Introduction
You want to know how scientific knowledge grows and how to separate genuine science from metaphysical speculation. Traditional epistemology tells you that science builds on inductive logic, accumulating confirmed observations from the bottom up. Karl Popper shatters this view by showing that induction is a logical myth. You can never verify a universal law by observing finite instances, no matter how numerous they are.
The book replaces verification with falsifiability as the ultimate criterion of demarcation. You learn that scientific progress does not proceed through cautious induction, but through bold conjectures tested by severe deductive attempts at refutation.
The following chapters examine the empirical basis, testability, simplicity, probability, quantum physics, and the ultimate corroboration of theories.
- How can basic statements be accepted without falling into infinite regress?
- What determines the precise degree of a theory's falsifiability?
- How do probability and quantum mechanics fit into deductive testing?
1. The Problem of the Empirical Basis
Empirical sciences cannot rest on immediate sense-perceptions or subjective experiences. You must recognize that every description uses universal names and has the character of a theory.
Empirical sciences cannot be reduced to immediate sense-perceptions or psychological experiences without falling into dogmatism or an infinite regress.
Karl Popper rejects the traditional view that scientific statements can be justified by perceptual experiences. Feelings of personal conviction interest the psychologist but do not touch logical connections.
Rejection of Psychologism
Objective science demands statements that are inter-subjectively testable. You cannot build a rigorous system on subjective feelings of certainty.
Otto Neurath and Rudolf Carnap advanced the debate by arguing protocol sentences are not inviolable. Yet they provided no rules for their deletion or acceptance. Without rules, any arbitrary system can be defended as an empirical science, effectively throwing empiricism overboard.
Protocol Sentences and Fries's Trilemma
Jakob Friedrich Fries exposed a trilemma that traps anyone seeking ultimate justifications. You face three unacceptable paths:
- dogmatism;
- an infinite regress;
- psychologism.
You escape this trilemma by recognizing that basic statements are accepted by convention. Testing must stop at some basic statement where investigators reach an agreement. You accept these statements temporarily as dogmas because they can always be tested further.
Basic statements function as singular existential statements that can contradict universal laws. They must refer to observable events involving macroscopic physical bodies.
Theory and Experiment
The naive empiricist view that science begins by collecting experiences or protocol sentences is false. A science needs points of view and theoretical problems. The theoretician must formulate questions sharply before the experimenter can test them.
Louis de Broglie predicted the wave-character of matter, which Davisson and Germer later confirmed. Other decisive tests include the Michelson-Morley experiment and the falsification of the radiation formulae of Rayleigh and Jeans and of Wien by Lummer and Pringsheim.
The acceptance of basic statements in science is analogous to the verdict of a jury in a trial. Scientific investigators reach an agreement on basic statements to apply a theoretical system. The verdict does not prove truth and can be revised when new tests demand it.
The empirical basis of objective science has nothing absolute about it. Science does not rest on solid bedrock. It is built on piles driven into a swamp.
You stop driving piles deeper not because you reached a natural base, but because you are satisfied they are firm enough to carry the structure for the time being.
2. Degrees of Testability
Theories possess degrees of testability or falsifiability depending on the size of the class of their potential falsifiers. A theory is more testable if it forbids a larger class of basic statements, thereby saying more about the world and providing more opportunities for refutation.
Theories possess degrees of falsifiability and empirical content determined by the size of their class of potential falsifiers and their dimension.
Potential Falsifiers
You can compare the degrees of falsifiability of different theories through the subclass relation between classes of potential falsifiers. If the class of potential falsifiers of theory x includes the class of potential falsifiers of theory y as a proper subclass, then x is falsifiable in a higher degree than y.
Empirical content is defined as the class of potential falsifiers, whereas logical content is the class of all derivable non-tautological consequences. The logical probability of a statement is inversely related to its degree of falsifiability and empirical content. The less testable a statement is, the more logically probable it becomes because it allows a wider range of possibilities and forbids fewer events.
Subclass Relations
Higher levels of universality and precision increase a theory's empirical content and thus its degree of testability. A theory with greater universality or precision rules out a broader or more specific class of events, making it more easily falsified than a narrower or vaguer theory.
Comparing planetary orbital laws: the hypothesis that all heavenly bodies move in circles is more universal and more easily falsified than the hypothesis that only planetary orbits are circles.
The range of permitted events is converse to empirical content, and operational measurement requires accounting for intervals of precision. Every measurement yields an interval rather than a strict point, requiring condensation bounds of the interval of precision for the statistical theory of errors to apply.
Millikan measured the elementary charge of the electron as e = 4.774 * 10^-10 with a range of imprecision of ± 0.005 * 10^-10.
Dimensions of Theories
The dimension of a theory, determined by its characteristic number relative to a field of application, dictates its degree of testability. The dimension corresponds to the minimum degree of composition a basic statement must have to falsify the theory, which equals the number of freely determinable parameters in its curves.
Kepler started with a circle-hypothesis, and its unequivocal negative result provided the first real success that allowed him to proceed further. The circle hypothesis is three-dimensional and requires at least 4 singular statements for falsification, whereas the ellipse hypothesis is five-dimensional and requires at least 6 singular statements for falsification.
The number of dimensions of a set of curves can be reduced either materially through specific point stipulations or formally by restricting the shape. Material reduction introduces individual names and coordinates, while formal reduction introduces universal names and preserves the general invariance of the equation definition under transformation groups.
3. Simplicity and Falsifiability
When people speak of an exposition or mathematical proof being simple or elegant, they express an extra-logical preference. Such uses relate to ease of understanding, beauty, or required training rather than formal logic.
Epistemological simplicity is fully resolved by equating it with a high degree of falsifiability and a paucity of freely adjustable parameters.
Exclusion of Aesthetic Simplicity
You must exclude aesthetic and pragmatic concepts of simplicity from the theory of knowledge. These notions focus on beauty or economy of thought rather than logical testability.
Paucity of Parameters
Hermann Weyl, along with Harold Jeffreys and Bertha Wrinch, noted that classes of functions used for scientific laws should be offered by mathematics based on their mathematical simplicity. This simplicity relates directly to having fewer parameters than the observations to be satisfied.
A simpler theory has a greater empirical content and imposes a higher degree of strictness. You prize simpler statements more highly because they tell you more and expose themselves to potential refutation more effectively.
Conventionalism and Complexity
Euclidean geometry is simpler than non-Euclidean geometry because it is falsifiable to a higher degree. When tested via physical objects like light rays, the Euclidean hypothesis permits direct falsification through measurements such as the sum of triangle angles adding up to 180 degrees. Non-Euclidean alternatives require additional units of measurement and tolerate greater variations, making them more difficult to falsify.
Conventionalists treat theories as conventions and choose the simplest based on arbitrary rules. By rescuing endangered systems through auxiliary hypotheses, they reduce the system's falsifiability to zero, which runs contrary to proper methodological rules. Systems protected in this manner are complex in the highest degree.
4. Probability and Physics
Subjective theories define probability as a measure of feelings of certainty or as a logical proximity of statements where content determines likelihood. You must reject these interpretations because they treat probability statements as non-empirical tautologies, which contradicts the actual application of probability in modern physics. Keynes treats probability as a degree of rational belief, whereas Laplacean theory defines probability via favourable versus equally possible cases.
Probability statements must be interpreted objectively as relative frequencies in n-free sequences and treated via methodological rules rather than metaphysics.
Subjective Versus Objective Probability
Richard von Mises defined probability as the limit of relative frequencies within a collective, governed by an axiom of convergence and an axiom of randomness. A collective is a sequence of repetitive events that can be continued indefinitely. Von Mises used the axiom of convergence to ensure a fixed frequency limit and the axiom of randomness, also known as the principle of the excluded gambling system, to guarantee that no systematic gambling method can alter the limiting frequency.
N-Freedom and Randomness
Popper rejects von Mises's convergence axiom and weakens the randomness axiom. Instead of excluding all gambling systems, you demand absolute freedom, which means the sequence is n-free for every n, remaining insensitive to ordinal and neighbourhood selections based on any number of predecessors. Generating periods like (A) for 1-free, (B) for 2-free, (C) for 3-free, and (D) for 4-free alternatives demonstrate how sequences with increasing n-freedom can be constructed mathematically. The minimum size of the generating period of an n-free alternative with equal distribution is 2n + 1 elements.
Probability in Physics
By deriving the binomial formula for sequences of adjoining segments of a random sequence, Bernoulli's theorem shows that as segment length increases toward infinity, the probability of encountering a fair sample approaches 1. Short segments show large fluctuations, while long segments exhibit quasi-convergent behaviour and statistical stability. The maximal value approached by the probability of segments having the property delta-p as n increases without limit is 1.
No finite number of observations or basic statements can contradict a probability statement because even the most improbable finite segment has a probability greater than zero of occurring. Therefore, probability statements are strictly existential hypotheses and empirically uninformative on their own.
To apply probability statements empirically, physicists make a methodological decision never to explain reproducible regularities as mere accumulations of accidents. Speculating that the law of gravity is merely the result of an accidental accumulation of particle movements over an immense cosmic period represents an unscientific, metaphysical use of probability.
Formally singular probability statements assign a probability to a single occurrence by defining it through its membership in a reference sequence. A statement like the probability of throwing five with the next throw is 1/6 becomes meaningful when linked to the reference class of throws with true dice.
5. Quantum Theory and Scattering
Heisenberg's uncertainty relations describe objective statistical scatter relations among mass aggregates rather than absolute limits on human knowledge.
Uncertainty Reinterpreted
You can separate mathematical Heisenberg formulae from Heisenberg's own subjective interpretations. Born's statistical framework treats a wave-packet describing a single particle as a formally singular probability statement. The derived formulas describe random scattering ranges among aggregates of particles rather than absolute boundaries on human knowledge or physical measurement.
Heisenberg attempted to eliminate unobservable magnitudes like electron orbits from atomic theory. However, the mathematical formalism of quantum mechanics allows precise calculations of past paths and position-cum-momentum coordinates during periods when observation is impossible.
Schlick and other physicists argued that if a position statement is not verifiable, no sense can be attributed to it, making it impossible to speak of an electron's path between observed points.
Statistical Interpretation
Born resolved the equivalence between particle theory and wave theory by showing that Schrödinger's wave equation determines the probability of finding a particle based on the square of the wave's amplitude. Experimental atomic physics deals exclusively with mass-phenomena and statistical questions.
You must maintain a clear distinction between physical selection, which screens particles, and measurement, which records them without selection. Physical selection screens out particles from a stream using apertures to isolate a range, altering the system and increasing scatter in complementary properties.
A Geiger counter records electrons arriving at a certain position, measuring spatial intervals through time without performing a physical selection of particles according to their positions.
Causality and Metaphysics
Precise non-predictive measurements and calculations of past paths are indispensable because you need them to test statistical frequency predictions. Without them, the statistical scatter relations would remain untestable and metaphysical.
Interpreting Schrödinger's waves as subjective waves of human knowledge imperils the objective character of physics and invites subjectivist probability theories.
Jeans argued that the particle picture shows indeterminate knowledge while the wave picture shows an indeterminate electron, concluding that the wave picture represents solely our knowledge of the universe.
When a single photon falls on a semi-translucent mirror and is observed to be reflected, the formal probability appears to jump from one-half to zero or one.
A semi-translucent mirror reflects part of the light and lets part through, where a photon's formally singular probability of passing is equal to its probability of being reflected.
The belief in strict causality is a metaphysical hypostatization of the scientific methodological rule never to abandon the search for laws. Neither determinism nor indeterminism can be proven from the logical situation of physics.
6. Corroboration and the Path of Science
Theories are corroborated by surviving severe tests, driving a deductive evolution of science through bold conjectures and refutations.
Corroboration Versus Probability
You cannot verify universal scientific hypotheses or make them probable through induction. Attempts by thinkers like Reichenbach to reduce the probability of a hypothesis to a truth-frequency of statements or events fail. Treating a hypothesis as a sequence of basic statements leads either to a probability of zero, a probability of one, or makes the probability hopelessly subjective based on the skill of the experimenter. Any attempt to validate inductive logic through a principle of induction leads directly to an infinite regress or apriorism.
Soothsayers and palmists offer cautious, imprecise predictions with extremely high logical probability, which is why you do not accord them positive corroboration. Contrary to inductivist views, a theory's corroborability stands in inverse ratio to its logical probability. Logical concepts like truth are timeless, whereas corroboration values require temporal subscripts because they relate to basic statements accepted up to a specific point in time.
The Positive Theory of Corroboration
Counting the number of corroborating instances is insufficient because a trivial hypothesis can accrue many instances. The true degree of corroboration is determined by how severely a theory can be tested, meaning simpler theories with higher falsifiability are corroborable in a higher degree.
Comparing the hypothesis that all crows are black with Millikan's electronic charge value shows that Millikan's hypothesis is better corroborated despite encountering fewer instances.
You must design the severest possible attempts to refute the hypothesis. Record the results of these refutation attempts as the empirical evidence report.
Science as Conjecture and Refutation
The historical advance of physics toward higher levels of universality is a deductive, quasi-inductive process driven by the refutation of theories through crucial experiments. Historical metaphysical concepts like atomism, the corpuscular theory of light, and the fluid-theory of electricity eventually made contact with testable science when presented in falsifiable forms.
Science is not a system of established knowledge or probabilities, but a reckless quest where bold anticipations are systematically subjected to refutation.
You should perform these methodological actions:
- Formulate bold ideas, speculative thoughts, and anticipations as means for interpreting nature.
- Expose ideas to the hazard of refutation through systematic, planned actions and experiments.
- Attempt to prove anticipations false using logical, mathematical, and technical armor in order to test them rigorously.
10 Key Ideas
- Karl Popper shatters traditional epistemology by showing that induction is a logical myth and replacing verification with falsifiability as the ultimate criterion of demarcation.
- Empirical sciences cannot be reduced to immediate sense-perceptions or psychological experiences without falling into dogmatism or an infinite regress.
- The naive empiricist view that science begins by collecting experiences is false because a science needs theoretical problems and points of view before an experimenter can test them.
- Theories possess degrees of testability and empirical content determined by the size of their class of potential falsifiers and their dimension.
- Epistemological simplicity is fully resolved by equating it with a high degree of falsifiability and a paucity of freely adjustable parameters.
- Probability statements must be interpreted objectively as relative frequencies in n-free sequences and treated via methodological rules rather than metaphysics.
- Heisenberg's uncertainty relations describe objective statistical scatter relations among mass aggregates rather than absolute limits on human knowledge.
- The belief in strict causality is a metaphysical hypostatization of the scientific methodological rule never to abandon the search for laws.
- The true degree of corroboration is determined by how severely a theory can be tested, meaning simpler theories with higher falsifiability are corroborable in a higher degree.
- Science is not a system of established knowledge or probabilities, but a reckless quest where bold anticipations are systematically subjected to refutation.