An excerpt from Embracing the Unknown: A Fresh Look at Nature and Science
Modern science has its foundation in the European Enlightenment, starting around four hundred to five hundred years ago. As part of the general drive toward intellectualism and reason, scientists challenged previously held doctrines, especially those that did not fit well with their experimental observations. In addition, fuel for the explosion of scientific knowledge came from development of a systematic approach to discovering truth about nature through the method described below.
Scientific terminology: Law, theory, model, and hypothesis
Words in science have specific meanings. Let’s first define basic terms that are used to describe results of scientific inquiry. This is a top-down list, flowing down in the degree of certainty.
A law in science is a generalization of a body of observations into a concise mathematical or other expression. At the time that it is deemed a law, no exceptions will have been observed. Because of this stringent requirement, there are very few laws in science. An example is Newton’s law of motion, which seemed ironclad for two hundred years. Einstein demonstrated that it is not true under all conditions, so it was disproven.
A theory is a concept that summarizes how nature works based on a set of observations. A theory generally starts out as a set of educated guesses. As scientists test the theory with experiments, and the supporting data are amassed, the theory gains strength. Alternatively, a theory can be disproven experimentally. An example of a famous theory is Stephen Jay Gould’s theory of punctuated equilibrium, a theory that explains data from fossil records relating to evolution of species, as described in chapter 7.
A model is a way to simplify complex concepts. These are generally derived from representative data, not all available data, and may take many forms and include mathematical equations, flowcharts, or physical models. For example, a physical model of the solar system captures the general concepts of planetary movement, but it is not precise.
A hypothesis is an educated guess. Very commonly science is driven by a hypothesis regarding some known phenomenon or unexplained experimental data. Generally, experiments are designed to either support or refute a hypothesis. For example, I observed that a butterfly can flap its wings and fly. My first hypothesis was that the butterfly’s wings push the air downward and that that is the main reason that it stays aloft. That hypothesis was then debunked by data showing that the major force keeping the butterfly in the air comes from negative pressure from above (due to the vortex caused by flapping its wings).
The scientific method: A cycle of refinement
The scientific method progresses by an ongoing cycle of experimental observation, development of hypotheses to explain the experimental data, and then further experimentation to prove or refute the assumptions of the hypotheses. In other words, good scientists develop hypotheses to explain their experimental data in a way that can be tested with further experiments so that their hypotheses can be either confirmed or refuted.
Asking good questions
The most important step in discovery is to ask good questions. This is true at whatever level of knowledge and to whatever depth of exploration you are interested in. While it is hard to strictly define what a good question is, you can usually recognize it when you see it. It is the type of question that builds upon your current knowledge to fill a gap in understanding. In some cases, the questions will provoke exploration into terrain that has yet to be explored. Prior knowledge is a key ingredient to driving good scientific questions.
Building on existing knowledge
The most common questions are directed to fill in chasms of understanding. Here’s a simple example. In the butterfly story, I did not understand the function of the wings. So I formulated a direct question based on the observation of how butterfly wings are shaped—what function does the shape of the butterfly’s wings serve? Armed with that question, my exploration consisted of searching available information sources on the Internet, and I discovered a simple answer to my question—when the wing is flapped, the wing’s shape facilitates formation of a vortex in the air above the wings, and the negative pressure from the vortex pulls the wing upward. Now, if I want to go to the next level of knowledge, I can ask better questions. For example, I may now be interested in how a vortex generates negative pressure.
Scientists often formulate questions to complete missing pieces with existing knowledge. Through thorough orientation to prior knowledge and a sharp analytical approach, they can decide which scientific facts are still uncertain and require more proof. Or the facts may have multiple possible explanations, and a good question may be one that could help decide which is the best explanation.
The hypothesis is an educated guess that is formulated to attempt explanation of an observed phenomenon or a proposed connection between different observations. There are many ways to generate hypotheses, and they usually include reasoning based on prior knowledge. Using an example from medical science, I could formulate a hypothesis that a chemical shown to lower blood pressure in mice would be effective in lowering blood pressure in humans.
Experiments are conducted to evaluate the truth of a hypothesis. They involve careful design with adequate measurement and control of experimental conditions. Carrying forward the example from above, a good experiment could be to test the chemical that is thought to lower blood pressure in a sample of people with high blood pressure. The experiment, known in medical research as a clinical trial, would involve two groups of patients—one receiving the active drug and the other receiving an inert substance that looks, feels, and smells like the active drug. Since blood pressure may be affected by many factors not controlled in the experiment, the patients would be divided into the groups in a randomized fashion. That way, chance variations in the uncontrolled variables would likely be distributed equally between the two groups. The outcome of the experiment would be a measurement of the change in blood pressure between the two groups.
Data from experiments are usually numerical measurements from some measurement device. This is not always the case, since sometimes observations are based on human experience that is acquired using the five senses. For numerical data, analysis is done using appropriate mathematical techniques, including statistical comparisons. For the above example, the outcome of the study would be determined following careful statistical testing to see if the difference in blood-pressure changes between the two experimental groups could be explained by chance alone.
How exact are scientific models?
It would be wonderful to study nature in its entirety and to measure it with exact precision. Unfortunately, it is not possible to do so. We simply cannot measure everything or create models that include all possibilities. Experiments produce data that are limited in two important ways: they are only as precise as the measurement tools, and they are based only on a small sample of possible conditions. So the measurements are approximations of reality. Mathematical models and other modeling techniques attempt to capture the essence of the data in a concise manner but only approximate the entirety of the data set. The models do not account for all the variability in the data. Approximation is done routinely in science to help make sense of a lot of information, so it serves an important function. However, the approximation is never as good as the real thing.
The distance from the earth to the sun of ninety-three million miles is an approximation. The actual distance varies over time, and there is a limit to how precisely we can measure it.
The number of neurons in the brain is one hundred billion, but that is just an approximation. We cannot measure the actual number, and there is significant variability between individuals.
For the purposes of simplifying calculations, physicists approximate the shape of the earth as a sphere, even though it is not perfectly spherical.
Truth is ever to be found in simplicity, and not in the multiplicity and confusion of things.
—Sir Isaac Newton (1)
As Newton stated in the above quote, scientists adopted reductionism as a preferred approach, firmly believing that the universe can be understood by a set of simple laws and principles. Newton’s laws of motion, Einstein’s law of conservation of mass and energy, and Maxwell’s equations of electromagnetic radiation are just a few examples in physics. Biologists have also embraced reductionism as the preferred approach to understanding life. Darwin’s theory of natural selection, genomics, and basic laws of cellular processes are examples from biology. Reductionism assumes that simple rules relating to components are powerful enough to explain the entire system. This idea carries with it a strong tendency toward causal thinking—the behavior of the whole can be completely predicted by the behavior of its parts.
Reductionism may work for simple systems, but it breaks down for complex systems. Emergence, as we will see later, is a common natural phenomenon whereby the whole is greater than the sum of its parts. For example, the function of the brain cannot be explained directly by knowledge of its component cells. Many natural systems are too complex to be properly understood when reduced. This poses serious challenges, since most scientists desire to provide the simplest models possible to capture the essence of the natural phenomenon, even if those models are not complete. Scientists are now developing broader approaches that include more details of the physical system in their models.
1. Frank E. Manuel, The Religion of Isaac Newton (Oxford University Press, Oxford, UK 1974).