints and therefore not in custody for Miranda purposes. Those findings are supported by evidence in the record and thus are binding on this Court. Inasmuch as no question of law was decided wrongly, the Court is required to affirm. In an attempt to avoid that result, the majority first contends that the suppression court misunderstood what constituted "an added restraint". No fair reading of County Court's decision can support that conclusion. Next, the majority improperly suggests that the burden of proof is on the People to establish that defendant was not in custody for purposes of Miranda (majority opn, at 102). Federal case law makes clear that the burden of establishing custody rests with the defendant. Finally, although the added restraints test adopted by the majority enjoys support in other jurisdictions, the most recent Supreme Court decision in this area suggests a different analysis in applying Miranda to prison questioning: whether the interrogation takes place under circumstances that, when viewed objectively, can be said to present a danger of coercion. In my view, such an analysis more adequately reflects the underlying purposes of Miranda and better accommodates the distinctive context of prison life and the unique needs of prison administration and order.

Nor can I join Chief Judge Kaye's opinion. She applies a per se rule. In Chief Judge Kaye's view, inmates are always in custody for Miranda purposes and therefore any questioning, no matter how benign or routine, must be preceded by warnings if it might elicit an incriminating response. That is so, Chief Judge Kaye would hold, unless some undefined public safety emergency exception, to be explained in the future, applies. The rule places an impossible burden on the effective management of the prison system.

Undoubtedly, prison inmates must be given Miranda warnings in some situations before being questioned. The test should be whether the interrogation takes place under circumstances which, when viewed objectively, can be said to present a danger of coercion. Because the suppression court here employed this test, as well as the added restraints test, and concluded that defendant was not in custody for Miranda purposes, I would affirm the order of the Appellate Division.

I

The majority supports its conclusion that a remittal for new fact finding is needed by asserting that County Court failed to apply the added restraints test correctly and that the record of the hearing court was insufficient to render a decision on that issue. Any fair reading of the court's decision and the record from the hearing proves otherwise.

County Court clearly stated its holding following the suppression hearing: "The Court thus does not find any indicia that the interview between Rhodes and the defendant [was] inherently coercive or entailed a measure of restriction upon the defendant's movement `over and above' that generally imposed by reason of his incarceration" (emphasis added). The Appellate Division affirmed, citing to a Federal case that likewise applied the added restraints test (see, United States v Conley, 779 F.2d 970, 974, cert denied 479 U.S. 830 [defendant's "freedom of movement cannot be characterized as more restricted than that of other prisoners"]). The majority now accepts that same test — i.e., custody will be found when the circumstances "entail added constraint that would lead a prison inmate reasonably to believe that there has been a restriction on that person's freedom over and above that of ordinary confinement in a correctional facility" (majority opn, at 100). Nevertheless, it claims it is not bound by the findings of County Court because that court erroneously required that the additional restraints be coercive (majority opn, at 102).

That reading of County Court's decision is possible only if one ignores the disjunctive language of County Court's holding, as the majority has done (see, majority opn, at 101). County Court did use a second test — whether the evidence showed that the interview was "inherently coercive" — but the two tests were clearly stated in the alternative. If any legal error occurred, it was to the benefit of defendant, for he was given the opportunity to prevail under the added restraint test or the "inherently coercive" test. The evidence in the record supports the court's findings under either test and the Appellate Division affirmed the suppression court's findings.^[1] The import of all of this is obvious: Having applied the right test and been affirmed by the Appellate Division, County Court's decision is not subject to further review by this Court unless there is no evidence in the record to support its conclusions (People v Centano, 76 N.Y.2d 837; NY Const, art VI, § 3 [a]; CPL 470.35).

Nor can the majority find support for its decision to remit in the fact that the hearing record is, in its view, incomplete (majority opn, at 102). When a court applies the wrong legal test to the facts, as the majority claims happened here, the accepted remedy is to remit for the hearing court's application of the proper test to the facts already on the record (see, People v Morales, 65 N.Y.2d 997, 998). That is an entirely different matter from what the majority does in this case: Remit for a new fact-finding hearing. Admittedly, the record is scant, but that is no more grounds for a new hearing than the failure by a moving party to make its case in any proceeding. Manifestly, when a full and fair hearing has been provided, as it was here, gaps in the record are to be construed against the party with the burden of proof — not treated as a basis for giving that party a second chance to make the case.^[2]

In discussing how the new suppression hearing should examine the evidence, the majority implies that it is the burden of the People to show that defendant was not subjected to added restraints and therefore not in custody (majority opn, at 102). That reverses well-settled law. The burden of proving custody is on defendant, as the courts below properly held (see, United States v Charles, 738 F.2d 686, 692). The point was made clear in Berkemer v McCarty (468 U.S. 420), the principal decision relied upon by the majority today. The Supreme Court stated there: "[R]espondent has failed to demonstrate that, at any time between the initial stop and the arrest, he was subjected to restraints comparable to those associated with a formal arrest" (at 441 [emphasis added]). The Federal rule stated in Berkemer and Charles is consistent with the rule in New York, which holds that the burden of persuasion on a suppression motion remains with defendant (People v Di Stefano, 38 N.Y.2d 640, 652). Thus, when reviewing the record the Court must determine if defendant established he was in custody at the time he was questioned.

When the burden is properly assigned, it is manifest that the evidence in the record supports the suppression court's factual conclusion that custody above and beyond incarceration — i.e., added restraints — was not proved. The sole witness at the suppression hearing was Michael Rhodes, the officer who questioned Alls. Except for Rhodes' testimony that he "took" defendant from his dormitory area to the basement recreation area, there is no evidence whatsoever of the circumstances surrounding Alls' movement. Defendant presented no evidence whether he felt compelled to go, whether he wanted to go somewhere private, whether Rhodes issued an order to him, or whether some other circumstance was involved (cf., People v Ripic, 182 AD2d 226, 231-232, appeal dismissed 81 N.Y.2d 776). Defendant likewise failed to make any record of the physical circumstances of the basement — for instance, whether public access to the recreation room was cut off, whether the room was physically isolated from the rest of the dormitory, or whether others were passing nearby during the course of the interview.

Thus, the record before the County Court Judge supported his conclusion that Officer Rhodes had applied no added restraints to defendant's freedom in procuring a place where he and defendant could speak with some privacy. Had some other Judge been presiding, perhaps a different determination would have been made from this evidence. But that is not the test. Our review is limited by the State Constitution to a determination of whether there is evidence in the record to support the suppression court's finding. There is, and therefore our review of the facts should be at an end.

But the problems with the record and appellate procedure aside, I find myself unable to accept the majority's added restraints test. While this approach properly avoids the per se rule of Mathis v United States (391 U.S. 1) — that incarceration itself is custody for Miranda purposes — I question how workable the test will be for the courts and for prison officials. For well-founded security reasons, prisoners are routinely subjected to certain restraints when they are being moved to a new location. At other times, for their own protection, inmates may be removed from their living area so they can speak freely about incidents under investigation without the risk of being overheard inculpating fellow inmates. In fact, prison officials are required by State regulation to investigate incidents "as quietly and routinely as possible" (7 NYCRR 251-1.1). Accordingly, as a matter of practicality and regulation, and for the safety of the inmate, officials will often be required to move them to less public settings for questioning. Arguably, they will be imposing additional restraints by doing so, yet courts and prison officials might rightly doubt whether such routine prison procedures, without more, create the kind of inherently coercive atmosphere Miranda was designed to combat. Thus, the added restraints test poses a dilemma: If it is applied literally, the difference between it and the per se rule of Mathis will be negligible; if it is not, the term "added restraint" becomes largely meaningless. The test, as applied over time, will likely be rendered overinclusive or hollow. Neither the rights of prisoners nor the needs of prison administration will be well served in either case.

II

I believe the Supreme Court's decision in Illinois v Perkins (496 U.S. 292) offers a better test for resolving this and other cases involving prison inmates.^[3]

A.

The problem is best defined by reexamining the basic principles underlying Miranda's requirement that a person must be advised of his or her rights before being subjected to custodial interrogation. The decision arose out of a concern that those interrogated while isolated in custody might incriminate themselves in the face of compelling physical or psychological pressure (Miranda v Arizona, 384 U.S. 436, 444-450). Accordingly, Miranda does not foreclose all questioning undertaken without appropriate warnings, but instead applies only to "custodial interrogation" (id., at 444; see also, Illinois v Perkins, supra, at 296). Following the Miranda decision, the courts were faced with the problem of defining what exactly constitutes custodial interrogation — and therefore implicates the concerns underlying the decision — amid the varied circumstances in which citizens encounter police. What emerged was the two-prong test of custody and interrogation. In determining whether someone was "in custody", the relevant inquiry is whether a reasonable person in the suspect's position would have believed he was restrained to a degree associated with formal arrest (Berkemer v McCarty, 468 U.S. 420, 442, supra; see also, California v Beheler, 463 U.S. 1121, 1125; 1 LaFave and Israel, Criminal Procedure § 6.6, at 105 [1991 Pocket Part]). The "interrogation" prong, on the other hand, was drawn broadly: Interrogation occurs whenever the words and actions of the police are likely to elicit an incriminating response (Rhode Island v Innis, 446 U.S. 291, 300-301).

Under that traditional analysis, all prisoners are technically "in custody", restrained to a degree associated with formal arrest and never at liberty to leave an interview conducted by prison officials. Thus, to analyze prisoner interrogation in traditional Miranda terms, as the Chief Judge does today, is to say that the custody prong is met per se whenever an inmate is interrogated by officials. Because the interrogation prong is so easily met — it is hard to imagine how prison officials can look into a complaint or incident without asking questions likely to elicit incriminating responses — adherence to the traditional test would require that formal warnings be given prior to the myriad inquiries that occur daily in prison life, no matter how casual, routine or nonthreatening the questioning may in fact be. That, of course, is a far more expansive application of the Miranda requirement than the courts have applied in nonprison settings (see, e.g., Berkemer v McCarty, supra; People v English, 73 N.Y.2d 20; People v Yukl, 25 N.Y.2d 585).

Such an outcome is inconsistent with both the underlying purpose of Miranda — curbing coercive interrogation by investigators — and with the plain language of the decision: "General on-the-scene questioning as to facts surrounding a crime or other general questioning of citizens in the fact-finding process is not affected by our holding" (Miranda v Arizona, 384 U.S. 436, 477, supra). Therein lies the dilemma posed by this case: On the one hand, general questioning and fact finding expressly fall outside the Miranda rule, for such inquiries serve important public purposes and do not implicate concerns of coercion. On the other, when the incident being investigated has occurred inside prison walls, the general questioning and fact finding necessarily take place in a custodial setting where concerns about coercion may be implicated.

B.

There can be no doubt that questioning by prison officials sometimes takes on the quality of a police-dominated interrogation, and Miranda should obviously apply in those cases. But at other times the questioning of an inmate has decidedly more in common with the sort of routine investigative questioning police engage in outside prison which does not require warnings (see, People v English, 73 N.Y.2d 20, supra [role that parole officer is playing at time of questioning is relevant to Miranda inquiry]; People v Yukl, supra [warnings not required to precede station house interrogation]). Plainly put, the traditional two-prong analysis is incapable of distinguishing between those two types of prison inquiries and will dictate suppression in circumstances far removed from the legitimate concerns that gave birth to the Miranda rule. Prisoners certainly retain the right to be free from the kind of coercive interrogations that Miranda warnings are intended to deter, but it is equally clear that the analysis of when Miranda will apply must necessarily reflect the distinctive context of life in prison. As the Supreme Court has said: "Fidelity to the doctrine announced in Miranda requires that it be enforced strictly, but only in those types of situations in which the concerns that powered the decision are implicated" (Berkemer v McCarty, 468 U.S. 420, 437, supra).

The idea that the traditional two-prong analysis is unworkable in a prison context is not merely a matter of common sense. It flows directly from the Supreme Court's most recent ruling in a Miranda case involving an incarcerated suspect. In Illinois v Perkins (496 U.S. 292, supra), the Court ruled that a trial court could admit into evidence a confession made by an inmate to an undercover agent posing as a fellow prisoner even though no Miranda warnings were given. Under the traditional analysis, Perkins was undoubtedly in custody and interrogated. In fact, the Supreme Court acknowledged that he was (see, at 297, 299). Yet, the Supreme Court concluded that no Miranda warnings were required. Rather than mechanically apply the traditional two-prong analysis, the Court instead looked to Miranda's original purposes and advanced another test: Whether the circumstances of the encounter "implicate[d] the concerns underlying Miranda" — in other words, whether the risk of coercion was present (at 296).^[4]

As the majority correctly points out, Perkins forecloses the use of the traditional two-pronged custody and interrogation test (majority opn, at 97-98). Thus, defendant's reliance on Mathis v United States (391 U.S. 1, supra), where Miranda was applied to a tax agent's questioning of an inmate, is particularly misplaced. Mathis was incarcerated; Perkins was incarcerated. Mathis was interrogated; Perkins was interrogated. If the traditional Miranda analysis operative in Mathis were still good law, the result in Perkins should have been suppression. Instead, the Supreme Court expressly distinguished the two cases on a basis unrelated to either the custody test or the interrogation test: The fact that there was no "possibility that the suspect might feel coerced" when the interrogator was a fellow prisoner (Illinois v Perkins, supra, at 299). Like the majority, I too conclude that the Perkins Court intended a different analysis to apply in prison settings when it said: "The bare fact of custody may not in every instance require a warning even when the suspect is aware that he is speaking to an official, but we do not have occasion to explore that issue here" (at 299).

C.

The Supreme Court's methodology for distinguishing Mathis in Perkins articulates a principled basis for determining when Miranda warnings should be required in prison cases. The test is whether a "danger of coercion result[ed] from the interaction of custody and official interrogation" (Illinois v Perkins, supra, at 297; see, Carr v State, 840 P2d 1000 [Alaska]). Because Miranda is a prophylactic rule and may require suppression even in the absence of actual coercion, the analysis does not turn on a case-by-case analysis of the voluntariness of the statement or the subjective state of mind of the inmate. Instead, it is the danger or possibility of coercion — not coercion itself — that triggers the Miranda requirement.

This is precisely the analysis that the suppression court stated and applied as an alternative to the added restraints test. I would hold that it was correctly applied and that there was evidence on the record to support the court's conclusions, as affirmed by the Appellate Division. Officer Rhodes was the sole witness at the suppression hearing, and what emerges from the record is a picture of an officer engaging in rudimentary fact finding about a possible rules infraction, uncertain whether Alls had even been a participant in the previous night's incident. Officer Rhodes testified that before coming to see Alls, he had been given only "limited" background information or "none" at all. So limited was his knowledge, in fact, that he had to ask Alls whether Alls had been involved in the fight. He did not raise the sodomy allegation because, he testified, "[Alls] didn't mention it and * * * I didn't even know it was an issue."^[5] Nothing in the record indicates that Rhodes went into the interview viewing it as part of a criminal investigation. In fact, he testified that at the conclusion of the interview he intended to handle the matter through the prison's internal disciplinary proceedings.

Beyond that, nothing about the conditions of the interview suggests a danger of coercion. Alls was questioned in his own residential unit, a dormitory-like building in a low security facility. Officer Rhodes testified that he "took" Alls to the building's recreational/kitchen area for privacy, but the record is silent as to whether Alls volunteered to go, expressed reluctance or had any reaction at all. There is no reason to assume from this that the move was made to intimidate Alls any more than we can assume it was made to protect Alls from being overheard should he "snitch" on one of his fellow prisoners. While prison regulations require inmates to go with officers when so ordered, the record does not indicate whether any order was given. Nor does the record show that the area where the questioning was conducted was physically restrictive, isolated or intimidating in any other way.

In short, while Alls was technically in custody and interrogated, the record falls well short of establishing that the danger or possibility of coercion was present in these circumstances. Given that, I conclude Miranda warnings were not required here.

III

Miranda remains a vital source of protection for all our citizens, including incarcerated individuals. Its vitality depends, however, on clear and principled application of its rule in cases where concerns about coercive custodial interrogation are truly implicated. Because I believe no danger of coercion existed in the circumstances under which the questioning occurred here, and no legal error was made by the courts below, I would vote to affirm the order of the Appellate Division.

Order modified, etc.

NOTES

[*] This analysis points up the error of County Court's heavy reliance, in reaching its result, on the fact that Rhodes did not advise defendant that he was not free to leave or to break off the interview. The converse was the truly relevant factor, namely, that Rhodes (backed by the full authority of prison regulations) may have directed defendant to go with him to an isolated location "without indicating in any way that he was free to depart" (Florida v Royer, supra, at 501).

[*] My colleagues overstate the result of a per se custody rule. Miranda warnings would not be required prior to casual, spontaneous or routine interchanges between prison officials and inmates (majority opn, at 100; dissenting opn, at 113, 119). Custody is only half the equation — there must also be "interrogation," as occurred here.

[1] All of these facts from the record were cited by County Court in support of its dual conclusions that there were no coercive circumstances or added restraints: (1) that defendant was not "restrained, shackled or handcuffed"; (2) that the questioning was brief; (3) that the setting of the basement was "office-like"; (4) that defendant was not advised that his immediate freedom of movement was curtailed; and (5) that there was no proof that "he was not free to break off the interview".

[2] People v Dodt (61 N.Y.2d 408) and People v Havelka (45 N.Y.2d 636) cited by the majority are not in point. In those cases, the Court addressed situations in which a party had either been denied a hearing or a fair opportunity to present its evidence. Defendant claims neither here.

[3] Though the added restraints test has been widely accepted in other jurisdictions, the Supreme Court, although presented with opportunities to do so, has never adopted the rule (see below). The Court has denied or dismissed certiorari in at least eight cases where it was used. In the absence of Supreme Court precedent embracing the added restraint test, I believe Perkins offers the clearest guidance on Federal law relevant to the prison Miranda issue at the present time.

Federal decisions adopting an added restraints test in prison Miranda cases include United States v Rogers (899 F.2d 917 [10th Cir], cert denied 498 U.S. 839); Leviston v Black (843 F.2d 302 [8th Cir], cert denied 488 U.S. 865); United States v Cooper (800 F.2d 412 [4th Cir]); United States v Conley (779 F.2d 970 [4th Cir], cert denied 479 U.S. 830); United States v Scalf (725 F.2d 1272 [10th Cir]); Cervantes v Walker (589 F.2d 424 [9th Cir]) (see also, United States v Willoughby, 860 F.2d 15, 23-24 [2d Cir]; United States v Cofield, 1992 US App LEXIS 8284 [6th Cir]). High courts of our sister States have reached a similar result (see, State v Post, 118 Wash 2d 596, 826 P2d 172; Whitfield v State, 287 Md 124, 411 A2d 415, cert dismissed 446 U.S. 993; State v Bradley, 236 Neb 371, 461 NW2d 524, cert denied 502 US ___, 112 S. Ct. 143; State v Bradley, 42 Ohio St 3d 136, 538 NE2d 373, cert denied 497 U.S. 1011, supra; State v Vickers, 159 Ariz 532, 768 P2d 1177, cert denied 497 U.S. 1033; People v Patterson, 146 Ill 2d 445, 588 NE2d 1175, cert denied ___ US ___, 121 L. Ed. 2d 73; Beamon v Commonwealth, 222 Va 707, 284 SE2d 591).

[4] It is difficult to understand Chief Judge Kaye's portrayal of Perkins as an "exception" to Miranda for two reasons: First, nowhere does the Supreme Court portray it as such. When the Court has intended a Miranda exception, it has said so explicitly (see, New York v Quarles, 467 U.S. 649, 656). Second, the Supreme Court's reasoning in Perkins is aimed not at drawing a necessary exception to Miranda (compare, New York v Quarles, supra), but at explaining Miranda and showing how Miranda and Perkins are in fact consistent.

[5] Alls told Rhodes that he did not initiate the fight, was not the aggressor but had acted only in self-defense. The jury, by its verdict, did not believe the justification defense.