The nineteenth century French philosopher Auguste Comte (1798-1857) invented a scientific philosophy he called "positivism." In Austria and Germany in the early twentieth century, a new positivism appeared, called logical positivism, associated with names like Ernst Mach, Moritz Schlick, Rudolf Carnap, and Otto Neurath. This obscure central European movement became unexpectedly decisive for philosophy written in English in the twentieth century, thanks to Adolph Hitler. Shortly after he came to power most of the logical positivists, including the Vienna Circle and the Berlin group, immigrated to English-speaking countries, especially the US and Britain. From Cambridge and London to Princeton, Harvard, Chicago, and UCLA, these transplanted positivists and their students were the germ-form of what eventually became so-called Analytic philosophy, still the norm in English-speaking universities.
The logical positivists bequeathed an important foundational myth. According to the myth, there are certain so-called Continental philosophers. They are obscure, confused, authoritarian irrationalists, like Martin Heidegger (whom the positivists love to hate), and also Hegel and Nietzsche. Then there are the good guys, the Analytic philosophersùsober scientific rationalists rigorously trained in methods of logic and conceptual analysis.
No Analytic philosopher today would call himself a logical positivist. But the way you get to be an Analytic philosopher of science these days is, first, by acquiring their prejudice against "Continental" philosophers, and then learning the reasons why one shouldn't be a logical positivistùreading Carnap, his critics, his patricidal sons; studying the formal methods to which Carnap powerfully contributed; and acquiring the prejudices which were an important part of his legacy.
Comte invented the modern idea of a "philosophy of science", as a philosophical demonstration of the essential unity of all the sciences. That unity requires science as a whole to be distinguished from something else, its mirror-image. The positivists called this constitutional Other to science "metaphysics". It is unverifiable, experience-transcendent, incapable of precise formulation. It was a scandal to the logical positivists that apparently intelligent, morally decent people took Nietzsche and Heidegger seriously. Their language is simply meaningless nonsense.
Unity of science, then, is a unifying positivist theme. Neurath edited a multi-volume Encyclopedia of Unified Sciences (which ironically contained the inconspicuous monograph by Thomas Kuhn that would be the undoing of positivism). Comte understood the unity of science in terms of his Law of Three Stages. He thought this was one of his most important discoveries. According to this "law", human history passes through, first, a religious, then a metaphysical, finally a scientific stage. Science is the climax of the rational mind, and humanity's historic destiny.
The later logical positivists would have none of this, and shunned the name of Comte. Yet they shared his extravagant estimation of science, as well as his ambition for a philosophical demonstration of its unity. And like Comte, they claim the right, as guardians of the treasure, the priesthood of Science, to pronounce on the "scientific content" ("meaningfulness") of anybody's discourse. Of course, if all that philosophy can discuss is what science says, which is all that strictly speaking makes sense, then ethics, art, beauty, justice, friendship, happiness, and all the other topics which have made philosophy interesting for the last 2500 years must be declared unsuited to serious discussion and off-bounds to "professional philosophers." Analytic philosophers persuaded themselves that they actually preferred these desert landscapes of arid technicality, and counted it proof of their rigor that no one outside their charmed circle understood what they were talking about. That accounts for the sterility of so much so-called Analytic philosophy.

Karl Popper (1902-94) was a sort of maverick fellow traveler among the Vienna Circle positivists. "Neurath used to call me 'the official opposition' of the [Vienna] Circle." A Vienna native, trained in psychology and physics, Popper knew most of the European positivists of his generation (as well as enjoying the friendship of scientists like Einstein and Schr÷dinger). He shared interest in many of the same philosophical questions as the positivists, though he seldom found their answers satisfying. At nearly every point his work offers an alterative to positivist philosophy.
Popper left Austria in 1937, first for New Zealand, then, in 1946, for Britain and his position at the London School of Economics. He was knighted in 1964. His first important book, The Logic of Scientific Discovery, was published in German in 1935. While in New Zealand he published The Open Society and Its Enemies (1945), and after settling in Britain, The Poverty of Historicism (1957), and Conjectures and Refutations (1963). Later works include Objective Knowledge, The Self and Its Brain (with Sir John Eccles), and Popper's intellectual autobiography, Unended Quest.
For the positivists, science is a body of universally valid knowledge arrived at by careful induction from neutral, pre-theoretical observations. First come the innocent observations, then slowly a theory is built up, with each step carefully verified. That is the method of all sciences, and accounts for their unity, setting science apart from any other discourse, whether of "metaphysics" or social criticism. Any discourse not constructed according to this inductive norm is unverifiable, insignificant, without meaning, and therefore not to be taken seriously.
Logical positivism enforced a disciplinary quietude on philosophy. Social values and political ideas are merely emotional, and quite unsuited to philosophical discussion. It is not surprising that logical positivism and its descendent form of Analytic philosophy flourished in Eastern Europe (Poland, Czechoslovakia, DDR, Bulgaria) during the Cold War. With the flagrant, though fortunately singular exception of the ebulliently political Otto Neurath, the positivists worked hard at the immunity of their research to politics or social criticism.
With Popper all that changes. One of his most important arguments, and a theme of his philosophy from first to last, is that "induction" is a myth. We never observe without some theory in mind, and no number of observations can ever, under any circumstances, render a universal statement (a law of nature) either probable or verified. The whole idea of observing some number of events, and "inducing" from them to a general statement is a fiction. There is no such inference, not in logic or psychology, and it has nothing to do with the progress of knowledge in science.
For Popper, scientific theories, however widely accepted or seemingly well-verified, never rise above the condition of sheer conjecture. They are bold, poetic hypotheses. Where they come from, the "method" by which they are arrived at, is of no importance. Their value to science lies in how we use them, not where they come from. The proper scientific use of a conjecture is to make it a target for refutation. The real work of science, as Popper sees it, is the flow of new conjectures and redoubled efforts to refute them.
Popper has a very convincing argument. No number of singular observations can "establish" or "verify" a generalization (no number of swan-observations proves that all swans are white). Yet a single contrary observation (one non-white swan) is enough to refute any hypothesis. While no scientific theory can ever be "verified" (to say nothing of "proved"), such theories can be falsified or refuted. The engine of scientific progress moves from bold conjectures to eventual refutation. It is not enough to say that we learn by our mistakes. The only discovery we can make is the discovery of our own mistakes.
For the positivists, discourse is intellectually respectable if it is scientifically verifiable, and what isn't verifiable is sheer nonsense. Again, Popper takes a different stand. Discourse is scientific when it makes a definite enough claim to admit of falsification or refutation should something not turn out as predicted. Unfalsifiable discourse is not scientific, to be sure, but that is no shame. It's like saying a zebra is not an antelope. They are different animals. So too with scientific and unscientific discourse. Quite important things may be said (in philosophy or social criticism, for instance) which are not "falsifiable" in the way a scientific conjecture must be.
Unfalsifiable metaphysics even has value in science. Nothing could be more monstrous to the positivist, but that's how Popper sees Darwin's theory of Natural Selection. Darwin's conjecture that species evolve as a result of "natural selection" culling the unfit and leaving better adapted descendants is not directly testable and cannot be definitively refuted. It is to that extent unscientific. Yet it is invaluable to science. "I do not see how, without it, our knowledge could have grown as it has done since Darwin." Its value lies in its capacity to guide research. It tells scientists to seek a mechanism of adaptation, rather than rest with the religious idea of design. It tells them to expect gradual change in species over geological time, and to look for it. And it tells them to expect contingency in the course of evolution, unlike the majestic unrolling of a Master Plan.
So Popper doesn't locate the exemplary rationality of science where the positivists do, in a mythical "inductive method," and he rejects their ideas of neutral observation and certainty. He also rejects their quietism. Popper wrote nearly as much about social criticism as about science. The Open Society and Its Enemies is a two-volume study of the political ideas of Plato, Hegel, and Marx. Popper exposes their latent authoritarianism, which he links to their philosophical ideas about reason and knowledgeùideas his philosophy of science refutes.
The Poverty of Historicism attacks the idea that history, or the future of societies, can be predicted. Popper argues strenuously against any form of historical predictability. There are no "laws" of society or history on par with the laws of nature, and claims (for instance in Marxism) to know the laws by which the social future must unfold are sheer dogma. Human societies do not form closed, law-bound wholes, while the social effects of changes in knowledge are inherently unpredictable.
To Popper, these arguments favor liberalism over the radical ideologies of both Left and Right. He explains a "liberal" as "not a sympathizer with any one political party but simply a man who values individual freedom and who is alive to the dangers inherent in all forms of power and authority." He thinks this is the politics most consonant with scientific rationality. What matters in science is not where a conjecture comes from (its pedigree), but whether it withstands refutation (its merit). The problem in science is not "How can we know the truth?" It is "How can we minimize the risk of dogma?" Likewise in political philosophy. Not "Who must rule?" but "How can we minimize the risk of bad rulers?" There are no foundations to our knowledge; instead, there is only the endless task of conjecture, criticism, and refutation. Likewise in politics. There is no advantage to a revolution that would supposedly set society on the right foundation. There is no such foundation, neither in science nor politics. The reasonable course is therefore piecemeal social engineering.
Philosophy of science can make stiff demands on a reader's scientific literacy. To read Comte you have to know something about all the disciplines of modern science, including those Comte invented, like sociology. To read Carnap you have to know advanced mathematical logic. Popper can be as technical as anyone. There are stretches of Conjectures and Refutations that will be unintelligible to anyone who has never studied formal logic and probability theory. He takes for granted a basic understanding of Einstein's Theory of Relativity, and the preparedness of readers to follow him in philosophical argument with the architects of Quantum Mechanics. Yet Popper is pretty gentle as philosophers of science go. He writes a plain, clear, unlabored, if slightly pedantic English, makes his technical points as lightly as I've seen it done, and has a lot to say that doesn't require technicality.
Popper's view of science is still capable of shocking some of our preconceptions. Induction is a myth, as is the idea of observational knowledge. Science never rises above sheer conjecture, and all we know for sure is what we don't know. Our best, most scientific knowledge is never more than provisional, and always incomplete. These ideas began as criticisms of logical positivism, and it seems likely that is how Popper will be remembered. Other ideas, such as his evolutionary theory of knowledge or the ontology of World 3, have fared less well. With logical positivism a spent force, Popper is little read in philosophy today.
Partly he was eclipsed by the work of T. S. Kuhn, whose Structure of Scientific Revolutions appeared simultaneously with Popper's Conjectures and Refutations in 1963. Kuhn and Popper agree on key points against the positivists. They reject the picture of the gradual growth of science through the accumulation of observations, and draw attention to the way observation is always compromised by preconceptions. Yet Kuhn sees an important difference between science in its occasional periods of revolutionary change, and the interregnum phase he calls "normal science", when the new "paradigm" or model of scientific work is accepted, and scientists settle down to extending it.
The effort to falsify fundamental conjectures is not characteristic of normal research. Anyone's failure to confirm an accepted result will more likely be deemed proof of ineptitude than a falsification of the paradigm. Theories in science are not abandoned merely because of observations which contradict their predictions. There also has to be an alternative at hand. Until then, theories tend to be immune to falsification. The growth of knowledge is less rigidly bound than Popper's elegant logic of falsification implies, and science has no monopoly on rationality.
Since Kuhn, philosophical ideas about science have been profoundly influenced by the rise of a new history and sociology of science, which is enjoying a period of immensely creative scholarship, associated with such names as Michel Foucault, Bruno Latour, and Toronto philosopher of science Ian Hacking. Their themes would be unlikely to garner Popper's approbation.
Popper agreed with the positivists that science is a unity, defined by the rationality of its method. The new work doubts this unity. There is no essential scientific method, or logic, or rationality. The many sciences, from linguistics and meteorology to paleontology and particle physics, have their own traditions of research and styles of reasoning. Nor are these philosophers enthusiastic about the careful distinctions Popper and the positivists drew to secure the purity of scienceùfussy distinctions between power and knowledge, discovery and validity, science and politics. The force of work by Foucualt, Latour, Hacking, and others today is to see all of these, both poles of the dichotomies, as "normal" qualities of modern technoscience.

Barry Allen teaches philosophy at McMaster University. He is author of Truth in Philosophy, and a new book, Knowledge and Civilization, which will appear in 2003.