The Scientific Method

Introduction to the Philosophy of Science

Science is often regarded as the most reliable way to gain knowledge about the physical world. But what, exactly, makes a “science” scientific? How does it differ from “pseudoscience” or mere speculation?

The Philosophy of Science is not about doing science, but about analyzing how science works. It examines the assumptions scientists make (like the idea that nature is uniform) and the logic they use to draw conclusions from data.

The Logic of Induction

Traditionally, science was thought to be based on induction. Induction is the process of moving from specific observations to general laws.

Observation 1: This swan is white.
Observation 2: That swan is white.
Conclusion: All swans are white.

However, the 18th-century philosopher David Hume famously pointed out a problem with induction. No matter how many white swans you see, you can never be logically certain that the next one won’t be black. We assume that the future will resemble the past, but Hume argued that this assumption cannot be proven by logic or experience—it is merely a “custom” or habit of the mind. This is known as the Problem of Induction.

Karl Popper and Falsificationism

In the 20th century, Karl Popper proposed a radical solution to the problem of induction. He argued that science doesn’t actually use induction at all. Instead, it uses Deduction and Falsification.

Popper suggested that the “Criterion of Demarcation”—the thing that separates science from non-science—is falsifiability. For a theory to be scientific, it must make specific predictions that could, in principle, be proven wrong.

Scientific: “Gravity causes objects to fall at 9.8 m/ $s^{2}$ .” (If an object fell at a different rate, the theory is debunked).
Non-Scientific (according to Popper): Astrology or certain versions of Psychoanalysis. If a theory can be “explained away” no matter what the evidence shows, it isn’t science.

For Popper, we can never “prove” a theory is true; we can only “corroborate” it by failing to prove it false.

The Hypothetico-Deductive Model

Modern science typically follows the “H-D Model”:

Formulate a Hypothesis: A tentative explanation for a phenomenon.
Deductively derive a prediction: “If my hypothesis is true, then if I do X, Y should happen.”
Test the prediction: Conduct an experiment or observation.
Evaluate: If Y happens, the hypothesis is supported (but not proven). If Y does not happen, the hypothesis is falsified (or perhaps the experiment was flawed).

Underdetermination: The Duhem-Quine Thesis

A major challenge to the simple “falsification” model is the idea of Underdetermination. Pierre Duhem and W.V.O. Quine argued that we never test a single hypothesis in isolation. We are always testing a “web of beliefs.”

If an experiment fails, it could be because:

The primary hypothesis is wrong.
The auxiliary assumptions are wrong (e.g., the telescope was slightly out of focus).
The measurement tools were miscalibrated.
The laws of physics changed (unlikely, but possible).

Because we can always “blame” an auxiliary assumption rather than the theory itself, evidence “underdetermines” the theory. We can often hang onto a theory in the face of conflicting evidence by making small adjustments elsewhere in our belief system.

Realism vs. Anti-Realism

One of the deepest debates in the field is between:

Scientific Realism: The view that the goal of science is to provide a true description of the world. Realists believe that atoms, electrons, and DNA actually exist exactly as science describes them.
Scientific Anti-Realism (Instrumentalism): The view that science is just a “tool” (instrument) for making predictions. Anti-realists argue that we shouldn’t necessarily believe that unobservable entities (like quirks or fields) “exist” in the literal sense—we should only care that the theories work to solve problems.

Conclusion

The scientific method is more than just a list of steps in a textbook. It is a complex logical structure that wrestles with the uncertainty of induction, the difficulty of falsification, and the philosophical question of what “truth” really means in a physical world. As we see in the next lesson, how these methods play out in history is often much messier than the “ideal” model suggests.