A The Unity of Science

The vision of all scientific knowledge as forming a grand, coherent whole is an­cient. The biologist E. O. Wilson dreams of a glorious unification—his word is consilience¹—of all knowledge in the sciences and the humanities.

4. A.1 Science and Nonscience: The Demarcation Problem

If there are limits to the kinds of knowledge that science can accommodate, how should we distinguish what is in its bailiwick from what is outside? Drawing the line between science and nonscience is the famous demarcation problem³ (Chapter 2). Karl Popper named the problem and argued that the distinction between science and nonscience should be made according to the criterion of falsifiability. A scientific statement can, in principle, be empirically tested in such a way that one conceivable test outcome would show that the statement is false. In other words, scientific statements are open to the possibility of disproof. We ac­cept that, since all knowledge is uncertain, falsifiability itself is also provisional, but we do the best we can. By saying that hypotheses are testable in principle, we mean that we might not be able to test them directly. “Water is formed by chem­ical reactions on the dark side of the moon” is a scientific statement; we can im­agine doable tests that would falsify it.

Nonfalsifiable statements, though not scientific, might stimulate scientific inquiries, or they might be meaningful in another domain of human experience; they are not necessarily “meaningless,” as the Logical Positivist philosophers said they were. Popper thought such statements were nonscientific, not nonsensical. He accepted that certain metaphysical concepts are inescapable, even in philos­ophy and science; we've said that we accept the principle of realism, that the ex­ternal world does exist, although we can't prove or falsify it.

Even with the standard of falsification as a guide, there are “gray areas.”⁴ We accept that “science is characterized by a fuzzy borderline.”⁵ In some instances, the testability of hypotheses is highly problematical, such as “the Astonishing Hypothesis”⁶ that human consciousness is created by the brain or the “Super Strings”⁷ that modern theoretical physics suggests are the most elemental building blocks of the universe. We encounter fuzziness when we try to tell sci­ence, such as evolution, apart from “almost-science” such as the search for extra­terrestrial intelligence (SETI).⁸ And fuzziness is present when we try to separate kinds of science from each other.

Fuzziness is unacceptable to certain philosophers who reject falsifiability as a standard for demarcation and insist on rigid distinctions between categories. Yet fuzziness at the borders does not mean there are no borders. It is good to distinguish between day and night, even though we can't pinpoint when the change-over happens. Similarly, if you rely on falsifiability as a standard for telling whether you're talking about science or something else, you'll usually be on safe ground. In any case, this does not answer the question of whether, within the wide boundaries established by falsifiability, science is a unified intellectual realm or a disorganized jumble of unrelated fields.

Why should it matter? Scientists go about their jobs in a variety of ways and have a variety of goals. If there were a fundamental pluralism of sciences, we could be faced with incompatible systems of thought, of ways of comprehending the world. A science unified by a set of shared premises, values, goals, and ge­neral operating principles would make for a more harmonious view of na­ture and a more straightforward task of understanding it than if science were disunified. The unity of science question has consumed much intellectual en­ergy, and philosophers have identified many possible unities of science⁹ without arriving at a definitive answer. We'll start by considering two alternative ways of thinking about the unity of science.

4. A.2 An Untidy Account?

An extreme interpretation of unity is that science aims to wrap up everything in one “tidy account.”¹⁰ Probably no one who knows anything about science would say that it is “tidy”—much of it is messy business even on a good day—but is it one account? Visions as grand as Wilson's are, in one way or another, forms of reductionism: the idea that we can understand a complex subject in terms of its most basic constituents. “Quarks” are subatomic particles that are realities for both physicists and psychologists, but physicists can use the concept of quarks to achieve their goals of scientific understanding and psychologists can't, in spite of the fact that people are just complicated arrangements of atoms and subatomic particles. The most “greedy”¹¹ or “uncompromising”¹² forms of reductionism can't solve the unity of science problem because not all facts are equally useful to all sciences. Some complexity is irreducible: we cannot comprehend why a com­puter gives the correct answer to a math problem based on our knowledge of the computer's electronics alone because math is a product of the mind and culture, and physics cannot explain them.

One interpretation is that science is unified by its major objectives (explana­tion and understanding of nature) and values (empirical testing, reproducibility, etc.) and that different branches of science have much in common from this per­spective. This is sometimes called a container model of unity¹³: all forms of in­quiry that share these objectives and values are within the container.

Another way of viewing the relationships among different sciences is the inte­grated causal model.

4. A.2 The Integrated Causal Model

Evolutionary psychologists Leda Cosmides and John Tooby¹⁴ were concerned about the disarray and lack of progress that plagued the social sciences and that finally broke out in the “Science Wars.”^15,16 The Science Wars represented a per­ceived split between culture and science that divided academia during the 1990s, and, as part of a plan to restore order, Cosmides and Tooby proposed that science is unified via a framework that they called conceptual integration—a consist­ency of principles throughout science. Their Integrated Causal Model holds that one branch of science cannot be based on principles that are flatly contradicted or forbidden by well-established knowledge developed in another branch. The natural sciences are already conceptually integrated: for example, although bi­ology is separate from chemistry and physics, biology is compatible with chem­istry and physics. Yet biology cannot be predicted from or entirely explained by chemistry and physics; conceptual integration is not a front for a reductionist argument. According to the Integrated Causal Model, psychology and the so­cial sciences should be compatible with each other and with the natural sciences; psychology should not, for example, create hypotheses of human behavior based on cellular mechanisms of learning that neuroscience has ruled out. Sociologists should not erect theories of group interactions that depend on interpretations of individual psychology that psychologists can show are not true.

The exchange should go both ways because science and culture are pro­duced by the human mind. Evolutionary psychology approaches the study of the mind in the context of its evolutionary history. The central hypothesis of evolutionary psychology is that our present mental capabilities owe more to adaptations acquired during the 2 million years that we spent as Pleistocene Age hunter-gatherers than to changes that have occurred during the few thou­sand years since the dawn of civilization. From this vantage point, psychological investigations can lead to insights into such socially meaningful phenomena as male jealousy,¹⁷ the perceptual organization of color,¹⁸ and human preferences for certain landscapes.¹⁹

Since the mind is shaped by its built-in information processing mechanisms, cognitive science could conceivably discover principles that limit the intellec­tual constructs of physicists or help us understand cognitive phenomena, such as biases, that confront science. Anthropologists could investigate how earth scientists arrive at consensus decisions about global warming. Or it might be that problems such as human consciousness remain forever beyond our grasp be­cause our ancestors’ survival never depended on the type of thought processes that would be required to comprehend consciousness.

The Integrated Causal Model readily accommodates the cross-pollination and hybridization that creates new scientific fields—“socio-biology,” “psycho­genetics,” “geo-chemistry”—and makes fixed classifications of individual sci­ences obsolete.

The container and connectionist models of scientific unity are not mutu­ally exclusive. Both recognize that relationships exist among different branches of scientific fields and that all fields of science are part of one giant endeavor. Another tie that binds fields of science together is the methodological unity that Karl Popper proposed.

4. A.4 Methodological Unity

We’ve seen that Popper’s Conjectures and Refutations is a program of trial and error; it is a formal way of learning from our mistakes. Proposing, testing, fal­sifying, and inventing better hypotheses is how science grows and tends to its knowledge base. For Popper, science is held together by this methodological unity,²⁰ which is founded on falsification. At first glance, falsification may seem to be an element of narrowly defining characteristics of a container model, but it is broadly applicable to all of science.

To appreciate the scope of Popper’s methodological unity, we need to keep in mind that falsification plays two separate roles in science: it is the key to the over­arching concept of demarcation, which is how we distinguish science from non­science; and it is the centerpiece of Conjectures and Refutations, which is how we conduct certain kinds of science on a daily basis. A statement in a particular field is scientific if it can be falsified (demarcation), even if the scientists in that field are not always engaged in testing hypotheses (conduct). Imagine a biologist out counting monarch butterflies in the American Southwest. Everything about her field of study, entomology, passes the falsification test for demarcation, but she might be occupied in taking a butterfly census, gathering data, not testing a hypothesis. Thus what she's doing at the moment is not within Popper's defini­tion of methodological unity of science as it relates to her conduct. Popper does not emphasize the distinction between the two roles of falsification, but it is im­portant when we come to distinguishing among different branches of science.

Popper took it for granted that science would constantly discover new phe­nomena to investigate. He “readily admits that only observation can give us knowledge concerning facts, but this knowledge does not establish the truth of any scientific statement.” Plain observations are “there-is” statements about the world (e.g., a physical object exists in such-and-such a place at such-and-such a time). Once we have observations, then we can invent and test hypotheses to explain them.

In summary, although Popper proposed that falsification is the essence of the methodological unity of science, he recognized that there are clearly scien­tific activities that do not adhere to the process of hypothesis testing at all times. Being mindful of the two roles of falsification makes sense of the differences be­tween kinds of science while simultaneously seeing them as part of the unifica­tion of science.

4.