This paper stands as a legacy of a prolonged debate that has left little of great value despite the best efforts a many very bright and industrious scholars. It was written for a volume called The Philosophy of Rudolph Carnap, in the Library of Living Philosophers series edited by P A Schilpp. Each volume in the series has a standard format, first an autobiography by the philosopher, then a series of essays on his or her work (mostly his) from invited contributors, then a reply from the philosopher. The idea is to get a definitive statement of the authors views in response to a range of comments and criticisms in the hope that there will be a minimum of loose ends when the philosopher dies, so people are not left wondering -- what would he have said in reply to this or that criticism?
Albert Einstein was invited to contribute to the Carnap volume but he declined on the ground that discussing positivism had become a waste of space. I am inclinded to agree and it is hard to justify going to the trouble to write about this chapter in the book. Still, some good may come of it, if only a better understanding of the way that unhelpful research programs in philosophy have wasted so much time and effort.
Background: the Vienna Circle of Logical Positivists
The most immediate influence on the philosopher-scientists of the Circle was the philosopher-physicist Ernst Mach (1838-1916) an empiricist in the tradition of David Hume whose mission was to purge science of metaphysics and place it on the firm "positive" foundations of sensation. His major followers were Moritz Schlick (1882-1936), Rudolph Carnap (1891-1970) and Otto Neurath (1883-1945). Schlick took over Mach's chair at the university in Vienna and he became the front man for the Vienna Circle which pursued Mach's positivism, with Russell's Principia their inspiration and Wittgenstein's Tractatus providing the program. This was essentially a war on metaphysics by application of the strict "verificationist" definition of meaning. They proposed that statements should be regarded as literally meaningless if they could not be confirmed or verified by evidence. The propositions of logic and mathematics were exempt from the requirement for verification on the understanding that they are true by definition and they do not pretend to convey information about the world.
The most obvious casualties of the verification principle were religious and moral principles, though there were others that were less obvious, including the principle itself and, more surprisingly, the laws of science. When these laws are stated in their strong form, along the lines of "All swans are white", they cannot be verified by any number of observations of white swans, simply because you cannot be certain that you have sighted all the swans in the universe. This dilemma, with the unsolved problem of induction, represented twin "skeletons in the cupboard" of positivism, but still, by 1928 the Circle had an official program and a vehicle, the Ernst Mach Society, with a mission for scientific philosophy as a basis for popular education and social transformation. Nowadays they would have a web site and a blog as well.
In the words of Malachi Hacohen (Popper's biographer) "Neurath persuaded the militantly secularist Free Thinkers, now numbering most circle associates among them, to establish the Ernst Mach Society. Schlick was elected president, apparently without his prior knowledge, and other circle members chief officers". The circle gained worldwide influence, indeed they institutionalised the philosophy of science with a series of conferences in the 1930s with the sponsorship of luminaries such as Bertrand Russell. Then the predominantly Jewish and left wing members of the circle had to scatter for their lives, like Popper himself, and they were dispersed far and wide by 1939.
Popper addressed a different problem from that of meaning and metaphysics because he was concerned with the difference between science and pseudo-sciences such as astrology which appear to be based on observations but are actually "unsinkable". His exemplar of science was Einstein's theory which might have been refuted by a particular set of observations on the eclipse of the sun. Inspired by this example Popper advanced his criterion of falsification to demarcate between testable statements on the science side of the line ("All ravens are black" which is logically refuted by the observation of a white raven) and various categories of statements on the other side of the line (morals, metaphysics, methodological principles and, incidentally, nonsense). Moral principles, aesthetics, metaphysics and methodology are meaningful and they can be rationally discussed in terms of their adequacy for their purposes, their consistency and their consequences, though such discourse was in principle outlawed by the positivists.
Hacohen gives a lot of credit to the logical positivists of the Vienna Circle for putting up with Popper's Steppenwolf-like activity, prowling on the fringe of the circle where Schlick, Carnap, Neurath et al huddled around their campfire, seeking warmth and consolation from the dying embers of the verification principle They accepted Popper's book which most of them perceived to have merit even if none of them really agreed with Popper's turn from induction to a theory of conjectural knowledge that is tested but never confirmed.
Publication of the book aroused intense disagreement among the members of the Circle and no less than three reviews appeared in their house journal because Reichenbach and Neurath insisted on writing critical reviews after Carnap signaled assent to the turn from verification to testing. Support from Carnap was short-lived and the positivists generally managed to convince themselves, and many others, that Popper's falsification criterion was concerned with meaning which was their own ruling obsession. This misunderstanding persisted for some decades because the positivist diaspora carried it to Britain and North America where it became well entrenched while Popper was "down under" in New Zealand.
Popper's paper has six sections: (1) Introduction, (2) My Own View of the Problem, (3) Carnap's First Theory of Meaninglessness, (4) Carnap and the Language of Science, (5) Testability and Meaning and (6) Probability and Induction.
"Writing about Carnap brings back to my mind the time when I first met him, at his Seminar, in 1928 or 1929. It brings back even more vividly a later occasion, in 1932, in the beautiful Tyrolese hills, when I had the opportunity of spending part of my holidays in prolonged discussions with Carnap and with Herbert Feigle, in the company of our wives. We had a happy time, with plenty of sunshine, and I think we all tremendously enjoyed those long and fascinating talks, interspersed with a little climbing but never interrupted by it. None of us will ever forget, I am sure, how Carnap once led us in a steep climb up a trackless hill, through a beautiful and almost impenetrable thicket of alpone rhododendrons, and how he led us at the same time, through a beautiful and almost impenetrable thicket of arguments whose topic induced Feigl to christen our hill 'Semantische Schnuppe', something like Semantical Shooting Star -- though several decades had to elapse before Carnap, simulated by Tarski's criticism, discovered the track that led him from Logical Syntax to Semantics."
Popper went on to say that Carnap was a captivating person, utterly absorbed in his problems, eager to listen to criticism, with the courage to change his mind even on fundamental aspects of his philosophy. It has to be said that Carnap cast a strange spell over his associates, friends and critics (like Popper) alike. Joe Agassi's comments on this topic are interesting and I will come back to that in some concluding comments.
Popper went to the mountains with a big manuscript in his kitbag, which in 1934/5 became Logik der Forschung (Logic of Scientific Investigation). The manuscript contained a lot of criticism of two books by Carnap; some of this criticism Carnap took on board and some he said exaggerated the differences between Popper's views and those of the Vienna Circle. That is a view that people like Richard Rorty have been pleased to accept as a justification for ignoring Popper's work, as though it is a dead end like logical empiricism. Amazingly, that rejoinder from Carnap resulted in Popper saying nothing about their differences for a decade after the publication of his own book. It seems that by the time he did speak up the damage had been done, Freddy Ayer's Language, Truth and Logic and the Continental diaspora had poisoned the philosophical wells of the western world.
2. Popper's view of the problem
In a nutshell: "The repeated attempts made by Rudolf Carnap to show that the demarcation between science and metaphysics coincides with that between sense and nonsense have failed... In all its variants demarcation by meaninglessness has tended to be at the same time too narrow and too wide: as against all intentions and all claims, it has tended to exclude scientific theories as meaningless, while failing to exclude even that part of metaphysics which is known as 'rational theology'."
Popper's perception of the need to specify the difference between science and (say) astrology was described above in the historical background. The key points are (a) his demarcation criterion had nothing to do with meaning (b) it was not designed to eliminate metaphyiscs in a stroke and (c) there are degrees of testability, so the line is not clear-cut and it shifts over time as theories develop and experimental technology improves.
3, 4 and 5. Carnap's various theories of meaninglessness
The Oxford Companion to Philosophy states that "Technical rigor was the hallmark of his important contributions to formal semantics, the philosophy of science and the foundations of inductive probability". Against that, it is my impression that his three major works stand like the masts of a sunken vessel, a warning to mariners to avoid the rocks that wrecked the good ship Carrnap. Popper spends a good deal of time on technical talk about Carnap's evolution from his early Tractatus-like view on meaning, through the language of science -- physicalistic, unified -- to "Testability and Meaning", a 1936 article where Carnap made some Popperian changes to his system but balked at the next decisive steps away from "justificationism" to a full-blooded fallibilism. That was probably the high tide of agreement between Carnap and Popper, later when Carnap moved on to the foundations of inductive logic there could be no raprochment.
6. Probability and induction
In this section Popper described how the full consequences of approaching confirmation as if it was just a weaker form of verification became apparent in Carnaps's two books on probability. This is all more of the same -- the appearance of rigor, new terms, tortured arguments, to what effect? What scientist in living memory has made any use of Carnap's logic of induction? It seems that some of the consequences of Carnap's wild goose chase are virtually paradigms of absurdity, presumably due to the structure of thinking that generated the paradox of the ravens discovered later by Carnap's younger colleague, Hempel. For example (and this is Popper's account, so be wary):
"He [Carnap] proposes that we accept (as probable) a principle to the effect that the evidence 'Sandy is clever' increases the probability of 'A is clever' for any individual A -- whether A is the name of a cat, a dog, an apple, a tennis ball or a cathedral. This is a consequence of the definition of 'degree of confirmation' which he proposes. According to this definition, any two sentences with the same predicate ('clever' or 'sick') and different subjects are inter-dependent or positively correlated, whatever the subject may be... I am far from certain whether he has realised these consequences of his theory, for he nowhere mentions them explicitly."
A concluding comment from Joseph Agassi, from a review of a collection of papers dedicated to Carnap, reprinted in Agassi's book The Gentle Art of Philosophical Polemic.
"I do not think that any of his doctrines are going to be cited for long: he said nothing original. I do not think any of the techniques he invented, his modal logic, his state-descriptions, his reduction sentences, his theoretical and observation languages, his inductive system, his c function, his lambda and the continuum of inductive methods -- all these are museum pieces and they never worked."