People v. Henderson: Prior Drug Possession Is Inadmissible Molineux Evidence Absent a Logical, Issue-Specific Nexus; Step-One Relevance Is Reviewed De Novo

I. Introduction

People v Henderson (2026 NY Slip Op 01627 [Ct App Mar. 19, 2026]) addresses a recurring prosecution tactic in narcotics cases: offering a defendant’s prior drug possession (or drug conviction) to prove knowing or constructive possession and/or intent to sell in a later, separate incident. The Court of Appeals reiterates that New York’s Molineux doctrine generally bars “propensity” proof and emphasizes that the People must show a logical connection between the prior act and a specific, material issue in the present case—something more than “he did it before.”

Background. A SWAT team executed a search warrant at a Rochester home in October 2019. Cocaine and paraphernalia were found in multiple rooms. Defendant was in the south bedroom with his girlfriend; his father was in the west bedroom; his grandmother was in the living room. Defendant was indicted on multiple counts including Penal Law § 220.16 (1) (possession with intent to sell), § 220.16 (12) (weight-based), and § 220.50 (2), (3) (drug paraphernalia).

Key evidentiary issue. Over defense objection, the trial court admitted Molineux evidence of defendant’s 2017 drug incident: police recovered cocaine from a vehicle registered to defendant, resulting in a plea to attempted criminal possession of a controlled substance in the third degree. At trial, the People argued the prior conviction showed “knowing possession and intent.”

Procedural posture. The Appellate Division affirmed (234 AD3d 1254 [4th Dept 2025]) over a two-Justice dissent. The Court of Appeals reverses, holds the Molineux ruling erroneous as a matter of law, finds the error not harmless, and orders a new trial.

II. Summary of the Opinion

The Court of Appeals holds:

  • Evidence of defendant’s 2017 possession of cocaine in his car was not “logically connected to some specific material issue” in the 2019 case beyond propensity, and thus failed step one of the Molineux analysis (citing People v Cass, 18 NY3d 553 [2012]; People v Molineux, 168 NY 264 [1901]).
  • The evidence was inadmissible to prove constructive/knowing possession because the prior incident and the charged incident differed in location, circumstances, theory of possession, and temporal proximity, rendering the similarity “only in broad respects” (drawing on People v Telfair, 41 NY3d 107 [2023]).
  • The evidence was inadmissible to prove intent because intent to sell was readily inferable from the contraband and paraphernalia and was not genuinely placed in issue by the defense, aligning this case with the general rule in People v Alvino, 71 NY2d 233 [1987] and the holding in People v Blair, 90 NY2d 1003 [1997].
  • The error was not harmless under People v Crimmins, 36 NY2d 230 [1975] and People v Arafet, 13 NY3d 460 [2009], because (without the propensity evidence) proof of defendant’s dominion and control over the drugs was not overwhelming, and there was a significant probability the jury might have acquitted.

The dissent would have upheld admissibility for knowledge and intent and, alternatively, would have found any error harmless given the split verdict and limiting instructions.

III. Analysis

A. The Governing Framework: Molineux’s Two-Step Analysis and Appellate Review

The Court re-centers New York’s distinctive approach to “other crimes” evidence. Under the Molineux rule, evidence of uncharged crimes or prior misconduct is inadmissible when it “tends only to demonstrate the defendant’s propensity” (People v Cass, 18 NY3d at 559; People v Molineux, 168 NY 264). While exceptions exist (intent, motive, knowledge, common scheme/plan, identity), the Court warns against letting exceptions “swallow the rule,” echoing its concern about “forbidden propensity inferences” (People v Telfair, 41 NY3d at 114).

A critical doctrinal clarification—stated explicitly here—is the division of labor between relevance and prejudice:

  1. Step One (Relevance to a non-propensity material issue): The People must identify a “material issue, other than the defendant’s criminal propensity, to which the evidence is directly relevant” (People v Denson, 26 NY3d 179, 185 [2015]). Whether that burden is met is treated as a question of law reviewed de novo (People v Cass, 18 NY3d at 560 n 3).
  2. Step Two (Probative value vs. prejudice): Only after step one is satisfied does the court balance probative value against prejudicial effect, reviewed for abuse of discretion (People v Cass, 18 NY3d at 560 n 3).
  3. Harmless error: If erroneously admitted, appellate courts apply harmless error analysis (People v Telfair, 41 NY3d at 114).

In practical terms, Henderson strengthens the gatekeeping function of step one: it is not enough that prior misconduct could be rhetorically linked to an element (e.g., “knowledge”); the prior act must be directly relevant to a specific disputed issue in the current proof.

B. Precedents Cited and Their Role in the Court’s Decision

The opinion is anchored in an extended line of Court of Appeals cases restricting prior-act proof in drug and weapon possession prosecutions. The cited precedents perform three main functions: (1) reaffirm the anti-propensity principle, (2) define the narrowness of exceptions, and (3) frame harmless error.

1. Core Molineux doctrine and the anti-propensity principle

  • People v Molineux (168 NY 264 [1901]): The foundational rule—other-crimes evidence is excluded unless tied to a specific issue beyond character, and courts must guard the presumption of innocence. Henderson quotes Molineux’s historical rationale and uses it to rebut the dissent’s broader admissibility approach.
  • People v Cass (18 NY3d 553 [2012]): Supplies the modern formulation—evidence must be “logically connected” and “directly relevant” to a specific material issue; also provides the step-one legal-question framing and de novo review language adopted and applied here.
  • People v Hudy (73 NY2d 40 [1988]): Emphasizes the “very real danger” that juries overvalue prior bad acts; Henderson cites Hudy to underscore why thin “similarity” is insufficient when the real value is propensity.
  • People v Sin (44 NY3d 455 [2025]) and People v Weinstein (42 NY3d 439 [2024]): Used to refute the dissent’s suggestion that denial of guilt alone opens the door; Henderson treats these cases as continuing the requirement of a logical link to a specific material issue and rejects “similar crime” reasoning as a substitute.

2. Defining the permissible “knowledge/possession” use: Telfair and constructive possession doctrine

  • People v Telfair (41 NY3d 107 [2023]): The central analogue. In Telfair, prior gun incidents were held inadmissible to show knowing possession of different guns in a later case, where the incidents had different circumstances and were remote. Henderson directly imports that logic: prior drug possession generally cannot prove knowing possession of different drugs in a different setting unless the connection is tight and issue-specific.
  • People v Manini (79 NY2d 561 [1992]) and People v Headley (74 NY2d 858 [1989]): These cases articulate that “mere presence” and “mere knowledge” do not establish constructive possession; Henderson uses them in the harmless-error discussion to show the weakness of the remaining proof once the prior-act evidence is removed.
  • People v Cota (199 AD3d 1237 [3d Dept 2021]): Cited for the same proposition regarding presence/knowledge; supports the Court’s conclusion that the case against defendant was not overwhelming.

3. Prior drug possession and “intent” in drug cases: Alvino, Blair, Valentin, Harvey, Crandall

  • People v Alvino (71 NY2d 233 [1987]): Establishes the “general rule” in drug sale cases: intent is often readily inferable from the sale itself and prior uncharged crimes are usually neither necessary nor permissible. Henderson applies that general principle to a possession-with-intent prosecution where trafficking paraphernalia made intent obvious. Alvino also provides the contrasting scenario: when a defendant affirmatively reframes the case (admitting possession but denying intent to sell), prior acts may become genuinely probative rebuttal. Henderson finds no such reframing here.
  • People v Blair (90 NY2d 1003 [1997]): The closest drug-case parallel. Blair held it error to admit a prior drug transaction where intent/knowledge were not truly in issue and the evidence served mainly to show propensity. Henderson treats defendant’s theory—“the drugs are my father’s”—as Blair-like: the 2017 incident did not refute that defense and instead suggested character.
  • People v Valentin (29 NY3d 150 [2017]): Cited as an example of the general inadmissibility of prior drug possession to prove intent/knowledge on another occasion, setting the doctrinal baseline that Henderson enforces.
  • People v Crandall (67 NY2d 111 [1986]): Reinforces that evidence is inadmissible when it only shows criminal “bent”; Henderson uses it to stress that similar wrongdoing does not automatically establish a non-propensity purpose.
  • People v Harvey (235 NY 282 [1923]): Quoted for the principle that when intent is “so apparent” from the charged conduct, introducing other offenses cannot be justified to prove intent.

4. The list of exceptions and the prosecution’s burden: Leonard and Denson

  • People v Leonard (29 NY3d 1 [2017]): Provides the enumerated exceptions (intent, motive, knowledge, common scheme/plan, identity) and the caveat that intent-based use may fail step one when intent is easily inferable.
  • People v Denson (26 NY3d 179 [2015]): Supplies the formulation that the People must identify a material non-propensity issue and bear the burden to show direct relevance. Henderson relies on Denson to reject the People’s broad, catch-all pretrial assertions.

5. Harmless error: Arafet, Crimmins, Smith, and the dissent’s reliance on split verdicts

  • People v Crimmins (36 NY2d 230 [1975]) and People v Arafet (13 NY3d 460 [2009]): Establish the two-part test: overwhelming proof absent the error, and no significant probability the verdict would differ. Henderson finds both prongs unmet given the contested constructive possession proof.
  • People v Smith (33 NY3d 454 [2019]): Cited for the proposition that where proof is not overwhelming, the error is not harmless.
  • People v Till (87 NY2d 835 [1995]): Invoked by the dissent to argue that a “discerning” split verdict suggests limited prejudice. The majority, however, focuses on the weakness of the non-Molineux proof.

6. Limiting instructions and discretion: Morris, Frumusa, Israel, Small

  • People v Morris (21 NY3d 588 [2013]), People v Frumusa (29 NY3d 364 [2017]), People v Israel (26 NY3d 236 [2015]), and People v Small (12 NY3d 732 [2009]): These authorities appear primarily in the dissent, emphasizing deference to trial courts on balancing and the mitigating role of limiting instructions. The majority does not dispute that limiting instructions matter; it concludes they cannot cure evidence that fails step one (i.e., evidence that is, at bottom, propensity).

7. The “close nexus” example: Hayward

  • People v Hayward (42 NY3d 753 [2024]): In a footnote, the majority offers Hayward as the counterexample where prior/uncharged conduct is admissible: an uncharged drug sale by the defendant occurred hours earlier in the same apartment where the charged drugs were found, creating a direct logical connection to constructive possession and intent. Henderson distinguishes Hayward to mark the boundary: proximity in time and place can supply the missing “specific material issue” nexus—here it did not.

C. Legal Reasoning: Why the 2017 Incident Failed Step One

1. Constructive or knowing possession

The Court treats the “main issue” as whether defendant had constructive/knowing possession of contraband spread across multiple rooms in a residence occupied by several people. The 2017 car incident was deemed irrelevant because it did not make it more likely—except through propensity—that defendant exercised dominion and control over drugs in a shared apartment in 2019.

The Court’s logic is comparative and fact-driven:

  • Different location: car vs. apartment.
  • Different circumstantial proof: drugs in a vehicle registered to defendant (with IDs and mail) vs. drugs in a residence with multiple occupants.
  • Different possession theory: individual possession vs. joint/constructive possession among several people.
  • Temporal distance: approximately two years.
  • Different packaging details: “little baggies” in 2017 vs. “vials” and other forms in 2019, plus unpackaged product and a large “cookie” elsewhere.

Importantly, the Court rejects the People’s attempted “similarity” anchors (e.g., claims that drugs and mail were “mixed in with” each other) as unsupported or too tenuous. Once those asserted parallels fell away, the remaining “commonality” was simply that defendant was accused of cocaine possession in both incidents—the forbidden propensity inference.

2. Intent

The Court treats “intent” as failing step one for a different reason: even if intent is technically an element, it must be a genuinely material issue that is not already obvious from the charged conduct. Here, intent to sell was “readily inferable” from the quantity and, especially, the presence of scales, packaging materials, and “workstations” for breaking down cocaine into consumer quantities—plus the absence of drug-use paraphernalia. This tracks the principles in People v Alvino and People v Harvey.

The Court also stresses the defense posture. Defendant did not claim “these drugs are mine but for personal use,” nor did he concede possession while disputing intent (the Alvino “confession and avoidance” scenario). Instead, defendant’s theory was non-possession: the drugs belonged to his father. In that posture, a prior possession conviction does not rebut the defense theory; it only invites the jury to decide “he’s the kind of person who has drugs,” which People v Blair condemned as impermissible propensity reasoning.

Because the People failed at step one, the majority does not proceed to step two balancing—reinforcing the idea that prejudice-balancing discretion cannot rescue evidence that lacks a non-propensity relevance nexus.

D. Harmless Error: Why the Verdict Could Have Been Different

Applying People v Crimmins and People v Arafet, the Court finds the error not harmless because the remaining proof of constructive possession was not overwhelming. It underscores that “mere presence” in an apartment where drugs are discovered is insufficient (People v Manini; People v Headley; People v Cota).

The Court characterizes the non-Molineux case against defendant as thin: defendant and his girlfriend were asleep in a room where contraband was not in open view. The majority rejects the dissent’s reliance on residue in shoebox paraphernalia, noting defendant was not charged with possessing cocaine in those shoeboxes. In that evidentiary context, introducing a prior cocaine conviction risked becoming the most intuitively powerful proof—precisely what Molineux aims to prevent— and thus created a “significant probability” the verdict was affected.

E. Potential Impact

1. A tighter “nexus” requirement for prior drug acts.
Henderson makes it harder for the People to use prior drug possession to prove knowledge/constructive possession in later cases involving different premises and multi-occupant settings. The decision signals that the “knowledge” label cannot substitute for a concrete explanation of how the prior act illuminates the contested factual question (e.g., dominion and control over a particular place or container).

2. Intent evidence is especially constrained in drug trafficking prosecutions.
Where packaging, scales, and trafficking paraphernalia make intent to sell obvious, Henderson reinforces Alvino’s general rule: prior drug acts are usually unnecessary and therefore impermissible. Expect trial courts to scrutinize whether “intent” is truly disputed or merely nominally an element.

3. Appellate posture: step one as a legal question.
By reiterating that step one is reviewed de novo, Henderson encourages more robust appellate correction of “exception creep.” Trial courts retain discretion at step two, but only after the People satisfy a genuinely issue-specific relevance showing.

4. Practical litigation consequences.
Prosecutors will likely focus on:

  • evidence closer in time and place (the People v Hayward model),
  • proof tying the defendant to the specific location or container (keys, leases, admissions, surveillance), and
  • alternative non-character uses (e.g., identity/common scheme) only where the factual pattern is distinctive and genuinely contested.

Defense counsel, conversely, will cite Henderson to challenge boilerplate Molineux motions that list every exception without a tailored theory, and to argue that “constructive possession” disputes do not automatically open the door to prior possession convictions.

IV. Complex Concepts Simplified

  • Molineux evidence: Proof that the defendant did something wrong on another occasion (uncharged misconduct or prior crimes). It is usually excluded because jurors might convict based on “bad character” rather than proof of the charged crime.
  • Propensity inference: The forbidden reasoning: “He did it before, so he likely did it again.”
  • Constructive possession: Possession without physical holding—requiring proof the defendant exercised dominion and control over the contraband or the place where it was found. Mere presence or awareness is not enough.
  • Step one vs. step two (Molineux): Step one asks: “Is this evidence truly relevant to a specific, disputed issue other than character?” Step two asks: “Even if relevant, is it too unfairly prejudicial?” Henderson emphasizes that step two discretion cannot be reached if step one fails.
  • Harmless error: Even if the judge made a mistake, the conviction stands only if the remaining proof is overwhelming and the mistake likely did not matter. Here, the Court found the mistake likely mattered.

V. Conclusion

People v Henderson reinforces a strict, issue-specific application of Molineux in drug possession prosecutions: prior drug possession (and even a prior conviction) is inadmissible to prove constructive/knowing possession or intent to sell unless it is directly relevant to a specific material issue beyond propensity. The Court’s reasoning draws heavily on People v Telfair (rejecting “broad similarity” as a gateway for knowledge) and on People v Alvino and People v Blair (limiting intent-based use where intent is obvious or not genuinely disputed). Finally, the decision underscores that when constructive possession is the contested fulcrum, an improperly admitted prior drug act can rarely be brushed aside as harmless.