Popper, K. R., & Popper, K. R. (1934). The Logic of scientific discovery . Routledge.
Karl Popper begins by agreeing with Hume-on-the-problem-of-induction that no amount of observed instances can logically justify a universal law, but he denies that this poses a threat to science, like logical-empiricism leads us to believe. According to Popper, the problem of induction only arises because it is mistakenly assumed that science proceeds by induction from observation to universal laws.
Popper posits that scientific theories are conjectures proposed independently of experience and are tested deductively through their observable consequences. Inductive justification is not required because no amount of positive evidence can verify a theory, it advances through falsification instead.
Karl Popper believed that scientific theories can only be falsified (proven to be false) and that scientific theories do not get inductively confirmed. This in contrast to logical-empiricism that focuses on confirmation-and-logic.
For Popper, science changes via a two-step cycle that repeats endlessly.
Stage 1: conjecture
- offer a hypothesis that might describe or explain some part of the world
- a good conjecture is bold that takes risks by making novel predictions Stage 2: attempted refutation
- hypothesis is subjected to critical testing in an attempt to show that it is false
- if a hypothesis is refuted, it is back to stage 1
A theoretical idea can be refined and modified via many rounds of conjecture and refutation. However, it is important to avoid ad hoc hypotheses
example: Hypothesis = Blue litmus paper turns red in clear liquids Auxiliary hypothesis = Water is a clear liquid Prediction = Blue litmus paper turns red in water Observation = Blue litmus paper does not turn red in water falsification
For Popper, an ad hoc hypothesis for this falsification = e.g.. = Blue litmus paper turns red in clear liquids not water
Section 1: Problem of induction
Whether inductive inferences can be justified is called the problem-of-induction, but can also be described as “the question of the truth of universal statements which are based on experience”, such as hypotheses and theoretical claims in science.
The problem with being “known by experience” is that it is clear that this can only be applied to singular statements and not universal statements. The truth of inductive inferences making generalizations means to reduce generalizations to singular statements that are then known by experience, which leads to the question of justification.
principle of induction
Hence, there is need for a principle of induction that would put inductive inferences into a logically acceptable form. According to Reichenbach (logical-empiricism) the principle of induction is necessary for the scientific method as it decides the truth or falsity of its theories (Part 1, section 1, p.4).
Reichenbach defines the principle of induction as
“the principle of induction as the means whereby science decides upon truth. To be more exact, we should say that it serves to decide upon probability. For it is not given to science to reach either truth or falsity … but scientific statements can only attain continuous degrees of probability whose unattainable upper and lower limits are truth and falsity” p. 29
Reichenbach’s principle of induction argues that it is not necessary for the justification of inductive inferences to show that the conclusion is true. For Reichenbach, it is not necessary for induction to always work, but inductive inferences justify universal laws to a degree of probability.
- An analogy Reichenbach makes is for a man suffering a disease, a physician can say “i do not know whether an operation will save the man, but if there is a remedy, it is an operation” (Reichenbach, 1938)1. This provides some justification even if one does not know whether the operation will succeed.
It’s important to note that Hume’s Principle of the uniformity of nature is a metaphysical assumption about how the world must be structured for induction to work, whereas Reichenbach’s principle of induction is a methodological rule governing how we should reason form observed cases to future predictions. → they are two attempts to the problem-of-induction.
For Popper, this principle of induction that applies a degree of probability requires its own justification, and that nothing is gained by claiming that the principle of induction is taken as “probable” as it leads to an infinite regress.
For Popper, inductive skepticism is no threat to the rationality of science (See falsification-as-a-criterion-of-demarcation to Popper’s solution to the problem of induction).
psychology of knowledge vs. logic of knowledge
Popper insists a strong distinction between psychology of knowledge and logic of knowledge =
- psychology of knowledge = empirical questions about human behavior and knowledge
- “the sun will rise tomorrow because it always has before”
- logic of knowledge = concerned with logical relations and whether certain inferences are valid and whether claims are rationally justified by evidence
- “what justifies the generalization that the sun will rise each day?”
Popper notes that philosophers mistakenly blur this distinction, especially when defending inductive inferences. The fact that humans psychologically tend to make inductive inferences does not provide logical justification for those inferences. Popper notes Hume-on-the-problem-of-induction showed that our confidence in induction is a product of habit but the psychological facts do not provide logical justification.
Section 2: Elimination of Psychologism
For Popper, the question of how a person gets a new idea is an interest to empirical psychology and is irrelevant to the logical analysis of scientific knowledge. Logical analysis of scientific knowledge is concerned with matters of justification or validity.
His view is that there is no logical method of having new ideas, and that there is a distinct difference between the process of conceiving an idea, and the methods and results of logically examining it.
Section 3: deductive testing of theories
Popper outlines the method of critically testing theories as followed =
- set forth a hypothesis (not yet justified in any way)
- conclusions are drawn by means of logical deduction
- the conclusions are compared with one another and other relevant statements, to find logical relations that might exist between them
This gives for lines along which the testing of a theory could be carried out =
- logical comparison of the conclusions among themselves = internal consistency of the system is tested
- investigation of the logical forms of the theory = determining whether the system is an empirical / scientific theory, or a tautology
- comparison with other theories = determining whether the theory would constitute as a scientific advance
- testing of the theory by the way of empirical applications of the conclusion = finding practical applications
Point (4) is done deductively as well, seeing if the predictions can be verified by comparing them with the results of practical applications and experiments. If the prediction is verified then we have no reason to discard the theory yet; if the prediction is falsified, then the theory from which the prediction came is also falsified.
Corroboration
Corroboration is the measure of how well a scientific theory has survived testing and attempts at falsification. A corroborated theory is not proven true, or even more likely to be true. It simply represents the best-tested option available at the moment.
It should be noticed that a positive decision can only temporarily support the theory, for subsequent negative decisions may always overthrow it. (p. 10)
Scientists choose between competing theories by comparing their degrees of corroboration. The theory that has survived the most severe tests (made the boldest predictions and withstood the most rigorous attempts at falsification) is preferred, even though we cannot say it is “true” or “probably true.”
Section 5: Experience as a method
Defining the idea of an empirical science should satisfy three requirements =
- it must by synthetic, so it represents a non-contradictory, possible world
- satisfy the criterion of demarcation i.e. must not be metaphysical
- must be distinguishable from other such systems as the one which represents our world of experience
Experience hence acts as a method to distinguish our would from other theoretical systems, and so the theory of knowledge (“whose task is the analysis of the method particular to empirical science”) can be described as the theory of the empirical method, i.e. experience.
Section 7: Problems of empirical basis
Popper holds onto the idea that science is an empirical enterprise and so the question of how experience can test theories remain a serious one.
Popper defines empirical basis as “problems concerning the empirical character of singular statements and how they are tested” (p. 21)
Problems with Popper
- This does not solve the problem of holism about testing (that logical-empiricism also struggled with)
- whenever we try to test a theory by comparing it with observations, we must make a large number of additional assumptions in order to bring the theory and observations into “contact” with each other, If an unexpected result occurs it is always possible to blame the assumption rather than the theory itself.
- Popper’s answer to the holist objection is that while logically, it is always possible to blame other assumptions involved in the test, a good scientist would not try to do this.
- This, however, moves away from describing a characteristic of scientific theories to describing a characteristic of scientific behavior
- we cannot be completely certain about the observation reports that we use to falsify the theory
- we assume that the observation report is certain and true, but that is not always the case.
- Popper insists that while we make conventional decisions to accept certain observation report, this is fundamentally different from making arbitrary decisions about which theories to accept or reject
- But it could be argued by critics that this “fundamental difference” is not defined
- Example: Copernicus’ heliocentric astronomical theory predicted that during the course of the year, the apparent size of both Venus and Mars should change significantly but with the naked eye, no such change of Venus’s size were observed. In this case, Copernicus’ theory should not be falsified because the observation was to blame
These points show that any system of hypotheses can be held onto despite falsification depending on the decisions made. Note that this does not necessarily cause problems for Popper’s differentiation between science and pseudo-science, as Popper claims that while scientific claims can be falsifiable by observation reports, a pseudo-science does not clash with any possible observation.
the problem with corroboration - the bridge building challenge
“Suppose we are trying to build a bridge, and we need to use physical theories to tell us which designs are stable and will support the weight that the bridge must carry… How does Popper’s philosophy treat the bridge-building situation? … suppose we have to choose between (1) a theory that has been tested many times and has passed every test, and (2) a brand new theory that has just been conjectured and has never been tested. Neither theory has been falsified. We would ordinarily think that the rational thing to do is to choose the theory that has survived testing. But what can Popper say about this choice? Why exactly would it be irrational, for Popper, to build the bridge using a brand new theory that has never been tested?”
Theory and Reality, Godfrey-Smith, p. 67
Popper refuses to say that when a theory passes a test, we have more reason to believe that the theory is true. Instead, he uses the notion of corroboration to explain why it is more rational to choose the bridge that has survived more tests at falsifying it.
The problem with corroboration is that it can be questioned that corroboration is simply just a new name Popper gave to confirmation. If so, Popper has given up one of his main difference from logical empiricists. If not, there seems to be no particular why we should choose a corroborated theory over a non-corroborated theory when we build a bridge; because corroboration is simply a “record” of past tests and gives no prediction for future tests. So, claiming that corroboration gives a rational reason to prefer one theory over another seems to smuggle the notion of induction in (that Popper was against)