Science and The Pursuit of Knowledge: Part II
In part I, we discussed the concept of knowledge and introduced the Justified True Belief theory as a possible framework for establishing how it is that we might be able to say that we know anything at all. We also introduced the notion that science doesn't actually ever prove anything. That is, due to the philosophical structure of science, we never can be absolutely certain about any of the knowledge attained through it. This may appear problematic, but it is the best that we can do and just because we cannot be 100% confident with our current scientific facts or theories, doesn't mean that we can't have something asymptotically approaching absolute truth. Now, let's explore precisely why this is and the limits of what we can know through the scientific method.
There Is No Such Thing As A “Scientific Proof”
Unlike math, the concept of a proof does not exist within science. That is, science doesn't actually ever prove anything in an absolute sense. While science is the best epistemological framework that has ever been devised to understand the world around us, it doesn't allow us to know anything with absolute certainty. However, this doesn't mean that we can't know anything and that the existing theories that we currently have or scientific consensus are dubious. The truth is that we can be very certain that these are correct, we just can't be absolutely certain.
The reason why mathematical proofs are absolute is that they exist inside of a highly structured world created by the mathematician through axioms (i.e., rules that are taken as self-evident). From these foundational rules, the mathematician then begins to build a world where each successive step must always obey these rules. New rules may be derived from the axioms, but if these new rules do not respect the axioms, they are summarily dismissed as every step of the process must remain self-consistent otherwise the emerging world no longer makes sense.
The fact that you can build a world with your own set of building blocks is why the concept of a proof can exist within math. Every step in the process logically follows from the previous step, which allows results to be framed in absolutes. Mathematics is the pinnacle of deductive logic. On the other hand, while scientific reasoning comes with a set of axioms, it doesn't get to build a world from these rules like mathematics does. The whole point of science is to try and describe the world around us as we don't understand how it was built.
This fundamental difference between mathematics and science is why the concept of “proof” doesn't exist within science. When you are attempting to deconstruct an incredibly complex, dynamic structure, you are forced to concede that you may not know something with absolute certainty even if you are able to describe its behavior with a well-defined model. At some point, your current model may be insufficient as a better model is created that supplants the old. This is precisely how science progresses.
In regards to logic, this concept is further elucidated by what is known as the problem of induction, which was first defined by the philosopher David Hume. He states that since inductive arguments, which scientific arguments are, don't guarantee the truth of their conclusions, this can lead to problems trusting induction. Essentially, science is attempting to validate induction through use of induction. Moreover, since the concept of an inductive proof does not exist due to inductive reasoning's inability to guarantee the truth of the conclusion, it becomes clear as to why science is also unable to form proofs or claim to prove anything with absolute certainty. This is why it is best to think of science as asymptotically approaching truth as hypotheses are dismissed in favor of ones best representing reality at some moment in time.
Science Has Been Wrong Before
The “science has been wrong before” argument is often used by those embracing anti-science positions to dismiss scientific results or impugn the scientific enterprise in general. That is, this somehow implies that all results are untrustworthy since science never proves anything. While it is true, this doesn't mean that the results shouldn't be taken seriously. This argument castigates the process of refinement (i.e., asymptotically approaching objective truth) that's integrated into science and tries to recast it as a negative in order to dismiss unfavorable scientific results.
Formally, the argument takes the following form:
P: Science has been wrong before.
C: Therefore, the results of [insert scientific topic of interest here] can be dismissed.
Further, beyond being used to dismiss unfavorable scientific results, this argument is also used as a way to rationalize scientifically unfounded positions (e.g., homeopathy, free energy, acupuncture, etc.). In this instance, the argument looks like:
P: Science has been wrong before and can't explain everything.
C: Therefore, it doesn't matter that [insert pseudoscience here] doesn't have any credible scientific evidence supporting its position.
These arguments are dishonest as the self-correcting mechanism within science is precisely why science is so successful. Change within science has very strict constraints. Newer models must acknowledge all previous facts and be able to reproduce all the results of the previous model. For example, in order for a new scientific theory to be accepted by the community, it must be capable of not only demonstrating that it can properly model novel observations that previous theories failed at, but it must also be able to reproduce everything in the previous theory.
It's also important to note that the only way we know science has been wrong in the past is by doing more science. This information wasn't intuited, a product of dogma, ideology, or some form of magical thinking. The sole way that we know science doesn't have it quite right is by scientists putting in the work and conducting more experiments. Often people using this argument don't realize this is the case, which essentially makes this argument self-refuting. In other words, if science is not to be trusted because it has been wrong in the past, then why should we trust the new science that tells us that it was wrong?
Now, let's analyze an example that I frequently encountered towards the beginning of the pandemic about the shelter-in-place orders recommended by the scientific community to mitigate community spread:
P1: Science has been wrong before.
P2: We once thought the Earth was flat. Newton was wrong too. Science clearly isn't right all the time.
C: Therefore, the shelter-in-place orders are ridiculous and we should just allow the virus to spread naturally to reach herd immunity.
Explanation: As explained above, using the fact that science has been wrong in the past isn't a valid reason to dismiss scientific results. While the SARS-COV-2 virus is a novel pathogen and the whole world is witnessing first-hand the ordered chaos of scientific progression, it is still the best system we have for providing us with accurate information about the pandemic. Yes, it needed to course-correct on more than one occasion during the pandemic (e.g., mask wearing), but this shouldn't be extrapolated to mean the scientific community is wrong about everything. That being said, to date, the best available evidence says that allowing unmitigated community spread in order to reach herd immunity is a terrible idea that would cause many unnecessary deaths [1].
The Correspondence Principle
As you are now aware, when science is wrong it doesn't throw out all of the previous scientific facts that have been learned. Instead, it creates a new model that not only accounts for all the previous learned facts, but explains the gaps in the previous model as well. This is why it is best to think of these instances when science fails as being “incomplete” instead of being just “wrong.” The word “incomplete” allows for nuance where “wrong” often conveys a meaning of absoluteness when this is clearly not the case. This concept of “incompleteness” is encapsulated in what is known as the correspondence principle.
Formulated by Niels Bohr in 1920, this principle reconciles quantum mechanics (i.e., the physics of the very small) with classical mechanics (i.e., the physics of everyday life). Specifically, it requires that the results of quantum mechanics be able to reproduce the known results of classical mechanics in the limit of large quantum numbers (i.e., for large orbits and for large energies). In other words, the physics of quantum mechanics in the macroscopic limiting case must agree with the physics of classical mechanics. By ensuring that the correspondence principle is adhered to, we are able to weave one continuous picture of reality all the way from the smallest scales to the largest.
While this principle was originally conceived to provide consistency between quantum and classical theories within physics, this concept applies to all areas of science [2]. Whether it is biology, psychology, chemistry, or another scientific discipline, new theories can only be adopted if this principle is sufficiently demonstrated. This is why there are certain scientific facts which have advanced to immutable fact and will most likely never be refuted even if the theory in which they are currently a part of is augmented in the future. The new theory that supersedes the original must account for these facts in addition to explaining additional phenomena that the original theory could not.
If you were to think of scientific theories as gigantic containers that hold humanity's knowledge about a specific aspect of our world in the form of some fluid, you can view the new theory as a volumetric addition to the old. That is, the new theory is an expansion of the old theory's container. All of the original fluid (i.e., knowledge) is still present, but there is now more room to hold additional knowledge.
Further, the impetus for this expansion is that there were new observations made or new fluid being poured into the container that evaded explanation or would simply pour out of the already full container. What is more, using this container analogy, it is easy to see why this is the only logical way for science to progress as it makes no sense to replace the existing container with a smaller container. If this were the case, you would lose knowledge in the process, which is antithetical to the scientific endeavor.
One of the best examples of the transition from an old scientific theory to a new one is Einstein's general theory of relativity (GR). Before Einstein, the prevailing gravitational theory was Newton's law of gravitation. Various observations were made by scientists that Newton's laws were unable to explain, but GR was able to. More importantly, beyond GR being able to explain these anomalies, it is able to describe all of Newton's results as well in the limit of weak gravitational fields, which indicates that GR obeys the correspondence principle. It was only after GR was confirmed experimentally to be able to do this that it was then accepted by the scientific community. If GR had been unable to reproduce or conflicted with results using only Newton's laws that we know are true, then GR would have never superseded Newton's theory as the prevailing gravitational theory. It would have been tossed into the dustbin with other failed theories that sounded great, but failed to reflect reality.
Conclusion
While true that science doesn't ever prove anything, this shouldn't deter us from trusting its results and acknowledging the facts that arise from its methods. Our current scientific theories cannot explain everything, but they have accomplished describing many of the wonders of the world around us. In the future, current theories will be supplanted with newer, more complete models that will confer us with even more knowledge about the inner workings of the universe. One day, if we’re lucky and sufficiently clever, we may develop a theory of everything. Hopefully this day doesn’t come too soon as being curious and searching for answers is a sublime experience for those who choose to dare.