No “Scene‑of‑Crime” Exception to Miranda: Handcuffs + On‑Scene “Investigatory” Questions Constitute Custodial Interrogation — People v. Robinson (N.Y. 2025)

Introduction

In People v. Robinson, 2025 NY Slip Op 05871 (Oct. 23, 2025), the New York Court of Appeals addressed whether police may ask “what happened?” and similar questions to a handcuffed suspect at the scene of a suspected crime without first providing Miranda warnings. The Court squarely rejected the notion—relied upon by the lower courts—that there exists a categorical “investigatory questioning” exception for on‑scene exchanges. Emphasizing that the presence of handcuffs carries “very substantial weight” in the custody analysis, the Court held that the defendant was subjected to custodial interrogation without Miranda warnings and that his inculpatory statement should have been suppressed. Nevertheless, the Court affirmed the conviction because the Miranda error was harmless beyond a reasonable doubt in light of overwhelming independent evidence.

The decision clarifies the interplay between two recurring doctrines: (1) the boundary between “street inquiry” and custodial interrogation, and (2) the scope of permissible on‑scene questioning following volatile events. It harmonizes older New York authority (especially People v. Huffman) with modern federal standards on interrogation (Rhode Island v. Innis) and reaffirms the Court’s recent guidance in People v. Cabrera on the custodial import of handcuffing.

Case Background

Just before midnight in Times Square in 2020, police observed the defendant, Savion Robinson, and another man fighting near Broadway and 46th Street. Officers intervened, handcuffed both men, and began questioning them about who started the fight and why. The complaining witness accused Robinson of stealing his bicycle and assaulting him; after several minutes of on‑scene questioning while handcuffed, Robinson admitted he punched the complainant. Officers had not administered Miranda warnings before eliciting the admission.

At a suppression hearing, footage from surveillance and body‑worn cameras showed the sequence: police separated and handcuffed both men, repeatedly asked Robinson variations of “What happened?” and “Let me hear your side,” and reassured the complainant, “we’re gonna figure this out.” Later, after reviewing video corroboration, police arrested Robinson. At a bench trial, Robinson was acquitted of second‑degree robbery but convicted of third‑degree robbery.

The trial court denied suppression, ruling the questioning “investigatory” rather than “custodial.” The Appellate Division affirmed, finding that a reasonable innocent person would understand that the police were still gathering information and, alternatively, that any error was harmless. The Court of Appeals granted leave.

Summary of the Opinion

Writing for a unanimous Court, Judge Cannataro held:

Robinson was in custody for Miranda purposes when he was handcuffed, surrounded by uniformed officers, questioned for approximately eight minutes, and given no indication that the restraints were temporary or that he was free to leave.
He was interrogated: officers asked express questions about the altercation and employed words and actions that they should have known were reasonably likely to elicit an incriminating response (Innis standard).
There is no categorical “investigatory questioning” exception to Miranda merely because police are still ascertaining whether a crime occurred or because the questions happen on‑scene. Huffman does not create such an exception and must be read consistently with Innis; where public safety is at issue, Quarles provides the relevant framework.
Although the statement should have been suppressed, the error was harmless beyond a reasonable doubt given overwhelming independent evidence (victim testimony, officer testimony, and surveillance/body‑cam video) and no reasonable possibility that the statement affected the outcome (Crimmins; Best).

Detailed Analysis

1) Precedents Cited and their Influence

Miranda v. Arizona, 384 U.S. 436 (1966): The bedrock rule that statements obtained during custodial interrogation are inadmissible unless preceded by Miranda warnings. Robinson applies Miranda straightforwardly: the key question is whether Robinson was (a) in custody and (b) interrogated.
People v. Cabrera, 41 NY3d 35 (2023): The Court reiterates Cabrera’s central teaching that handcuffing is not a per se rule for custody but will, in “very few circumstances,” not amount to custody; handcuffs must receive “very substantial weight.” Robinson’s facts mirror Cabrera—nighttime encounter, multiple officers, handcuffing precedes questioning—underscoring that Robinson’s freedom was restrained to a degree akin to formal arrest, with no indication of a brief or limited restraint.
Rhode Island v. Innis, 446 U.S. 291 (1980): Defines “interrogation” to include not only express questioning but also police words or actions that officers should know are reasonably likely to elicit an incriminating response. Robinson relies on Innis to treat repeated “What happened? Let me hear your side” questions to a handcuffed suspect as interrogation; the foreseeability of an incriminating response was high.
People v. Huffman, 41 NY2d 29 (1976): Often invoked to justify on‑scene, “investigatory” questions. Robinson clarifies that Huffman never created a categorical scene‑of‑crime exception; its analysis turned on exigent, volatile circumstances that today would be evaluated under the public‑safety exception (Quarles), not as a carve‑out from Miranda’s interrogation framework.
New York v. Quarles, 467 U.S. 649 (1984): Recognizes a public‑safety exception to Miranda for questions reasonably prompted by concern for safety (e.g., “Where is the gun?”). Robinson indicates that Huffman’s volatility discussion aligns with Quarles, and that safety exigency—not the mere fact of on‑scene questioning—determines whether Miranda can be temporarily bypassed.
United States v. Newton, 369 F.3d 659 (2d Cir. 2004): Notes that conditioning handcuff removal or leniency on a suspect’s cooperation is coercive and indicative of custody. Robinson echoes Newton, emphasizing how defendants reasonably perceive that continued restraint depends on answering police questions.
People v. Rodney P. (Anonymous), 21 NY2d 1 (1967), and People v. Shivers, 21 NY2d 118 (1967): Cited historically in Huffman to frame the “street inquiry” versus “custodial interrogation” distinction. Robinson preserves that distinction but reaffirms that the dispositive factor is custody, not the “investigatory” label.
Matter of Kwok T., 43 NY2d 213 (1977): States that routine street inquiry into suspicious conduct usually does not involve a significant deprivation of freedom. Robinson embraces this principle but holds that handcuffs and sustained questioning change the calculus: freedom is significantly restrained.
People v. Berg, 92 NY2d 701 (1999): Restates Miranda’s New York application; Robinson cites Berg for the basic rule that unwarned custodial statements are inadmissible.
People v. Paulman, 5 NY3d 122 (2005): Clarifies standard of review—custody is a mixed question of law and fact, and appellate review is deferential if supported by record. Robinson finds the record does not support the lower courts’ non‑custody conclusion, especially in light of Cabrera’s guidance.
People v. Crimmins, 36 NY2d 230 (1975), and People v. Best, 19 NY3d 739 (2012): Set out the harmless‑error standard. Robinson applies the two‑part inquiry: (1) overwhelming evidence of guilt, and (2) no reasonable possibility that the error contributed to the verdict.

2) The Court’s Legal Reasoning

a) Custody

The Court applies New York’s custody test: whether a reasonable person, innocent of wrongdoing, would believe they were not free to leave, and whether police restricted freedom to a degree associated with a formal arrest. Several facts proved decisive:

Robinson was handcuffed and surrounded by multiple uniformed officers for about eight minutes.
He was searched and questioned while the complainant repeatedly accused him of theft and assault in his presence.
No officer communicated that the restraints were temporary or would soon be removed.
The scenario resembled Cabrera: handcuffing pre‑questioning at night with multiple officers—conditions ordinarily amounting to custody.
The circumstances suggested that release from restraints might depend on cooperating—an inference of coercion noted by federal courts (Newton).

The Court reiterates that while handcuffs are not per se custody, they carry “very substantial weight,” and “very few circumstances” will fall outside custody once handcuffs are applied. The record did not support a non‑custody conclusion.

b) Interrogation

Applying Innis, the Court finds interrogation in two ways:

Express questions aimed at the incident (“What happened?” “Let me hear your side.” “Why’d you go after him like that?”) were designed to elicit admission or explanation.
Police words and actions—informing Robinson they had seen him strike the complainant and heard the complainant’s version—created a context where an incriminating response was reasonably foreseeable.

Because both custody and interrogation were present, Miranda warnings were required before eliciting the admission.

c) Rejecting an “Investigatory Questioning” Exception

The Court explicitly rejects the lower courts’ reliance on the “investigatory” label:

Interrogation is “investigatory” by nature; the label does not alter Miranda’s requirements.
New York jurisprudence distinguishes between coercive custodial interrogation and permissible street inquiry. The difference turns on custody, not the timing or setting of the questions.
Huffman does not stand for a scene‑of‑crime or “ongoing investigation” exception to Miranda; it concerned extremely volatile circumstances closer to Quarles’s public‑safety rationale, not a blanket permission to question handcuffed suspects without warnings.

d) Harmless Error

Although the admission should have been suppressed, the Court concludes the error was harmless beyond a reasonable doubt:

The People introduced compelling independent evidence: corroborated surveillance video, body‑worn camera footage, and unrebutted testimony from the victim and Officer Sasso.
Given this record, there was no reasonable possibility that the unwarned “I punched him” statement affected the bench verdict convicting of third‑degree robbery.

3) Impact and Prospective Significance

a) For Police Practices

No scene‑of‑crime carve‑out: Officers should not assume that on‑scene status or the fact‑finding posture permits questioning a handcuffed suspect without Miranda.
Handcuffs are a custody pivot: Once handcuffs are on, the safest practice is to Mirandize before asking about criminal conduct unless an immediate public‑safety concern justifies limited questions under Quarles.
Public‑safety exception is narrow: It must be tied to urgent safety risks (e.g., “Where is the weapon?”), not general narrative building (“What happened?”).
Body‑cam recordings cut both ways: They document safety justifications but also capture custodial conditions and interrogation—fuel for suppression motions if Miranda is omitted.

b) For Prosecutors

Expect more suppression litigation over on‑scene statements obtained while handcuffed.
Charging and trial strategy should not depend on unwarned on‑scene statements; prioritize independent corroboration (video, eyewitnesses).
When invoking harmless error, be prepared to show both overwhelming evidence and no reasonable possibility the statement affected the outcome, especially in jury trials.

c) For Defense Counsel

Robinson + Cabrera provide a robust framework to argue custody where handcuffs are used before questioning, even at the scene and even if officers say they are “figuring things out.”
Use body‑cam timelines to demonstrate sustained restraint, accusatory context, and foreseeable elicitation under Innis.
Challenge reliance on “investigatory questioning” and redirect analysis to custody and foreseeability; distinguish Quarles unless an immediate safety threat is documented.

d) For Trial Courts

Apply Cabrera’s instruction that handcuffs carry “very substantial weight”; resolve close custody calls in light of Robinson’s warning that few handcuffed scenarios are non‑custodial.
Avoid treating Huffman as a free‑standing exception; assess any on‑scene questions under Innis and, where asserted, under the narrow Quarles public‑safety exception.
When finding harmless error, apply Crimmins rigorously: the “overwhelming evidence” prong and the “no reasonable possibility of contribution” prong are distinct.

Complex Concepts Simplified

Custody: Think “arrest‑like restraint.” Would a reasonable person who did nothing wrong feel free to leave? Are restraints (like handcuffs), officer presence, and circumstances similar to a formal arrest? Handcuffs almost always mean “yes.”
Interrogation (Innis): Not just direct questions. If police words or actions are reasonably likely to make a suspect incriminate himself, that’s interrogation. Repeated “What happened? Tell me your side” to a handcuffed person counts.
“Investigatory questioning”: A label without legal force. The real questions are: Was the person in custody? Were police words or actions reasonably likely to elicit incrimination?
Public‑safety exception (Quarles): A short‑term allowance for questions needed to remove immediate danger (e.g., location of a gun). It doesn’t cover general fact‑gathering or narrative building.
Harmless error (Crimmins/Best): Even if a constitutional error occurred (e.g., Miranda violation), a conviction can stand if there’s overwhelming independent proof and no reasonable chance the error impacted the verdict.

Practical Checklists

For Officers

Before questioning about a crime, ask: Is the person handcuffed or otherwise restrained? If yes, treat as custody and give Miranda warnings unless a true public‑safety emergency exists.
If handcuffs are used briefly for safety, explain their temporary nature, minimize questioning to safety concerns, and remove restraints as soon as feasible if you intend to continue general inquiry.
Document any safety exigency contemporaneously (body‑cam narration can help) if relying on Quarles.

For Counsel

Map the encounter minute‑by‑minute from body‑cam: when handcuffs were applied, what was asked, and what was said about safety or release.
Identify statements by officers or circumstances suggesting release depends on cooperation (Newton), underscoring coercion.
Argue that Huffman does not excuse Miranda; reserve Quarles for genuine, documented safety threats.

Key Takeaways

No categorical “investigatory” exception: On‑scene questioning of a handcuffed suspect is custodial interrogation requiring Miranda warnings absent a narrow public‑safety rationale.
Handcuffs matter—immensely: Following Cabrera, handcuffs carry “very substantial weight,” and “very few” handcuffed scenarios will be non‑custodial.
Interrogation is broad under Innis: Repeated prompts to “tell your side” to a restrained suspect foreseeably elicit incrimination.
Huffman narrowed and harmonized: Its volatility rationale aligns with Quarles; it does not create a scene‑of‑crime carve‑out.
Harmless error remains a backstop: Robust independent evidence can preserve a conviction even where Miranda is violated, but the state bears a demanding burden.

Conclusion

People v. Robinson delivers a clear doctrinal message: Miranda’s protections do not yield to the “investigatory” aspirations of on‑scene policing once a suspect is handcuffed and subjected to questions reasonably likely to elicit incrimination. By integrating Cabrera’s emphasis on handcuffs, Innis’s interrogation standard, and Quarles’s limited safety exception, the Court eliminates any lingering ambiguity that Huffman created a scene‑of‑crime exception. The decision will recalibrate on‑scene police practices in New York, promote earlier administration of Miranda warnings when handcuffs are applied, and sharpen the litigation focus on custody and foreseeability rather than labels. Even as Robinson affirms the conviction under the harmless‑error doctrine, its core suppression holding marks a significant clarification in New York’s Miranda jurisprudence—one likely to guide police training, inform prosecutorial decisions, and frame defense suppression strategy in street‑encounter prosecutions for years to come.