Science is Not Self-Correcting

The Reconquista Initiative

Presents…

Science is Not Self-Correcting

 Let’s be blunt: the idea that science is self-correcting is bullshit. Indeed, it is a claim pushed by science-fetishists and the proponents of scientism in a self-serving attempt to give science an allure and prestige that it simply does not deserve. Furthermore, even if we ignore the fallacious personification inherent in the idea that “science” itself is self-correcting—after all, science is not some ‘thing’ or some ‘entity’ capable of action—the fact is that science is not only not self-correcting, but it’s correction method is not in any fundamental way unique or special when compared to other disciplines. And we can demonstrate this fact with a simple example.

Imagine, for a moment, a particular situation that has no doubt previously occurred in reality at some point:  namely, a scientist performs an experiment, but his experiment is flawed in some way that skews his results and renders them false, but not obviously so; or even imagine that the scientist himself has purposely but subtly falsified his experiment in order to achieve the results that he desired. Now imagine that this experiment is reported to the scientific community and accepted. It is even referred to from time-to-time in other experiments or studies, but not that often. In essence, it is a relatively obscure experiment, but nevertheless one which has been accepted and acted upon. And as is often the case with such studies, it is not replicated. No one checks the results. The error in the experiment remains undetected. Now ask yourself: has “science” self-corrected this experiment? Obviously not, and it would be absurd to suggest otherwise. In fact, if the experiment sat untouched and unchecked from now until eternity, then at no point would science self-correct it. And if no experiment was conducted which contradicted the results of the first experiment, then there would be no evidence against it, and thus again, science would not correct the faulty results of the first experiment, neither directly through replication nor indirectly through counter-evidence.

Thus, the only way that the experiment would be corrected would be if an actual human scientist examined the experiment, replicated it, and hopefully caught the error—a fact which is itself not guaranteed. But note that “science” did nothing here, a human did. But even more importantly, note that the human scientist did nothing different or more “correcting” than any other profession or field would do. Indeed, the scientist simply used the appropriate tools and techniques of his particular discipline to check the work of a fellow scientist. But the point is that an accountant or mathematician or engineer could correct another accountant’s or mathematician’s or engineer’s work in much the same way that a scientist corrects a fellow scientist’s work: namely, by using the tools and techniques appropriate to their field to check their colleague’s work. In fact, a philosopher could check the work of another philosopher for faulty reasoning or an incorrect deduction in much the same way as the scientist corrects another scientist. And the similarity here is even closer than you think, for if we are dealing with forensic or historical sciences, such as archelogy or evolutionary biology, where the scientist’s “experiment” is often little more than trying to infer the best explanation of the evidence at hand, much like a Detective does with the evidence at a crime scene, then a future scientist’s correction of such an “experiment” is little more than a correction of the initial scientist’s reasoning or a challenge to his facts. But such a type of correction is little different from a philosopher correcting a first philosopher’s inference to a best explanation about some philosophical matter or a second Detective double-checking the inference made by a lead Detective concerning a particular case under investigation.

So what is the point of all this? The point is to realize that science is not self-correcting, or if it is, then many other disciplines known to Man are self-correcting as well, for they correct themselves in the same way that science does:  namely, by individuals using the tools appropriate to their specific discipline to check the work of other individuals for errors or faulty reasoning. And so, if so many disciplines are self-correcting, then there is nothing special about science. Thus, again, the point here is to show that not only is science not self-correcting in any literal sense, but also that there is nothing special about science’s correction process, for it is the same one that numerous other disciplines use to correct themselves as well.

And note that if the science-fetishist tries to argue that in science, very prominent and popular experiments are indeed often replicated, and so, in this practical way science is somehow “more correcting” than other disciplines, well, there are three responses that can be given to this point. First, the practical fact—admitted for the sake of argument—that scientific results are checked more often than other disciplines has nothing to do with science being self-correcting or with science having a unique way of correcting itself, and so this is a largely irrelevant point. Second, it is rather questionable if other disciplines are not subject to the same amount of correcting as science is; for example, popular philosophical arguments are often challenged, argued against, and corrected. So the claim that scientific results are more often checked than other disciplines is in no way certain, and clear data would be needed to show that this is the case. And finally third, we have some data which tentatively shows us that scientific experiments are rarely replicated, so the claim that they are is, again, questionable. For example—and this is just one example, so it needs to be taken with skepticism—note that a study by Matthew C. Makel, Jonathan A. Plucker, and Boyd Hegarty, titled “Replications in Psychology Research: How Often Do They Really Occur” and published in November 2012 in the Perspectives on Psychological Science Journal found that there was only a 1.07% replication rate of previous scientific experiments. So even if these researchers were incorrect by a factor of ten or twenty, and thus the replication rate was 10% or 20%, that rate would still be horrendously low for a discipline, namely psychological science, which supposedly prides itself on self-correction and replication. So the idea that science is a “more” self-correcting discipline than others is, once again, questionable at best.

Now, if someone tries to claim that science is unique in that its correction occurs through the use of experiments, and thus it is a more accurate and accountable process than other disciplines that correct themselves, the fact is that this objection also does not hold water. First, corrections in math, for example, are as accurate if not more so than an experiment, so some disciplines are arguably even more accurate than scientific correction. Second, scientists, being humans, are also susceptible to error, and so a replicated experiment, given its human factor, is prone to error as well; and this, of course, means that there is no guarantee that the replication is correct either. So the replication of a scientist is as potentially prone to error as that of a mathematician, engineer, or philosopher. And this is especially the case for the historical sciences were the science is largely done by inference to the best explanation, and thus is little different from other disciplines. And so while a scientist’s use of a replication experiment is a good way for that scientist to check a fellow scientist’s original experiment, and it is the appropriate way for him to do so given that experiments are the tool that a scientist is supposed to use to conduct his work, the use of such replication experiments is by no means the most accurate or fool-proof way of checking someone else’s work, and so even in this respect the sciences are not particularly unique or special in comparison to other disciplines.

But, one may ask, what if the objection is made that the corrective method of science must be, somehow, special, for it leads to results that are usually agreed upon by the wide consensus of scientists. Well, how a scientific consensus is reached, and whether that consensus is solely or even mainly based on the state of the evidence, rather than other factors like social pressure, career considerations, and so on, is a topic all its own, but in this case, all that will be mentioned is that science, again, is not unique in reaching a consensus, for disciplines like math, accounting, engineering do so as well. Furthermore, science often reaches consensus because, unlike philosophy, science, in the quest for knowledge, simply cheats in that it accepts a number of easily questionable assumptions and then operates as if those assumptions are not questionable. For example, present science embraces methodological naturalism, an idea which automatically restricts the types of theories that are deemed scientific and thus automatically creates a type of consensus by narrowing the range of options that a scientist can choose from. Additionally, scientists largely ignore the question of whether scientific realism or anti-realism is correct, but a person’s views on which one of these positions is correct will have a large effect on the evidentiary weight that a person gives to the scientific theory under consideration. And so if science had to drop its presently operative assumptions, and if certain social pressures and career worries were removed, one wonders just how consensus reaching the scientific community would be. And while this is speculation, the point is that the consensus in science is not necessarily based solely on an objective assessment of the evidence, but on a number of outside pressures and restrictive axioms which create the conditions where a consensus can be banged into place much more easily than if the conditions were not there.

And so, the long and short of it is this: science is not self-correcting, and the way that it corrects itself is in no significant way different from the other disciplines. Science, therefore, is not particularly special in its correction method, and so it is time to put the myth that science is self-correcting to bed.

Anno Domini 2016 11 22

Non nobis Domine, non nobis, sed nomini tuo da gloriam