So far in this series, I’ve been using the word “science” as if its meaning were self-evident. And in a sense, it is: science is the investigation of the natural world. Insomuch as every society acquires knowledge of the natural world, every society has science. It seems to me this is the usual sense of the word “science”; what one might call its naive, or non-philosophical, sense. Another, closely related naive sense of the word is that offered by the Victorian biologist T. H. Huxley, who said that science is “organized common sense.” To be reasonable, in even a very mundane and everyday sense, is to be scientific, and vice versa.
Both of these senses of the word capture something intuitive and important. Science is indeed the investigation of the natural world. Even if the boundaries of the natural world are a bit hazy and require extensive philosophical inquiry to sort out, most definitions would probably include things like stars, rocks, trees, grasshoppers, and photons. That is, they would include all of the “dead stuff,” the inert objects, around us, and at least most of the living things as well. Science is organized common sense in the sense that the practices and assumptions that we call science have been so influential over the last several centuries that they have changed our sense of what it means to “know” something, or to speak “sensibly” about it. As I’ve tried to show in previous articles, even such a basic category as “knowledge” is historically variable. So in the sense that science has become a kind of paradigm case or exemplar of the category of knowledge, it is indeed, for many of us, just organized common sense.
From a philosophical point of view, however, neither of these descriptions are quite adequate. In the first case, if every society has science, then the question why knowledge of the natural world has advanced so rapidly over the last few centuries becomes difficult to answer. They all have science in the sense that they all involve inquiry into the natural world, but they do not all have the same, or equally effective, kinds of science. Something distinct happened in Europe over the last several centuries that allowed knowledge of the natural world to proceed faster and penetrate deeper than it had in other times and places, and eventually to spread to very different societies as well. What was it? What is distinctive about science—not as a set of theories, but a set of assumptions and practices that underlay those theories—as we have it today? What made them so effective that they could change the grounds of knowing itself? In this and the next article, I’d like to sketch a “virtue theory” of science—my own attempt to provide a satisfactory “demarcation criteria” for separating science (in the second sense) from non-science, on the basis of an inventory of what virtues a research activity can have that help it qualify as scientific. I hope to satisfy both the philosophical demand for conceptual clarity, and the historical demand for an empirical basis. In other words, I’ll try to be both normative (what should the word “science” mean?) and, in the subsequent articles, explanatory (how did this sense of the term in fact come about?)
A good place to begin such a discussion might be with an observation: “science” is a contested word. There are a lot of people trying to get their activities labeled “science,” and themselves labeled “scientists,” and these claims are not all of equal validity. Some clearly pass: Lawrence Krauss’s claim to be a scientist, for instance, would be pretty tough to dispute. As a cosmologist whose work involves a lot of physics and math, and who is employed at a major research institution, his claim is quite credible. Others aren’t so clear. Consider, for instance, digital humanities. This is an approach to texts, often literary or historical, that takes advantage of recent developments in computer technology to search for large-scale patterns across hundreds or thousands of documents. Once those patterns have been identified, they can be explained. The idea is that this approach could provide a more naturalistic, uniform understanding of texts—one less dependent on the idiosyncrasies of their authors, or the scholars who study them. Should this count as science? Perhaps. This claim doesn’t seem to be as secure as that of a physicist, but it seems much more secure than that of a homeopathist or an astrologer. Another contested area is sociology. Beginning in the mid-nineteenth century, and continuing to our own time, Marxists have frequently claimed that their research and advocacy was simply a scientific approach to political and economic activities currently governed by ideology and wishful thinking. The advance of science, of course, included a science of human social relations, and that’s what they had to offer. There probably are not many people, apart from Marxists, who think that is a credible claim.
Just waving the word “science” around shouldn’t be enough to secure our assent, and critics are right to point that out. In order to sort out these conflicting claims, we need to do a bit of philosophy.
From these examples, it seems reasonably clear that the claim to scientific status is also a claim to prestige and authority. When someone says that their work is science, they’re saying that you ought to believe in and value it. They’re distinguishing their work from that of non-scientists, people who don’t have the same epistemic rights. Similarly, criticisms of science (either in particular cases, as with anthropogenic climate change, or more generally, as with postmodernism) frequently involve accusations of overconfidence on the part of the scientists: either that they aren’t living up to their own standards, or that those standards are not as secure as claimed. After all, scientists are only human, and they’re subject to the same failings and limitations as the rest of us. Just waving the word “science” around shouldn’t be enough to secure our assent, and critics are right to point that out. In order to sort out these conflicting claims, we need to do a bit of philosophy.
So the attempt to differentiate science from non-science clearly has high stakes; it’s important to get this right. I think a good way to do this is to look at the paradigm cases—areas of research that virtually everyone counts as scientific—and to expand from there. Physics, as we’ve already observed, is clearly in, and to such an extent that (it seems to me, just based on observation) when people say “science” what they often mean is “physics.” There is often a presumption that a research activity is scientific just to the extent that it employs the assumptions and methods characteristic of physics research, especially empirical observation, mathematical equations, and experiment. More about this later. I don’t think it’s really accurate to argue that science is reducible to physics, however, and the main reason I would give for that is that there is another research activity that I think most of us would want to count as science, but employs different assumptions and methods: biology.
Let me expand on this. When, in 2016, physicists at the LIGO observatories detected gravitational waves, it was hailed as a triumph for Einstein’s theory of relativity, and for physics generally. It was what we expect of good, solid physics, and three researchers earned the Nobel prize as a result. Their success involved, in the first place, empirical observation. Einstein’s prediction that gravitational waves would result under certain circumstances could grant them only a hypothetical reality until they were actually observed—an event important enough to justify the construction of a pair of facilities at a cost of hundreds of millions of dollars. Similarly, contemporary criticisms of multiverse and string theory are often made on empirical grounds. Both theories posit the existence of entities that cannot now be observed, and may never be observed—so, the criticism runs, they aren’t really physics. You’re only doing physics when you’re talking about entities that you can now observe, or have some reasonable prospect of observing in the near future. Observation takes precedence over theory.
Observation, while necessary, is not sufficient for physics. The key distinction is between qualitative and quantitative observation. A qualitative description is something like “happy,” “angry,” or “sad,” “heavy” or “light,” “good” or “bad,” “red” or “brown,” etc. It’s our normal language of description, and part of what makes language so slippery is that qualities are matters of interpretation. An object that is heavy for one person might be light for another. I might think that Donald Trump is a terrible president, while another person might think he’s merely a bad one. Qualitative description draws implicitly on a shared set of assumptions. When those assumptions are not, in fact, shared, these descriptions are open to all sorts of misinterpretations. As we will see in future articles, the transition from qualitative to quantitative descriptions in physics was an important one, and its chief importance lay in the precision that quantitative description makes possible. While the meaning of “bad” is open to a wide variety of interpretation, the meaning of “2” is precise and consistent. When physicists describe natural events, they almost always do so in quantitative terms. Einstein’s field equations, for instance, are just that—they’re equations. A well-formulated physical theory is a mathematical description of an empirically observed (or observable) event. Unless both of these elements are in place, a theory in physics hasn’t really reached maturity.
But it’s not enough to have only these two elements. Ideally, observation is of a certain, special kind: the kind that occurs under precise, controlled, repeatable circumstances, or, in other words, an experiment. Our naive intuition, for instance, tells us that heavy objects fall faster than lighter ones, and that’s what Aristotle had said as well. For the roughly nineteen centuries between Aristotle and Galileo, it was true, as far as anyone knew. But Galileo’s famous experiment, where he is said to have dropped two differently weighted balls off the leaning tower of Pisa, showed that this wasn’t true. They hit the ground at the same time. You or I or anyone else can repeat this experiment today, under more or less the same physical conditions that Galileo did. If you want to control for wind, heat, and other such factors, you can do the experiment in a lab. The Apollo 15 astronauts even repeated it on the moon.
This element of repetition under controlled circumstances distinguishes an experiment from other types of observation. Consider, for instance, the stellar redshift, which provides the principal evidence for the Big Bang. When we examine the light spectrum of most of the stars that are visible to us, we find that spectrum is disproportionately toward the red end of the color scale. This is generally interpreted as a consequence of the Doppler effect—the stretching of light (or sound) waves as an object moves away from an observer. The stars appear redder than they are because they are moving away from us. If the stars are getting farther away as we move forward in time, it stands to reason that they get closer together as we move backward in time, and all at the same point if we move all the way back. There must, then, have been an original, cosmic explosion that brought the visible universe into being, and that’s one of the bases (there are others) for supposing that the universe began in a “Big Bang.” The difference between the redshift observation and Galileo’s experiment, is that the redshift observation is passive, while Galileo’s was active. In other words, we just make this observation about stars, record, and interpret it, but we can’t really control it in the way we can falling balls. We can’t bring a star into the laboratory and subject it to different conditions in the way we can with falling balls, and that limits our ability to “poke at it,” so to speak. It limits our freedom of action with respect to it.
There are other elements to physics research besides empirical observation, mathematical equations, and experiment, but these should be enough to draw the distinction I promised earlier with respect to biology. All of these elements are present, to a certain extent, within biology, but biology isn’t based on them in the way that physics is. This has to do with the subject matter itself. To a certain extent, the object of study imposes limits on the methods of inquiry as well as opening new opportunities.
Consider, for instance, Darwin’s finches. During the 1830’s, the young Charles Darwin was the naturalist aboard the HMS Beagle. He collected numerous specimens during his voyage, including various species of finches (small birds) from the Galapagos Islands. He later made an important observation: there were about a dozen species of finches on the islands, all very similar, yet with distinct shapes of their beaks. What could explain the difference? The dominant theory at the time was that species had been created according to fixed, ideal plans, which were not variable. Evolutionary ideas had been proposed, but rejected, in an earlier generation—including by Darwin’s grandfather, Erasmus—but in the 1830s were seen as discredited. But the small variations in the finches’ beaks made Darwin wonder: did it really make sense to say that God had created 15 very similar species on this one island, and that they never afterward changed? A better explanation, he eventually realized, was that species were mutable. The differences in the shape of the beaks of the finches were brought about by the conditions on the islands: broad, heavy beaks were ideal for getting through the tough hides of fruits to the seed and flesh within; longer, more slender beaks were better for hunting insects; a midrange beak helped the finch to eat plant leaves. As one food source become more plentiful, the finch with the corresponding beak shape would become more numerous. Species were not fixed, but mutable. In fact, a “species” is just a name we apply to very similar, particular organisms. There is no species as a preexisting category that we discover in the world. Rather, we discover particular organisms, we observe similarities, and we group them under similar categories that we then call “species.” What actually exists is the particular organisms; the classification is something we apply to them. The classification is useful and important (one can hardly do biology without taxonomy), but it’s not something that exists “out there.” (Here we can see the influence of nominalism in science, discussed in previous articles.)
On the basis of this and similar observations, Darwin rehabilitated evolutionary theory in the subsequent decades. He reasoned that if small changes in organic structures can occur over short periods of time, large changes could occur over a longer period as well. A dozen generations might account for the differences in beak shapes (assuming uniform common ancestors), so perhaps a couple thousand could account for differences like warm and cold blood, a lateral or an upright posture, and so on. This insight remains central to modern evolutionary biology.
If we take Darwin’s finches as a paradigm case for biology, I think it illustrates some important differences between biology and physics. This will support my contention that science is not just physics, because evolutionary biology is clearly science. If its methods are substantially different from those of physics, some of which we’ve already described, then there are at least two rather different activities subsumed beneath this one word.
So, how are Darwin’s finches, and hence biology more generally, different? Well, the element of empirical observation is the same. Darwin had to go out and observe the finches just like Galileo had to observe the falling balls. If we’re just sitting back and thinking in the abstract, it’s not really science. Science is about what you can observe—and, preferably, have observed. When we get to mathematics, however, I think we can see an important difference. Darwin’s explanation of the different shapes and sizes of the finch beaks was not an equation—not the way that relativity and quantum mechanics in physics are—and they don’t have to be. Events in evolutionary biology are, in principle, quantifiable: it’s conceivably possible to count up all the finches on the Galapagos islands, to measure their beaks, to quantify their lifespan and mating habits, and so on, but it’s not necessary to do that. Darwin’s explanation of speciation does not rise or fall based on its adherence to an equation. It can rest on more vague terms, such as “population increase,” “better-adapted,” and “successful.” It can incorporate, in other words, a greater element of qualitative as opposed to quantitative description.
If we look at our third item, experiment, we can see an important difference here as well. We can (and have) confirmed the reality of microevolutionary changes (such as with beak sizes) through experiment. We have not, and may never, confirm the reality of macroevolutionary changes (such as the switch between cold and warm blood) experimentally—not because they did not occur, but because the timescales involved are so enormous that they do not seem to be the kinds of things that can be observed in this way. We might be able to observe them indirectly (through the fossil record, for instance, or perhaps through computer simulations—raising additional, and highly interesting, philosophical questions), but we have not yet. Some of Darwin’s critics in the nineteenth century (and in our time) thought that this was an important point against his explanation, but working evolutionary biologists generally do not. The reason is that it just seems unreasonable to demand experimental confirmation of an event not amenable to experiment in the first place. It’s not a deficiency of the theory, it’s an appropriate shift in methodology brought about by the difference between events characteristic of biology and physics—namely, that the former can take millions of years to play out, while the latter often occur quite quickly. (Larger-scale physical processes are usually studied by separate sciences—astronomy, geology, cosmology, etc.) That being said, if experimental confirmation could ever be brought to bear on macroevolutionary processes, that would be an important point in Darwin’s favor, and a noteworthy event in the history of biology.
On the basis of what has been argued so far, I’d like to advocate a distinction between types of sciences. In the nineteenth century, the German philosopher Wilhelm Windelband proposed a distinction between what he called the “nomothetic” and the “idiographic” sciences. A nomothetic science is law-seeking. Its subject matter is more or less invariant. Physics and chemistry are both nomothetic sciences. A quark is a quark is a quark; if relativity is true, it is (presumably) just as true here as in the Andromeda galaxy, and will be just as true a billion years in the future as it was a billion years in the past. Nomothetic sciences involve, in other words, an assumption of nearly complete uniformity. (If the Big Bang theory is true, the uniformity cannot be absolute, since Big Bang cosmology presumes the coming into being, and hence the change of both matter and the laws of physics—but even in this case, the uniformity would be very nearly complete.) A successful explanation in physics explains, not particular objects, but what conditions apply to all objects invariably. Similarly, chemistry tells us what elements are and how they interact, not how particular elements of carbon or gold or lead interacted on particular occasions, but how all such elements interact on all occasions. The point is to subsume variation beneath a general formula.
Idiographic sciences, on the other hand, are concerned with the description of particular objects. There is no presumption of uniformity specific to that science itself. Biology is a paradigm case. With Darwin’s finches, for example, variation in food source explains the variation in the shape and size of one body part of one type of animal on one archipelago. Variations in the shape, size, and color of other parts of their bodies (feathers, for instance) may be a result of sexual selection, or they may be spandrels, i.e., changes that confer no adaptive benefit, but are secondary consequences of other changes that do. Variations with other species, or in other times and places, may be the result of altogether different processes. The extinction of the dinosaurs, for instance, was certainly an important event in evolutionary biology, but is not explicable in terms of (primarily) biological processes. The explanation is that a meteor struck the earth, resulting in catastrophic changes in the environment, and with that a mass die-off. Biological processes are involved, but the primary cause is the meteor strike—an astronomical rather than a biological event. Finally, there is no guarantee that the processes studied by evolutionary biology are fundamentally similar on other planets, if there is life on them at all. Perhaps a very different set of processes could bring about life on other planets. If that were the case, it would not invalidate or impugn Darwin’s discoveries in any way. The reason is that they are specific to the earth and the living things on it. They do not pretend to explain, and do not have to explain, what happened anywhere else.
Geology and cosmology are other idiographic sciences. Geology is concerned with the description of particular events on our particular planet. The geologist who wants to explain the formation of the Himalaya mountains does not have to consult conditions on Mars or Venus. The explanation of the particular event on earth is right or wrong in itself, not with respect to its ability to explain events elsewhere. In cosmology, the goal is to explain the origins and history of our one universe, irrespective of the existence of other universes or their histories.
So, to sum up, the claim to be doing “science” is a claim to prestige and authority. For this reason, it’s important to distinguish between legitimate and illegitimate claims to that term, and in order to do that we need to have a clear idea of what science really is. We need what philosophers of science call a “demarcation criteria.” A good place to start would be (like a good scientist) with observation: which subjects enjoy the best claims to be scientific in our time? The answers are, I would like to suggest, physics and biology, which are much more distinct, methodologically, than is generally realized. In order to account for this distinction, we can make use of Wilhelm Windelband’s distinction between nomothetic (general and law-seeking) and idiographic (particular and descriptive) sciences.
With this distinction in place, we can then take a “virtue” approach to their description. We can, in other words, isolate and describe particular aspects of research that can help us determine whether it qualifies as scientific in either the nomothetic or the idiographic sense of the term. We’ve already explored three (empiricism, quantification, and experiment.) In the next article, we’ll explore others, and I’ll try to bring all of these elements together into a proposal for demarcating science (in the peculiarly modern sense) from non-science.
This essay is part of a series; the previous essay can be found here.
Daniel Halverson is a graduate student studying the History of Science and Technology. He is also a regular contributor to the PEL Facebook page.