Kuhn's Structure of Scientific Revolutions

Last modified by Malcolm R. Forster on October 30, 1998.

Kuhn’s book is about scientific revolutions—the grand changes in theory that take place periodically in the history of science. The choice of the word ‘revolution’ is significant because it invites the analogy between scientific change and political change (e.g., the French revolution).

Butterfield as a Precursor to Kuhn

Kuhn's Structure of Revolutions

Kuhn's view of scientific revolutions centers around his notion of a paradigm. He introduces it in the first few pages of SSR, but revamps the notion in the postscript in the light of charges of vagueness. I will use his later definition.

Paradigm as Disciplinary Matrix, has 4 components:

  1. Symbolic generalizations: E.g. Newton's laws of motion. (In what we call ‘theory’.)
  2. Metaphysical presumptions. E.g., Atoms as "billiard balls", or light as a wave, or light as particles. (In what we call ‘theory’.)
  3. Values: E.g., the accuracy of prediction (pp.184-85), puzzle solving success, simplicity. (Not usually thought of as changing much.)
  4. Exemplars: Textbook or laboratory examples that students learn. E.g., harmonic oscillator (simple pendulum), Keplerian orbits, or random mating models in population genetics. (Introduces the idea of tacit knowledge.)


  1. Exemplars help introduce the notion of tacit knowledge, learned by doing science, like the ability to read x-rays, rather than applying rules), and cannot be written down or articulated. (Even logic and mathematics is an example of this.)
  2. Tacit knowledge is neuropsychological: "so much past experience is embodied in the neural apparatus that transforms stimuli to sensations." (p.195) Note that he is suggesting that tacit knowledge affects the way scientists see.
  3. (p.188) Exemplars give a theory empirical content. This makes perfect sense if exemplars are models (= theory + auxiliary assumptions), because we need auxiliary assumptions to make predictions from a theory.
  4. A disciplinary matrix is different from a ‘theory’ not only in its inclusion of values and tacit knowledge, but also because it refers essentially to a scientific community, a sociopolitical entity, rather than an item of ‘knowledge’.

Two Kinds of Science

Normal Science: Science as it is practiced within a single paradigm. Models are constructed as the solutions to puzzles under the guidance of a theory, but the theory itself is not criticized or blamed for failures (e.g., Leverrier-Adams example—discovery of Neptune).

Revolutionary Science: The transformation from one paradigm to another. Contrasts with normal science.

Normal and Revolutionary Science are Different

Contrasts with a Popperian Picture of Science

Popperianism and Bayesianism assume that the same methodological principles that apply to normal science and revolutionary science. Kuhn denies this:

  1. Kuhn’s argues: Normal science does not aim at novelty:
    (1) Novelties of fact or theory (i.e., discoveries?) lead to the end of normal science.
    (2) Normal science does not aim at its own demise.
    Therefore, (3) normal science does not aim novelties of fact or theory and, when successful, finds none.
  2. It’s not clear that Kuhn is right about (1). Why should a new discovery (e.g., x-rays, pulsars, radio waves, HIV virus) necessarily lead to the demise of normal science? Perhaps he should only assume that they may. Will the argument still work?

  3. The falsificationist’s folly is to think that the falsification of a model is also a falsification of the background theory. Anomalies (= mismatches between the current best model and observation, or between accepted models) are not sufficient. Crisis (= failure at prediction) is sufficient. E.g., Ptolemaic astronomy failed not because it could not remove anomalies, but because the anomalies, once removed, would reappear when new observations were made.
  4. In normal science the theory is not tested. Even in the face of crisis, the theory is not immediately rejected—one also needs a better alternative. (A leaky roof is better than no roof at all.) Here Kuhn (p. 80) points out that "science students accept theories on the authority of teacher and text, not because of evidence." But is that really the only explanation of why this is true? This one of the ways that Kuhn shifts from an internal (= non-sociological) view of science to an external (=community-based) view of science.
  5. Expectations obscure our vision. Kuhn cites a psychological experiment in which subjects are shown ordinary playing cards mixed up with some anomalous cards, like a black four of hearts. The results show that subjects initially see what they expect to see (either the four of spades, or the four of hearts). Therefore, science is subjective.


Scientists who work in different paradigms live in different worlds.

KUHN: "...after discovering oxygen Lavoisier worked in a different world." (p.118) "after Copernicus, astronomers lived in a different world." (pp.116-7) "until that scholastic paradigm was invented, there were no pendulums, but only swinging stones, for the scientists to see. Pendulums were brought into existence by something very like a paradigm-induced gestalt switch." (p.120) We could add our own examples. After evolutionary theory, human beings became just another primate. After the "darkenings of the sun" were explained, they became solar eclipses. After atomic theory, water became H2O. And so on.


  1. Scientists who work in different paradigms live in different psychological worlds; i.e., have different beliefs about the world. [PLAUSIBLE]
  2. Scientists who work in different paradigms live in different external physical worlds. [IMPLAUSIBLE]

Theory-Ladeness of Observation

"What scientists observe depends on what they believe." (Visual gestalt analogy)

  1. No scientist can ever observe something that contradicts their theory?
  2. Scientists who accept rival theories can never observe the same thing?

Kuhn cites many examples:

  1. Galileo sees a pendulum, where Aristotle sees a (slowly) falling stone.
  2. "Looking at a contour map, the student sees lines on paper, the cartographer a picture of a terrain." (p.111)
  3. "Looking at a bubble-chamber photograph, the student sees confused and broken lines, the physicist a record of familiar subnuclear events." (p.111)
  4. "Looking at a bubble-chamber photograph, the student sees confused and broken lines, the physicist a record of familiar subnuclear events." (p.111)
  5. Herschel’s discovery of Uranus, which he first thought was a star, and then a comet.

Kuhn’s argument for relativism:

  1. Kepler’s sun is at rest and Ptolemy’s sun is moving.
  2. Therefore, Kepler’s sun cannot be the same as Ptolemy’s sun
  3. Therefore, the object that Kepler observes is not the same as the object that Ptolemy observes.

Fallacies of Equivocation

A fallacy of equivocation occurs when an argument uses an ambiguous word in two different ways. E.g.,

P1. We have the right to abstain from voting.

P2. We should do what is right.

C. Therefore, we should abstain from voting.

When we eliminate the ambiguity, the argument is either invalid or unsound.

P1. We are entitled to abstain from voting.

P2. We should do what is correct.

C. Therefore, we should abstain from voting. [INVALID]

P1. We are entitled to abstain from voting.

P2. We should do what we are entitled to do. [FALSE]

C. Therefore, we should abstain from voting.

P1. It is correct to abstain from voting. [FALSE]

P2. We should do what is correct.

C. Therefore, we should abstain from voting.

In Kuhn’s argument, there is an equivocation between what Kepler and Ptolemy believe about the sun, and the object of their beliefs.

  1. Kepler believe that sun is at rest and Ptolemy believes that the sun is moving.
  2. Therefore, the object of Kepler’s beliefs and the object of Ptolemy’s beliefs cannot be the same object. [INVALID]
  1. The object of Kepler’s beliefs is at rest while the object of Ptolemy’s beliefs is moving. [FALSE]
  2. Therefore, the object of Kepler’s beliefs and the object of Ptolemy’s beliefs cannot be the same.

Meaning Variance

Incommensurability: "The normal-scientific tradition that emerges from a scientific revolution is not only incompatible but often actually incommensurable with that which has gone before." (p.103) ‘Incommensurability’ is vague in at least two dimensions: (a) Are values the same across revolutions? (b) (Meaning variance) Are all terms of one theory translatable into the other?

The Slippery Slope to Relativism

    1. Newtonian mass is not the same thing as Einsteinian mass.

B. If Einstein’s theory is true, then Newtonian mass does not exist.

    1. It is extremely unlikely that Einstein’s theory is true, so Einsteinian mass does not exist either.
    2. No scientific theory refers to anything real.
    3. The real world is of no concern to science.
    4. Relativism. "The world of the scientist is a world given and structured by theory, and as theory changes, so does the scientist’s world." (From Musgrave 1979, p. 338)

Musgrave recommends that one avoid stepping onto the slope in the first place. Kuhn is not so careful.

There is no such thing as a neutral observation language

Kuhn refuses to accept the reasonable idea that Aristotle and Galileo agree on the facts (described without reference to the notion of a ‘pendulum’) but interpret them differently. "How could it be so in the absence of fixed data for the scientist to interpret?" (p. 122)

NOTE: If models from two different traditions are incompatible because they make incompatible predictions. Then they cannot be incommensurable in all their terms.

Proof: Let A and B be two models in different theories, and suppose that A Þ  O and B Þ  not-O, where O is a prediction. Then the terms in O are in A and in B, and they mean the same in both contexts.

How communication is restored by conversion to the new paradigm, according to Kuhn. His answer is surprisingly conventional: a) the claim to have solved crisis-provoking problems, b) the claim to novel predictions, and c) the claim to simplicity. However, note his emphasis on claims, for Kuhn is thinking of the process as a sociopolitical one, rather than a process of logic.

Kuhn’s theory of science is a description rather than a prescription of science. "Some critics claim that I am confusing description with prescription, violating the time-honored philosophical theorem: ‘Is’ cannot imply ‘ought.’" (p.207)

The sociological turn

One of the most serious influences of Kuhn’s book has been to provide support for those who want to turn away from the philosophy of science towards the sociology and psychology of scientific communities.

One argument goes like this (ask the class to evaluate the argument):

  1. Either the philosophy of science is the right way to study science, or the sociology is the right way to study science, but not both.
  2. Kuhn showed that the philosophy of science is wrong way to study science.
  3. Therefore, the sociology of science is the right way to study science.

Just because an argument is bad, does not mean its conclusion is false. There is noting wrong with studying the sociology or the psychology of scientific communities. But there are some questions it does not address. For example, Kuhn is unable to explain the sense in which his theory of science describes what scientists should do, rather than merely what they actually do. It’s not that philosophers are denying that scientists are doing things right. The point is that any such claim must be assessed relative to a particular set of goals (e.g. truth), and it should be explained how their practice leads to such goals. Sociologists do none of this, so there is role for philosophers of science.