What Is This Thing Called Science?

Citation

Chalmers, A. F. (2013). What Is This Thing Called Science? (4th ed.). Hackett Publishing Company.

Chapter Summary

The fourth edition is a 304-page book, co-published with the University of Queensland Press, with Hackett holding rights in North America. Hackett notes that the first edition appeared in 1976, that the book has been translated into eighteen languages, and that it is known for its accessibility to beginners and for its value to advanced students. After prefaces to all four editions and an introduction, there are fifteen chapters, an epilogue (carried over from the third edition) and a postscript, then notes, a bibliography and an index of names. Hackett prints a full table of contents with the headings inside each chapter, so the summaries below follow those headings; each chapter also ends with a "Further reading" section.

Chapters 1–3: Science as knowledge derived from the facts of experience; Observation as practical intervention; Experiment

The book begins with the commonsense view that science is built on the facts of experience. Chapter 1 sets this view out ("seeing is believing") and then examines it: visual experiences are not determined solely by the object viewed, observable facts are expressed as statements, and observation statements are fallible. Chapter 2 treats observation as practical intervention rather than passive and private seeing, using Galileo and the moons of Jupiter. Chapter 3 turns to experiment: relevant rather than merely any facts, the production and updating of experimental results, historical examples of transformed experimental bases, and whether experiment is an adequate basis for science.

Chapter 4: Deriving theories from the facts: induction

This chapter is on inductivism. Its sections introduce "baby logic", ask whether scientific laws can be derived from the facts, discuss what makes a good inductive argument, and set out further problems with inductivism together with its appeal.

Chapters 5–7: Introducing falsificationism; Sophisticated falsificationism, novel predictions and the growth of science; The limitations of falsificationism

Three chapters follow Popper's alternative to inductivism. Chapter 5 presents falsifiability as a criterion for theories, degrees of falsifiability, and the relation to progress. Chapter 6 moves to relative degrees of falsifiability, ad hoc modifications, boldness, novelty and background knowledge, and compares inductivist and falsificationist views of confirmation. Chapter 7 gives the limitations: problems arising from the logical situation, inadequacy on historical grounds (with the Copernican Revolution as the case), and weaknesses of the demarcation criterion and Popper's response.

Chapters 8–9: Theories as structures I: Kuhn's paradigms; Theories as structures II: research programs

These two chapters treat theories as structured wholes. Chapter 8 covers Kuhn: paradigms and normal science, crisis and revolution, the function of normal science and of revolutions, the merits of the account, Kuhn's ambivalence about progress, and objective knowledge. Chapter 9 covers Lakatos's research programs, methodology within and across programs, novel predictions, testing the methodology against history, and problems with it.

Chapters 10–11: Feyerabend's anarchistic theory of science; Methodical changes in method

Chapter 10 reviews the story so far and then Feyerabend's case against method, his advocacy of freedom and a critique of his individualism. Chapter 11 argues against universal method and for changes in method, using the move from naked-eye to telescopic data as an example of a change in standards, and discusses piecemeal change of theory, method and standards.

Chapters 12–13: The Bayesian approach; The new experimentalism

Chapter 12 introduces Bayes' theorem and subjective Bayesianism, shows applications and gives a critique. Chapter 13 returns to experiment: experiment with a life of its own, Deborah Mayo on severe testing, learning from error, and the new experimentalism in perspective, with an appendix on meetings of theory and experiment.

Chapters 14–15: Why should the world obey laws?; Realism and anti-realism

Chapter 14 considers laws as regularities, laws as characterisations of powers or dispositions, and thermodynamic and conservation laws. Chapter 15 takes up global anti-realism (language, truth and reality), standard objections and the anti-realist response, scientific realism, conjectural realism, idealisation and structural realism.

Chapters 16–17: Epilogue to the third edition; Postscript

The contents list gives the Epilogue to the third edition only a further-reading section. The Postscript is the new feature of the fourth edition: it uses Chalmers's research on the history of atomism. Its sections cover confirmation by arguments from coincidence, philosophical versus scientific knowledge of atoms, Perrin's experiments on Brownian motion and the theory-dependence of observation, nineteenth-century chemistry, realism versus anti-realism again, the point that strongly confirmed theories are never completely discarded, approximate truth and levels of reality.

Key Concepts

Chapters 1–4

• Inductivism: the view that scientific knowledge is derived from observable facts by induction.
• Observation statements: statements of observable facts, which the book treats as fallible and as dependent on more than the object viewed.
• Experiment: the production of relevant facts, treated in its own chapter.

Chapters 5–9

• Falsifiability: whether a theory can be refuted by observation, used as a criterion for theories.
• Ad hoc modification and novel prediction: the ideas used to compare theories in the sophisticated falsificationist account.
• Paradigms, normal science and revolutions: Kuhn's account of how sciences change.
• Research programs: Lakatos's units of appraisal, with methodology within and across programs.

Chapters 10–17

• Methodological anarchism: Feyerabend's case against a single method.
• Bayesian reasoning: probabilities updated by Bayes' theorem, in its subjective form.
• Severe testing: Mayo's account within the new experimentalism.
• Laws as regularities or as powers: two ways of understanding laws of nature.
• Scientific realism and anti-realism: whether scientific theories describe an independent reality.
• Arguments from coincidence: the form of argument used in the Postscript on atomism.

Critical Analysis

Strengths

• A clear historical arc. The contents trace one line of argument from the commonsense empiricist view of science, through inductivism and falsificationism, to Kuhn, Lakatos and Feyerabend, and then to Bayesianism, the new experimentalism, laws and realism. Several chapter titles frame a position through its difficulties ("The limitations of falsificationism", "Further problems with inductivism"), which encourages the reader to see each view as a response to the one before.
• Fine-grained contents. Every chapter breaks into short, named sections and ends with "Further reading", so a reader can find a topic quickly and use the book for revision.
• Cases from the history of science. The headings mention Galileo and Jupiter's moons, the Copernican Revolution, the telescope, Perrin's experiments on Brownian motion and nineteenth-century chemistry, anchoring the philosophical positions in episodes.
• Revision history. Hackett says the fourth edition improves the text throughout in light of Chalmers's teaching experience, reader comments and developments in the field, and that the chapter-length postscript on atomism gives a worked example of the themes. Hasok Chang's comment on the previous edition, quoted by Hackett, praises the new chapters as an insightful, concise treatment of recent developments.

Limitations

• A particular selection of positions. The chapters are organised around the Popper–Kuhn–Lakatos–Feyerabend debate and later developments; scientific explanation and the philosophy of the special sciences, for example, have no chapter of their own by title, and laws appear in a single chapter (Chapter 14).
• Quoted praise concerns the previous edition. The only published comment on Hackett's page, from Hasok Chang, concerns the previous edition, so readers have no independent assessment of the new postscript.

Where It Fits in the Course

Chalmers's book is a compact, chronologically ordered introduction to scientific method and theory change. Samir Okasha's Philosophy of Science: A Very Short Introduction is a shorter introduction, and Peter Godfrey-Smith's Theory and Reality follows a similar century-long arc from logical empiricism through Popper, Kuhn, Lakatos and Feyerabend to Bayesianism and scientific realism, with a glossary and reading suggestions. Chalmers can serve as an alternative to Godfrey-Smith or as a second look at the same debates from a different author. Rosenberg and McIntyre's Philosophy of Science: A Contemporary Introduction is a fifteen-chapter textbook organised by topic, with a block of four chapters on explanation, laws and causation, where Chalmers has a single chapter on laws.

For further reading, Chapter 8 prepares the ground for Thomas Kuhn's The Structure of Scientific Revolutions, and the realism and anti-realism material in Chapter 15 leads to Bas van Fraassen's The Scientific Image, a statement of constructive empiricism.

How to Use This Book

Hackett describes the book as accessible to beginners, and no prerequisite is listed. It suits a first course in philosophy of science and also students from the sciences who want to see where the received picture of scientific method comes from.

A one-semester plan is to read Chapters 1–7 as a unit on the inductivist and falsificationist pictures, Chapters 8–11 on Kuhn, Lakatos and Feyerabend, and Chapters 12–15 on contemporary alternatives. The Epilogue and the Postscript are optional but valuable; the Postscript in particular shows the book's themes applied to a single historical case, the atom. Short courses can read Chapters 1, 4–8 and 15.

Work through the sections as short readings and use the "Further reading" at the end of each chapter to reach primary texts by Popper, Kuhn, Lakatos and Feyerabend. Pairing each of Chapters 5–10 with a primary text by the thinker it discusses is a good way to test the book's summaries.