The Demarcation Problem

What is science, exactly? What do disciplines like physics, astronomy, chemistry, geology, and biology have in common, and what distinguishes them from nonscientific disciplines like mathematics, philosophy, and history? Do the natural sciences share a distinctive set of features not shared by pseudosciences like astrology, alchemy, and phrenology? As mentioned in the first chapter of this book, it is not easy to formulate a precise definition of ‘science’ that includes all and only the activities and fields of study regarded as science today. The eminent 20th-century philosopher of science Karl Popper dubbed this the “problem of demarcation,”Karl Popper, The Logic of Scientific Discovery (London: Routledge Classics, 2002), 11, originally published in German in 1934 and first translated to English in 1959. and he devoted much of his career to an ultimately unsuccessful attempt to solve it.

Unfortunately, philosophical views of science are often incongruous with the actual practice of science. Historically, at least, philosophers frequently mischaracterized how science works, and Popper was no exception. Conversely, scientists themselves sometimes adopt simplistic or outmoded philosophical conceptions of science. For example, many scientists today embrace Popper’s demarcation criterion, which says that “falsifiability” is a defining characteristic of scientific hypotheses. This demarcation criterion is inadequate for several reasons, as we’ll see. Yet, Popper’s mistaken definition of science persists in the scientific community and is even taught in some college-level science textbooks.For example, in their otherwise excellent textbook The Sciences: An Integrated Approach, 9th Edition (New York: Wiley, 2023), James Trefil and Robert M. Hazen write: “The central property of scientific ideas is that they are testable and could be wrong, at least in principle. … As we stated above, a central aspect of the scientific method is that every scientific statement is subject to experimental or observational tests, so that it is possible to imagine an experimental result that would prove the statement wrong. … Such statements are said to be falsifiable. Statements that are not falsifiable are simply not part of science.” (p. 17) We’ll consider Popper’s view shortly. First, an examination of several older attempts to demarcate science will be instructive.


bust of Aristotle
AristotleThis marble bust of Aristotle is an ancient Roman copy of a bronze original, which was created by the Greek sculptor Lysippos in 330 BC. Image source: Wikimedia Commons (public domain)
384 - 322 BC

From the medieval period into the early years of the scientific revolution, Western approaches to the study of nature were shaped by ancient Greek philosophy. Aristotle’s writings in the Organon were especially influential. Organon, a Greek word meaning “instrument” or “tool,” was the title given by later editors to a collection of Aristotle’s writings on logic and methods of reasoning. In this impressive body of work, Aristotle develops a sophisticated system of deductive logic and a detailed methodology for reasoning carefully about the world.

Aristotle’s methodology can be summarized, with some oversimplification, as an attempt to expand our knowledge of the world using deductive reasoning. Deductive reasoning is the sort of reasoning that can be represented using formal logic, and it has an important feature: it is truth-preserving. This means that if you start with true premises (true assumptions) and use logical deduction correctly, you are guaranteed to arrive at true conclusions.For further explanation, see this page of my e-book Skillful Reasoning: An Introduction to Formal Logic and Other Tools for Careful Thought. According to Aristotle, we begin with knowledge of simple truths about nature, truths which we can apprehend directly with our senses. From our knowledge of these “first principles,” as he called them, we then logically deduce many other facts about the world, expanding our knowledge via logical demonstrations, or proofs.

This approach to the study of nature, which eventually came to be known as deductivism because of its reliance on deductive logic, was not intended as a demarcation criterion in the modern sense: Aristotle wasn’t trying to demarcate science from pseudoscience. Nonetheless, insofar as it delineates a specific methodology for the study of nature, Aristotelian deductivism can be viewed as defining something akin to science. Studying nature in the way Aristotle prescribes might even be considered a kind of ancient science, which I’ll call Aristotelian science. Moreover, although Aristotle was not concerned with the demarcation of science per se, he did offer two criteria by which to distinguish what he considered scientific knowledge (Greek: epistêmê) from other kinds of knowledge and from mere belief or opinion:

  1. Unlike mere opinion, Aristotle claimed, scientific knowledge can be obtained with absolute certainty. This is because, on his view, scientific knowledge is proved by strict logical deduction from “first principles” that are obviously true.
  2. Scientific knowledge involves causal explanations, whereas the kind of knowledge employed in the crafts does not. A carpenter, weaver, or goldsmith need not understand the causes of natural phenomena to master his craft, or so Aristotle maintained.Philosopher of science Larry Laudan characterizes Aristotle’s view this way: “In his highly influential Posterior Analytics [one of the works included in the Organon], Aristotle described at length what was involved in having scientific knowledge of something. To be scientific, he said, one must deal with causes, one must use logical demonstrations, and one must identify the universals which ‘inhere’ in the particulars of sense. But above all, to have science one must have apodictic certainty. It is this last feature which, for Aristotle, most clearly distinguished the scientific way of knowing. What separates the sciences from other kinds of beliefs is the infallibility of their foundations and, thanks to that infallibility, the incorrigibility of their constituent theories. The first principles of nature are directly intuited from sense; everything else worthy of the name of science follows demonstrably from these first principles. What characterizes the whole enterprise is a degree of certainty which distinguishes it most crucially from mere opinion.” Moreover, Laudan continues, “Aristotle sometimes offered a second demarcation criterion, orthogonal to this one between science and opinion. Specifically, he distinguished between know-how (the sort of knowledge which the craftsman and the engineer possess) and what we might call ‘know-why’ or demonstrative understanding (which the scientist alone possesses).” Laudan concludes: “Coming out of Aristotle’s work, then, is a pair of demarcation criteria. Science is distinguished from opinion and superstition by the certainty of its principles; it is marked off from the crafts by its comprehension of first causes.” Larry Laudan, “The Demise of the Demarcation Problem,” in R. S. Cohen and L. Laudan (eds.), Physics, Philosophy and Psychoanalysis: Essays in Honor of Adolf Grunbaum (Boston: D. Reidel Publishing Company, 1983), 112-113.

Unfortunately, Aristotelian science and its deductivist methodology are not adequate descriptions of how science works today, for several reasons. First, modern scientific methodologies don’t begin by apprehending “first principles,” nor do they rely primarily on logical demonstrations or proofs (though logical deduction still plays an important role, as we’ll see shortly). Contrary to the first criterion above, moreover, modern science never yields absolute certainty. All scientific theories are amenable to revision or replacement in light of new evidence. Indeed, even the most thoroughly-tested and well-confirmed scientific theories may turn out to be false. (As an example, consider how Einstein overturned Newton’s laws in the aftermath of the Michelson-Morley Experiment. Similarly, the two best-confirmed theories of contemporary physics—quantum field theory and Einstein’s general theory relativity—are logically inconsistent with each other, so we know at least one of them must be false.) Furthermore, modern scientific explanation doesn’t always involve the notion of causation, contrary to the second criterion above. We’ll see why when we revisit the topic of scientific explanation later in this chapter. For now, suffice it to say that some aspects of modern science are concerned with description rather than with explanation of any kind.

Curiously, despite these glaring incongruities between Aristotle’s conception of science and the new approaches developed by Galileo, Newton, and others, some of the most brilliant thinkers of the scientific revolution—including Galileo and Newton themselves—initially tried to conceptualize and justify their work in deductivist terms. For example, Galileo claimed that his conclusions could be logically deduced with absolute certainty in the Aristotelian way. Even Newton occasionally justified his reasoning in Aristotelian language, though he explicitly abandoned Aristotle’s requirement of giving a causal explanation for the force of gravity—a defection for which he received sharp criticism from many of his scientific contemporaries.

The surprising persistence of Aristotelian deductivism during the scientific revolution illustrates an important point raised above: philosophical views about the nature of science, including views held by scientists themselves, are often incongruous with the actual practice of science. Scientists typically view their own work through the lens of the prevailing philosophical account of what science is supposed to be, even when their work directly contravenes that philosophical picture. Even the greatest pioneers of the scientific revolution, Galileo and Newton, tried to characterize their innovative work in terms of the old, Aristotelian methodology, which remained the dominant view well into the 17th century.As philosopher of science Larry Laudan observes: “Galileo claimed to know little or nothing about the underlying causes responsible for the free fall of bodies, and in his own science of kinematics he steadfastly refused to speculate about such matters. But Galileo believed that he could still sustain his claim to be developing a ‘science of motion’ because the results he reached were, so he claimed, infallible and demonstrative. Similarly, Newton in Principia was not indifferent to causal explanation, and freely admitted that he would like to know the causes of gravitational phenomena; but he was emphatic that, even without a knowledge of the causes of gravity, one can engage in a sophisticated and scientific account of the gravitational behavior of the heavenly bodies. As with Galileo, Newton regarded his non-causal account as ‘scientifical’ because of the (avowed) certainty of its conclusions.” Larry Laudan, “The Demise of the Demarcation Problem,” in R. S. Cohen and L. Laudan (eds.), Physics, Philosophy and Psychoanalysis: Essays in Honor of Adolf Grunbaum (Boston: D. Reidel Publishing Company, 1983), 114. Similarly, scientists today sometimes characterize their work in terms of outmoded philosophical views, as we’ll see.