Concept
Williamson and Collateral Statements
Statements against interest
Williamson and Collateral Statements
A statement against interest earns belief from what it costs the speaker. Rule 804(b)(3)(A) turns that idea into a test: would a reasonable person in the declarant's position have made the statement only if believing it true, because when made it was so contrary to the declarant's interests?1 The earlier sections applied the test to single remarks, so the unit under test was never in doubt. Confessions are rarely so tidy. A suspect questioned after arrest tells a story, and the story mixes remarks that damage the speaker with remarks about who planned the crime, who did the worst of it, and who deserves the blame. The rule speaks of a “statement” and does not say whether the test applies to the whole narrative or to each assertion inside it.
Williamson v. United States answers that question for the federal courts. The Supreme Court held that the exception covers only the remarks within a confession that are themselves genuinely self-inculpatory, so a remark that points at someone else does not come in merely because it sits inside a damaging narrative. The reason lies in the mechanism that drives the whole exception. The cost that guarantees sincerity attaches to a particular assertion. A sentence admitting the speaker's own part in a crime carries that cost, while a sentence placing the larger share of the blame on a partner costs the speaker nothing and may lighten the speaker's own load. If the admission lent its credibility to everything near it, a declarant could blend true admissions with false accusations and have both believed. Testing the remarks one at a time prevents that.
The subsections below follow the rule through its applications. The first sets out the facts and holding of Williamson. The second shows why a remark that mentions another person is not automatically excluded, since context can make it costly to the speaker. The third and fourth compare blame-shifting statements made to police with statements made to friends and relatives. The fifth applies the same reading when a defendant offers a third party's statement to clear the defendant, and the sixth sets out the competing positions that Williamson resolved. In a criminal case a statement that exposes the declarant to criminal liability must also meet a separate corroboration requirement, which the next major section takes up.2
Williamson's Narrow Reading
The word “statement” in Rule 804(b)(3) controls how a confession that mentions other people is analyzed, and Williamson fixed its meaning for the federal courts. The Supreme Court held that the exception does not admit non-self-inculpatory remarks, even when they appear within a broader narrative that is generally self-inculpatory.3 The facts show why the question arose.
Reginald Harris was caught with 19 kilograms of cocaine, enough to subject even a first-time offender to a minimum of twelve and a half years in prison.4 He made two statements to DEA Special Agent Donald Walton. In the second he said that he had lied about the source of the drugs because he was afraid of Williamson. Harris freely implicated himself, yet he did not want his story recorded and refused to sign a written version. Walton testified that he had promised to report any cooperation by Harris to the Assistant United States Attorney, though he said Harris was promised no reward or other benefit.5 Harris later admitted that the non-self-inculpatory things he had said in his first statement were false.6
At Williamson's trial the District Court found Harris unavailable and allowed Walton to relate what Harris had said. The court reasoned that Harris's statements clearly implicated himself and so were against his penal interest, that Harris was unavailable, and that corroborating circumstances ensured their trustworthiness. The statements implicating Williamson were therefore admissible.7 In effect the trial court treated the confession as one unit: because the confession as a whole hurt Harris, the parts pointing at Williamson rode in with it.8 The Eleventh Circuit affirmed without opinion, and the Supreme Court granted certiorari.9
The case turned on what “statement” means. Read as an extended declaration, the word would admit a whole confession so long as the narrative in the aggregate sufficiently inculpated the declarant, self-inculpatory and non-self-inculpatory parts alike.10 Read as “a single declaration or remark,” it would cover only those remarks within the confession that are individually self-inculpatory.11 The Court chose the second reading, and its reason lies in what the exception assumes.
People seldom “make statements which are damaging to themselves unless satisfied for good reason that they are true,”12 and that assumption works through cost. A remark that damages the speaker gives the speaker a reason to be accurate. A remark that points at someone else costs the speaker nothing, so the credibility earned by the first remark has no way to pass to the second. The Court put the danger directly: “One of the most effective ways to lie is to mix falsehood with truth, especially truth that seems particularly persuasive because of its self-inculpatory nature.”13 Harris supplied the proof, since the false remarks in his first account were the ones that did not incriminate him. Remarks that excuse the speaker are the weakest of all, because people are most likely to make them even when they are false, and nearness to damaging remarks does not make them more plausible.14
The same caution applies with special force when the remark implicates another person. A district court may not simply assume that a remark is self-inculpatory because it belongs to a fuller confession, and the arrest statements of a codefendant have traditionally been viewed with special suspicion, a point the Court drew from earlier decisions such as Lee v. Illinois.15
The Advisory Committee Notes point in two directions. They say that whether a statement is in fact against interest depends on the circumstances of each case: a statement admitting guilt and implicating another person, made in custody, may be motivated by a desire to curry favor with the authorities and so fail to qualify, while the same words spoken to an acquaintance would have no difficulty qualifying.16 The Notes also endorse the position of Dean McCormick's treatise, and the Court acknowledged that this language is not particularly clear and that some of it points the other way.17 Justice Kennedy argued that the text fairly expresses a policy of admitting collateral statements. The Court disagreed: the policy expressed in the Rule's text points clearly enough in one direction that it outweighs whatever force the Notes may have.18 The Notes' contrast between custody and conversation with an acquaintance survives as a way to judge motive, and later subsections use it.
Harris's own confession shows where the line falls. When he said he knew there was cocaine in the suitcase, he gave up his only possible defense to possession, lack of knowledge, so that remark would clearly have been admissible.19 The parts that implicated Williamson did little to subject Harris himself to criminal liability. A reasonable person in Harris's position might even think that implicating someone else would reduce his practical exposure, at least at sentencing, because small fish in a big conspiracy often get shorter sentences than the people running the whole show, especially if they help the authorities catch the big ones.20 The speaker and the occasion were the same for both sets of remarks. Only the cost to Harris differed.
The Court did not decide how each of Harris's remarks should come out. It sent the case back for the remark-by-remark inquiry, because whether a statement is in fact against interest depends on the circumstances of each case,21 and it left Williamson's Confrontation Clause challenge undecided.22
The narrow reading does not strip the exception of value against accomplices. A self-inculpatory remark can reach the defendant directly when other evidence supplies the link: “I was robbing the bank on Friday morning,” coupled with testimony that the declarant and the defendant drove off together that morning, is evidence that the defendant also took part.23 A remark that never mentions the defendant can also incriminate the defendant through logical inference or the operation of law.24 What the reading forecloses is admitting the defendant-implicating part of a jointly incriminating statement merely because the declarant implicated himself too. A co-criminal may implicate the accused to share or alleviate his own blame, to win prosecutorial or custodial favor, or for some other self-seeking reason, and that part of the statement is then not against his interest.25
Context Decides Self-Inculpation
Williamson does not ban every remark that mentions another person. The Court said that whether a statement is self-inculpatory “can only be determined by viewing it in context,” and that even a statement neutral on its face may be against the declarant's interest.27 Context is part of the test itself. Neither a remark's grammar nor its mention of a third party settles the question.
Context matters because the guarantee of sincerity is the cost the declarant bore, and that cost depends on what listeners can do with the words. A remark that names no crime can still hand investigators evidence, or place the speaker beside the people suspected of the crime, and a reasonable person who saw that would not volunteer the remark unless it were true. The opposite holds for a remark that puts the weight of a crime on someone else, which gives the speaker a reason to say it whether or not it is true. The inquiry therefore runs on the particular remark in the setting where it was spoken.
The Court gave two examples of remarks that look harmless and are not. “I hid the gun in Joe's apartment” may not confess a crime, but if it is likely to help the police find the murder weapon, it is certainly self-inculpatory. “Sam and I went to Joe's house” might be against the declarant's interest if a reasonable person in the declarant's shoes would realize that being linked to Joe and Sam would implicate the declarant in their conspiracy.28 Each remark mentions other people, yet in each the cost falls on the speaker. The Court added that other statements giving the police significant details about the crime may also be against the declarant's interest, depending on the situation.29
Justice Scalia, who wrote separately, rejected the simplistic view that only an outright confession to an element of a crime can qualify. In his example a declarant makes two statements that confess no element of any crime and then says, “I then went inside, robbed the bank, and shot the security guard.” The first two statements are obviously against the declarant's penal interest in context, and he had no doubt that a trial judge could properly admit them.30 Every detail fits the story that proves the robbery, so a reasonable person would not recite those details against himself unless they were true.
Naming another person does not disqualify a remark either. A lieutenant in an organized-crime operation who describes the inner workings of an extortion and protection racket, naming some of the other actors, inculpates himself on racketeering or conspiracy charges, and some of those remarks could be admitted.31 The limit comes from the setting. Naming another person in a context where the declarant is minimizing culpability or criminal exposure bears on whether the remark meets the standard, and the inquiry is always whether the particular remark, and not the extended narrative, satisfies the Rule. Labels such as “collateral neutral” and “collateral self-serving” only cloud the waters.32
After Williamson, statements against interest by third parties can still be admitted against the defendant when the statement does not mention the defendant directly but logical inferences or the operation of law make it incriminating to the defendant. A statement that mentions the defendant may also come in if a reasonable person in the declarant's position would realize that being linked to others implicated the declarant in another crime.33 Admission under the hearsay exception does not end the analysis in a criminal case, because a statement exposing the declarant to criminal liability must also be supported by corroborating circumstances that clearly indicate its trustworthiness.34
An exam stem that quotes an accomplice's words and names the defendant invites the question whether Williamson excludes them. The answer begins with what each remark cost the speaker, since the mention of another person decides nothing.
The first remark fails for the reason the Court gave about Harris's accusations against Williamson. Implicating someone else did little to expose Harris, and a reasonable person in his position might think that it would decrease his practical exposure.35 The Court also said that even the confessions of arrested accomplices may be admissible if they are truly self-inculpatory, rather than attempts to shift blame or curry favor.36 The second remark therefore survives even if the speaker was in custody, because its content costs him something. Federal courts have most often admitted third-party statements that inculpate a defendant when the statement neither seeks the favor of law enforcement nor shifts blame, and those two conditions organize the next two subsections.37
The question under Rule 804(b)(3) is always whether the particular remark was sufficiently against the declarant's penal interest that a reasonable person in the declarant's position would not have made it unless believing it true, and the Court said that question can only be answered in light of all the surrounding circumstances.38
Blame-Shifting and Currying Favor
An arrested suspect who talks to the police about other people has little left to lose by admitting a role of his own and may gain by pointing at someone else. A remark that serves the speaker equally well whether it is true or false cannot rest on the guarantee of sincerity that justifies the exception. That is why blame-shifting remarks made to the police after arrest usually fall outside it.
Williamson spelled out the incentive. The parts of Harris's confession that implicated Williamson did little to subject Harris himself to criminal liability, and a reasonable person in his position might think that implicating someone else would reduce his practical exposure, at least at sentencing.40 The Court pointed to the Sentencing Guidelines provisions on minor roles (§ 3B1.2) and on assistance to the authorities (§ 5K1.1) for the point that small fish in a big conspiracy often receive shorter sentences than the people running it, especially if they help catch the big ones.41 Both rewards are open to a blame-shifting arrestee, and both arrive whether or not the story is true.
The Court also recalled an older caution. The arrest statements of a codefendant have traditionally been viewed with special suspicion, and a district court may not assume that a remark is self-inculpatory merely because it belongs to a fuller confession, a point that applies with special force when the remark implicates someone else.42
Justice Ginsburg's concurrence shows these incentives at work in Harris's account. Harris was caught with enough cocaine to expose even a first offender to a minimum of twelve and a half years, and she concluded that his arguably inculpatory remarks were too closely intertwined with his self-serving declarations to be ranked as trustworthy.43 To the extent any of his statements incriminated him, they supplied only marginal or cumulative evidence of guilt, and the whole projected the image of a person “striving mightily to shift principal responsibility to someone else.”44 She went further than the majority, which thought that some of Harris's confession would clearly have been admissible.45
The opinions converge on one point. The Advisory Committee Note says that a statement admitting guilt and implicating another person, made in custody, may be motivated by a desire to curry favor and so fail to qualify.46 Justice Kennedy would examine statements made to authorities for that motive,47 and Justice Ginsburg found it in Harris's account. A statement made to curry favor with the police does not satisfy the exception.
Justice Kennedy, joined by the Chief Justice and Justice Thomas,48 would have admitted collateral statements related to a genuinely disserving remark, but he added a separate limit for statements made to authorities. That limit reaches the precise words against penal interest as well as the collateral remarks, because a declarant may think that a statement of guilt to authorities serves his interest as a way to obtain more lenient treatment, or simply to clear his conscience.49
Courts therefore sort police statements along a spectrum. At one pole the government has explicitly offered leniency, such as reduced charges or a lighter sentence, in exchange for the admission. The declarant then has a significant motive to obtain favorable treatment, and the entire statement should be inadmissible. At the other pole the government made no promise of leniency and the declarant was told that he had a right to remain silent and that his words could be used against him, and courts have not required exclusion, as in United States v. Garcia.50 Drawing that line keeps the leniency concern from swallowing the exception, which would otherwise exclude every statement against penal interest made to police, a result neither the Rule nor the Note contemplates.51
The second scene shows why custody alone cannot decide the question. Courts have not treated being in custody as conclusively establishing that a statement is self-serving when the statement was voluntarily given, strongly incriminates the declarant, and shows no sign of currying favor.52 The variable that moves the answer is motive, which the court reads from the setting and from any promise of leniency. Currying favor is also only one of two filters. A declarant who received no promise can still lose the exception for any remark that shifts blame.53
The embezzlement illustration from earlier in this chapter shows the same analysis on a second set of facts. Ellen described her scheme with Harvey to a law enforcement official after she had agreed to testify against him, which suggests she may have been currying favor, and she appears to have tried to shift most of the blame to Harvey. Her declarations are therefore not sufficiently disserving to be admitted against him.54
Statements to Private Confidants
A private listener removes the hope of leniency that clouds an arrest statement. After Williamson, courts treat the identity of the listener as a central fact, and they give greater weight to the point that a statement made to private individuals, rather than to law enforcement authorities, was not likely made to curry favor.57 That change of setting explains how a remark naming another person can fail in a police station and succeed in a living room.
A late-night confession shows the contrast. A declarant tells a friend, “Sam and I robbed the convenience store, ditched the car, and split the money.” Courts have found a statement of this kind sufficiently disserving to prove Sam's part in the robbery as well as the declarant's. The reference to Sam costs the declarant something in its own right, because it indicates inside knowledge of the crime and shows that the declarant was involved not only in a robbery but in a conspiracy.58 The same statement made to police officers upon arrest would not be admissible against Sam, given the likelihood that at that point the declarant is currying favor by identifying an accomplice.59 The words are identical in both settings, and only the listener differs.
The embezzlement example makes the same point with a single declarant. Ellen's account of her scheme with Harvey failed when she gave it to a law enforcement official after agreeing to testify against him.60 Change the listener. Suppose that when the auditor finds discrepancies, Ellen confides in her best friend, Ben, that she has been involved in an embezzlement scheme and describes how she and Harvey carried it out and covered their tracks. Ellen is unavailable at Harvey's criminal trial, so the prosecutor calls Ben. Her statement was not an effort to please the authorities, and it tends to disserve her because it places both Ellen and the accused in a scheme they carried out together.61
Statements made to private individuals, particularly in informal settings, also generally avoid a Confrontation Clause violation under Crawford v. Washington.62 The two points reinforce each other. A speaker with no purpose of feeding a prosecution is not currying favor with anyone, so the same fact helps under the hearsay exception and under the Confrontation Clause.
A defendant offered such a statement may argue that confiding in a trusted friend is not against interest, because the declarant expected the friend to keep the secret. Courts have not accepted the argument. A relationship of trust and confidence could make the declarant less aware that the statement might hurt him, but the possibility of disclosure appears to be enough.63 The rule asks whether the statement had a tendency to expose the declarant to liability, and a tendency does not vanish because the declarant hoped the listener would stay quiet.64
When the statement is offered in a criminal case as one that exposes the declarant to criminal liability, it must also be supported by corroborating circumstances that clearly indicate its trustworthiness.65 Those circumstances fall into two groups: corroborating evidence, such as forensic evidence supporting the declarant's account, and circumstantial guarantees of trustworthiness, such as a statement made near the time of the event or spontaneously with no sign of a litigation motive.66 The next major section takes up both.
Decisions applying Williamson draw the same line inside private statements. They admit statements to friends or colleagues that attribute blame evenhandedly to the speaker and the defendant, and they exclude statements that shift blame by attributing worse conduct to the defendant and exonerating the speaker.67 In United States v. Smalls, the declarant's statement implicating the defendant was nontestimonial, so Rule 804(b)(3), applied remark by remark under Williamson, governed its admissibility, and the statement came in.68 Walter v. State went the other way, holding that blame-shifting statements that do not implicate the declarant and the defendant equally fall outside the exception.69
That contrast marks the limit of the private-confidant rule. If Ellen began describing things Harvey did alone, those remarks would not disserve her, because they would only shift blame. Williamson does not let non-self-inculpatory remarks in because they sit within a self-inculpatory narrative, and the Court explained that “[s]elf-exculpatory statements are exactly the ones which people are most likely to make even when they are false; and mere proximity to other, self-inculpatory, statements does not increase the plausibility of the self-exculpatory statements.”70 A private setting removes the motive to please the police. It leaves the speaker every ordinary reason to look less culpable.
An exam fact pattern that sets a confession in a friend's kitchen invites the conclusion that everything said there comes in. The correct analysis credits the setting first, since it answers the curry-favor objection, and then sorts the remarks into those that tie the declarant to the crime and those that assign solo conduct to someone else.
Exculpatory Statements After Williamson
The remark-by-remark reading applies whichever side offers the statement, because Rule 804(b)(3)(A) asks only what the statement cost the declarant when made and says nothing about whom it helps.72 The majority rejected the view that a statement means an extended narrative, admissible whole once the narrative is sufficiently inculpatory, and adopted the narrower view that the exception reaches only remarks that are individually self-inculpatory.73 Justice Kennedy, concurring in the judgment, drew out the consequence for an accused. If a declarant said, “I robbed the store alone,” only the portion saying “I robbed the store” could be introduced by a criminal defendant on trial for the robbery.74 He found that result extraordinary and saw no justification for it in the Court's opinion.75
Justice Scalia approached the problem from the other side. The inquiry is always whether the particular remark, rather than the extended narrative, satisfies the Rule, and context can make a remark that confesses no element of a crime obviously self-inculpatory.76 The two opinions explain why courts divide over clearing language. Kennedy's example treats the exculpatory tail of a confession as lost, while Scalia's approach leaves its fate to the circumstances in which the remark was spoken.
The same filter applies to both sides because the guarantee of sincerity comes from the cost the declarant bore, and that cost attaches to the assertion that carries it. A remark saying that the defendant had nothing to do with the crime costs the declarant nothing by itself. A declarant may make it out of loyalty, or to shield someone who matters to him, whether or not it is true. A clearing remark therefore gains no credibility from sitting beside an admission, and it must show a cost of its own or be tied closely enough to the admission to share its cost.
A co-conspirator's plea allocution shows the filter at work. Rule 11 plea proceedings exist to ensure that a plea is knowing, voluntary, and grounded on a proper factual basis,77 and statements made there can qualify under Rule 804(b)(3) only if each is separately parsed and is itself self-inculpatory. In the case at hand a co-conspirator who had pleaded guilty and then invoked the privilege against self-incrimination at the defendant's trial had said that he never supervised the defendant and never asked the defendant to travel to smuggle drugs. Those remarks did not inculpate him or expose him to liability, so the guarantee of trustworthiness was missing for them.78 The remarks failed because of what they said, and other remarks in the same allocution remained free to qualify.
A highway stop poses the harder version of the question. Officers find marijuana in suitcases behind the sleeper bunk of a truck driven by two men, Garvin and Torrens. After warnings, and outside Garvin's hearing, Torrens tells an officer that the marijuana belongs to him, that Garvin had nothing to do with it, that he wanted to get rich quick, and that he took a chance, lost, and is ready to do the time. At Garvin's trial Torrens refuses to testify, and the trial court excludes the portion exonerating Garvin.79 Torrens was not blaming anyone else while offering to help convict that person, so the curry-favor concern from the preceding subsection does not decide the case. The question is whether the sentence clearing Garvin cost Torrens anything.
Courts disagree. In State v. White the New Jersey Supreme Court admitted exculpating remarks that, in light of the surrounding circumstances, subject the declarant to criminal liability or that, as a related part of a self-inculpatory statement, strengthen the incriminating effect of the declarant's exposure. In United States v. Vegas the Second Circuit excluded a declarant's remarks that certain people had not supplied him with drugs.80
Each side has a coherent argument. A court inclined to admit can say that a declarant who insists that no one else was involved gives up the chance to shift blame, so the denial is the other face of an admission that he acted alone. A court inclined to exclude can say that the denial adds nothing to what the claim of ownership already cost and is exactly what a loyal friend would say whether or not it was true. Phrasing therefore matters. A court is more receptive when the clearing words simply restate the admission, as in “it is mine alone,” and less receptive when they form a separate assertion about what another person knew or did, because that sentence must then justify itself.
The easy case is a plain third-party confession that never mentions the defendant. In Gray v. State, a Maryland prosecution of a husband for murdering his wife, the court held it error to exclude a statement by the wife's lover to a friend, “I took care of her,” which seemed to confess the murder.81 Nothing in the remark needs to be severed, because all of it is self-inculpatory. It helps the husband only by logic: if the lover killed her, the husband did not. In a criminal case a statement of this kind must still be supported by corroborating circumstances that clearly indicate its trustworthiness, a requirement the next major section takes up.82
Courts show particular skepticism toward clearing statements by declarants who have already been convicted. United States v. Albert excluded a statement by a convicted co-offender made at a plea proceeding.83 The concern follows from the mechanism of the exception. A person already convicted faces little new exposure by confessing, so the cost that is supposed to vouch for the words may be missing, and the declarant may speak freely to help someone else.
Competing Views of Collateral Statements
Everything turns on the word statement, because a declarant who admits a crime usually says other things in the same breath. A remark that accompanies the damaging one without itself being damaging is a collateral statement, as when a declarant says “I shot the bank teller” and adds “John Doe drove the getaway car.”87 The rule's text does not say whether such remarks travel with the admission, and the opinions in Williamson record a long debate over the answer.88 For federal courts the majority settled the matter by reading the exception to reach only the remarks that are individually self-inculpatory.89 The rejected alternatives still repay study, because the reasons given for and against them explain the shape of the majority's rule and show where its limits fall.
The debate produced three positions, and each answers differently the question of how far the declarant's own cost vouches for the words around the damaging one.90
- Dean Wigmore took the broadest position. The statement may be accepted “not merely as to the specific fact against interest, but also as to every fact contained in the same statement,” because it is made under circumstances fairly indicating the declarant's sincerity and accuracy.
- Dean McCormick took a middle position. Collateral statements of a neutral character come in, and collateral statements of a self-serving character are excluded.
- Professor Jefferson took the narrowest position. The reliability of a statement against interest comes only from the disserving fact stated, so the exception should be confined “to the proof of the fact which is against interest.”
The middle position sorts collateral remarks by their relation to the declarant's interest. In a criminal case a self-serving statement is one that tends to reduce the charges or mitigate the punishment the declarant might face, such as one robber's statement that the other robber was the triggerman.91 The words “John and” in “John and I robbed the bank” are neutral, subject to the possibility of conspiracy charges, while “John, not I, shot the bank teller” is to some extent self-serving.92 The Advisory Committee's Note, by its reference to and apparent incorporation of McCormick, contemplates excluding a collateral self-serving statement and admitting a collateral neutral one.93
The Court chose the narrowest position. Justice Kennedy read the opinion as adopting the extreme view that no collateral statements are admissible under Rule 804(b)(3).94 Justice Scalia, who joined the relevant parts of the Court's opinion, described the same choice from the other side: the Court rejected reading “statement” as an extended declaration in which self-inculpatory and non-self-inculpatory parts both come in so long as the whole is sufficiently inculpatory.95 The broadest position therefore lost, and the live contest was between the narrowest position and a middle one.
Justice Kennedy, writing for himself, the Chief Justice, and Justice Thomas, concurred only in the judgment.96 The Rule's silence, he reasoned, no more incorporates Jefferson's position than it does McCormick's or Wigmore's, so he looked for outside guides.97 The Advisory Committee's Note speaks to the problem directly: a third-party confession “may include statements implicating him, and under the general theory of declarations against interest they would be admissible as related statements.”98 His other two guides were the common law of the exception and the presumption that Congress does not enact statutes that have almost no effect.99 The second guide carried practical force, since the exclusion of collateral statements would cause the exclusion of almost all inculpatory statements.100
Kennedy did not conclude that everything near a confession comes in. The Note's reference to balancing the self-serving against the disserving aspects of a declaration suggested to him that neutral collateral remarks are admitted and self-serving ones are excluded.101 His approach first finds a fact against penal interest and then admits the statements related to it, subject to two exclusions. The first is a collateral remark so self-serving as to be unreliable, such as one that shifts blame to someone else for a crime the declarant could have committed. The second is the whole statement when the circumstances show that the declarant had a significant motive to obtain favorable treatment, as when the government made an explicit offer of leniency.102 He would have left these fact-bound judgments to the trial judge under Rule 104(a), with wide discretion.103
Justice Scalia answered the charge that the narrow reading guts the exception. A remark can be self-inculpatory without confessing an element of a crime, as in his example of two statements that confess no element of an offense and yet are plainly against the declarant's interest in context. Naming another person does not disqualify a remark, as when an organized-crime lieutenant describes the inner workings of a racket and names other participants.104 Scalia also objected to the vocabulary of the middle position. Labels such as “collateral neutral” and “collateral self-serving” are manufactured categories that cloud the waters, and the question is always whether the particular remark, rather than the extended narrative, meets the standard of the Rule.105
The approaches overlap more than their labels suggest, and on those facts they produce the same result. They part ways on a remark that costs the declarant nothing without serving the declarant either. Kennedy's approach admits the remarks related to the damaging one unless they are so self-serving as to be unreliable, so such a remark rides in with the admission.106 The majority's reading requires every remark to qualify on its own, so a remark of that kind stays out unless context makes it self-inculpatory.107 The practical difference is where the burden sits. Under the majority's rule the proponent must show what each remark cost the declarant, and proximity to a confession does not carry a remark in.
Justice Ginsburg, concurring in part and concurring in the judgment, would have gone further than the majority. She concluded that Harris's statements did not fit the exception even in part, because his arguably inculpatory remarks were too closely intertwined with his self-serving declarations to be ranked as trustworthy.108 In her account, to the extent any of the statements incriminated Harris, they offered only marginal or cumulative evidence of his guilt, and as a whole they projected “an image of a person acting not against his penal interest, but striving mightily to shift principal responsibility to someone else.”109 The majority took the opposite view of one part, because some of Harris's confession would clearly have been admissible.110 Their disagreement concerns severability, meaning how far a remark that costs the declarant something can be separated from the self-serving setting in which it was spoken. Ginsburg's worry and Kennedy's limit for statements to authorities rest on the same concern: a statement made to curry favor cannot satisfy the exception.
The majority's reading binds federal courts because it construes the federal rule itself, and a state court interpreting its own against-interest exception may read it differently. Where a state has adopted the Federal Rules of Evidence, Williamson is influential but not binding, and the states have split, with some following the narrow reading and others rejecting it.111 A rule adopted by a state is that state's own law, so only its courts can say authoritatively what it means. A federal question is therefore answered under Williamson, while a state-law question first asks which view the jurisdiction has taken.