The Cheapening of Forensic Science*: How Philosophical, Practical, and Legal Definitions of Science Shape our Discipline
Bad science gets into court unless real science defines itself
Although, as scientists, we each think we know what it is we do, we rarely consider what does constitute “science”? The classic example is that astronomy is science but astrology is not. But why? What are the reasons something is or is not ‘science’?
Science is typically thought of as experiments conducted in laboratories by people in white lab coats. They predict future effects through replication of standardized events. But some sciences don’t deal with the future, for example, geology. Geologists can’t reproduce past events, such as continental drift, and they are unable to accurately predict future ones, like earthquakes. These are historical sciences. Many sciences are in this category: Astronomy, Archaeology, Paleontology, and...Forensic Science*.
Historical sciences can only test hypotheses about what could have produced the facts we see — they cannot test the events that did produce those facts. We weren’t there, e can only see the remnants of the activities. And yet, especially in forensic science*, the question will always be about the actual causes. Why can’t we get at the causal events?
Time is asymmetric: It moves in only one direction. We cannot undo the past and even if we did, we’d leave evidence of doing it (which is why it’s hard to plant evidence). This means the past is overdetermined by the present — there is much more evidence available than is needed to reach conclusion.
But this means that the future is underdetermined — many factors go into an auto crash but any one possibly could have prevented it, Good tire pressure, for example. You can’t predict the exact causal factor. Because time is asymmetrical, you have to wait for the results to see if you’re right.
Therefore, experimental science works with minimal clues to determine the future effect; it must reduce false negatives and false positives. Historical science, on the other hand, works with too many clues, some irrelevant, to determine the past cause; they must sort and categorize.
The view of empirical science is based on Popper’s concept of falsifiability, that you can’t prove something, only disprove it, and the more a theory is not refuted, the more confidence is gained in it. This empirical view is heavily echoed in the Daubert decision and this affects how forensic science* (being a historical science) is ushered into the courtroom.
But science is more than testing facts, a la’ Kuhn’s view of “normal science”: The daily puzzle-solving of science using agreed-upon core theories. Only rarely are these theories actually challenged and, when they are, a revolution results, such as occurred with the shift from Newtonian to Einsteinian physics.
These core theories are crucial for the definition of science. If an astronomer fails, the results are checked and the method revised reviewing old data, instrument calibrations, etc. But If an astrologer fails, there is no corrective mechanism (no puzzle-solving tradition) to revise the methods. Astrology has too many vague sources of difficulty, most out of astrologer’s control, with no guiding core theory.
The hazard is throwing out the baby with the bathwater. If empirical science is the only standard, then you accept navigation and surveying but you exclude geology, taxonomy, and evolution … and forensic science*.
Science must embrace both the empirical and the historical.
If Popper’s falsifiability was the only criterion, then most normal science would fail because we’d reject too many ‘good’ theories. Falsifiability is a very restrictive concept. Ultimately, all theories are unprovable, scientific or pseudoscientific. In Kuhn’s viewpoint, it is better to think of theories as “extended research programs,”not static events. Science is continually investigated and adjusted. This conflicts with what the law wants—a single definite answer. Additionally, these are scientific rules we’re discussing but when we go to court we have to abide by a completely different set of non-scientific rules. The law may not care about the delicate philosophical intricacies of the definition of science; in fact, it almost certainly doesn’t. A lack of understanding between scientists and legal experts (try using the word “error” sometime, like “standard error of the mean”) amplifies this difference between scientific rules and legal rules.
It is a mistake to look at a discipline in a ‘snapshot’ view — no simple rule will suffice, a la Popper. Science is more complicated than that, although to a lawyer, it looks like a pretty simple approach (part of why it was adopted in Daubert). It is imperative to consider how a theory or technique has developed over time, how it deals with problems, how it resolves them. A pseudoscience is a discipline that is less progressive over time than its rivals because it has no corrective mechanism, no central theory to refine it when errors are found.
So, is forensic science* a science? Just asking for some friends.
“...science can sometimes get it wrong...Science is not always value-free, and most importantly,... when lawyers accept the validity of an established scientific paradigm uncritically, they do risk making the same mistake as scientists.”
Michael Hoeflich, Dean, U of Kansas Law School