Teaching students the nature of science
My former student Finn, when asked the cause of some chemical property or reaction, would often answer by saying, “SCIENCE!” This was pronounced in stentorian tones with expansive gestures, which, together with his jumbo-sized hairdo, created an impression.
It was tongue-in-cheek, of course; he did not literally believe that Science was some sort of mysterious and powerful agent causing physical phenomena–it was just a humorous way of admitting he didn’t know an answer. We sometimes do encounter, though, people who speak of Science as though it were some sort of deity to be reverenced rather than understood.
The popular TV personality Bill Nye the Science Guy says: “I really believe in science. It is a faith. It is a reverence akin to religion. But as we always say, it’s different from religion in that, as near as we can tell, it exists outside of us. It has an objective quality, the process of science.”
This perspective on science is seen rather commonly in the media. The model of science presented is akin to a religion in which a priestly caste of scientists conveys revelations about the god Science to people, and the duty of the laity is blind faith and obedience. This is not, obviously, the way science actually works. Instead science gains the authority it has by virtue of the strength of evidence and arguments that scientists produce.
Every educated person should have a good understanding of the nature of science. This requires understanding that there are different kinds of scientific knowledge. I teach my students that there are four types of scientific knowledge: observations, laws (generalizations), theories, and models. Each allows for a different level of certainty and demands a different degree of assent. As Aristotle noted: “For a well-schooled man is one who searches for that degree of precision in each kind of study which the nature of the subject at hand admits” (Nicomachean Ethics, 1094b3).
Observations are perceptions originating in the senses, often aided by instruments. In science, we make both qualitative and quantitative observations: colors, textures, shape, heat, light, and so on are qualitative; measurements of mass, volume, temperature, spectral lines, etc. are quantitative.
Observations, whether scientific or not, produce a high degree of certainty. They are the facts that serve as the foundation of all knowledge. Although everybody realizes that mistakes occasionally occur, we cannot help but believe the evidence of our senses unless we have some reason to suspect that appearances have deceived us.
We are all in the position of relying for most of our knowledge on other people’s observations, and this is perfectly reasonable in most cases. In this regard, we have long trusted scientists’ observations more than others’, because the profession has developed recording and publishing standards that guard the reliability of observations. But scientists, like ordinary mortals, sometimes make mistakes. And just as ordinary people sometimes exaggerate the fish’s length, scientists too can have incentives to fudge things.
Observations underlying the biology, chemistry and physics taught in high school are quite reliable, since introductory classes tend to present long-established science rather than the latest findings. But students need to be aware of the unfortunate fact that many results published in scientific journals cannot be replicated, and that the problem appears to be growing, especially in areas of science that are more susceptible to ideological influences, such as psychology and environmental science. Students–and science teachers!–need to assess the reliability of sources and build the habit of critical judgment.
Laws are generalizations from observations, often expressed mathematically. It is a marvelous and beautiful thing that so many patterns are discernible in nature, that the universe appears orderly and intelligible, and that complex phenomena can often be described with strikingly simple equations.
In the early years of modern science, scientists such as Newton, Lavoisier and Mendel were full of confidence that the patterns they discovered were universal laws, and that the establishment of natural science on such sure and certain foundations would result in a new era of more rational living. The burgeoning of scientific knowledge between the mid-17th c. and our own day shows how very successful their approach was.
Eventually, however, it became evident that the validity of various scientific “laws” was circumscribed to particular conditions. Newton’s Laws needed to be modified by the theories of Special and General Relativity; Lavoisier’s law of conservation of mass turned out to work for chemical reactions not nuclear reactions; etc.
The basic issue here is called the Problem of Induction. The logical process of induction involves identifying a general principle from a number of particular cases. If scientists have only ever observed white swans, it cannot be said with utter certainty that all swans are white; there may well be black swans that have never been observed. This is not to say that inductions are invalid, only that they are always subject to modification if new evidence appears.
In the 20th century, it became less common for scientists to call their generalizations laws, since so many scientific laws had turned out to need modifications. Nowadays, the better sort of scientific article expresses a scientist’s results cautiously and modestly. This epistemic humility has not, however, percolated through every science-adjacent field. Schoolteachers and journalists often present scientific knowledge as though it were all as certain as geometry theorems; this is inaccurate and does science a disservice.
Theories are coherent systems of principles that explain the causes of observed phenomena. When an explanation is first proposed, it might be referred to as a hypothesis, such as Avogadro’s hypothesis that the properties of gases depended on the number not the identity of the particles in a sample. As hypotheses receive experimental confirmation–or at least survive experimental attempts to falsify them–and are connected to other hypotheses, the set of related ideas eventually coheres into a theory. Examples are the atomic theory, the kinetic-molecular theory, quantum mechanics, evolutionary theory, etc.
Students need to be reminded more than once that a theory in science is not a mere opinion, as the word is used in common parlance. Instead a scientific theory is the result of reasoning about evidence. As such, theories are not dubious in the way opinions are.
Nor are theories facts in the way observations are. Reported observations tend to be true or false, whereas theories require more nuance to evaluate. A theory can be proven incorrect by the discovery of contrary evidence or mistakes in reasoning, but it cannot be proven true in any absolute sense, since it is a construction of human reasoning that rests ultimately on induction. We say rather that a scientific theory is supported by significant evidence or is speculative, that it has broad explanatory power or is narrow in scope, that it is compelling or persuasive or creative or unconvincing or improbable.
The degree of certainty that we have about any particular theory depends on the strength of the evidence and arguments for it. We are highly certain of atomic theory, for instance, but not of string theory. And we are more certain of some parts of theories than others; for instance, in evolutionary theory, we are more sure about the operation of natural selection on genetic variations than we are about how variations come about in the first place.
For an analysis of models as a type of scientific knowledge, please see my article on Teaching Scientific Modeling. Briefly, I will point out here that models are simplified representations of an entity or process. They might be very simple, such as a ball-and-stick model of a molecule, or highly complex, such as models used in meteorology or economics to predict the weather or stock market behavior. By reducing a system to its most important features, modeling makes it possible to wrap our minds around complex phenomena. Models are extraordinarily useful, but students need to grasp that they are not reality itself; as simplifications, they can only ever be approximately true.
Teaching and Thinking in All Four Modes
Teaching science with accurate epistemic standards is important so that students realize that science isn’t just a body of knowledge to learn; it is also and fundamentally a method of reasoning from evidence. Most of all, we want them to discover that they are capable of evaluating the strength of the evidence and the arguments for themselves, and eventually to participate in the process of doing science.
For this to happen, the teacher needs first to understand the various types of scientific knowledge and the level of certainty they afford. I hope that this article has been of service in this regard. Second, the teacher needs to be willing to make the effort to give nuanced explanations. Students need to be challenged to think about the tentative aspects of real science rather than just settle for easy simplifications. To this end, I start out my introductory science classes by going over the material which I have presented in this article. Then I reinforce these ideas throughout the course.
I hope that my students will thoroughly absorb the idea that science, objective as it is, will never be settled. There are always more discoveries to be made and better theories and models to be created–and they can be the ones to do it!