The National Association of Scholars will focus more of its attention in the next several years to the modern crises of science—the linked crises of reproducibility, misuse of statistics, sloppy research procedures, and frequently politicized groupthink. We will publish a series of reports dealing with aspects of these crises, as well as shorter pieces by distinguished scientists. We are delighted to inaugurate this series with Edward R. Dougherty’s summary of the epistemological roots of the modern crisis of science—and the ever-sharper imperative for working scientists to apprehend the philosophy of science, as a prerequisite to their scientific practice.
Edward R. Dougherty is Robert M. Kennedy '26 Chair and Distinguished Professor of the Department of Electrical and Computer Engineering, and Scientific Director of the Center for Bioinformatics and Genomic Systems Engineering, at Texas A&M University. His publications include Epistemology of the Cell: A Systems Perspective on Biological Knowledge (co-author Michael L. Bittner, 2011) and The Evolution of Scientific Knowledge: From Certainty to Uncertainty (2016).
Mention philosophy of science and many people, including scientists, will think that you are referring to arcane reflections unrelated to the pragmatic business of the world. This is especially so if you mention “epistemology,” the theory of knowledge. But they could hardly be more mistaken. Scientific epistemology has developed over the centuries to characterize what we mean by the knowledge of empirical events. A scientist aims to gain knowledge. To achieve this aim, the quest must be guided by an appreciation of that which he wishes to pursue. Quoting Albert Einstein, “Science without epistemology is – insofar as it is thinkable at all – primitive and muddled.”
Most people have the commonsense Aristotelian notion that one has empirical knowledge if an idea in the mind possesses concordance with the thing outside the mind to which the idea pertains. This view presupposes that one knows the meaning of the words “idea”, “thing”, and “concordance.” Epistemology gives meaning to these terms and in doing so shapes our thinking about phenomena.
With Isaac Newton, science became essentially what it is today: the “idea” in the mind is a mathematical system constituting the knowledge, the “thing” is a measurement, and “concordance” is a relation between the mathematical system and the measurement that facilitates testing the theory. It took three centuries, from Newton’s Principia through the development of quantum theory, to gain a deep appreciation of these terms and to recognize the subtleties. Only by serious contemplation on the evolution of scientific thinking through the centuries can one truly appreciate these subtleties.
Especially troublesome is the elusive character of scientific “truth.” The “truth” of a theory somehow lies in its concordance with Nature. But we don’t observe Nature directly; rather, we observe phenomena, not the underlying thing-in-itself. More perplexing still, not only is Nature perceived through the filter of our senses, the perceptions themselves and the ideas to which they lead depend on the structure of mind, on what Immanuel Kant called “categories of the understanding.” Our minds impose structure on ideas as a prerequisite for ideas to be thought.
Understanding is limited to our mental categories but science is not limited to them. Consider causality, one of Kant’s categories. Whereas Aristotle had regarded the determination of cause as basic for scientific knowledge, Newton (following Galileo) rejected causality as part of science. For Newton, mathematics constitutes scientific knowledge and that knowledge is not subservient to human understanding. Thus, his famous phrase: “Hypotheses non fingo.” (“I frame no hypotheses.”)
With quantum theory in the early part of the Twentieth Century, the full implication of Hypotheses non fingo is realized. Whereas classical terms such as “particle,” “wave,” and “force” had their origins in pre-scientific perceptual experience, with quantum mechanics the observations do not fit into everyday understanding.
There is the mathematical theory, the relations between the mathematics and the observations, and the experiments that test predictions of future events deduced from the theory. Nature may not be understandable, but mathematics can provide predictions. Mathematics is understandable because it is a product of human intellect, but it may not describe the phenomena in a manner compatible with pre-scientific physical intuition. Does light consist of particles or waves? Take your choice. It really doesn’t matter. The theory allows you to predict future behavior. Scientific knowledge is functional.
Owing to its mathematical form, science goes beyond understanding, but at a price: a theory’s validation lies outside of pure reason in the domain of prediction, whose analysis requires difficult statistical theory. Naïve concepts of objectivity and subjectivity must be discarded. A theory is accepted or rejected, always provisionally, on the basis of whether or not observations are sufficiently close to predictions derived from the theory. Sufficiency is based on statistical criteria. There is some range of measurements within which a theory is accepted. Two scientists may disagree on the acceptance region. Thus, one might accept a theory while the other rejects it. Science is inter-subjective: all can agree on the mathematical formulation of a theory and understand the criteria of acceptance, but they need not agree on the criteria.
There is significant subtlety in scientific epistemology and one might ask, Can’t I just apply the concepts of model and validation and get on with my work? Perhaps, if you can do the mathematical modeling, rigorously relate the model to measurements, and perform the necessary experiments. However, none of this is straightforward, especially when dealing with complex systems, such as in biology, where there are tens of thousands of potential variables in a single cell and it is experimentally unfeasible to test more than a few basic predictions, if that.
We meet the crisis of the Twenty-first Century: the human desire to study complex systems for which the measurement processes themselves are intricate and often very noisy. The requirements of science may not only be extraordinarily difficult to meet, they may be impossible to meet unless focus is extremely narrowed, perhaps so much that the resulting theory is so crude that accurate predictions cannot be made. We are confronted by the impossibility of scientific knowledge regarding aspects of Nature vital to human well-being.
It is mandatory that the epistemological issues related to this conundrum be studied by those hoping to gain scientific knowledge regarding high-dimensional systems, such as the human cell, or systems whose behavior must be modeled across large time scales, such as climate. Science faces an epistemological crisis greater than the one engendered by quantum mechanics. Those issues were resolved within the basic Newtonian epistemology, whereas it does not appear that contemporary scientists will be so fortunate when it comes to their epistemological quandary. Those who wish to pursue fundamental knowledge must profoundly comprehend the path that has got us where we are and prepare for the long, steep road ahead.