by Bill James
The only thing is, I didn’t purchase no goddamned watermelon. Motive, means and opportunity have nothing to do with real evidence. Real evidence that I purchased a watermelon is like a sales receipt for a watermelon that has my fingerprints on it, a check that I wrote to the grocery store for that amount on that date, and a videotape of me carrying a watermelon out of the store. There’s a half-eaten watermelon in the refrigerator and watermelon rinds in the garbage; you got me. Motive, means and opportunity, you’ve got squat.
The same applies to criminal cases. There is a place where motive, means and opportunity are useful in understanding a crime. MMO is a kind of “directional arrow” which, very early in the case—in the investigative stages, when the case is in the hands of the detectives—can be used to focus the investigation. There is a circle of fifteen people surrounding the victim. Of those people, maybe four had some kind of grudge against him and two others stood to profit from his death, so that’s six of the fifteen. Two of those six have solid alibis for the time of the assault, so there’s no opportunity there, four left. There’s a bullet hole in the victim and one of those four has been seen carrying a weapon which has now disappeared, bingo.
Motive, means and opportunity are useful in telling the investigators where they should look for evidence, at the beginning of the case. But if a prosecutor is arguing at the end of the investigation that the accused had the motive, means, and opportunity, what she is essentially saying is that she doesn’t have a clue what happened.
OK, so what does work?
Let us begin by asking this question: is there any one piece of evidence, in any case, which is entirely sufficient to sustain a conviction?
No, there is not. A confession? People confess to murders that they didn’t commit every day of the year, and sometimes they figure out enough facts about the case to convince people that they actually did it. A videotape? They’ve got fifteen cameras pointed at a running back every Sunday, and at least thirteen of them won’t tell you whether he stepped out of bounds or not. What if it’s somebody who looks a lot like me? Didn’t you ever know two people who looked just alike? If you have a common face, like mine, there are hundreds of people who can easily be mistaken for you.
The most convincing single piece of evidence that we see in real cases is DNA from semen in a rape/murder. But even that evidence, by itself, is not sufficient to convince beyond a reasonable doubt, because it remains possible that the victim had consensual sex shortly before she died with someone other than the person who murdered her.
But certainly, DNA evidence in a rape/murder case is very, very damning evidence. Suppose that it takes 100 points’ worth of evidence to convict a person beyond a reasonable doubt. How many points do you give to the DNA evidence?
The answer is fairly apparent: it has to be about 80 points. It has to be in that range. It doesn’t take two or three or four facts of that nature to convict somebody; it just takes one, plus a little bit of a shove. Such a fact is not proof by itself, but it is so close to being complete proof that the rest of the proof must almost inevitably follow like the dirt that comes up with a carrot when you pull the carrot.
On the other hand, a great deal of the evidence which is introduced in crime cases is so tangential that it would be impossible to score at even one point. Crime books … and this is really what I am interested in; not the trial, but the crime itself and the whole universe of admissible and inadmissible evidence which surrounds it … crime books are normally composed almost entirely of “evidence” that is so tangential that it bears no weight whatsoever. A book about Rabbi Neulander points out that he served shrimp at his wedding reception—an odd choice for a prospective Rabbi. It is evidence against him, sort of. It paints a picture of a Rabbi who is less than 100% committed to his belief system, thus it might be more believable that he went bad, as opposed to Rabbi Leibowitz across the hall, who served whitefish.
This is evidence, I suppose, but one hundred facts of this nature would not be sufficient to convince anyone that Rabbi Neulander was a murderer. Books arguing that Sam Sheppard was innocent or that Richard Hauptmann or Julius Rosenberg was innocent will go out of their way to tell stories about the time that Sam or Richard or Julius stopped to help an old lady with a flat tire. They are essentially arguing that “this is not a bad person; this is a good person. This is not the kind of person who would pass nuclear secrets to the Lindbergh baby.”
What do we do with such evidence as this? How do we state the relationship of these alleged facts to the guilt or innocence of the accused?
I would suggest that we can approach this problem by working the evidence in a seven-step process.
1) State the fact itself in a way that is unambiguously true.
2) State that which tends to be proven by the fact, as if this was known to be true.
3) Put the statement of fact proven by (2) in a “standard evidence” form (a statement of evidence, as opposed to a statement of fact).
4) Establish the value of the statement of evidence (3) with reference to a standard set of values for such evidence.
5) Make an estimate of the extent to which statement (3) is unproven.
6) Make an estimate of the extent to which the statement (3) is irrelevant.
7) Discount the value (4) by the extent to which the statement is unproven (5) or irrelevant (6).
We begin by making statements about the case that are unambiguously true. The unambiguously true statement is not that Lizzie Borden attempted to buy poison on the morning before the murders, but that it is alleged that Lizzie Borden attempted to buy poison on the morning before the murder.
This is certainly a damaging fact, if true, but … why is it a damaging fact? We are moving on to (2) here, trying to state that which is proven by the fact. The fact is damaging because it tends to prove that Lizzie Borden was involved in a previous effort to murder her parents.
The range of things that are proven by evidence is much narrower than the range of things that are evidence itself. DNA in semen is devastating evidence, because it proves that the accused had sex with the victim near the time of her death. The same thing might be proven in ten thousand other ways. A bartender may report that he saw you leaving the bar with your arm around the victim. This proves the same thing, only less definitively. The victim’s downstairs neighbor may report that she heard noises coming from the apartment that sounded like two people having sex, and 20 minutes later she saw someone who looked kind of like you leaving the apartment. Your fingerprints may have been found on a wine glass in the apartment.
If the court has the accused’ DNA as a starting point, none of these other facts add much to the case against him. They all serve to establish the same fact that there was a sexual encounter between you and Hermione just before Hermione died. Whereas the universe of relevant facts (1) is incomprehensibly large, the universe of things proven by those facts (2) is much smaller.
We can make that universe yet smaller (3) by placing the statement in the form of a type of evidence that normally appears in (let us say) a murder case. In this case, Lizzie Borden was involved in a previous effort to murder her parents becomes the defendant had been involved in prior acts of violence toward the victim.
At this point our protocol calls for us to check the potential value of this statement of evidence against a list of standard values of such statements. Unfortunately, since no such list of values exists, we’re going to have to wing it. If the defendant has been involved in prior acts of violence toward the victim(s), what is the value of such evidence, on a scale in which 100 points is proof beyond a reasonable doubt?
This statement of evidence would appear in the table of values as one of a set of three, for which I will suggest values:
1. The defendant had been involved in prior acts of violence toward the victim(s) (35 points).
2. The defendant had made threats against the victim(s) (25 points).
3. The defendant bore malice toward the victim(
s) (15 points).
These three, of course, often appear together, and when they appear together—that is, the accused has made threats against the victim and has also been involved in prior acts of violence toward the victim—they can go a long way toward convincing me that the accused is guilty of the crime. If B is found murdered and there is clear, convincing evidence that A had a long history of violence toward B and had threatened to murder B, I am more than 50% convinced that A is good for the crime. Let us say that the potential value of (1) and (2) combined is about 60 points.
But what is it in this case? First, how certain is it that Lizzie was in fact involved in a previous act of violence (attempted poisoning) directed at the victims?
Not at all certain … in fact, it seems unlikely that it is true. Abby Borden told her doctor/neighbor that she thought she was being poisoned, it is true, but
1. The doctor dismissed her comments out of hand, and
2. Thorough and professional autopsies were performed on the victims by the leading experts in the field, finding no evidence of poisoning.
A pharmacy clerk did claim that Lizzie had attempted to buy prussic acid on the day before the murders, it is true, but people come forward in cases that excite the public’s imagination and say all kinds of things, many of them untrue—in fact, most of them untrue. The clerk did not know Lizzie before the attempted purchase; he stepped forward in the investigation in response to rumors. He appears to be just an excitable gossip who pushed his way into the case. Think about it this way: setting aside the murders, could you convince a jury beyond a reasonable doubt that Lizzie Borden was trying to poison her family? You couldn’t come close. We’re not within a mile of clear proof that Lizzie Borden was involved in an effort to poison her family.
The 6th stage of the analysis—relevance—does not apply a further discount in this case. We would discount the fact for relevance if, for example, the previous violence was remote in time from the murder, or if it was of a distinctly different nature than the murder. The 60 points for (1) and (2) combined would apply, for example, if an abusive husband had repeatedly told his wife “if you leave me I’m going to kill you.” But suppose that the accused and the victim had once had a shoving match. A shoving match is very different from a murder, so that’s more like 10 points than 35. If it was a shoving match that occurred two years ago, if there were no incidents between the victim and accused in the intervening two years, it might be 2 points. In this case, we’re talking about murder-on-murder (a poisoning plot vs. a hatcheting, true, but murder is murder) and about events that occurred on the day before the murders. It’s entirely relevant, if it’s true.
So what value do we place on this evidence? We can’t “confirm” that Lizzie was trying to poison the family unless the family was actually being poisoned, which, in my opinion, they probably were not. The 35 points would apply in a case where there was a well-documented history of violence by the accused toward the victim—an exceedingly common feature of real murder cases. We don’t have anything like that here. We don’t have Lizzie beating up her stepmother, pushing her downstairs, threatening to kill her if she doesn’t do this or that … we don’t have it. We don’t have Lizzie being aggressively mean to other people. While it is well documented that Lizzie did not particularly care for her stepmother, the events of the first week of August, 1892, arise out of a blue sky in terms of a documented history of violence. Pondering the massive gap between the actual evidence and the evidence necessary to convince a skeptic that Lizzie Borden was attempting to poison her family, I would discount the charge by about 70%, and score these facts at about 10 to 12 points.
The case against Lizzie can be summarized into thirteen unambiguously true statements:
1) Her stepmother complained, 24 hours before the murder, that someone might be trying to poison the family.
2) A drugstore clerk testified that Lizzie attempted to purchase a poison, prussic acid, the day before the murders.
3) Two other persons confirmed that Lizzie was in the drugstore on that day.
4) Lizzie denied that she was in the drugstore that day.
5) Lizzie did not like her stepmother.
6) Lizzie stood to profit by the crime, by inheriting a substantial fortune.
7) Lizzie’s emotional affect after the murders was flat, and seemed to some people inappropriate to the situation.
8) Lizzie burned a dress three days after the murders which the prosecution claimed was the dress she had been wearing at the time of the murders.
9) Bridget (the maid) said that she heard Lizzie laugh from the landing, near where Abby had been murdered, as she unlocked the doors for Andrew Borden at about 10:45 AM.
10) Lizzie said that she was in the barn at the time of the murder, in the hayloft.
11) Two police officers said that there was a thick layer of dust on the floor of the hayloft, and it didn’t appear that anyone had been there for some time.
12) The house was locked, and no one else is known to have been in the house, other than Lizzie and Bridget and the victims.
13) No murderer was seen coming or going from the house.
There are a lot of other known facts about the case … it would take me several thousand words to summarize them briefly. There’s a hatchet with a newly broken handle, the head of the hatchet apparently rolled in ashes. There’s a relative who was staying in the house, a sister who was out of town, etc. What we’re trying to do is to shrink the universe of facts so that we can get our arms around it. None of those things are evidence against Lizzie any more than they are evidence against anyone else, and I’m going to focus here on the facts that relate to Lizzie. Lizzie’s alleged lesbianism is about as relevant as the shrimp at Fred Neulander’s wedding reception, and is certainly less relevant than the well-documented fact that Lizzie taught Sunday school.
I earlier evaluated the “poisoning” evidence at about 10 to 12 points … let’s say 12. I think that these thirteen unambiguously true statements can be summarized into six prosecution contentions, or “statements of evidence”:
1) That Lizzie was involved in an earlier effort to poison the family.
2) That Lizzie disliked her stepmother, Abby Borden.
3) That Lizzie stood to profit by the crime.
4) That Lizzie’s actions after the murder show consciousness of guilt.
5) That many of Lizzie’s statements about the case are untrue.
6) That no one except Lizzie had the opportunity to commit the crime.
Let me back off and answer a question to which I have, to this point, just assumed that the answer was apparent. Why the hell are we doing this?
We are doing this because the evidence in crime cases—like Lizzie Borden’s—is often very confusing, and it is often extremely difficult to judge whether the evidence against a defendant is or is not sufficient to conclude that he or she is guilty beyond a reasonable doubt. I have read books about people who were convicted on what seemed to me to be very flimsy evidence, because the prosecutor was much better at selling his evidence than the defense lawyer was. I have read books about people who walked free despite what seemed like overwhelming evidence. These things would happen less often if people thought more clearly about evidence. I am trying to propose a structured system to enable us to think more clearly about the evidence in criminal cases, so that, in theory, injustice might be avoided, and justice might be delivered, a little more often.
A trial is rather like a basketball game at which no one keeps score, but at the end of the game the audience is asked to vote on which team has played better. All games are supposed to be on a neutral court. I am trying to propose, because I am too ignorant to understand that this is impossible, a system of keeping score.
One of the dirty little secrets of crime stories is that some people follow them not because they are morbid and obsessed with the salacious details of horrible crimes, but because they are anxious to see justice. Crime stories are about the search for
justice. I am trying here to offer my ideas about how this process could work better. It makes my blood boil when I read about some poor sap being convicted of a crime because he had the motive, the means and the opportunity. I think that somebody should point out that this is, when you think about it, a really stupid concept. The world will reject these ideas; lawyers will scoff at them, judges will ignore them, juries will never hear of them and will probably be legally prohibited from giving them any weight if they happen to hear of them. That’s OK; I’m going to say what I have to say. I can live with being ignored.
Now, back to the six statements of evidence against Lizzie, having already dealt with the first:
2) Lizzie disliked her stepmother, or the accused had malice toward the victim.
This means almost nothing in my view. We all have people we don’t like. If one of the people you don’t like is murdered, it is a very long step from “George didn’t like Fred” to “Fred is dead; George must have killed him.” I would give a well-documented dislike of the victim a potential evidentiary weight of 15 points, but
a) That would be in the case of an unusually clear and well-documented hostility, and
b These 15 points are not to be a redundant accounting of the values given to threats or prior violence.
While it is well documented that Lizzie did not like Abby Borden, she loved and revered her father, Andrew, and is not known to have spoken ill of him. The Borden house was not a happy one—you have four adult women, ages 25 to 65, living in a relatively small house with a sour old miser—but there is tension of some kind in most families. The animosity between Lizzie and Abby was not of a type or intensity that is normally predictive of violence; they just didn’t really like each other. That’s fairly normal. You can’t convict people of murder based on the normal conditions of everyday life. I would score this evidence at approximately 2 points, giving us about 14 points in the case.