- Length
- 544 pagesRoutledge Classics, 2002, from the 1959 English text
- Reading
- DemandingReadable with attention; assumes no other book
- Availability
- Many editionsIn print in several competing editions
The Logic of Scientific Discovery
The book that made falsifiability the popular test of whether something is science, written by a schoolteacher on the edge of the Vienna Circle who thought its members had the wrong criterion and the wrong problem. It reached print only after his uncle cut it in half.
- Date
- 1934
- Author
- Karl Popper
- Region
- Vienna
- Language
- German
- 1959 English, Karl Popper His own translation with extensive new appendices; in copyright Attested
Where it stands in the argument
- Position
- Reply — Answers a named earlier work
- Form
- Treatise
- Method
- Analytic — Clarification of language and logical form
- Transmission
- Translated — Reached its later readers in another language
- Standing
- Live — Still argued in current philosophical work
- Years back to what it answers
- 195 years
Cut in half by the author's uncle to fit a publisher's page limit; the full original manuscript appeared forty-five years later.
What the book says
A book that accepts the sceptical argument about induction in full and denies that science needs induction at all. Popper's starting point is the logical asymmetry between confirming and refuting a universal statement: no number of white swans establishes that all swans are white, but one black swan settles it. Science, on his account, does not proceed by accumulating instances and generalising; it proceeds by proposing conjectures - which may come from anywhere, including dreams and myths - and then trying as hard as possible to destroy them. A theory is scientific not because it can be verified but because it forbids something, so that a stated observation would refute it; a theory compatible with every possible outcome tells you nothing. This is the demarcation criterion, and Popper is explicit that it separates science from non-science rather than sense from nonsense: metaphysics is not meaningless, only unfalsifiable, and has often been the seedbed of testable theories. The more a theory forbids, the better it is, so bold and improbable conjectures are preferable to safe ones, and a theory that has survived severe testing is corroborated rather than proved.
What it argues
That the growth of knowledge is a process of conjecture and refutation, and that certainty is not available and not needed. Since no universal statement can be established by observation, the question is not how theories are proved but how they are tested; and testing works because a genuine theory rules things out. From this follow the book's characteristic positions. Bolder theories are better, because they forbid more and are therefore more exposed. Corroboration is a report on past testing and confers no probability of future success. Basic observation statements are themselves theory-laden and are accepted by decision rather than given by experience, which Popper compares to a structure built on piles driven into a swamp - driven deep enough to carry the building, not down to bedrock, because there is none.
What it answered
Hume on induction, and the Vienna Circle on meaning. Popper argued that verifiability fails twice over: it cannot be met by any universal scientific law, and it wrongly declares metaphysics meaningless when the real distinction is between the testable and the untestable. His other targets were named repeatedly in later work - Marxist history and psychoanalysis, which he took to be unfalsifiable in practice because they can accommodate any evidence, a charge that made him famous and that specialists in both fields dispute.
How it argues
By logic, and much more technically than the popular reputation suggests. Long sections deal with probability, with the interpretation of quantum mechanics and with the structure of basic statements, and the famous demarcation criterion occupies a fraction of the book. Popper writes clearly and combatively, and the later English appendices, added over twenty-five years, are where a good deal of the interesting argument now sits.
The moment
Vienna in the early 1930s, with Popper teaching in a secondary school and on the fringes of a Circle that discussed his manuscript without admitting him. The publication history is part of the book. Schlick and Frank accepted the work for their series in 1933; Springer capped it at fifteen signatures; Popper submitted extracts from two volumes and was still over; and his uncle Walter Schiff, a professor of economics, cut the manuscript by about half and left his own mark on the opening. Popper left for New Zealand in 1937 and never returned to live in Austria.
How it travelled
In two versions and two languages, with a gap of a quarter century. The German book was read within the Circle and by physicists; the English translation of 1959 made Popper internationally famous and is the version almost everyone means. Falsifiability escaped into general use and appears in court judgements, in science policy and in popular argument, usually in a cruder form than the book supports. The full original manuscript, twice the length, was finally published in 1979.
The strongest objection
That no theory is ever refuted on its own. A prediction is derived from a theory together with a mass of auxiliary assumptions about instruments, initial conditions and background physics, so a failed prediction shows only that something in the whole package is wrong - and the scientist is free to revise an auxiliary rather than abandon the theory. This is the Duhem-Quine problem, and it means falsification is never the clean logical operation the asymmetry suggests. Kuhn added the historical version: scientists routinely retain theories in the face of anomalies, and are right to, since a young theory has anomalies by definition. Popper's later work accommodates some of this with talk of methodological rules against evading refutation, which is a policy rather than a logic.
Why read it
The first two chapters, which set out the problem of induction and the demarcation criterion and are as clear as philosophy of science gets. The material on probability and quantum theory is for specialists. Popper's own Conjectures and Refutations of 1963 is the more readable statement of the same position and is where most people should start.
- First printed
- 1934
- First in English
- 1959
- Digitised
- No
The book is an abridgement of a much longer manuscript, Die beiden Grundprobleme der Erkenntnistheorie, written 1930-33 and published only in 1979.
Popper's own English of 1959, with new appendices, is the standard and is in copyright. The German is also in copyright in most jurisdictions.
Where it stands
Live and contested. Falsificationism as a complete account of scientific method has few defenders since Kuhn and Lakatos, but the underlying arguments - about induction, corroboration, the theory-ladenness of observation and what makes a theory testable - are all active, and Popper remains the standard opponent in every discussion of scientific method.
On quoting this book
In copyright in both languages in most jurisdictions. Nothing is quoted.
Contested
How much the 1959 English differs from the 1934 German, and whether Popper's later additions shift the position, is a real question given that the two texts are separated by twenty-five years and a doubling in length. How much of the book is Popper and how much his uncle's compression is unresolved, and Popper's own account of the abridgement changed over time. Whether the demarcation criterion works on real historical cases is the standard objection from historians of science. And whether the charge against psychoanalysis and Marxism is fair, or trades on a caricature, is disputed by practitioners of both.