Concept
Inference and Assumption
Relevancy
Inference and Assumption
Rule 401 sets the standard, any tendency to make a fact more or less probable, but it supplies no procedure for deciding whether a particular item meets it.1 The procedure is to write out the reasoning, and four terms organize it. The evidence is the item a party offers. The fact of consequence is the proposition that matters to the action and that the evidence must ultimately help prove.2 An inference is a step of reasoning from the evidence toward that fact. An assumption is the generalization about how people and things behave that makes the step reasonable.3
Naming the assumption matters because the argument is won or lost there. An opponent who sees the assumption knows which link to attack, and a judge who sees it can ask whether a reasonable juror could accept it. A bare announcement that evidence is relevant or irrelevant hides that work. The first subsection builds a chain from evidence to fact of consequence. The second applies the method to evidence about a different time, event, or person than the one in dispute, where the assumption holds only as far as the two situations resemble each other.
The Chain of Inference
Suppose Polly, a parrot, disappears from a fenced yard, and her owner offers proof that the alley gate was found open. The open gate supports an inference that an intruder came into the yard and took Polly. That inference holds only if we assume that people open gates, because gates do not open themselves, so the relevance of the gate depends on the validity of the assumption.4 Laid out in order, the reasoning looks like this:
| Step | Gate example |
|---|---|
| Evidence | The alley gate was found open. |
| Assumption | People open gates; gates do not open themselves. |
| Inference | An intruder came into the yard and took Polly. |
| Fact of consequence | An intruder took Polly. |
The assumption need not be certain. A truck passing in the alley might have shaken the latch loose, but gates are usually opened by people and not by passing trucks. Evidence stays relevant even when it does not raise a fact to the level of certainty, and as long as the assumption underlying an inference is reasonable, the logic essential to relevance is established.5 The rule's own wording points the same way, because it asks for any tendency and not for proof.6
Flight from a crime scene shows the same reasoning in a criminal case. A defendant who flees may have done so out of consciousness of guilt, or may be innocent and have fled to avoid being apprehended for some other reason entirely. Still, the premise that people who flee are more likely to be guilty than people who do not is at least plausible, and as long as some plausible chain of reasoning leads to the desired conclusion, the evidence is probative of that conclusion. For that reason, most evidence seriously offered at trial has some probative value.7
The gang evidence takes several steps to reach the disputed fact. It supports an inference that Defendant had a motive to silence the victim, and that inference depends on the assumption that members of the same criminal gang have a motive to protect each other.8 Most judges probably will make that assumption. The chain then runs on: a person with a motive is more likely than a person without one to act on it, which makes it more probable that Defendant struck the victim. Each link rests on a generalization a reasonable juror could hold, so the evidence is relevant even though the chain has several links. A later section takes up the separate question of how much a reasonable but weak assumption is worth.
Every chain that begins with testimony has one link that comes before all the others. Testimony proves a fact only if the witness is testifying correctly, so the first inference is that the witness is right about what the witness reports. That inference connects the testimony of all witnesses to the facts they describe, and it must be supported by a showing that the witness has personal knowledge of those facts. The showing is usually enough for the judge to infer that the witness might be correct.9
The judge treats that link as established once personal knowledge appears, because credibility is for the jury. A judge who held evidence irrelevant because she doubted the witness would decide the credibility question herself, and she should not.10 After the evidence comes in, each juror makes her own decisions about the validity of the assumptions in the chain when she decides what weight to give it.11
A trial court and an appellate court can judge the same assumption differently, as State v. Jaeger shows. The defendant was convicted of murdering his 19-year-old girlfriend, and he claimed her death was a suicide. The trial court excluded evidence that she may have attempted suicide years earlier, reasoning that proof of an attempt at a younger age is not probative of whether the same person would commit suicide years later.12 Written as a chain, the earlier attempt supports an inference that she was prone to self-harm, and that inference rests on the assumption that a person who has attempted suicide once is more likely than one who has not to do so again. The trial court rejected that assumption. The Utah Supreme Court reversed and held the evidence relevant.13 The result fits the any-tendency standard, because the generalization need not hold for every person, or at full strength after many years, to make the disputed fact somewhat more probable.
A chain breaks when a link has no assumption behind it that experience supports. Suppose a plaintiff injured on a building's stairs offers proof that she once fell down a different flight of stairs years earlier. The earlier fall is remote in both time and place, and nothing connects it to the cause of this fall, so without more the evidence is irrelevant. The proponent can repair the break with a connecting circumstance, such as a showing that the plaintiff wore the same high-heeled shoes on both occasions.14 The test is whether each link has a supporting assumption.
A chain can run long, and can run in opposite directions for the two sides, and still hold at every link. Knapp was tried for murder and claimed self-defense. The defense offered his testimony about a rumor he had heard that the sheriff beat and killed a man. That testimony makes it more probable that someone told him so, which makes it more probable that the sheriff killed the man, which makes it more probable that Knapp had reason to be worried, which makes it more probable that he had reason to defend himself. The prosecution answered with medical testimony that the man died of an illness caused by alcoholism and not of anything the sheriff did. The court held that this testimony was logically relevant too. It supports an inference that Knapp never heard the rumor he claimed to have heard, which makes it less probable that he had reason to fear the sheriff, which makes it less probable that he acted in self-defense.15
Length has a cost, though. Each inference that falls short of certainty lowers the probability that the final conclusion is correct, in proportion to the uncertainty of that inference. This is the product rule for combining probabilities, applied here to inferences from a single fact and not to different facts.16 A long chain therefore proves less than a short one, but Rule 401(a) asks only whether the evidence has any tendency, so length alone does not defeat relevance.17 How much a thin chain is worth is a question of probative value, which a later section takes up.
Similar Events and Remoteness
Many relevance problems involve evidence about an event or person not directly at issue, and the admissibility question is whether that evidence is similar enough to the matters directly at issue to affect the probability of a fact of consequence, or so different as to be irrelevant.22 In the chain of inference, the weak link is the step from the other event to the disputed one. That step rests on an assumption of likeness, that whatever explains the one helps explain the other, and likeness comes in degrees.
Two measures test it. Similarity asks whether the circumstances of the other event match those of the disputed one. Proximity asks how close the two events are in time and place.
A department-store case shows the difficulty when similarities and differences pull in opposite directions. A teenager on a high school gymnastics team falls in the store and sues for negligence, claiming the floor was slippery. The store says she tripped over her own feet. She offers proof that an hour later an elderly man who has difficulty walking fell on the same spot. The store objects that his physical ability is so different from hers that his fall says nothing about why she fell. The question is difficult. The falls share a spot and happened only an hour apart, but the two people differ greatly in physical ability.23
Reverse the roles and the answer comes into focus. Suppose the elderly man is the plaintiff and offers proof that a gymnast fell on the same spot an hour later. The gymnast's greater ability does not make the evidence irrelevant. If anything, the fact that even a young gymnast fell there suggests more strongly that the floor was slippery.24 A difference between two people therefore does not decide the question by itself. What matters is whether the difference supplies a separate explanation for the other event, as the elderly man's difficulty walking would, or makes a common cause more likely, as the gymnast's agility does.
All the measures operate at once in a stairway problem. A plaintiff hurt on a building's stairs claims that debris caused her fall. Proof that people fell on stairways in other buildings fails, because nothing suggests those falls shared a cause or circumstances with hers. Proof that the stairs were used at other times fails, because it is close in place but not in time, and the condition of the stairs on the day of the fall is what matters; timing is crucial to relevance. Proof that she fell on an unrelated flight years earlier fails for the reason given in the last subsection. Proof that the lobby of the same building was littered and uncleaned when she fell is different, because it is close in both place and time and makes it more probable that other common areas, including the stairs, were neglected too.25
Both halves of the hypothetical turn on the same assumption of likeness. Where the other event shares the cause, place, and time of the disputed one, the assumption is strong. Where it shares only a general resemblance, the assumption is empty and the probability stays where it was.
Time weakens the link the way distance does. A threat made the day before a killing may bear strongly on who committed it and why, but the same threat made ten years earlier probably adds little, because intentions and circumstances change over a decade. The kind of evidence is the same in both cases, and the variable that moves the answer is time proximity. Where the link is thin but real, the weakness goes to probative value, which a later section covers. Where nothing connects the events at all, as with the unrelated earlier fall, the evidence fails at the relevance step.