Clues are not necessarily evidence, and evidence is not necessarily proof, but recently I’ve heard clues presented to juries as if they were evidence and in closing arguments as if they were proof. I suspect this confusion of concepts pervades the justice system, and I also suspect the source of the error is law-school textbooks.
Judges decide what can be presented in court as valid evidence for juries to consider. As I understand it (and I am not a lawyer), before a trial begins, the judge holds an “evidentiary hearing” during which both sides argue the validity and legality of the evidence (wiretaps, fingerprints, DNA, crime scene photographs, etc.). Witness testimony is also vetted in the evidentiary hearing, based on pre-trial affidavits. Typically, certain types of witness testimony are excluded from the trial, including various types of hearsay.
In these hearings, the judge’s evaluation of the evidence is based on a number of “rules of evidence,” which come partly from English common law, from prior judicial decisions (precedents), frrom Supreme Court decisions, and in some situations from state and federal statutes. However, these precedents, high court decisions, and statutes are almost all derived themselves from English common law—unwritten law, that is, tradition.
Judges learn about English common law from law textbooks such as McCormick On Evidence, which is so well-respected as a source of information about evidence that Supreme Court decisions have cited it.
Part of the problem, though, is that textbooks aren’t “gospel.” Textbooks can and do perpetuate errors. As I have demonstrated in an earlier post, McCormick’s On Evidence was cited as the authority for the Supreme Court’s decision that firmly established “beyond a reasonable doubt” as the standard of proof of guilt, even though the textbook’s claim that the phrase was firm common law by the late 18th century is not documented (i.e., it gives no evidence that the phrase comes from English common law).
Modern textbooks are especially unreliable. I know, because I was once a college teacher and an editor of books for academics. Most modern textbooks are simply revised and updated editions of older ones, many of whose authors are now dead. Academic publishers no longer invest in quality editorial services. Many textbooks aren’t even fact-checked; instead their publishers rely on “peer review,” which means only that a friend of the textbook author gave the manuscript a read-through and thumbs up.
Clues are discrete, little facts that provide investigators with hints. Clues aren’t evidence. Every mystery writer knows the difference between clues and evidence. We don’t sprinkle evidence throughout a story; it would spoil the surprise. We sprinkle clues—hints, suggestions—many of which are intentionally misleading.
Clues in criminal investigations are like scientific data: some clues are meaningful data points, but most clues are just meaningless noise.
Sherlock Holmes’ great talent is identifying the meaningful clues. It isn’t every flake of tobacco that points him to the culprit, but only the rare pipe tobacco that few men in London smoke.
Modern cops aren’t all Sherlocks, though. CSIs may vacuum up all the tobacco flakes at a crime scene (not the cops) and then identify their source (perhaps). A police detective must then analyze all the clues (data) and decide which may help him to track down the culprit. Unlike Sherlock, most cops aren’t really very good at this analysis.
Clues are for developing hypotheses, not for proving a case. A certain type of pipe tobacco may lead a detective to hypothesize that a male pipe-smoker could be a suspect, but it isn’t proof of guilt. Connecting all the relevant clues may lead a detective to a suspect, but then the detective must gather solid evidence of the suspect’s guilt. The flakes of tobacco may or may not be among the evidence. The flakes may turn out to be noise. Just because a certain pipe smoker at some unknown time left a flake of tobacco at a crime scene does not mean he is the culprit.
Trial of Baseball Star
CNN/Tru TV’s In Session is covering the trial of Jim Leyritz for vehicular homicide. It provides a perfect example of a judge not understanding the difference between a clue and evidence.
A key piece of evidence for the defense is a videotape of the intersection where the accident occurred, which was taken by a traffic-enforcement camera. This, in conjunction with cell-phone evidence that suggested the victim had been texting while driving, was evidence for the defense that the accident was at least partly the fault of the victim.
The cell-phone evidence shows the precise time at which the victim was on the phone, because cellular technology is precise about the times and durations of each phone’s communications. The videotape’s time stamps, however, were not well-calibrated: traffic cameras are designed only to show that a traffic violation occurred (usually running a red light), where it occurred, and which vehicle committed the violation.
The defense argued that the time stamp on the video was wrong; the accident could have occurred seconds or a minute or two after the victim was proven to be on her cell phone. If so, she may have been distracted and not operating her vehicle with sufficient care when she entered the intersection.
In the evidentiary hearing a few days ago, an expert witness testified that the video time stamps not only could be many minutes off, they undoubtedly were.
Unfortunately for the defense, the judge clearly could not tell the evidence from the noise in this situation. He questioned the witness at length about the video time stamps, despite her repeated assertions that there was nothing in the technical systems that could prove with any certainty when the relevant frames were actually taped.
The judge even went so far as to argue with her that—if nothing else—the individual frames could be used to determine the speed of the vehicles. He pointed to pedestrians in several frames, extrapolated a typical walking speed, and compared the walking speed with the speed of the vehicles. When the expert explained that the number of frames per minute on the videotape were not calibrated either, he said it didn’t matter because of the pedestrians in the background.
The expert tried to point out that there was no way of telling what the walkers were doing between frames (traffic cameras take only a few frames per minute). She said they could have sped up, paused, or turned around briefly, for instance. The judge refused to accept (or maybe could not understand) what the expert was saying. To him, the speed of the perhaps inebriated pedestrians was evidence.
Most of the trials covered on cable TV contain clues presented as evidence. For instance, in the Scott Peterson trial, the lead detective read a list of clues on the witness stand that had caused him to suspect Peterson of murdering his wife; among these was an indentation in a duvet on the couple’s bed.
In the first Raynella Dossett-Leath trial, a neighbor testified that the defendant had waved at her from her car on the morning in question and had never done that before. Friends of the deceased testified that he was in a very good mood before he died, and this was presented as evidence against the defense theory of suicide. The deceased’s heir testified that her father did not believe in cremation and that made the defendant’s decision to cremate her deceased husband suspicious.
In both trials these highly questionable facts were clues that apparently led the detectives to suspect the defendants, but they were not evidence of any crime, let alone the guilt of the defendants.
In the Juan Mendez, Jr., trial, male DNA from the defendant’s family was found at the crime scene. (DNA is always presented as solid evidence these days.) In the Mendez trial DNA was not only presented as evidence, it was presented as positive proof that the defendant was guilty, even though he had lived in the house previously and his son lived in the house at the time of the murders.
If that isn’t noise, I don’t know what is. That kind of DNA ought not even to be a clue. The only relevant DNA in such situations is unidentified DNA or DNA from blood stains on a murder weapon or a victim.
In the Mendez case, the DNA ought not to have passed muster in the evidentiary hearing. IMHO: The judge who admitted it into the trial ought to be removed from the bench as incompetent.
How can I say this, when I’m not a lawyer? Well, because I have brain. The Florida grand jury also had a brain, apparently: they refused to indict Mendez because of the insufficiency of the DNA evidence. The prosecutor then lowered the charges to second-degree murder so he could indict Mendez on his own and try him before a six-person jury. Then he sent the samples back to the lab for more testing; in the interim the sensitivity of the tests had increased, so the lab was able to identify new DNA where before they had found none. As a result, the prosecutor was able to submit the new “evidence” of Mendez-family male DNA. The lab had not identified new clues, let alone evidence of the defendant’s guilt.
I’m not a lawyer and I’m even less a physicist. Lawyers don’t believe in systemic entropy, and physicists don’t use the term except to mean a specific physical state that’s capable of being described mathematically. I am, though, a bit of a computer geek, so I use the term in the sense of loss of data in communications and the tendency toward chaos that arises out of complexity.
The American justice system is collapsing under the weight of its chaotic complexity. The more laws that are written and the more judicial decisions that are written interpreting and twisting those laws, the less law and order we have.
During the English Civil Wars of the 17th century, the rebels called for “root and branch reform” of government, because it had degenerated into a state where life was “nasty, brutish, and short.” (Yes, scholars, I know that Hobbes was a monarchist.) IMHO the American justice system needs root and branch reform now, too. It’s nasty and brutish, but, even worse, it’s endless, like the law suit in Dickens’ Bleak House.
Congress apparently wishes to be the purveyors of the needed change. Recently it passed the National Criminal Justice bill, which will fund (from what monies?) a committee to study the problems in the justice system. As the source of our mare’s nest of laws, I hardly expect any good to come of any such Congressional effort. The feds can’t force much of anything on the states, which have the responsibility for crime control and punishment.
Our only hope, I suspect, is the U.S. Supreme Court. The criminal appeals process needs to be used to strike down many of the state statutes and centuries of judicial precedents that have warped the system. And the very first of these precedents I would like to see go is the Supremes’ decision in Daubert v. Merrel Dow Pharmaceuticals, Inc., 509 U. S. 579 (1993), which established junk forensic science as legitimate: the decision relies on “peer-reviewed” publications as the gold standard for science!
A small step in this direction was taken by the more recent Supremes’ decision in Melendez-Diaz v. Massachusetts, No. 07-591 (2009), which requires lab technicians who conduct tests to testify in court, rather than simply submitting a report to the prosecution. That it should take until 2009 for this to be required is horrifying, though.
How many people are in prison now because of junk-science lab tests that were never cross-examined in court? Raynella Dossett-Leath is one of these: the prejudiced ME in her case sent a vial of the deceased man’s blood to the TN state crime lab for analysis. The lab returned a report saying the blood contained a near-fatal level of barbiturates. In court only the ME testified about this report, not the lab technician. Worse yet, by that time the lab had lost the blood sample and cannot now prove either that the test results were valid or that they even tested the correct blood sample.