Philosophy of science
“”We should never forget that our reasoning process can go wrong. This is why you always want to check your careful reasoning for flaws. Make sure that no fallacious, instinctive reasoning slipped in. Try to find errors, even look for evidence that you're wrong. After all, how hard is it to confirm your beliefs if that's all you're trying to do? Anyone can find some evidence for anything! If you want to prove yourself right — try to prove yourself wrong. Because if you try to prove yourself wrong — and can't? Then it's a really good indication that you're right. Want to be even more assured? Invite others to prove you wrong. And remember — admit it when you are wrong.
|—David K. Johnson, The Big Questions of Philosophy
Falsifiability is a deductive standard applied to scientific theory — a working framework for explaining and predicting natural phenomena — to have its falsity demonstrated by overwhelming evidence through experiments or observations. The ability to evaluate theories against observations is essential to the scientific method, and as such, the falsifiability of theories is key to this and is the prime test for whether a proposition or theory can be described as scientific. Put simply, if there's no way to falsify a hypothesis, it's probably not much of a hypothesis.
Philosophy of science
Falsifiability is one of the bedrocks of science.
All scientific knowledge and theories are based on two things: observation and consistent logic. A theory is a logical explanation for observations. A good, scientific theory also proposes a set of new observations that could test a theory's power to explain. Once technology, time or funding catches up with the theory, these observations can be made, which can either support or invalidate the theory. This ability to be tested, and the potential for the theory to be invalidated by the experiment, is the essence of falsifiability.
It has been argued, most notably by Karl Popper, that the scientific method demands that a theory must at least in principle be falsifiable in order for it to be valid as science. This requirement was Popper's solution to the demarcation problem, or what is and what is not science. Popper's view is not widely acceptable in contemporary philosophy of science. However, that does not mean that falsifiability is not important. Falsifiability is a virtue in a scientific theory. Evolution, for example, is theoretically falsifiable — "fossil rabbits in the Precambrian", as J. B. S. Haldane once said — whereas intelligent design is not, mostly because it makes no predictions that can actually be tested.
To determine if an idea is scientific, we use a process:
- A theory is a well substantiated explanation for some aspect of nature; it is different from a hypothesis.
- A hypothesis or conjecture is an idea that a researcher believes may be true. The researcher can test this idea using the scientific method.
- Scientists constantly investigate even highly supported theories.
- If evidence is found that contradicts a theory, the theory must be discarded or revised.
Naive falsificationism and the Duhem-Quine Thesis
The Duhem-Quine Thesis is something of a conflation of the ideas of Pierre Duhem and W.V.O. Quine. In short, it states that it is impossible to test or falsify a hypothesis in isolation for two reasons. Firstly, the hypothesis relies on a number of supporting assumptions. For example, does the exposure of the Piltdown Man fraud falsify the validity of that fossil alone or all of evolutionary theory?
Secondly, a discrepancy between theory and data does not necessarily falsify the theory. For example, in the early 19th century, scientists discovered discrepancies between the orbit of Uranus as predicted by Newton's theory of gravity and the orbit which was actually observed. However, this discrepancy between theory and evidence was not considered a falsification of the theory; in time, the discrepancy was resolved through the discovery of Neptune. In the late 19th century, a similar discrepancy was discovered in the orbit of Mercury. This time, however, this discrepancy did lead to the falsification of Newtonian gravity; the discrepancy could only be resolved when Einstein proposed his amended theory of gravity. Scientists will typically not consider a theory as falsified simply because of the existence of discrepancies between theoretical predictions and observations, even if those discrepancies remain unexplained for a long time; in fact, attempting to account for such discrepancies is what motivates a lot of scientific research.
The Duhem-Quine Thesis is often contrasted with or considered to modify "naive" or "Popperian" falsificationism.
Lakatos and research programs
Imre Lakatos further extended Popperian falsification and the Duhem-Quine Thesis with his concept of "research programs." Lakatos defined extensively developed theories and techniques in a field as the "hard core" of a research program. Around this hard core is a "protective belt" of auxiliary hypothesis and modified or ad hoc assumptions added by scientists to protect the core of the program from falsification. While the protective belt is auxiliary and ad hoc, this does not necessarily translate to "bad" in Lakatos' view. If the protective belt leads to the discovery of new facts made by novel predictions, this leads to what Lakatos called a "progressive research program." A "degenerative research program" is one in which the protective belt grows but leads to no new discoveries, merely acting as a way to cover up the flaws of the hard core.
Until the twentieth century, Newton's laws of motion were:
- a) scientific and
- b) believed to be true.
Newton's laws allowed us to make specific predictions regarding such things as the trajectories of artillery projectiles or the orbits of planets. These predictions were observed to hold most of the time, but in some cases such as the orbit of Mercury, there were observations that contradicted predictions based on Newton's laws. This led to their modification and replacement by relativity, which, rather than being a complete rejection of Newton's laws, was a clarification and refinement that allowed them to hold true in a greater range of observable circumstances. It was the specific predictions made by Newton's laws that allowed scientists to test them, and eventually replace them. This happened because the predictions made Newton's laws falsifiable.
Is it falsifiable?
A simple procedure can sometimes be used to determine whether or not a hypothesis or conjecture is falsifiable. One could begin by asking someone to give them an observation that, if witnessed, would disprove their hypothesis. If this question cannot be answered, then the hypothesis is likely unfalsifiable and cannot reasonably be assumed to be correct without overwhelming evidence. In addition, a good test of a theory is that it can be used to make predictions about some future event. For example, Einstein's ideas about relativity predicted specific things that would be observed during a total solar eclipse. When the eclipse came, the predictions were confirmed, something which strongly supported his theory.
Logically, the two applications are the same thing. A theory useful in practice is not defined by what it allows, but by what it disallows, because that is where its predictive power lies. For example, Newton's theory of gravity says that the force of universal attraction cannot act in any other way except in accordance to the inverse square formula. If a theory is able to explain any conceivable observation (e.g. Goddidit), it cannot make any useful predictions. Conversely, if a theory says that an event cannot happen, then even one single observation of it would falsify the theory — and practical uses of the theory can rely on the fact that this event will never happen.
“”The invisible and the non-existent look very much alike.
|—Delos Banning McKown
A science stopper is a hypothesis that makes no testable or useful predictions and therefore prevents any science from being performed based on that hypothesis. It is usually viewed as a very bad thing and a consequence of poorly formulated ideas that fail to grasp concepts such as falsifiability and methodological naturalism.
Real science takes no heed to science stoppers. Peer-reviewed journals, universities, and institutions that fund scientific research pay no attention to pseudoscientific theories, much to the infuriation of proponents of young earth creationism and so-called 'intelligent design'. In fact, science is a highly conservative activity that has a very high bar for even taking notice of, let alone accepting new hypotheses and theories.
The falsifiability of creationism is crucial to determining whether or not it is science, because only falsifiable theories can be scientific. The concept of creationism is, for the most part, not falsifiable; The parts of it that are falsifiable have already been falsified countless times.
A hallmark of science is that a future experiment could always potentially prove an idea to be false. However, many creationist ideas are not falsifiable, even in principle.
Deductive argument for falsifiability
Scientific theories can usually be thought of as a series of statements and deductions which infer whether or not some observation will be made. This view is commonly called the hypothetico-deductive model. Now let us apply this model to the concept of falsifiability. Let's say some theory deduces that some observations will be made. This theory thus becomes verifiable, but not necessarily falsifiable. Now, let's analyze what happens in two cases. In the first case, if the observations deduced by the theory aren't made, the theory gets rejected. In the second case, the observations which are deduced by the theory aren't made, and yet, the theory isn't negated. The latter case fails the test of falsifiability. Not only that, this theory also violates the basic rules of logic — it proves both the theory and its negation. Thus, this theory is logically inconsistent. If we consider accepting logically inconsistent theories as futile, this argument makes a case for the criterion of falsifiability.
Let T be a formal theory which denotes a scientific conjecture in the context of the hypothetico-deductive model. Let O be a set of first-order propositions denoting observations which can be deduced from the formal theory T.
A scientific conjecture T is said to be consistent if and only if: .
Falsifiability can be expressed formally as the following condition:
Falsifiability is thus a necessary condition for a scientific theory being consistent.
- Disproving Evolution
- Disproving Intelligent Design
- Methodological naturalism
- Threats to scientific validity
- Not even wrong
- Science was wrong before
- Popper, Science, and Pseudoscience. Crash Course Philosophy.
- Chapter 2, 29:10 (Audiobook)
- Free dictionary - falsifiable
- Hansson, Sven Ove, "Science and Pseudo-Science", The Stanford Encyclopedia of Philosophy (Fall 2021 Edition), Edward N. Zalta (ed.)
- Gordin, Michael D. (2021) (in English). On the Fringe: Where Science Meets Pseudoscience. Oxford University Press. ISBN 9780197555767.
- A New Look at Falsification in Light of the Duhem-Quine Thesis, Andrew Lewthwaite
- Kuhn vs. Popper; the Philosophy of Lakatos, Antimatter blog
- Imre Lakatos. Falsification and the Methodology of Scientific Research Programmes, Part 3 in The Validation of Scientific Knowledge