Science and the Law: Uncomfortable Bedfellows, Jed Rakoff
Articles,  Blog

Science and the Law: Uncomfortable Bedfellows, Jed Rakoff


– It’s my distinct pleasure to welcome, Judge Jed Rakoff, to Dartmouth to give this year’s Timbers Lecture. I’d like to start first by thanking the Dartmouth Lawyers Association, and all the people who have worked on this Timbers Lecture
to make it possible, including the Rockefeller Center staff, and many other people, so thank you very much. Judge Rakoff was born in Philadelphia, and attended Swarthmore College where he now sits on
their Board of Managers which is their equivalent
of our Board of Trustees. After Swarthmore, Judge Rakoff went to Oxford University to study South Asian history, and wanted to pursue a career in writing musicals, so he went to Harvard Law School thinking that he could do
the musicals on the side. I don’t know if any of the musicals have actually come out. After Harvard Law School he was Law Clerk to Judge Freedman of the United States Court
of Appeals Third Circuit. Judge Rakoff went onto practice law as assistant United States attorney as well as a litigation partner in several prestigious firms before 1996 when he became United States District judge for the Southern District of New York, and he has also been a lecturer at Columbia Law School since 1988. In addition to over
100 published articles, and 400 published opinions Judge Rakoff has three books, one on RICO, one on corporate sentencing guidelines, and one on modern federal
jury instructions. He’s now working with others to revise the federal judges manual
on scientific evidence. I want to add three examples to give you some sense of Judge Rakoff, and how opinionated he is, and outspoken he is. In 2002 Judge Rakoff declared the federal death penalty unconstitutional holding that it quote: “Creates an undue risk “of executing innocent people.” Now this decision was
overturned on appeal, but it still will, we hope, have an effect on the law. More recently Judge Rakoff ordered the Department of Defense to release the names of the prisoners being held at Guantanamo, and this time the government complied with the order and didn’t appeal it. More personally, I want to mention that I met Judge Rakoff planning the MacArthur Project on Law in Neuroscience, and he provided much wise advice from the very beginning. He also put in countless hours, and let us use his courtroom for the launch of our project, and he’s continued to be
an invaluable resource to the neuroscientists and philosophers who need to learn something about the law before they can say anything which might be helpful. And I’ve learned as much, if not more, from Judge Rakoff on those issues as from anyone else in the project. So I look forward to
learning even more today as he speaks to us about “Science and the Law:
Uncomfortable Bedfellows.” Please join me in welcoming Judge Rakoff. (applause) – Well, I want to thank Walter for that very kind introduction. If my wife were here she’d want rebuttal time at this point. Also, thanks for mentioning the musicals. I promise a full song and dance. I’m very honored, genuinely honored to have been asked to
give the Timbers Lecture because I knew Judge Timbers quite well. He was in the Court of Appeals for the Second Circuit which includes New York, and Vermont, and Connecticut, and I appeared before him many times. He was always extremely gracious. He always asked very insightful questions. Towards the end of his career he actually decided one or two cases in my favor. That’s when I knew he was getting old, but he was a very great judge, indeed, and it’s so wonderful that this lecture has been named after him. I also want to mention that, although, this is my first visit
to Dartmouth College it is not my first visit to Hanover. My three daughters when they were younger each went to an overnight camp about 17 miles up the road in Fairlee, Vermont. The camp was named Camp Aloha, and that’s because it was founded by Hawaiian missionaries who traveled all the way from Hawaii to Vermont so that they could convert the heathen, or perhaps the Presbyterians, but every summer my wife, and I would go up to Camp Aloha, and visit for parents’ weekend, and by the end of parents’ weekend all the parents were in
desperate need of a drink, and so the natural place
we came was to Hanover. You can figure out why that might be. My topic is Science and the Law. I am going to refer in passing to some of my cases that have now been completed, but I do want to make clear that nothing I’m saying here today is being said in my capacity as a judge, and so I will try to make my comments as injudicious as possible. But I will start with the
obvious point which is that law and science
are now yoked together to a really unparallel extent. Just in the past few months I’ve had cases in which in a drug case the chemistry of the sweat patch that was used to detect cocaine use was questioned. I’ve had cases involving file swapping where the technology involved in what was
ultimately a copyright case was very much under dispute as to how it operated whether it was effective or not. I’ve had cases just, again, just in the past few weeks where psychiatric measures for detecting post-traumatic stress syndrome in personal injury situation has been questioned as not being very good science. I’ve had cases where the determination of economic loss in securities fraud cases has been the subject of heavy debates about the different forms of regression analysis used, and it goes on and on. It is, of course, a reflection of the growing importance of technology and science in our society as a whole, but it has very much
infiltrated the courts. The courts have kind of a love-hate relationship to science. This really goes back
decades, if not centuries. Law is by its nature a bit messy. Basically, the way the law resolves everyday disputes, is by applying general principles to specific facts. Identifying the general principles is often not that hard, but finding the facts
can be very difficult. The law looks, therefore, or is attracted to science as a tool, if you will, of helping find the facts
with greater certainty. On the other hand lawyers have used the opportunity to present science as an opportunity to introduce into court a lot of pseudo science. A lot of soft science, a lot of … To use that unfortunate
term, “junk science.” Judges have often felt sort of burned, and misused at the way science has crept in to some of our cases. Let me give a somewhat broader elaboration of that love-hate relationship. In Medieval England there were really only a couple of ways in the origins of Anglo-American Law the facts could be found. The jury was actually in those days a fact finding body and went out, and tried to work as a kind of detective, but it was very unsatisfactory. Eye witness testimony was often, as it is even today, very unreliable. And what the Medieval lawyers tried to do was, therefore, narrow each and every case through what is called the Démarrer system to a single, legal very narrow issue, an issue of pure law that could be decided almost irrespective of all, but the most undisputed facts. That approach was very unsatisfactory. It really gives lawyers the bad reputation they still have to this very day of being more interested in legal niceties than in justice. And eventually it was discarded. On the religious side the way facts were found was, for example, through the use of ordeals. There was, if you remember, back to your Medieval history the Ordeal by water. The accused person would be thrown into the river. If the person was guilty the pure waters would reject the person, and you would know then that he was guilty because he floated. If on the other hand
the person was innocent the pure waters would embrace him, and he would sink to the bottom. Now there were downsides to this system, but not everyone drowned, and it didn’t matter if you drowned because you were pure and innocent, and, therefore, you would go to heaven so all would be well. There was, also, particularly in the early renaissance,
the use of torture. The problem with torture is it’s not a very accurate way to obtain information because the torturee, if you will, wants to give the torturer what the torturer wants to hear whether it’s true or false. His whole point is to get out of the pain by saying what he thinks the torturing person wants him to say. Of course you know no one would use torture today, that goes without saying. But the point is that none of these ways of finding the fact
were very satisfactory, and what the Anglo-American system ultimately came up with in preference to all these techniques was the adversary system. Basically, the idea was that each side would have a champion who would cross-examine the other sides witnesses who would scrutinize the
other sides documents, and that by subjecting each side to this searching inquiry by someone who was a champion of
the other point of view the truth would emerge, and I think that on the whole that system has worked reasonably well. We’ve learned in recent years that it perhaps hasn’t worked nearly as well as we thought. Walter made reference to my death penalty decision. What that was based on was the fact that here we had what are
really dozens of cases of people who were convicted beyond the reasonable doubt. Cases were Appellate Courts had often in rejecting the appeals said: “Oh, the evidence of guilt was “absolutely overwhelming. “There is no doubt at
all that this defendant “committed this heinous crime.” And then lo and behold, in those relatively rare cases where DNA was available it turned out that they had the wrong guy. My opinion said in effect that since it’s not much good to be exonerated if you’re dead the death penalty did not at courts due process and had to go. I was promptly overruled
by the Court of Appeals which proved to me that
law is not a science, but the point I’m trying to get at is not so much that decision, but the fact that the adversary system, good as it is, is not a perfect way of
ascertaining the truth. Long before DNA people
kind of instinctively recognized that eye witness testimony, and cross-examination,
and things of that sort while good were not perfect, and so beginning in the 20th century they looked to something else, another magic bullet. Torture hadn’t worked, ordeals hadn’t worked, but maybe science would work because science, after
all, represents certainty. Now, actually, most
scientists today would say that science is about
probabilities, not certainties, but the public thinks of it more in terms of certainty, and the result was an attraction of the legal system to finding some “scientific” way of arriving at the truth. This, regretfully, led to some pseudo science coming into play as well. Sometimes it wasn’t accepted by the legal system, but even then it tended
to creep in indirectly, and let me give you an example of that which is the polygraph, the lie detector which was invented in
the early 20th century, and by 1923 the courts had decided that it was inadmissible because it was not really good science, and that’s true today
in every jurisdiction, both state and federal except New Mexico. I’m not quite sure why New Mexico, but it’s been kept out of trials, but it has crept in indirectly because it is regularly
used by law enforcement. The justification that has been given for its use by law enforcement. In fact, there was a
whole article about this in the “Wall Street Journal”
just about a month ago is that most people
think lie detectors work, and, therefore, if you administer a lie detector to them they will confess, they will tell the truth
because they’re afraid the machine will show that they’re lying, so whether it really works or not doesn’t matter because people think it works. And there have been some decisions including from some federal courts that have approved the
use of lie detectors by, for example, probation officers who are technically in
arm with the judiciary on this ground that if you can fool all
the people all the time it will help you get at the truth. Now, I personally find this argument rather repellent, frankly. It, basically, what you’re saying is let’s lie to everyone. Let’s pretend lie detectors
work when they don’t, and that way we will fool them into telling us the
truth by lying to them. But that is the argument that is used. Now on the science, just so you know where that stands. The theory of the lie detector is that when you lie you become more anxious, and that that shows up in the form of increased heart rate, increased sweating, increased respiration,
breathing, and so forth. That theory has never been tested, never been tested, and if you just think about it from common experience you might think well, there are a lot of practiced liars who maybe don’t get very anxious when they’re lying. There are lots of people who probably get very anxious taking a lie detector test whether they’re lying or not, and, therefore, their
pulse rate will go up, and they will start sweating not because they’re lying, but because they’re taking this test, in a very important moment in their life, and there are lots of pretty
good studies that show that blood pressure, and respiration rate, and pulse rate and sweating all vary with a hundred different factors, some external, some internal that have nothing to do with whether you’re lying or not. In 2003 the National Academy of Sciences did a very thorough study of this, and they concluded that while the lie detector did detect lies slightly better in chance it had a very high error rate, and that it was based on weak science, and that it was much more believed in than the evidence warranted. Now the Supreme Court, I should say, reviewed the same literature
a few years earlier, and did not take quite
such a negative view. They took the view that
the scientific community was polarized as they put it, and that some scientists believe that lie detectors were only slightly better in
chance in detecting lies, but there were some people who believed that it was as good as 87% accurate, but you know, even a 13% error rate is a huge error rate
when we’re talking about assessing a key witnesses credibility, or determining whether or not someone has committed a crime. So that’s even on the best sense a pretty high error rate, but not withstanding all that, as I said, it has continued to be used. I might add even assessing error rate in the case of polygraphs
is extremely difficult, and a lot of the studies
have been criticized as not being very good even in assessing error rate. I’m mentioning all this because it creeps into the courtroom. It’s not only that it’s
used by law enforcement, but it creeps into the courtroom as well in ways that in my view
cause a lot of mischief, and I will give you one example from my own court. Some of you who are
familiar with Ground Zero will know that there is the Millennium Hotel, that is right next to where the two World Trade Towers used to stand, and everyone in that hotel was evacuated, of course, on 9/11, and then a few months later the security employees, the private security
employees of the hotel were allowed back in the hotel to recover the clothes
and other belongings that had been left by the guests when they had had to evacuate. One of the guards, a fellow
named Ferry, F-e-r-r-y, reported to the FBI that he had found on the 51st floor, way up near the top of this hotel, in a room occupied by an exchange student named Abdallah Higazy, a copy of the Quran and in the safe in that room a pilot’s radio, the
kind of radio that’s used to guide planes from the ground, and he reported that to the FBI, and they immediately, and understandably began to wonder whether this had been used to help guide the planes on 9/11 that crashed into the World Trade Center. They interviewed Mr. Higazy, who was still here as an exchange student, and they discovered that
he was a former member of the Egyptian Air Force, and that made them more suspicious, and they asked him: “Did you have a pilot’s radio “when you were staying in the Millennium?” And he said flatly: “No.” Well, that made him extremely suspicous because here the security guard who had no apparent motive to lie had told them that he had found the pilot’s radio in the
safe in the very room that Mr. Higazy had
been occupying on 9/11. So, they arrested Mr.
Higazy on what’s called the “Material witness
warrant” which allows you to hold someone briefly, so you can present them, in this case, to a grand jury for investigative purposes when there’s some risk of flight, and pursuant to that warrant he was brought before me, and I expressed some doubt about how strong the government’s grounds were for holding him even on a
material witness warrant, but I agreed to hold him for a few days. And right there in my court, Mr. Higazy kept demanding that he
be given a polygraph test to prove that the radio was not his. I explained to Mr. Higazy
that polygraph testimony is not admissible in Federal Court, and that’s because the
tests are unreliable, but he kept insisting. The government, interestingly, the Assistant United States Attorney also said as part of that hearing that they didn’t think there’d be much point in giving a polygraph test to Mr. Higazy because
terrorists had been trained how to beat this “scientific” instrument which was an interesting
commentary in itself along the efficacy of polygraphs, but after he left the courtroom overruling his lawyer he insisted that he be given a polygraph test, and the
FBI which does conduct these tests said: “Okay.” So he was brought into the polygraph room of the local FBI headquarters. It was just him and the polygraph operator because according to
the polygraph operator the machine is too “delicate” and will be affected if
anyone else is in the room. So his lawyer had to remain outside, although he was told he could go and see his lawyer if he wanted to at any time. As soon as he started
taking the polygraph test he was asked about the radio, and he kept saying it wasn’t his radio, and the FBI examiner said: “I can see from this
test that you’re lying. “You better come clean.” And this went on for three hours at which point Mr. Higazy was somewhat hysterical, and he then blurted out: “Yes, I guess maybe it is my radio.” At that point the FBI
agent stopped the music, said: “We got a confession. “We want to put it down in writing.” He started to prepare a written confession at which point Mr. Higazy finally realized maybe he ought to talk
to his lawyer outside. He went to talk to his lawyer, and the lawyer stopped the music. So he was then brought back before me. Now he was charged criminally, and the actual criminal charge came down on January 11th, 2002. Three days later an American Airlines pilot who had been also staying
at the Millennium Hotel walked to the place where they were holding the clothing for the people who had been evacuated, and he got his clothing and he said: “Where’s my radio?” And, of course, the whole
thing then unraveled, and it turned out this
American Airlines pilot had had the radio in his room on a completely different floor, and it was completely
legitimate that he had it. It had never been in
Mr. Higazy’s possession, had never been in Mr. Higazy’s room, and the security guard had lied basically because he thought he
wanted to get revenge for 9/11 on anyone of Arab ancestry. Well, the charges were dropped against Mr. Higazy. The guard was criminally prosecuted. I ordered, though, because I was bothered by the polygraph side of this an investigation into what had occurred in the polygraph room, and I got a report back from the Department of Justice that said that the operations in the polygraph room were in accordance with
standard FBI practice. Now I’m not sure whether that means that it is the FBI agent actually believed in the efficacy of the matter which he was conducting the polygraph test, or whether, again, this is used in effect as a tool to try to get confessions. Mr. Higazy does have a civil suit pending before a different judge against the FBI, and I don’t
want to comment on that, but my point is that if there’s one thing that example shows as it
does in others I could cite it is that polygraph lie
detection is not science. It’s not remotely science, and when it creeps into the courtroom it can create a great deal of mischief. Now a different kind of example, and one where the courts have been much more receptive to
scientific evidence, and have allowed it into
the courtroom properly is psychiatric evidence. This sort of reached its high point in the 1950’s when psychoanalysis was very much accepted among enlightened people in the United States, and a good example of how this came in is the insanity test. The original defense of
insanity was formulated by the British courts in 1853 in a case involving an
attempted assassination on the British prime minister by a guy named M’Naghten, and the standard that that test formulated which was then generally adopted was that you’re legally insane only if you don’t know right from wrong. A test that’s fairly narrow, and was even more narrow
in its application. And that was very heavily criticized by the psychiatric community. In 1954 in a case by a very well respected American judge, David Bazelon, in a case called “Durham” a new test was promulgated which quickly was accepted by
virtually all courts, and that was that you are legally insane in terms of a given criminal charge if your misconduct was quote: “The product of mental disease or defect.” Now what actually happened because there really is no well-established scientific test for what is a
mental disease or defect. The analogy by Judge Bazelon in effect was it’s like measles. You see it, you figure out what causes it, and you figure out the cure, but it’s not like that. The result was that for many years in many cases you had on the defense side a very well credentialed
psychiatrist saying: “The reason he committed that crime “was he is suffering from “the following mental disease,” and you had an equally well credentialed psychiatrist
on the other side hired by the prosecution who said: “It’s absolutely clear
that he is not suffering “from any mental disease.” Both the experts had had the same access to the defendant. Both had had the same
access to his records. So whatever else this was it wasn’t the kind of science that the courts had in mind when they were seeking something that would give them greater certainty rather than less. It was also extremely difficult for courts to evaluate between
these two psychiatrists as to who they ought to believe. In 1981 using the Durham test the defense won an acquittal
by reason of insanity for John Hinckley who is the fella who attempted to assassinate
President Reagan. Some states then changed the law, and actually overruled Durham, and substituted either no test at all, no legal defense of insanity, or a defense going back
to the M’Naghten rule, but either where that didn’t happen, and it didn’t happen in the Federal courts judges became much more skeptical about the admissibility and usefulness of psychiatric testimony. It appeared to many of us
that we were essentially just delegating to the psychiatrists a determination that they were in not much better position to evaluate than everyday jurors. Putting this in broader perspective the problem that confronted the courts say as of around 1990 was that on the one hand courts had been burned
and were very skeptical about the admissibility
of scientific evidence because so much pseudo science had crept in either
directly or indirectly, but on the other hand
the courts recognized that in an ever increasingly technological society science was gonna inevitably come into the courtroom more and more, so the problem was how to separate the good science from the bad science for lack of better terms. And the Supreme Court in a case called: “Daubert verus Merrell
Dow Pharmaceuticals” in 1993 took that issue on directly, and what they held was which is now bodied in rule 702 of the Federal rules of evidence was that judges had to act as gatekeepers that this was not a test for the jury, this was a test for judges. It was not a test for scientific panels. It was a test for judges, and that the judges had to determine whether scientific evidence would come in, and to do that the scientific evidence had to be three tests. First, it had to be
based on sufficient data. Second, it had to be based on reliable methods, and third, the reliable methods had to be applied reliably to the data. Now, of course, that as
stated in those vague ways this was far from a self-executing test, but the Supreme Court
went a little further, and they said it
specifically to constitute proper scientific admissible evidence for criteria have to be met, and if these four criteria are not met it’s not real science. And the four criteria were first that the methodology had to give rise to predictions that could be tested so that if the theory was wrong it could be shown through a falsification. Secondly, that the theory, and the application to the theory had to be peer reviewed, had to be duplicated, not only duplicable, but actually duplicated by others who had carried out the same experiments, and reached the same results. Third, that it had to
have general acceptance in the scientific community, and fourth, that it had to
have a known error rate. The first three, those of you who are into the philosophy of science like Walter will recognize that that reflects the teachings of Karl Popper who was a very famous philosopher of science in the mid 20th century probably around the time when all the Supreme Court judges went to college, and, in particular, falsification is the cornerstone of the Popperian view of science. The fourth item, the known error rate really reflects, as I say, the more recent scientific view that we’re dealing with here in many, many parts of science with probabilities, not with certainties. So that’s all fine and good, but how do you apply it in a real case? Well, I had such a case just a few years ago. It was called the Ephedra litigation. The Ephedra, as some of you know is a plant. In the 1990’s it was
combined with caffeine, and “sold” in a pill form which was supposed to
help with weight loss with increased energy, and with improved athletic performance. It was, if you will, the
precursor to steroids. Although, as far as I know, Robert Clements never used Ephedra. The problem with Ephedra, or the seeming problem was of the millions of people who did take it, a fair number dropped dead from heart attacks or strokes. The result was that lawsuits were brought throughout the United States, about 500 in all, saying that Ephedra had caused the strokes and the heart attacks suffered by the plaintiffs in most cases, and through something called the Multidistrict Litigation Law these were all combined before me into one giant case for pretrial purposes. So the question before me was: Does Ephedra cause heart
attacks and strokes? Now causation in the law, so called legal causation is a fairly straight
forward kind of thing. In the tort context in
context of personal injury the causation exists where an antecedent event substantially and foreseeably causes the personal injury. The example, sometimes give is: if Mr. Pig builds his house of straw, and the wolf comes and blows it down, and eats Mrs. Pig the estate of Mrs. Pig can sue not only Mr. Wolf, but also Mr. Pig because in the grim forest of Pigland it was foreseeable that building a house of straw would not protect you
against ravenous wolves. To put it in its most basic form this is common sense causation, and there can be disagreements, and there are disagreements as to whether a particular factor as a matter of common sense causes a particular result, but it’s something within the juries ordinary experience. By contrast, scientific causation is
much harder to determine, and clearly no ordinary
juror is in a position to determine whether Ephedra does or does not cause
heart attacks or strokes. So how does the science deal with it? Well, the ideal way to
deal with it is with a study going forward, a predictive study. You have two groups of people. You take Ephedra, you don’t take Ephedra. Some of you dropped dead of heart attacks or
strokes, but now we know. So, needless to say ethically you can’t carry out that kind of experiment on people. You can carry it out on rats, but the problem with
carrying out animal studies is first the dosages are
usually very much higher than the dosages that a human would have, and, secondly, the mechanisms, the physiological mechanisms that operate with animals
are not necessarily the same as operate with humans. Now there’s a third alternative, and that’s what’s called an Epidemiological study, and the idea is if you have enough data you can go back in time, and you can look at
those who took Ephedra. You can look at similarly situated people who did not take Ephedra. And you can figure out who had more heart attacks and strokes, and if you can show to a very high level that there is a correlation between taking Ephedra and heart attacks and strokes after separating out all
the confounding factors then you may have a basis
for imputing causality. Those are very good
studies when done properly. There are lots of pitfalls that had to be very carefully dealt with, but most judges agree that a really good epidemiological
study is admissible under the Daubert test
that I mentioned earlier. But in the Ephedra case there were no epidemiological studies. Why were there no epidemiological studies? Well, congress has said that you don’t need FDA approval
if you’re marketing a pill that is a, quote:
“natural substance,” and, therefore, the
manufacturer has no motive to do an epidemiological study. By contrast, if you
have a prescription drug where the FDA has to approve it in advance the pharmaceutical company has a very strong motive to do an epidemiological study which are expensive studies, but it will help you get the FDA approval, but here since you
didn’t need FDA approval there was no need, and, in fact, no such study was done
by the manufacturers. On the other hand after a lot of reports of heart attacks and strokes after about 10 years
the FDA banned Ephedra, and the test for that ban is much lower than causality. The theory is if the pill is not serving any obvious important medical purpose, and Ephedra was not. If there’s even a threat of a risk the FDA can ban it. Once it was banned there was no motive, or no strong motive for
public interest groups, or universities to do
epidemiological studies. So the question for me was what do you do in allowing or not allowing science into the courtroom when the best evidence, the epidemiological
study is not available. Do you just throw up your hands, and say: “Well, that’s the end of the case.” Or do you see if you will adjust rule 702 in the Daubert test to the particulars of that situation. What I ultimately held after having a long two week hearing where many experts testified, and after I spent a lot of time getting into the scientific literature was that on the one hand no expert would be allowed to testify at the ultimate trial that Ephedra caused heart attacks or strokes, quote: “To a reasonable degree
of scientific certainty.” which is the “terminology” used in the most extreme situations, but then on the other hand experts based on the animal studies based on certain other evidence that was put before me could opine that it was more likely than not that Ephedra was a contributing cause to heart attacks and strokes with those people who had either high blood pressure or
a genetic predisposition along lives that would be affected by the mechanisms that possibly were applicable in Ephedra’s case provided that those experts also made clear to the jury that none of this had
been proven definitively. So it’s a kind of halfway house, and it had the immediate good effect and all these cases then settled pretty much at the midpoint of where they might have been expected to settle depending on what evidence I had allowed in or not. But the broader point I’m trying to make is that while the Supreme
Court has given us a test for the admissibility
of scientific evidence. It’s not very easy to apply this test to any given situation. The law has to be adjusted, or at least interpreted in light of all the broad facts of the case in order to then determine what scientific principles should be brought to bear in determining the
admissibility of science. I give you this somewhat extended example because there are much bigger problems that are about to confront
the court in this area, and I want to end by just making quick reference to some of them. First, there is forensic science. This is what you all see on CSI. There are growing challenges, very substantial challenges, that just that fingerprinting
is not science, and shouldn’t be admissible. That hair sampling is not science, and shouldn’t be admissible. That blood samples are not science, and shouldn’t be admissible. There’s already been numerous decisions holding that handwriting comparisons are are not science, but that’s a little
bit different, I think. That doesn’t carry the aura of science in the way that fingerprinting
or blood samples do. Now I’m not predicting how those challenges will come out. There is one case already that has held that fingerprinting is not science, and is not admissible. There are other cases going the other way. I just want to flag for you the kind of issues that are coming down the pike. Another kind of issue
that’s coming down the pike relates to what Walter
mentioned neuroscience. Neuroscience is the science of the brain, basically, as studied through brain scans, and it’s been a very fast developing area. At this point I think there’s
very little neuroscience that will make it into the courtroom if it were offered today, in particular, because there’s been
very little peer review. Very little duplication of the studies that have occurred, but there is some very
important potential science that may make it into
the courtroom very soon. To give just one example, a huge percentage, over 40% percentage of
all federal criminal cases involve narcotic addicts
in one way or another. The crime itself is not always possession, or distribution of narcotics. It’s often bank robbery, or fraud, or embezzlement, or a
hundred other things, but the driving force is addiction, and that’s true in the states as well. One of the problems that
has bedeviled the courts is that most of these defendants after they’ve served some time in prison are sent to drug rehabilitation centers, and a year later after they’re released they go right back to
using cocaine or heroine. There are important neurological studies that suggest that the reason for that is that the use of
cocaine changes the brain in ways you can actually
see on the brain scans, and in ways that slowly go back to normal, but it takes five to ten years before they’re back to normal, so a one year program just won’t do it. Now that raises all sorts
of difficult legal issues. Do we really want to send someone to an in-patient drug, or rehabilitation program for 10 years? But my only point is that a lot of that neurological evidence is just on the verge of coming into play in the federal system. I’ll give you one last example. This I saw just in the “Washington Post” about three days ago. There is now being offered purportedly as DNA evidence, evidence of supposed genetic predispositions towards violence, genetic predispositions towards certain other traits. This is being packaged as DNA evidence because of the good reputation that DNA deservedly has as hard science, but, in fact, what’s really been offered here is a hypothesis about the relationship between certain genes, and certain behavioral traits. And from everything I know this science they are still extremely
hypothetical, extremely soft, and it would not be
admissible at this point under a real Daubert rule 702 test, but some judges have, nevertheless, let it in, quote: “For what it’s worth.” And I’m worried about that because what that suggests to me is that the huge difficulties that are imposed on judges in determining what is good science and bad science may be too difficult, and in some cases, judges may just be
throwing up their hands. That would be very unfortunate, I think, because I think for all our failings, for all our lack of sophistication in these scientific areas, judges are the best people to kind of meld together the science and the law. They are as the title of my talk indicates “Uncomfortable Bedfellows: “Science and the Law,” but no twin beds are available, and I think sooner or later we’re going to have to figure ways to get them together. Thank you so much. (applause) – We have about 30 minutes for questions, and Judge Rakoff will acknowledge them. – Yes sir. – [Voiceover] Thank you,
I really liked your focus on the adversary process
as a way of finding facts as it leads to the question that I’ve been struggling for a long time, and that’s the proper allocation of roles between the parties
presenting the evidence, and the judge bringing
their own background especially in a pre-trial situation preliminary injunction
you have to decide on the reasonable probability
success on the merits, and you don’t have a
lot of time to do it in. Do you have the parties to bring in expert witnesses of a
battle of expert witnesses? Is it your own expertise,
your clerk’s expertise? Do you appoint a special master? Is that appropriate that you have an independent source or a separate, or non-adversarial source of information either from your own experience, however broad or biased, or whatever that might be, or do you have to sort of pretend you don’t know anything, and rely on the experts. What do you think is the right process for finding these sort
of scientific things? – [Judge Rakoff] What I do is I have the parties spend a day in my court first presenting what they both agree is the basic scientific
evidence that is undisputed. Then I let them go at each other on what is disputed, but that way I have the
fundamental grounding in a way that both sides have agreed to. I have experimented with appointing a scientific expert. I find that very unsatisfactory. It’s really almost impossible
in any disputed area to get someone who you can be confident is totally neutral. It makes much more sense to have the two extremes go at each other, and then you determine
where the truth lies. Nevertheless, some judges have gone the route of appointing special scientific experts. I do, ultimately, think that the adversary system is one of the great inventions in the history of truth detection. It’s not perfect, but I like to have the parties go at it, and I think it educates me enough that I can usually find my way through to an answer. In terms of consulting
things outside the court the only thing I would
ever feel comfortable in consulting with be a treatise. If there some acknowledged treatise that I would take a look at, so. Sir. – [Voiceover] You
mentioned briefly the FDA, but one question I have is based on personal experience. A decade of those epidemiological studies
of a particular compound that I’ve done with colleagues shows … And this is a lot of peer reviewed stuff that there are harmful effects. However, the CDC, on the basis of analogy between this chemical and another chemical decided in 1950 that it was safe, and what has happened
is the CDC’s decision has persisted in begin taken as, quote: “Science” by people who don’t want to consider science, and I wonder how often
that kind of problem arises in your experience with the law, and what you do about it? – [Judge Rakoff] It does arise. We give under our legal system an awful lot of deference to administrative agencies, and to other people
based on their expertise, but the truth is that many of these groups have access to crime. I think courts have slowly moved towards giving a little less
deference as a result of that. Also, and really very much
implicit in your question science is constantly changing, and I think one of the
hard things for judges is to make sure that
after they worked so hard five years ago to learn the science of X, and now they have a new case involving X that they don’t forget the science may be considerably
different five years later, let alone 30 years later
than it was back then. Walter made mention there’s a judge’s manual on scientific evidence, and it’s kind of interesting. There was not such manual until 1994 even though there were judge’s manuals on a host of other things, but it was only then that enough science was coming into the courtroom
to justify a manual, but it became to the extent
that a manual ever can, it became a best seller. Actually more of those have been distributed by far than any other manual
given to federal judges because judges really needed it. Now, six years later that 1994 manual was already out of date because some of the science had changed, and there were other new science areas that were now eligible, if you will, for admissibility that
hadn’t been back then, so a bunch of chapters had to be added. And now I’m part of a small group of judges and scientists that are working on the third edition because eight years later, once again, it’s both out of date and not as complete as it now needs to be, so it’s a constant battle, but one that we’re trying to work on. Sir. – [Voiceover] Your mention
of fingerprint evidence, and blood sample evidence led me to wonder whether there would be asymmetrically admissible evidence. Evidence that might be admissible by the defense in a
criminal case, for example, but not the prosecution. For example, where I defense lawyer, and there was a good
set of fingerprints left by the perpetrator that anybody looking at could see were not the
fingerprints of my client I’d feel surly used if I were not able to get those into evidence, and similarly if there was a blood spatter that the police believed
was from the perpetrator and it was not the same
blood type as my client I would want to be able to get that into evidence. It seems to be that sometimes there can be really good evidence to rule out something even if we wouldn’t necessarily have good evidence to rule it in. – [Judge Rakoff] Well, I’m not convinced. I should say by way of my own background that I was a federal
prosecutor for eight years, and then I was a criminal defense lawyer for 16 years, so I guess I’m two to one in favor of the defense. The rules of evidence it seems to be should be neutral among the parties, and no more so than when you’re dealing with science, or quasi scientific evidence because it’s going to carry inevitably in aura that non-scientific
evidence does not carry, and will be very weighty no matter how you look at it, and I see no reason why junk science should come in for the defense if it can’t come in for the prosecution. I knew I had hit the right note there. Time for the rebuttal, right. – [Voiceover] I just wonder if it isn’t a scientific principle that
for some kinds of tests false positives are a much greater risk than false negatives, and for other kinds of tests false negatives are a higher risk than false positives and whether it is in good use of science to take advantage of that, so the chances that somebody’s fingerprints that are left that don’t by anybody’s examination even vaguely resemble my clients. That seems to me to be a good deal different kind of science than someone opining: “Well, there’s the 14 points “of similarity here, “and on the basis of
14 points of similarity “we can conclude with certain percentage,” but those two are very
different kinds of evidence. – Yeah, but they’ll also be offered under different rules of evidence. If you wanted someone to present a fingerprint, and have the jury do its own lay persons comparison between that fingerprint, and the one that was found at the scene of the crime, your client’s fingerprint, and the one that was found at the scene of the crime. Without anyone offering any opinion just putting it before the jury, and then an argument on summation saying: “Look, you don’t have to be a scientist “because this is so extreme “look how totally different they are.” That might come in. That doesn’t implicate rule 702 at all, but if you offered someone who would opine: “This could not possibly be. “I have scientifically determined “that this couldn’t … “And I’m an expert on fingerprints, “and this could not possibly be “the fingerprint of the defendant.” I don’t see that that person should have a lesser standard than
the person who says” “This definitely is the
fingerprint of the defendant.” It should be the same
scientific principles. So I think it’s apples and oranges. Yes sir. – Hi, I guess I just
wonder about your cofidence in the adversary system as truth seeking may make good use of science. I’ve heard people talk about the use of scientific evidence in
civil courts in Europe, and it makes me wonder. For example, I think juries often find adversarial use of science to be highly discreditable. Both sides can come up with their scientists or experts to
testify opposite claims on almost any matter, so it looks like you can find a scientist to say anything about any
topic, some scientists. – [Judge Rakoff] That’s why we have Daubert and rule 702. That was exactly the problem, and it still is some state jurisdictions where any well credentialed scientist was allowed to testify
regardless of whether the science itself was reliable, and in those cases you’re right. Scientists are human beings, and for the proper amount of money they will say anything
that you want them to say, or at least some scientist, somewhere can be found to say that. So that was a very serious problem. I’m not suggesting that that problem has been totally cured, but the theory, at least, which I think has some weight is that if the judge as gatekeeper keeps out anything that’s not good science then you’re going to not see anything like the kinds of battles that you’re referring to. – [Voiceover] I wonder
what you think of the way in which German courts deal with this. Namely, their courts have … Courts have appointed scientific experts that have been appointed in conjunction with the scientific
academies in those countries so they have much more of a confidence, and a kind of neutral or
impartial scientific epertise available to the court apart from the adversary system. – Well, of course, my familiarity with the German system is immense, that is to say zero, but this was tried in American courts. The American Academy of Arts and Sciences set up a panel of scientists that were available to all federal judges on a whole wide range of areas. The one case I experimented involved someone from that panel. It was a question similar to the Ephedra kind of question whether X in a toxic tort situation X causes Y. They presented me with three possible alternative experts. One who was very expert in the area, but who had already
written a paper saying: “Yes, causation is there.” One who is also very well
credentialed in the area who had written a paper who said: “It doesn’t cause this result,” and one who knew nothing about the area. So I picked the one who knew nothing about the area, and he proved to know
nothing about the area. I’m a little more skeptical, therefore, that the, quote: “True,” neutral expert can be found in any of these sort of on the frontier areas where these issues are often joined, but having said that I’m not familiar with
how it works in Germany, and maybe it works much better there. Sir. – [Voiceover] I’m in full agreement that the adversarial system is the absolute worse system there is
except for all the others. It seems to me that when an issue is going to the jury if there are two experts who express very firm opinions that
are diametrically opposite it often comes down to an issue of burden of proof, and so one question I have is what is your experience in how well juries understand the concept of burden of proof? – [Judge Rakoff] Well, there are two parts to that question I want to respond to. The first, though it wasn’t exactly the question you put, but a question that’s often put is can juries deal with
scientific evidence at all, and when it’s more than
the most simplistic kind, and to my experience, and I talk to juries
after every single trial to try to get a sense of
what’s going on with them so I can adjust in future cases is that the burden is on the judge to absolutely require
the lawyers to present the scientific evidence in simple language in as clear a way as possible in a way that without
being overly simplified will be something that any reasonably intelligent person can follow, and when a judge does that I think juries are actually pretty good, and being able to follow
the scientific evidence, and make, if you will,
scientific judgments. On the burden of proof I think the burden of proof works very well in criminal cases. I think juries take very seriously the requirement of proof beyond a reasonable doubt. In civil cases you’ve got
a complicated problem, and it’s not just the burden of proof. The first problem you have is instead of a jury of 12 in many jurisdictions now
can be as small as six. There are even a few jurisdictions where it can be by six, and by majority vote, so it would really be four jurors, and I think in any complicated case that raises the risk of a mistake. If you have 12 people you
have to reason together, and come to a unanimous decision. An awful lot of learning is gonna be brought to bear. An awful lot of prejudice will fall by the wayside when you’re dealing with
a much smaller number, and with a dynamic where consensus is not required I think
it’s more difficult. Lastly, though, and most responsive to your question. In civil cases, on the whole I think
jurors are pretty good about following burden of proof, but economic ideology also clearly plays a
role in the jury system. There are clearly cases where the evidence favors the big company with the big insurance company standing behind it, and there’s little Miss Jones, the one time orphan who is now a widow, and she’s gonna win no matter what. So I mean that’s a price we pay for the jury system. I think it’s a relatively
modest price to pay, but there will be cases like that. Yes sir. – [Voiceover] In the
discussion you focused a lot on the scientific problems that you see with some of the criminal technologies that are used. Georgia, several years ago decided that it would draw a distinction between the evidence in criminal cases, and the evidence in civil cases. In the criminal case side it decided to use the old rule, so called Frye doctrine, and in the civil cases it decided it would use the more modern rule which notwithstanding what the Supreme Court said has now become the more strict rule with regard to admissibility. – [Judge Rakoff] Yes, there is that. You’re right to point out that irony that started out as purportedly more permissive, but it’s become a much stricter rule. – [Voiceover] Yes, and
my question to you is … And by the way, this went to the Georgia Supreme Court, and just a few months ago it ruled that that distinction was not an unconstitutional line to draw. Not by ruling that there was some … – Rumor has it that almost nothing is unconstitutional in Georgia, but I wouldn’t know. – [Voiceover] I would agree with that. And they didn’t rule that on the basis that there was a substantively good reason for drawing the line. They ruled the opposite. They ruled that there was no substantively good reason to say that
the line couldn’t be drawn, so they were deferring to the legislature to a great deal. My question to you is in the context in a federal court sytem can you think of any justification because the courts have
yet to draw this line, but it’s always there
available to think about to be more lenient in one kind of case, and more strict in another about the admissibility of evidence between criminal and civil cases. – No, I cannot. I recognize, and this was the whole point of the way I interpreted the DNA stuff in my death penalty decision that our fact finding processes are far from perfect, and that in the criminal context a mistake can have far
more dire consequences than it can have in a civil case, but having said that the object of the rules of evidence is to assist in determining the truth. We deal with questions
where truth is not involved where we are trying to make sure that the consequences of
mistakes are not permitted with things like the burden of proof, things like the Fifth Amendment, things like the presumption of innocence, and so forth, and they may or may not be sufficient, but when we are talking about evidence, and
admissibility of evidence, and scrutiny of evidence, and the standard to be applied to evidence it seems to me we are focusing on how best in our poor way can we arrive at the truth, and it doesn’t seem to me to make sense to have one standard of truth for the criminal side, and a different standard of
truth for the civil side, so I would not personally
be in favor of a split approach there. – [Voiceover] Would you like to go on the Georgia Supreme Court? – The weather is great I’ll bet. – [Voiceover] Just 43. Yes. – [Voiceover] I want to
go back to the question of juries because I’m
a little confused about your attitude, you seem
to have faith in them being able to figure out the science, and actually get it straight, but I thought part of
the reason for keeping the science out of the courtroom, the science that doesn’t pass these rules is to avoid the jury
getting inordinately swayed. They might have some probative value, but there’s prejudicial value as well that might override it. – It’s a good question, but
it’s a question of talk. The judge in the Ephedra case I spent two weeks in hearings, and then many, many weeks
reading the literature. You’re not going to be able to have a jury do that, so the idea is, okay, the judge is the gatekeeper. He keeps out the really pseudoscience, but then what does come in which is much more tight, much more attuned to the specific issue in the case is presented with the help
of the judge to the jury in as coherent a fashion as possible, and the jury then decides who’s right. I think that is an appropriate allocation of resources. The jury could do what the judge did if you want to keep him
around for three months on a case that otherwise
would take one week, and no one wants to do that. And even if we did we’d
have a rebellion of jurors so it’s really a function of time in my view as much as anything. – [Voiceover] So it’s
more a question of making the court system function efficiently than any fear that jurors
are going to be misled. – I think so. Now I have inordinate faith in jurors. There are other judges who will give you a different view, but when I talk to jurors after each case there’s always two, three, four jurors, at a minimum, who are damn smart, and could clearly master any of this stuff if given the amount of time, and opportunity to do so. It’s not gonna be possible
to do so, however, and that’s why we’ve made this allocation. Other questions? So I thank you all so very much. (applause)

One Comment

Leave a Reply

Your email address will not be published. Required fields are marked *