Concept
Mental and Emotional State
Statements of mental, emotional, or physical condition
Mental and Emotional State
Some lawsuits turn on what was going on inside a person’s head. A court may have to decide whether a wife’s affection for her husband had cooled before a third party interfered, whether a testator resented the child who now contests her will, or whether a defendant feared the informant who had befriended him. Rule 803(3) supplies a way to prove such things. It admits a statement of the declarant’s then-existing state of mind, such as motive, intent, or plan, and a statement of the declarant’s then-existing emotional condition, such as mental feeling, and the factfinder may take the statement as true of that condition.1
The exception works for minds because a person’s words about her own mind are unusually good evidence of it. A declarant who says what she is thinking, what she is feeling, or what she plans to do is reporting something she perceives about herself; a declarant who says what she remembers doing is reporting something else, and that last category is the one that stays out.2 The rule’s text draws the same line, first with the word then-existing and then with its carve-out for statements of memory or belief offered to prove the fact remembered or believed.3 A speaker who says “I am furious at my landlord” reports a feeling that exists as she speaks. A speaker who says “I was furious at my landlord last month” reports a memory of a feeling that has passed, and the exception does not cover it.
Four questions organize the rest of this section, in the order they arise. When is a person’s mental condition itself the thing the case turns on? What does then-existing demand? How far can a present condition be stretched to show the same condition a day earlier or later? And what should a court do when a statement of feeling carries a factual accusation inside it?
State of Mind in Issue
The substantive law often makes legal rights and liabilities hinge on a particular state of mind or feeling: the intent to steal or kill, the intent that a paper take effect as a deed or will, the maintenance or transfer of a spouse’s affections. When it does, the person’s mental or emotional state is the ultimate object of inquiry. The jury is not asked to infer earlier or later conduct from it; the state of mind is an operative fact on which the claim or defense depends. A state of mind may be proved by what a person does, but what the person says is often a primary source of evidence.4
Courts admitted such statements long before the Federal Rules codified the practice. In Hillmon the Supreme Court collected earlier decisions that received a wife’s letters to her husband or to third persons, written before any misconduct on her part and with no ground to suspect collusion, to show her affection for him and her reasons for living apart from him in an action for criminal conversation.5 It also quoted a Massachusetts judge in a will contest over testamentary capacity, fraud, and undue influence. Intention, purpose, and mental peculiarity, the judge reasoned, are mainly ascertainable through language, so “statements and declarations, when the state of the mind is the fact to be shown, are therefore received as mental acts or conduct.”6 Hillmon is best known for a different use of intent evidence, which a later section develops; the cases here involve a single step from the words to the condition they describe.
Many of these statements never needed an exception. When an out-of-court statement is used as circumstantial proof of the declarant’s state of mind, the hearsay rule is not implicated, because the statement is not offered to prove the truth of what it asserts. The children yelling “Barney, Barney, Barney!” in Lyons Partnership v. Morris Costumes, and the defendant in United States v. Parry who told his mother he was working with a narcotics agent, are examples of that use.7 Other statements describe the speaker’s mind outright: “Hey, I think that’s Barney!”, “I think she’s an undercover cop,” “I hate him.” Offered to prove what the declarant believed, knew, or felt, these seem to be hearsay, and they come in under the longstanding exception for statements of the declarant’s state of mind.8
The defense in United States v. Harris shows why the distinction rarely matters in practice. It tried to establish that Harris knew Steward was an informant and played along only out of fear of what would happen to him if he refused.9 Knowledge and fear were the whole defense, so Harris’s own statements about what he thought and felt were the natural proof, and it hardly matters whether a court files a given statement under non-hearsay or under the exception. The same economy applies when conduct is offered to show an abnormal mind. If a litigant objects that abnormal conduct can be simulated and so is assertive, the short answer is that the evidence would then be admissible under the present exception, and courts should spend little effort deciding whether the evidence is nonhearsay or hearsay showing an abnormal state of mind.10
What the proponent does with the words decides how far the exception reaches, and a will contest shows the hinge. Where the issue is undue influence, a testator’s declarations are competent to prove her feelings for her relatives, but incompetent as evidence of her conduct or theirs.11
The first offer asks the jury to find a feeling. The second asks it to find a fact about Norman’s past that the testator merely remembered or believed, and that is the use the memory-or-belief clause withholds.12 The subsection on embedded factual assertions, below, takes up what a judge does when both uses ride in a single sentence.
The Then-Existing Requirement
The word then-existing in Rule 803(3) controls timing.14 It points to the moment the declarant spoke, so the test is whether the condition the words describe existed at that moment. Two things do not matter: whether the condition also existed when the events in suit occurred, and how long the condition lasted. “I hate my husband, Norman” reports a feeling that is present as the sentence is spoken. “Yesterday, after a big fight with Norman, my hatred of him grew to new heights” reports a feeling that belonged to an earlier day, and the exception does not reach it.
The line follows from why the exception exists. The rule’s categories depend on the declarant testifying, in effect, about something she perceives about herself, such as what she is thinking, what she is feeling, and what she plans to do; a statement of what she remembers doing is the category that stays out.15 Backward-looking statements are the problem because they are out-of-court accounts of something remembered from the past. They stand in for the live testimony a witness would give under oath in front of the factfinder, and admitting them would undermine the need for that testimony.16
The two remarks share a speaker, a subject, and an emotion, and they differ only in the time to which the feeling belongs. They also show why length is beside the point. A hatred nursed for twenty years can be proved by a remark that says “I hate him” when it is made, and a flash of terror lasting a few seconds qualifies if the speaker says “I am terrified” while feeling it. A remark such as “I hated him for twenty years” fails despite the duration of the feeling, because it looks back. Whether a feeling expressed at one moment supports an inference about the same feeling earlier or later is a separate question, which the next subsection takes up.
Verb tense is a clue to the timing gate and does not replace it, because a present-tense sentence can still look backward. A statement phrased as a present belief about a past event reports a belief that is real when the words are spoken, yet the proponent usually wants the jury to treat that belief as proof that the event happened. The memory-or-belief clause exists to catch exactly that use.17
On an exam stem, the timing gate and the memory-or-belief clause therefore work as a pair. The gate asks whether the words describe a condition that exists at the moment of speaking, and the clause asks whether the proponent is using those words to establish an outside fact about the past. A statement must clear the first and avoid the second. The section on the memory and belief limit develops the second half through Shepard, where a statement that looked toward the past could not be used to prove what happened.
Continuity in Time
The statement must describe a state of mind or feeling that exists when it is made, but the evidentiary effect of the statement is broader than that moment, because states of mind continue over time.19 The requirement controls what the words report. Continuity controls what the jury may infer from them. The inference is what makes the exception useful, since a case rarely turns on what the declarant felt during the sentence offered in evidence; it usually turns on what she intended or felt when something else happened, a day or a week away.
The inference runs both ways. A declarant who says on Tuesday that he intends to go on a business trip the next day has given evidence of that intention on Tuesday and of the same purpose on Wednesday, when he is on the road.20 Continuity may also look backward. When a will has been mutilated by its maker, the maker’s later statements of a purpose inconsistent with the will are received to show an intent to revoke at the time of the mutilation. Likewise, whether a payment or conveyance was meant as a gift may be shown by statements of intent made before the transfer, at the time of it, or after it.21 In each case the words describe only the declarant’s condition when she spoke, and the jury connects that moment to a neighboring one.
This subsection answers a different question from the last one. The then-existing requirement asks what the words describe; continuity asks how far the described condition may be stretched as evidence of another moment. The steps come in that order, and the first must be met before the second begins. “I hate my husband,” spoken on Tuesday, passes the first step, and continuity then lets the proponent argue that the hatred was still there on Wednesday. “Yesterday I hated my husband” fails the first step, and no continuity reasoning rescues it, because continuity enlarges what a qualifying statement can prove and cannot turn a backward-looking statement into a present one.
A hearsay objection to the Tuesday remark fails, because the remark fits the exception. The opponent’s better argument is that Tuesday’s feeling does not show Wednesday’s. That argument asks whether the state of mind continued to the material time, and the trial judge decides it.22 The judge may require some reasonable indication, in light of all the circumstances including proximity in time, that the state of mind was the same when it matters. How long a state of mind lasts depends on the particular attitude or feeling and on its cause, so the judge looks at both.23 A flash of rage over a spilled drink says little about the following week. A settled plan to retire to Arizona, voiced in March, says a good deal about the declarant’s purposes in April.
Intervening events, such as the new discovery in that scene, are the facts that most often break the chain between the moment of speaking and the moment that matters.
Embedded Factual Assertions
People seldom announce a feeling without explaining it, and the explanation is where the trouble starts. Rule 803(3) admits a statement of the declarant’s then-existing emotional condition and excludes a statement of memory or belief offered to prove the fact remembered or believed.25 A sentence such as “I’m terrified because he threatened me” therefore falls on both sides of the line: the fear is within the grant, and the threat is a remembered event outside it. Courts have developed a two-stage response, a default and a safety valve.
The default is to admit the statement for the purpose the exception serves and to forbid every other use. When a statement of mental state also asserts something about another act, and the truth of that assertion coincides with a disputed issue, the normal practice is to admit the statement and direct the jury to consider it only as proof of the state of mind and to disregard it on the other issues.26 That direction is a limiting instruction, a judge’s explanation that evidence admitted for one purpose may not be used for another. The compromise rests on what each alternative costs. Excluding the whole statement would discard evidence that is often a primary source of proof of a mental condition.27 Admitting it for all purposes would let any past accusation enter the case by clinging to a feeling, and the memory-or-belief clause would stop meaning anything.
The embedded assertion in that scene is the plaintiff’s account of what the other driver did, which is exactly the fact the jury must decide on liability. The instruction keeps that account from carrying weight on liability while leaving the fear available on damages. Whether the instruction costs anything depends on what else the jury has heard. Where substantial evidence has been admitted on the other act, the embedded assertion adds little, and probably little harm results.28 If the defendant has conceded the texting, or several eyewitnesses have described it, the jury has nothing new to misuse and the instruction is close to a formality.
The harder case arises when the embedded assertion is the contested issue and nothing else in the record proves it, because then the instruction asks the jury to hear the one fact it most needs and refuse to use it. Compliance with such an instruction is probably beyond the jury’s ability and almost certainly beyond its willingness.29 For that reason the trial judge keeps a safety valve. Where the mental state is provable by other available evidence and the danger that the jury will misuse the declaration is substantial, the judge should exclude the statement entirely or prohibit the witness from giving the reasons for the state of mind.30 The first condition lowers the price of exclusion, since the jury can still learn of the feeling from other sources. The second raises the benefit, since the harm from misuse is large. The less drastic remedy keeps the feeling and drops the cause: the witness may say the declarant reported being hopeful or afraid but may not repeat the accusation that came with it.
Shepard v. United States supplies the classic picture of why an instruction can fail. The Government there offered the dead wife’s declarations as proof of an act committed by someone else, that she was dying of poison given by her husband.31 That fact, the Court said, the Government was free to prove, but not by hearsay declarations. It would not do to say that the jury might accept the declarations for the light they cast on the existence of a vital urge and reject them to the extent they charged the death to someone else; such discrimination was “a feat beyond the compass of ordinary minds.”32
The Court’s image of the clang explains the mechanism. The “reverberating clang” of the accusatory words, it said, would drown all weaker sounds.33 Evidence of a will to live is a weaker sound than an accusation of poisoning, so a jury that hears both will carry the accusation into its deliberations whatever it has been told. The Court also stated the balancing idea behind the safety valve: when the risk of confusion is so great as to upset the balance of advantage, the evidence goes out.34 The facts of Shepard involve the use of a statement to prove a past act, and a later section returns to the case as the source of the memory-or-belief limit. What carries over to embedded assertions is its reasoning about what a limiting instruction can and cannot accomplish.
The two scenes share one structure, a statement of feeling with an accusation attached, and they come out differently because of the other proof and the stakes. In the crash case the accusation concerns a fact the jury will decide anyway on other evidence, and the plaintiff’s fear of driving is hard to prove any other way, so admission with an instruction costs little. In the partner case the feeling is well proved elsewhere, so the remark adds almost nothing on the issue it legitimately serves, while the accusation adds a great deal on an issue it may not serve. The exception applies identically in both scenes, and the difference lies in the judge’s balancing. Statements that express fear of a particular person press this problem harder, and the section on the memory and belief limit treats them separately.
Two errors are common when a stem pairs a feeling with an accusation. One treats the accusation as poisoning the entire statement and excludes the feeling along with it, which loses evidence the exception was written to admit. The other treats admission of the feeling as admission of everything in the sentence, which erases the memory-or-belief clause. The correct answer sits between them and depends on steps three and four, because the judge’s power to exclude or redact is triggered by the combination of an available alternative source of proof and a real danger of misuse.