Sunday, August 9, 2020

The Nature of Evidence

In my daily professional life I am an experimental particle physicist. I currently analyze data taken with the ATLAS detector on the Large Hadron Collider at CERN laboratory near Geneva, Switzerland. I probe data that includes collisions of billions of protons to look for patterns that indicate the existence of some unknown phenomena or that measure the properties of a particular physical process. I am currently studying the Higgs Boson that was discovered in 2012 and measuring certain aspects of its character to try to understand if it exactly fits our expectations or whether there may be deviations from our standard model. 

In my personal life I have twice been a member of a jury in both criminal and civil cases. As a juror, I was asked to analyze data presented by witnesses and experts to determine the probability that an event in the past occurred in a certain way. Both my profession as a scientist and my civil duty as a juror required me to examine evidence and draw the most logical and reasonable conclusions based on the evidence. However, the type of evidence available in the different circumstances have very different natures. 

There are some people I talk with, primarily religious skeptics, who don't seem to understand the difference between scientific data and legal/historical evidence; the data from a scientific experiment that allows the investigator to draw reasonable conclusions, and the evidence presented in a courtroom that also leads to reasonable conclusions about what occurred in the past. These data are of a very different type but both lead to reasonably certain conclusions. Scientific data cannot directly tell us what occurred in the past, for direct scientific inquiry requires reproducible experimental results. Past events are not reproducible. Legal historical evidence that is used in a courtroom consists of direct evidence and circumstantial, or indirect, evidence. 

The primary type of direct evidence is eyewitness testimony. Indirect evidence is any evidence that relies on some inference to relate it to a particular event. Forensic evidence and other material evidence from a crime scene is all circumstantial evidence, for even if you have DNA evidence or fingerprints, an inference as to what events produced such evidence must be made. Although DNA evidence is often considered the most reliable evidence because it is "scientific," it is still circumstantial evidence since it must be interpreted. Historically, multiple eye-witness testimony has been considered as the most reliable method for determining the truth about what events actually occurred. 

Either direct, circumstantial, or some combination of both types of evidence is required to lead to a conclusion about a past event. In the United States, the conclusion must be "beyond reasonable doubt" in a criminal case, and must satisfy the "preponderance of the evidence" in a civil case. There is no such thing as a conclusion beyond any doubt in either scientific or in legal/historical inquiry. 

Although the types of evidence differ, a scientific approach can be applied to the question of whether or not a particular past event occurred. There are certain aspects of a scientific approach that are applicable to both scientific and historical evidence. For instance, all the evidence should be considered. The investigator cannot arbitrarily discard evidence that may not seem to fit with some preconceived notion. Any conclusion must be able to explain all of the evidence, not just a fraction of the evidence. Any hypothesis that does not adequately explain all the known evidence, whether direct or circumstantial must be discarded. 

When applying some of these principles to ancient events, all of the evidence is indirect evidence since we don't have direct eyewitness testimony. Even documents written by eyewitnesses are technically indirect evidence since the eyewitness himself cannot be interrogated. The historical indirect evidence then falls into two basic categories, either written testimony or archeological findings. The value of the written testimony can be determined by ascertaining the reliability of the writer, which in turn can be determined from the internal consistency of the writings and from the external confirmation of other writers and of various archeological findings. Some written testimony comes from previous oral traditions, which can also be tested as to its reliability using well defined criteria. 

In addition, bias, prejudice, and pre-conceived notions must be avoided in order to have reliable conclusions and results. In our physics experiments, we go to great lengths to minimize any bias toward one conclusion or another, particularly when looking for some unknown phenomena or particle. We do this by performing all of our data analysis on computer simulated data (often called Monte Carlo simulations) without ever looking at the real data. We keep the real data "blinded" until we are completely convinced that we have optimized our analysis methods with no bias toward what might actually be contained in the real data. Once the simulations have been checked and rechecked, an independent panel of physicists within the collaboration gives permission for the analyzers to "unblind" the data to see if any new phenomena is actually observable. It is truly an exciting moment when the data is finally unblinded and observed and we see if anything unexpected appears. Physicists have joked that we could raise money to do our experiments by selling tickets to the "unblinding" moment and offer betting odds and betting opportunities as to whether or not something previously undiscovered will be seen.

The evidence for the life, death, and resurrection of Jesus can be analyzed using the scientific approach applied to the historical evidence to determine whether or not the claims of Christianity can be validated. I will do that in my next posts. However, those who don't understand the difference between reproducible scientific data and legal/historical data may continue to claim that the historical evidence supporting the claims of Christianity is insufficient likely because they do not understand how to properly interpret various kinds of data correctly and without bias.


  1. No puedo esperar por el próximo post, bendiciones!...

  2. In the conduct of experimentation at CERN I wonder if you could describe the methods that avoid bias. Is the experimental hardware and software set up by the same team members under the same "tolerances" and then checked by another team? How id confirmation bias avoided, since people naturally have certain hopes for a ground breaking result? Repeating the experiment is both complex and expensive I presume yet necessary to establish a high degree of statistical significance to the results. How are these tradeoffs made?

    In historical document analysis do you believe the interpretation by many qualified scholars from both biblical and secular scholars from appropriate disciplines should be included in reaching a consensus opinion about the reliability of the Gospel accounts of the resurrection? Are these independent scholarly studies in some sense like independent scientific teams interpreting the results of a given experiments data set?

    “Every document, apparently ancient, coming from the proper repository or custody, and bearing on its face no evident marks of forgery, the law presumes to be genuine, and devolves on the opposing party the burden of proving it to be otherwise.”

    ― Simon Greenleaf, An Examination of the Testimony of the Four Evangelists, by the Rules of Evidence administered in Courts of Justice

  3. "In addition, bias, prejudice, and pre-conceived notions must be avoided in order to have reliable conclusions and results."

    This is probably The Hardest thing for people to do. Look beyond their bias, prejudice, and pre-conceived notions. I find myself often guilty of this. I Try to cover these failings with 5 words, But..I...Could...Be...Wrong.