Time to defend like a wolverine!
A Defense of Science: An Essay of at Least One Part
Part 0 – Introduction and Glossary
So-called “Mainstream Science”, that is, science as it is currently practiced by the vast majority of credentialed and experienced professionals, is deeply misunderstood by a large proportion of the general public. The extent of this misunderstanding has contributed to a widespread decrease in trust in scientists. Additionally, increase in the misuse of science—both the process and the findings—is a societal cancer that has abetted this trend.
Those who work in and those who support mainstream science (hereafter called “science” because, as we’ll eventually see, that’s what it is) need to counter these toxic trends. Such defense may be accomplished with simple observations, e.g., how people rely on phones which would not exist without a process like science. But these simple approaches need to be backed up by more in-depth arguments. Since science itself relies on both accuracy and precision (which aren’t the same thing), these arguments must perforce rely on the same. This starts with an agreement on the meaning of certain words. It’s important to note that words can mean different things in different contexts. For instance, what does “charge” mean? If we’re at the grocery store, it means a method of payment; if it’s a basketball game, it’s an offensive foul; if it’s wartime, it’s a frontal assault against an enemy position; and in science, it could mean “electric charge”. As this essay is a defense of science, it’s safe to assume that this is the context for the meaning of the words under discussion, so they might not mean the same thing as the general perception.
This glossary will not be alphabetical. The alphabet doesn’t care about the process of science. There will be an attempt, instead, to create a logical flow that will allow a subsequent word to build upon a previous word. First, a few foundational definitions:
Science: (Let’s make sure we’re on the same page) Easily argued to be the most powerful and practical knowledge-generating means humankind has developed. It is a systematic and impartial approach to producing operational (i.e., we can do something with it) explanations of the many phenomena found in the universe. There is a fundamental assumption that in order for science to be able to fully tackle an explanation, there must be an underlying cause that is itself amenable to scientific examination.
Nature: The aggregate of all phenomena that can be observed and analyzed by a scientific approach.
Supernatural: Anything that is, at best, only superficially amenable to scientific analysis and can’t be explained at an underlying level by the scientific process. Science is incapable of determining whether a phenomenon is supernatural. If there is a true supernatural event, the most honest and correct statement science can make is “We don’t have an explanation for it at this time.” Note that saying there is no current explanation is not equivalent to saying there will never be an explanation. Science would be perpetually befuddled by a truly supernatural occurrence.
Pseudoscience: It kind of walks like science and kind of quacks like science but it ain’t science. There is some underlying principle or methodology that doesn’t follow a truly scientific approach. Distinguishing a pseudoscience from actual science can be challenging. There will be one or more misconceptions in a pseudoscience that may require someone with some scientific expertise to identify. Understanding the ideas below can help a person determine when they are dealing with a pseudoscience.
Let’s now get into the nitty-gritty by starting with a definition of “observation” and going from there.
Observation: Originally, a sensory experience that can be measured in some way—certainly all initial observations by people were made this way but now technological advances have allowed us to expand the spectrum of possible observations beyond the capabilities of the five traditional senses. A scientific observation should be able to be specified and articulated in some way.
Experiment: A systematic investigation of an observation. For our foraging forebears, an observation might be “Those berries I’ve never seen before look and smell tempting!” The experiment would involve eating a berry and waiting for any effects. If it passes that test, then others may experiment with them to see if there are any benefits or harms to the elderly or to the young, to the women or to the men, etc. Credible experimentation performed nowadays is very systematic, often requires lots of time and very expensive and sophisticated equipment, and is painstakingly detailed with no hiding of results.
Precision: How similar separate measurements of a phenomenon are to each other. As much as reasonably possible, proper experimentation relies on repeatability. Suppose someone counts the number of tempting berries on a bush. Let’s say they count 38. Other people count the berries that look tempting and come up with the following results: 39, 38, 38, 39, 37, 38, 38, 39, 38, 37. “Precision” relies on how similar the measurements are to each other. Considering that all counts were between 37 and 39, it’s likely that 38 is a rather precise number dependent only on the relative “temptation” provided by a berry. If the counts were 38, 21, 52, 8, 65, 58, 31, 43, 28, 60, and 12, although the average of these counts is close to 38 there clearly is widespread disagreement. Saying there are 38 tempting berries is not precise.
Accuracy: How close a measurement is to the actual answer. In the case of the number of berries on the bush described above, if there are actually 38 nutrition-laden berries on the bush then both counts are very accurate. To be more specific, the first round of measurements gave a result that is both precise and accurate while the second round of measurements gave an average result that is only—and perhaps merely fortuitously—accurate. The so-called “actual answer” may not be known (though guesses, estimations, or predictions often can be made). But if repeated experimentation consistently yields precise results, then it is very likely those results are accurate.
Claim: A generalization based on observation and/or experimentation. “Everyone should eat these berries as they taste good and provide nourishment” is an example. Claims are capable of being wrong. For instance, maybe the berries give some people diarrhea. When a claim is shown to be wrong, it must either be further investigated and modified (“Don’t eat these berries if you’re allergic to other types of berries”) or discarded (“You know what? These berries have long-term negative effects and are poisonous after all. Don’t eat them.”). Changing a claim based on evidence is a strength of science, not a sign of flip-flopping. This last observation can be easily misconstrued by the general public and lead to acceptance of pseudosciences.
Fact: A well-supported and accepted observation or claim. We can suppose the observation that “Those berries over there are red” can be accepted as a fact (depending on the berry and whether the observer has color-blindness or is otherwise demonstrably unreliable). The claim “All berries are red” cannot be accepted as a fact because there are ready examples that contradict the statement. Like claims, things that are considered facts nowadays could be abandoned later in the light of compelling evidence, e.g., the “fact” that the Earth is motionless being replaced by Heliocentrism.
Evidence: Relevant facts that support the case for a more encompassing claim. “Those berries over there are red” might be a fact but it’s highly unlikely that it’s evidence in support of the claim that the Earth’s average surface temperature is increasing at an unprecedented rate.
Hypothesis: (We’re digging deeper into the process of science) A tentative and workable explanation behind one or more connected facts; the adjective “workable” means that this explanation can be investigated via observation and experimentation. A hypothesis should have the capacity to be shown to be incorrect. A hypothesis is a starting point in a systematic investigation. It is a bit like a claim with the emphasis on the “Why?”, e.g., “Why are those berries red? I think it’s due to the presence of a sufficiently high proportion of anthocyanin molecules. Let’s investigate!” (At this point, we have left our foraging forebears in the dust.)
Hypothesis Testing: A statistical means of determining if a claim or hypothesis is invalid (note the specific use of the word “invalid” rather than “valid”). It often involves analyzing data to determine the average and the standard deviation (a rough measure of how far afield a single measurement can reasonably drift from the average). For instance, suppose 10 red berries are examined and the proportion of anthocyanin molecules in each berry is determined:
|
0.000936
|
0.000973
|
0.000958
|
0.000956
|
0.000946
|
|
0.000929
|
0.000942
|
0.000968
|
0.000971
|
0.000945
|
The mean and standard deviation are 0.000952 ± 0.000015. Now suppose the hypothesis being tested is that the proportion of anthocyanin molecules in a berry must exceed 0.000995 in order to appear temptingly red. The mean, 0.000952, is inconsistent with that but that is, by itself, not a sufficient statistic for a conclusion. In conjunction with the standard deviation, though, this measurement becomes extremely relevant. While the difference between the hypothesis and the mean is 0.000995 – 0.000952 = 0.000043, the number of standard deviations by which the measurement differs from the hypothesis is 0.000043/0.000015 = 2.9, which might be large enough to indicate the hypothesis isn’t correct. Further investigation would be needed. The larger the difference is in terms of the number of standard deviations, the more likely the hypothesis is incorrect. But even if the result was more in line with 0.000995, it doesn’t mean that 0.000995 must be true (maybe the true minimum proportion is 0.0009947869304106531). The evidence can be accurately characterized by saying it’s consistent with the hypothesis or it supports the hypothesis—it doesn’t prove the hypothesis.
Law: A well-accepted universal claim for which there is no theory (see later). As an example, Isaac Newton’s First Law of Motion says the motion of an object will be unchanged unless acted on by an external push or pull (aka “force”). This does not explain why the motion would change if there were an external force on the object. It says it does and that is it.
Postulate: A foundational assumption used in subsequent explanations, e.g., the speed of light in vacuum is always 2.9979 x 108 m/s. Albert Einstein didn’t prove this--he used this as a postulate, a starting point in his development of the Theory of Special Relativity. The resulting predictions made by this theory have been consistent with experimental observations, supporting the validity of the postulate.
Theory: (This is the big one.) A robust and detailed explanation for one or more facts/claims which provide an avenue for further investigation and is capable of being proven wrong; here, “robust” means that the theory can deal with variations in expectations, i.e., it’s not rigid or dogmatic or limited to an extremely narrow confine. Theories often build upon hypotheses via experimentation and may include postulates and facts and laws as part of its makeup. A theory is an extremely powerful idea in science. Einstein’s Theory of Special Relativity and Darwin’s Theory of Evolution by Natural Selection are famous examples.
Proof: A word that has only a negational meaning in science. While claims and hypotheses and theories can be shown to be false, nothing is ever proven true. Either the observations and claims are consistent with the hypothesis/theory or they are not. If anyone, even a practicing scientist, says a scientific theory is proven, they are not literally correct. In such a situation, the word “proof” is used in the context that scientists are very satisfied that repeated investigation has supported the theory’s ideas and they don’t have questions about the theory’s overall validity.
Model: (This is the biggest one.) A substantive representation of nature or subsets thereof. Models are comprised of facts and laws and postulates and theories. As theories are never actually proven, they can only build up to what scientists think is happening. No matter how well a theory is ensconced in science, none of them are 100% proven to be a perfectly accurate account of what is actually occurring in nature. No scientific theory is ‘The Truth™’. But if the theories and facts and laws and postulates are all consistent with each other and with experimentation, then the model of reality built from them is sturdy and scientists and engineers can use it to generate reliable expectations for natural phenomena.
An example of such an encompassing model is Uniformitarianism. This model says that the underlying processes we see going on today have been going on since the origin of the universe. The word “underlying” is extremely important. Water pouring through Niagara Falls is a process but the uniformitarian model doesn’t claim that Niagara Falls have been around for billions of years. The underlying processes involve gravitational forces and the properties of flowing water. In uniformitarianism, those underlying processes are assumed to have been behaving under the same rules for billions of years.
Possible Upcoming Essays
The Practice of Science
The Malpractice of Science
At the Intersection of Science and Society: Religion
At the Intersection of Science and Society: Politics
At the Intersection of Science and Society: The Future
(Maybe others depending upon the exertions of my fevered imagination)
