Concept
The Hillmon Doctrine
Statements of mental, emotional, or physical condition
The Hillmon Doctrine
Rule 803(3) lists intent and plan among the states of mind it covers.1 A statement of what the declarant means to do describes a condition that exists as the words are spoken, so it fits the exception much as “I hate him” does. An explicit description of the speaker’s state of mind counts as hearsay when it proves what the declarant believed, knew, or felt, and a longstanding exception covers it.2 A statement offered only as circumstantial proof of a state of mind is not hearsay at all.3 The hard question is whether a statement of plan can prove more than the plan, namely that the declarant later carried it out. Mutual Life Insurance Co. v. Hillmon answers yes.
In most state-of-mind cases the mental condition is the very fact in dispute. Intent to steal or kill, intent that a paper take effect as a deed or will, and a spouse’s affections are examples, and in each the statement proves the condition directly.4 Hillmon adds a step. The declarant’s intent is not what the case is about; it serves as circumstantial evidence of what the declarant did next. When the performance of a particular act is in issue, the actor’s intention to perform it may be shown.5
The inference has two links. The words show that the declarant held a plan when speaking, and a person who expresses an intent is more likely to have acted on it than a person not shown to have had that intent.6 The conclusion is a matter of probability. People abandon plans, and the jury weighs that risk.7
The direction of the statement explains why the doctrine fits the text of the Rule. Rule 803(3) excludes a statement of memory or belief offered to prove the fact remembered or believed,8 and a statement of plan looks forward. Four subsections follow. The first presents the case in which the Supreme Court announced the doctrine. The second explains why intent is evidence of conduct and where the inference weakens. The third takes up plans that depend on another person, where the doctrine becomes contested, and the fourth asks whether such evidence needs independent support.
Hillmon and Its Holding
In Hillmon, the Supreme Court held that a declarant’s statement of present intent or plan, such as an intention to travel to a particular place, is admissible to prove that the declarant later acted on it. The case was an insurance dispute, and the declarant was not the insured. He was a third man, Frederick Adolph Walters, whom the insurers said had died in the insured’s place. The holding makes sense once you see how his plans became a link in the proof.
In each suit the plaintiff alleged that Hillmon died in March 1879, while the policy was in force, and that the insurer had refused to pay.9 The insurers denied that he was dead. They alleged that Hillmon, John H. Brown, and others had conspired to defraud them, had obtained the policies, and had afterward falsely represented that Hillmon was dead and that a dead body they had procured was his, when in fact he was alive and in hiding.10
The plaintiff’s evidence was that Hillmon and Brown left Wichita, Kansas, about March 5, 1879, traveled together through southern Kansas looking for a cattle-ranch site, and camped at Crooked Creek, where Hillmon was accidentally killed on the night of March 18 and his body was taken away and buried.11 Someone had indeed been killed at Crooked Creek; the parties disagreed about who. The insurers offered evidence that the body was Walters’s, not Hillmon’s.12
The insurers’ evidence about Walters was this. He had left home and his betrothed in Iowa in March 1878, had stayed in Kansas until March 1879, and had written regularly to his family and betrothed. His last letters came from Wichita in the first days of March 1879, and nothing had been heard from him since.13 His disappearance, standing alone, did not put him at Crooked Creek. The insurers needed to show that he went there with Hillmon, and his letters were, in the Court’s words, “the natural, if not the only attainable, evidence of his intention.”14
One letter said that Walters would stay in Wichita until the early part of the next week and then leave, “as I am going with a man by the name of Hillmon, who intends to start a sheep ranch.”15 The plaintiff objected to testimony about the contents of such a letter as incompetent, irrelevant, and hearsay. The trial court sustained the objection,16 and the letters stayed out.17
The Supreme Court reversed. It found two errors that required a new trial: excluding the letters and unduly restricting the defendants’ challenges.18 The insurers’ first theory for admitting the letters, that they were memoranda made in the ordinary course of business, failed because the letters were “clearly not such.”19 The Court accepted a different ground. The two letters were competent evidence of Walters’s intention at the time he wrote them, which was “a material fact bearing upon the question in controversy.”20
The Court relied on its earlier decision in Insurance Co. v. Mosley, which admitted declarations of mental feeling when the feeling was itself material.21 Its argument for trustworthiness was practical. After a person’s death there can hardly be another way to prove an intention. While he lives, his own memory of his state of mind at a former time is no more likely to be clear and true than a bystander’s recollection of what he then said, and it is less trustworthy than letters written at the very time and under circumstances that preclude any suspicion of misrepresentation.22 A person’s own statements are, in short, often the primary evidence of his state of mind.23
The Court then fixed the use the letters could serve. They were not narratives of facts someone else had told Walters, and they were not proof that he actually left Wichita. They were evidence that, shortly before he was said to have left, he intended to go, and to go with Hillmon. That intention made it more probable both that he went and that he went with Hillmon than if there had been no proof of it.24 Given the mass of conflicting testimony about whose body lay in Hillmon’s camp, the Court thought the evidence might properly influence the jury.25
That passage separates what the letters assert from what they help prove. They assert a present plan. The departure and the companion lay in the future when Walters wrote, so the letters could not assert either as an accomplished fact. The jury reaches them by inference, and the inference needs only the ordinary showing that the fact is more probable with the evidence than without it.26
Naming Hillmon matters, because it turned the letters from evidence about Walters alone into evidence about a second person. The Court found support in Hunter v. State: “If it was in the ordinary train of events for this man to leave word or to state where he was going, it seems to me it was equally so for him to say with whom he was going.”27 A reference to the companion is as natural a part of announcing a trip as any other incident of it.28
The letters were admitted because they point forward. Suppose that after the trip began Walters had written, “I have left Wichita with Hillmon the sheep trader.” That letter narrates a completed fact and offers it to prove that the fact occurred. Rule 803(3) excludes exactly that use, because it carves out a statement of memory or belief offered to prove the fact remembered or believed.29 The first letter reports a plan the jury can weigh; the second asks the jury to take the writer’s word that something happened. When an exam stem gives you a declarant’s remark about a trip, ask first whether the words describe what the declarant means to do or what the declarant has done.
Two cautions keep the holding in proportion. The ruling made the letters available to a jury and sent the case back for a new trial; it did not decide whose body lay at Crooked Creek.30 And although Walters was never heard from again, the doctrine does not depend on the declarant’s absence. In nearly every case admitting such statements the declarant has been unavailable, and that need probably encouraged courts to admit them, but neither the decisions nor Rule 803(3) require unavailability.31
Rule 803(3) does not say in terms that intent may prove the intended act. The Advisory Committee indicated that Hillmon would continue, and courts now routinely admit statements of intent for that purpose.32 The next subsection explains why the inference is strong enough to admit and where it weakens.
Why Intent Proves Conduct
The Advisory Committee Note closes the gap in the text. It states that “[t]he rule of Mutual Life Ins. Co. v. Hillmon . . . allowing evidence of intention as tending to prove the doing of the act intended, is, of course, left undisturbed.”34 This subsection explains why that inference is sound enough to admit, how it differs from the ordinary use of a state of mind, and which features of a statement make it strong or weak. Every example here concerns the declarant’s own conduct; a plan that depends on another person raises a harder question for the next subsection.
The ordinary use of a state of mind makes it the thing in dispute, and the Hillmon doctrine reverses that arrangement. The Ninth Circuit called it “an extraordinary doctrine” for exactly this reason: it does not require that the declarant’s state of mind be an actual issue in the case, because it uses that state of mind inferentially to prove some other matter that is in issue.35 Walters’s intention to leave Wichita with Hillmon was not what the litigants were fighting about. The fight concerned whose body lay at Crooked Creek, and the intention mattered only as a step toward showing that Walters had gone there.36 Compare two offers. A testator’s declared wish that a niece receive the house, offered in a will contest over testamentary intent, proves the wish, which is the fact in dispute. A traveler’s declared plan to leave for a city, offered to show that the traveler arrived there, proves the plan only as a stepping-stone to the trip.
Two links connect the words to the conclusion, and only the first is something the statement itself asserts. The statement asserts that the declarant holds a plan as the words are spoken, and that assertion avoids the memory problems of narratives about the past, because the declarant is reporting something happening in the declarant’s own mind. The second link is an inference the jury draws. A person who expresses an intention is more likely to have done the act than a person not shown to have had it, and that is all the relevance standard of “more probable than without the evidence” requires.37
The doctrine can look backward at first. It admits a statement that carries two weaknesses, sincerity and the survival of the plan, and it excludes a flat statement that the declarant already did the act, which carries only one. Two considerations answer the puzzle. In the cases that settled the doctrine the evidence usually had substantial convincing power, and the need to resort to statements of intent was often great.38 The direction of the statement also marks the line that protects the hearsay rule. Treating a statement about a past event as proof of the event, merely by calling it a present state of mind, would swallow the exclusion of narratives of the past, and Rule 803(3) draws that line itself with its memory-or-belief carve-out.39 Under it, “I’m going to the café at three” may prove the visit, while “I went to the café at three,” offered for the same visit, may not.
Admission does not make the inference strong. A declared intention is much less likely to be carried out than a declared state of mind is to be held, so a statement of intent to harm someone proves ill will far better than it proves that the declarant did the harm.40 Between present intention and later act, a myriad of contingencies may intervene, and the possibility of failure goes to weight, argued to the jury, and not to admissibility.41 The judge asks whether the plan makes the act more probable than it would be without the statement. The jury decides how much the plan is worth.
Elapsed time moves the inference first, because it depends on the plan lasting from the moment of speaking to the moment of acting. Each added day gives the declarant more chance to change course and events more chance to defeat the plan. The usual limit of remoteness in time that attends every statement of mental state therefore applies here too.42
Specificity is the second limit. A plan to do one thing at one time names the very act the jury is asked to infer, so holding that plan is a good reason to expect that act. The probative value of a statement of intent depends on how likely it is that the intent was carried out,43 and a general threat or boast describes a disposition more than a plan. Using it to prove a particular act on a particular day asks the jury to reason from the kind of person the declarant is, and that drifts toward character reasoning.
Sincerity is the third pressure point. Rule 803(3) states no condition that a statement be made under circumstances suggesting earnestness, although some formulations of the exception impose one.44 Even without a formal condition, a remark that sounds like a joke or a bluff gives the jury less reason to believe the plan existed, so it supports the inference less than an ordinary remark to a friend or relative.45
None of these limits turns on whether the declarant can testify. In almost every reported case the declarant was dead or missing, and the resulting need probably made courts more willing to admit the evidence, but neither the decisions nor Rule 803(3) condition admissibility on unavailability.46 The reliability comes from the nature of the statement, a plan held at the moment of speaking, and a plan held in that way is the same whether the speaker later dies or lives. Necessity persists too when the declarant is alive, because a witness’s later memory of an earlier state of mind is no more likely to be clear and true than a bystander’s recollection of what the witness then said.47 A stem that adds that the declarant is available therefore leaves the analysis unchanged, and a stem that adds that the declarant is dead does not rescue a vague or remote statement.
Intent Involving Another Person
Rule 803(3) admits a statement of the declarant’s then-existing intent, and the last subsection used that grant to carry a jury from the declarant’s plan to the declarant’s own later act.49 A joint plan is harder. The bookkeeper who says on Thursday morning, “I’m meeting Sam at the café at 3,” easily proves that the bookkeeper went. Offered to prove that Sam went, the same words ask the jury to infer the conduct of a person who never spoke, and the Rule, which speaks only of the declarant’s own state of mind, does not say whether that second use is allowed. The cases and committee materials below try to answer.
The leading illustration is United States v. Pheaster. Larry Adell, a sixteen-year-old, told friends, “I am going to meet Angelo in the parking lot to get a pound of grass.”50 His friends’ testimony was offered against the defendant Inciso to show that the meeting took place. Under the ordinary doctrine the remark makes it more likely that Larry went to the lot. The government wanted a second inference as well, that Angelo went too.
The Ninth Circuit named the objection before answering it. It divided the remark into two statements. The first is Larry’s statement of his own intention. The second is an implicit statement of Angelo’s intention, because unless Angelo habitually frequented that lot, the meeting could happen only if Angelo meant to be there.51 The first statement describes the declarant’s own mind, the subject the exception covers. The second describes someone else’s mind, and the reasons for trusting a person’s report of a present feeling do not reach it. The layering resembles the problem Rule 805 addresses: when one statement contains another, each part must satisfy an exception of its own.52
The court made the point concrete with an altered remark. Had Larry’s friends testified that he said, “Angelo is going to be in the parking lot of Sambo’s North tonight with a pound of grass,” no state-of-mind exception or any other would admit it, and the court observed that this was in effect at least half of what the testimony attributed to Larry.53 In the real remark the declarant announces a plan of his own. In the altered remark he announces another person’s presence, a plain assertion about a third party’s conduct. The real remark carries that assertion inside a statement of the declarant’s plan, and its force against Angelo comes from the half the exception was not built to admit. An answer that rejects the whole remark because it mentions another person overlooks the half that is the declarant’s plan, and one that admits the altered remark because it sounds like a plan overlooks whose mind it describes.
Despite that awkwardness, the court admitted the testimony, because the authority for the application was impressive, beginning with Hillmon itself.54 In Hillmon the Supreme Court approved a statement of intent to infer not only the declarant’s future act but the future act of another.55 Hunter v. State supplied the reasoning: naming a companion is as natural a part of announcing a trip as naming the destination.56
People v. Alcalde, a California murder case in which the victim said she was going out with Frank on the evening she was killed, applied Hillmon to uphold use of the statement against Frank. The court listed prudential considerations: the declarant should be dead or otherwise unavailable, the statement should be relevant and highly trustworthy, and other evidence should exist from which guilt could be inferred.57 The two decisions do different work. Hunter explains why the companion’s name is not a separable piece of the plan, and Alcalde lists the circumstances in which a court felt comfortable relying on it, although Rule 803(3) itself does not require unavailability.
The Ninth Circuit answered the reliability worry by looking at how plans fail in general. Even when no one else is needed, a myriad of contingencies may frustrate a plan, and the need for another person’s cooperation adds one more contingency, a difference of degree rather than kind.58 The possible unreliability of the inference goes to weight, which the opposing party may argue to the jury; it is no ground for excluding relevant evidence. The defendant also called Larry, a sixteen-year-old involved with narcotics, an unreliable declarant, and the court found nothing in the circumstances of his statements to suggest a reason for doubt.59
The objection keeps its force as a matter of weight. If completing a plan requires the continued inclination and ability of someone besides the declarant, the likelihood of completion is arguably substantially lower.60 Hillmon is usually read as approving the use of one person’s stated plan to infer another’s act,61 although the parties there agreed that Hillmon did go to Crooked Creek, so the Court did not confront the cooperation problem squarely.62
A recurring version of the problem is a homicide prosecution of Frank, in which a witness says the victim said on the morning of the killing, “I am going out with Frank tonight.” The words tend to prove the victim’s own acts, and they also tend to prove that Frank went out with the victim, a fact very much in issue. Courts have generally admitted such statements. The added danger has led some to attach conditions: a limiting instruction confining the evidence to the declarant’s conduct, a requirement of independent evidence of the defendant’s conduct, use of the statement only to explain the declarant’s intent, or a limit to cases in which the declarant is dead or unavailable, the statement is shown to be serious, and the event was realistically likely.63 The independent-evidence condition is the subject of the next subsection.
Pheaster was decided on common-law principles before the Federal Rules took effect, and the court consulted Rule 803(3) only for light on what the common law then was.64 The committee materials it consulted frame the question that later courts faced under the enacted text. The Advisory Committee Note says that Hillmon is left undisturbed.65 The House Judiciary Committee approved the Rule as submitted but stated its intention that it be construed to limit Hillmon, so that statements of intent prove only the declarant’s own future conduct and not another person’s.66 The Senate and Conference Reports are silent.67 Pheaster read the Advisory Committee as presuming that Hillmon would carry over in full force, and the House as an effort to cut back on what it too took to be the prevailing common-law view; it nevertheless held Larry’s statement admissible.68
A court deciding the question under the enacted Rule cannot sidestep the text, and United States v. Houlihan shows how one court resolved it. The victim was found shot, and the evening before he had allegedly told his sister that he was going out “to meet Billy Herd.”69 The government offered the remark through the sister as circumstantial evidence that Herd killed him, and the district court, facing a question of first impression in the First Circuit, admitted it. Its reasoning was textual. Rule 803(3) does not by its terms limit the class of persons against whom a statement of intent may be admitted, so the Rule codifies Hillmon as written, and the court would not graft on a limit that Congress could have enacted and did not.70
The textual argument gains force from the House Report’s own wording. The committee approved the Rule as submitted and then stated a hope about how it would be read. A limit stated only in a report leaves the Rule’s words untouched, and with the Advisory Committee pointing one way, the House the other, and the Senate and Conference silent, the history cannot show what Congress as a whole intended. Houlihan therefore held that conflicting history sends the analysis back to the text.71 Courts that have decided the question under the Rules have opted for admissibility.72
The extension has critics. One treatise suggests that proof of subsequent conduct through declarations of state of mind might be limited to conduct that did not require the substantial cooperation of persons other than the declarant, but it also recognizes that courts have not imposed that limit.73 The practical worry behind the criticism is easy to state. A jury that hears a dead victim say “I’m meeting the defendant” is likely to treat the words as direct evidence that the defendant was there, the defendant had no chance to cross-examine the declarant, and a limiting instruction cannot help, because using the words against the other person is the entire purpose of offering them.
The Corroboration Requirement
Rule 803(3) admits a declarant’s statement of then-existing intent or plan and says nothing about whose later conduct the statement may be used to prove.75 The preceding subsection followed that silence to the result most courts reach, that a declarant’s plan to do something with another person can show the other person took part. Whether that use needs a safeguard is a point on which courts have divided. Some admit the statement against the other person only when independent evidence connects it to that person’s conduct, and others admit it with no such showing.76 A complete answer describes both camps and the reasons each gives, and it does not announce a single national rule.
The corroboration view begins with the Fourth Circuit’s decision in United States v. Jenkins. That court held that a declarant’s statement of intent cannot prove the later conduct of a third party, but may be admitted to show why the third party acted as he did when independent evidence establishes that the third party in fact engaged in the alleged conduct. The Second Circuit stated the condition more broadly in United States v. Nersesian: “declarations of intentions or future plans are admissible against a nondeclarant when they are linked with independent evidence that corroborates the declaration.” A federal district court in Illinois applied the same idea in United States v. York, holding inadmissible testimony that a victim said she intended to meet the defendant at a particular time, because the statement could not be corroborated with particularity.77 The formulations differ. Jenkins looks for evidence that the third party acted, Nersesian for evidence that corroborates the declaration, and York for corroboration specific enough to fit the stated plan, so a court applying the approach must say which kind of independent proof it demands.
The reason for the condition lies in how little the statement can carry when it is the only proof. Used against another person, a statement of intent asks the jury to accept that the declarant sincerely held the plan, that the declarant followed through, and that the other person was both willing and able to take part, and the words themselves speak only to the first of these. A plan that needs a second person is less likely to have been completed than one the declarant could finish alone, because completion then depends on that person’s continued inclination and ability as well as the declarant’s.78 Independent proof that the declarant actually did the declarant’s own part shows that the plan went forward, and it turns the other person’s role from speculation into a reasonable inference. The condition is one of several safeguards courts have tried for this situation, along with a limiting instruction confining the evidence to the declarant’s conduct, a showing that the stated intent was serious and the event realistically likely, and a limit to cases in which the declarant is dead or unavailable.79
The opposing camp starts with the Ninth Circuit’s decision in Pheaster, which approved testimony that a victim told a friend, shortly before he disappeared, that he was going to meet a person with the same name as the defendant, and which did so without requiring corroborating evidence.80 Pheaster conceded that such testimony could be unreliable and treated that as a matter of weight for the jury, since the inference from present intention to later act is only an inference.81 The district court in Houlihan added a textual argument. Rule 803(3) does not by its terms limit the class of persons against whom a statement of intent may be admitted, so the court read the Rule to codify Hillmon as written and refused to graft on a limit that Congress could have enacted and did not.82 It conceded that the Second and Fourth Circuit approach may seem practical and fair, and it called that approach little more than judicial policymaking.83
What does the split change on an actual record? The answer depends on whether anything besides the statement shows that the plan went forward, and the clearest way to see it is to start with a record in which nothing does.
The two camps converge once independent proof appears. Hillmon itself sits in that position, because the parties agreed that Hillmon did go to Crooked Creek, and the Supreme Court therefore had no occasion to confront the cooperation problem directly.84 A New York decision shows corroboration arriving from a different direction. In People v. James the question was whether James attended a meeting at which exam answers were handed out in advance, and the corroboration came mainly from his results: he got every question right on the exam the day after the meeting, and when the results were invalidated and a new exam was given shortly afterward, he failed it.85 That pattern is independent evidence of advance knowledge, so it supports the claim that James was at the meeting without depending on the declarant’s word.
Two cautions keep the topic in proportion on an exam. The corroboration condition is a judicial gloss in the circuits that adopt it, and the text of Rule 803(3) contains no such language, which is exactly the point the courts refusing it press.86 The condition also attaches only to use of the statement against the non-declarant. Using the same remark to show that the declarant went to the meeting follows the ordinary doctrine from the earlier subsections, and the bookkeeper who says “I’m meeting Sam at three” supplies proof of the bookkeeper’s own trip without any outside support. An answer to a stem about a joint plan therefore identifies the third-party use, states that courts disagree about whether independent evidence is required, describes the corroboration approach and the Pheaster approach with the reason for each, and then asks whether the record contains independent proof that the plan went forward.