It is widely believed that the methods of science somehow “prove” scientific facts. You hear reference to “scientific proof” in the news media, and even scientists will use the word “proof” informally to describe a very strong set of evidence.
Some theories, such as the theory of gravity,
F = M * m * G / x^2
(Force equals mass of the object, times the mass of the second object, times the gravitational constant, divided by the square of the distance)
are so reliable that we say that this theory “explains” the fact, as if this theory were an inherent property of nature, and not just our assumptions about it. Thus, chalk falls to the ground because there is this equation written on the board. In ordinary language, our everyday experiences (observations, facts, phenomena) of the external world are the logical consequence of arguments whose premises are NATURAL LAW. What is the difference between a theory and a NATURAL LAW? A natural law is theory with so few exceptions that we’re willing to give it complete acceptance. If I drop my pen and it falls and I ask you why, you say it’s because of the law of gravity, as if the theory exists prior to and independently of the fact. We actually believe that the theory of gravity is part of the intrinsic structure of the universe, and so it follows that in this particular case, I drop the pen, and, it falls. However, if you look more closely at how scientific reasoning works, it becomes clear that even the “laws” of nature are never proven.
In fact, logically, the proof goes in the other direction. That is, hypotheses are predictions about nature that are logically derived from theories. This kind of inference is known as deduction. A deductive inference is one in which the conclusion about particulars follows necessarily from general or universal premises.
For instance, a famous deductive inference goes: “Socrates is a man. All men are mortal. Therefore, Socrates is a mortal.” In this argument, “All men are mortal” serves the same role as a scientific theory. “Socrates is a mortal” serves the same role as a hypothesis would. The state of Athens tested this hypothesis by sentencing Socrates to death.
Deductive inferences are well-understood in philosophy. Deductive inferences are either valid or invalid. They succeed entirely or fail completely, and there are no shades of grey in between.
So, as a generalization about nature, theories serve as premises or assumptions from which specific facts (observations, phenomena, hypotheses or, to use really fancy words, empirical test implications) can be deduced as a conclusion.
An argument consists of one or more premises, an inference and a conclusion.
Premises are themselves are either assumed to be true, or else they must be the conclusion some other argument which requires its own premises and inference. Ultimately, then, every method of inference (argument, justification, explanation) requires some premises which themselves remain unproven. This is a consequence not of anything particular about science but of logic alone. If you’ve heard of Godel’s incompleteness theorem of mathematics, this is what it boils down to.
So, as I said before, logic is never a source of truth. Valid logical inferences are not truth-creating, but merely truth-conserving, transferring the assumed truth of the premises to the conclusions.
When scientific creationists claim that evolution by natural selection is “just an unproven theory,” they are right in the trivial sense that no scientific theories are ever proven. Scientific theories are established by a different kind of inference called induction.
An inductive inference is one in which a generalized conclusion follows from the truth of particular premises. Thus, Newton’s Universal Law of Gravitation was derived by observing numerous objects falling with a given velocity with respect to time, mass, etc. Inductive inferences are either strong or weak, and there is a lot of grey area in between that is NOT well-understood in philosophy. In fact, all inductive inferences are INVALID because it is not logically necessary (nor empirically certain) that all members of any class can be characterized by observing only some of them. The fact that scientific progress requires this invalid inference– that only observation and experiment may decide upon the acceptance or rejection of scientific laws and theories– is called the PROBLEM OF INDUCTION in philosophy. David Hume described the problem of induction by saying, just because the sun has risen every morning for as long as I can remember, is no guarantee that it will rise tomorrow.
Let’s review how scientific inference works. By the “scientific method” as commonly taught in high schools, science begins with some practical problem or disturbing fact, from which we INDUCE or generalize a theory.
From that theory, we DEDUCE specific hypotheses or empirical test implications, about situations which are members of the same class, but have not yet been observed. If these test implications are true, then, by induction, the theory is stronger. If an empirical test implication is false, the observation is called an ANOMALY. If you have anomalous observations, you can either reject the theory, or you can qualify it with AUXILIARY ASSUMPTIONS. Auxiliary assumptions explain why the theory is true, but only appears to be false under certain conditions. For instance, an auxiliary assumption about wind resistance has to be added to Newton’s law of gravity to explain why feathers fall more slowly than rocks. If a hypothesis deduced from the auxiliary assumption is true, then, by induction, the theory grows stronger.
One factor contributing to the strength of an inductive inference or theory is its POWER. The power, or scope, of a theory is the number of independent empirical implications that are confirmed by observation. Another desirable property of a theory is known as PARSIMONY, or simplicity. That is, a theory or inductive inference is stronger if it makes fewer independent assumptions. William of Ockham summarized this principle in the fourteenth century in a famous dictum known as OCKHAM’S RAZOR: “Pluralitas non est ponenda sine neccesitate”, which translates as “entities should not be multiplied without necessity.” The anthropologist Marvin Harris summarized power and parsimony by saying, “The theory that explains the largest number of facts while making the smallest number of independent unverified assumptions will be given priority.”
Power and parsimony are important concepts in the philosophy of science, but let me return to the main point, which is that scientific inference actually works in the backwards direction to how we commonly conceptualize scientific explanation. That is, the theories that are put forward as explanations for empirical observations are, in fact, the conclusions of inductive inferences for which empirical observations are the premises. Understanding how and why induction is problematic forces us to admit that all scientific theories are inventions rather than discoveries, which is an essential step in thinking critically about science.
The fact that all scientific theories (indeed, all logical arguments) necessarily include assumptions that are irreducibly irrational creates a never-ending opportunity for scientific progress.
In The Structure of Scientific Revolutions, Thomas Kuhn described some of the irrational forces at work in the history of science. Sometimes a theory is so successful that it attracts a large enduring number of scientists from the field. Kuhn called these socially popular theories PARADIGMS. For example, behaviorism (the theory that only reinforcement contingencies matter; brain and mind don’t) was a paradigm that dominated psychology for much of the 20th century. A paradigm is not just a theory however. They are also sociological movements. Scientists invest many years of their lives, and institutions like UH invest a huge amount of money into particular scientific methods. Paradigms are often not just theories but a set of methods that are accepted as “scientifically sound.” If for some reason, the basic premises of a paradigm are called into question, human beings will resist accepting that all of their training and all of their equipment are obsolete. Scientists sometimes poke fun at religion by pointing out how Church officials refused to look through Galileo’s telescope. In fact, we are often guilty of doing the same thing to our colleagues.
Under the shadow of a strong paradigm, the only way a scientist can get funding and maintain a career in science is to propose experiments that confirm and extend the paradigm. So, rather than inventing new theories which are usually more risky, scientists will often follow in the footsteps of someone else’s outstanding achievement. This is what Thomas Kuhn called NORMAL SCIENCE. Normal science is the process by which large numbers of the members of a discipline systematically test, one after another, all of the empirical hypotheses that are entailed by a paradigm (e.g., filling in another element on the periodic table, or sequencing another gene). Normal science is contrasted with REVOLUTIONARY SCIENCE, which involves throwing out the old theory and explaining the same facts with a new theory that explains the same facts while making new predictions.
It is of interest the difference between normal and revolutionary science correspond roughly to the difference between deduction and induction. It also roughly to the difference between science and philosophy, in that scientists (even the revolutionary ones) spend most of their time doing observation (testing the hypotheses or conclusions deduced from theories) while philosophers spend most of their time examining fundamental assumptions (the premises from which hypotheses are derived.)
Kuhn argues that normal science has the advantage of focusing the attention of the scientific community on an important problem, but it has the disadvantage of shutting out numerous other opportunities for discovery. I think the lesson to be drawn from Kuhn’s analysis is that we need to look at our own work like anthropologists. That is, it is easy to see how people in distant times or places get fixated on arbitrary beliefs, but it’s more important to examine whether there are more powerful alternatives to our own beliefs.
There is another point of view in scientific philosophy called the CRITICAL SOCIAL SCIENCE perspective, which approaches the economy of science from an anthropological point of view. The key question to a critical social scientist is, “Who Benefits?” from the dominant paradigm. As a scientist who studies the efficacy of EEG biofeedback for psychological disorders, I take a critical social science view of how poorly funded research on biofeedback is compared to studies of prescription drugs. However, when we look at how little research is done on possible therapeutic applications of certain non-prescription drugs like marijuana and LSD compared to studies aimed at finding negative effects, many of us take a critical social science view of the influence of the “War on Drugs,” and the massive police-prison-industrial complex it maintains.
So, the fact that every argument requires premises that remain unproven-that every scientific theory involves assumptions that are only supported by the invalid inference known as induction-helps us to see the irrational, social, and political dimensions underlying the agenda of research. It may also help us to see how science, far from being a cold, detached, objective inquiry, requires ethical reflection and action at the most basic level.
As Heisenberg said, what we know about nature depends upon what we ask.
Leave a Reply