The Nonessential Tension: How Dogmatism Complicates Normal Science

by Shaun Terry

I do not think that Popper, in what we have read from him, was necessarily trying to answer whether or not sciences have adopted any methods or assumptions that remain consistent over time. That said, I think that Kuhn uncovered an interesting, albeit complicated, problem.

Kuhn says that when we engage in normal science we frame scientific activities using paradigms. Paradigms come with specific values and tools, which are used to argue for the virtues of their respective paradigms — a kind of dogmatism. As those tools are used to solve the kinds of problems that are commensurate with their paradigms, inconsistencies emerge. Eventually, those inconsistencies can become central foci in new paradigms.

However, new paradigms do not simply absorb old ones. Some aspects are lost, including the attention paid to particular problems that might not be suited to newer paradigms. By refocusing a science, it is likely that some questions will have been ignored and some will have been dismissed. When Kuhn talks about how science has treated gravity, it demonstrates that an important concept can be misunderstood, as well as that an important question can be reignored after a shift to a new paradigm.

Dogmatism in normal science can lead to bias in the selection of data used in research. If some data serve the tools and concepts of a paradigm, then those data are likely to be preferred over data that do not serve the tools and concepts. Something similar could be said about the models and tools that are developed, as well as the questions and experiments with which scientists engage. This could lead to distortions in understandings of scientific questions.

In economics, as an example, the predominant paradigm generally dictates that economic growth is preferable. With some qualifications, we presume that GDP should be maximized, we presume that companies should maximize profits, and we presume that people prefer higher wages and prefer to increase consumption. These presumptions in economics practices (as opposed to economic practices) can quickly and easily become a kind of closed circuit: the paradigm presumes things about human behavior, so we choose data that fit with the paradigm; then, the model predicts presumed behavior(s), and decisions made by all relevant participants are based on the model and on the presumptions of the paradigm; finally, the outcomes are consistent with the paradigm’s presumptions, demonstrating the paradigm’s value. Put in realistic terms, if the prevailing economics paradigm tells us that people naturally prefer to maximize GDP growth, then economists might develop a model that shows what people would do, keeping in mind people’s natural inclination to maximize GDP growth. This model, then, might be used to inform policy decisions, and business owners, investors, and people who otherwise make decisions in powerful institutions might assume that GDP growth should be maximized. Then, GDP might grow. It is then proposed that the market, made up of people, has rationally reacted to its situation by maximizing GDP growth; therefore, people must naturally prefer to maximize GDP growth. Of course, it might not necessarily be that people naturally want to maximize GDP growth. It may instead be that dogmatism leads to self-fulfilling prophecy and distorted understandings along the way.

Despite this problem with dogmatism, if Popper had been right — that a single falsification of a theory devalues it in science — then perfectly useful theories might have been discarded on the basis that their results were not always exactly as predicted. Instead, we seem to have gained a good deal from theories that sometimes made predictions that were not absolutely accurate. To this point, paradigm shifts seem to be reflective of incommensurability.

There does seem to be a subjective component in how we conduct science and in how we conceive of paradigms. If we accept results under one paradigm and accept those results under another paradigm, then the explanations for these results seem to necessarily differ under each paradigm, even if each paradigm has been seen as useful at some point or another. Paradigm shifts depend on a number of factors, but surely, limitations on humans’ perspectives play a role, leading to different forms of bias.

I once watched a video of Roland Fryer speaking about education reforms that he led in some of Houston’s public schools. I cannot recall how to find the video, so for the sake of this paper, I will describe what happened and allow you to imagine it, as the details are not as important as the idea. In his talk, Fryer talked about outcomes in his reform program’s schools. He said that a research assistant had brought him data, but that the results were so extreme that he was unsure of their accuracy. This led him to double-check the results, and given his explanation, it seems that he normally does not double-check results because he assumes that there is no need. Essentially, Fryer had an expectation of results such that a case, in which the results lay outside what he had expected, led him to alter his research process. Maybe this is a rare occurrence, and it may be worth noting that Fryer’s change was subtle, but I believe that the implication is important. How often do economics expectations lead to changes in research processes after unexpected results?

This is not an indictment of Fryer; the point is that biases guide sciences and that this complicates things. Even if dogmatism in sciences were necessary to achieve practical, useful results — which Kuhn might agree with — it does seem that science is not the objective undertaking that many people claim and that many would prefer. Science does not seem to provide real answers to big questions so much as it leads us to flawed, albeit somewhat educated, estimations. Those estimations can seem (and might even be) helpful in coming to decisions, but they, in and of themselves, are not solutions and they do not necessarily lead us to any solution(s).