Rudolf Carnap (1891-1970) was a German-American philosopher who rose to prominence with the Vienna Circle in the 1920s and 30s. As a refugee from the Nazis, he later became a major figure in North American analytic philosophy. His association with the Vienna Circle earned him the label “logical positivist” and the opprobrium that now comes with it. Yet much of that is undeserved, especially as recent developments have put his austerely technical philosophy (much of which must remain undiscussed here) into new light.
Carnap kept faith with the promise of the mathematical logic that had been invented by his teacher Gottlob Frege and developed by Bertrand Russell: that it rendered human reasoning transparent. Unlike traditional syllogistic logic, their new logic was not limited to analysing sentences into combinations of mere subjects and predicates; it could also recognize relations and therefore was able to give formal descriptions of reasoning that was previously inaccessible. For instance, given “2 is greater than 1” and “3 is greater than 2”, one can deduce “3 is greater than 1” only if “greater than x” is analysed into a relation and one of its relata. (The new logic also inspired logicism, the programme of reducing arithmetic to logic in order to explain knowledge of the former in terms of knowledge of the latter.) Exhibiting the logical transparency of ordinary, natural language claims posed the considerable challenge of translating potentially ambiguous natural-language sentences into a counterpart in a rigorous formal language. But Carnap thought it was worth trying, for the ultimate prize was to bring clarity and precision into philosophy itself – and thereby allow it to join the sciences and become a discipline of practical use.
When logical positivism first emerged in Central Europe, after the First World War, it scandalized established philosophers by dismissing much of the substance of contemporary philosophy. It was seen as a revolutionary movement, linked by some to socialism. Much of what the logical positivists rejected would hardly count as philosophy today (some was proto-Nazi ideology). But they also repudiated a major tranche of traditional philosophy by denying that philosophical intuition can ever transcend logic or experience.
Core to logical positivism was the principle of “verificationism”. According to it, for a statement to be meaningful, it must be either analytic and justifiable a priori, or synthetic and justifiable a posteriori. Any meaningful sentence must therefore either express a tautology (spelling out logical truths or conceptual connections that are determined by the rules of the language in question and therefore deductively provable) or make a factual assertion (where it is possible in principle to adduce intersubjectively available empirical evidence for or against their truth). Kant had argued that it was possible to have synthetic a priori statements – whereby fundamental empirical truths are determinable by pure reason. The Vienna Circle rejected this: pure reason may only give us logical and mathematical truths. It is impossible to learn anything substantive about the world, simply by thinking in an armchair.
Importantly, the “metaphysical” statements now rejected were deemed to be only “cognitively meaningless”: unable to express truths or falsehoods and therefore unable to express knowledge. They were not held to be meaningless altogether. A host of statements that are neither tautologous nor backed by empirical evidence may be central to a believer’s religious experience, say – but this is down to the “emotive meaning” of the statement, not its truth content. Likewise, art and poetry were to be valued for their aesthetic and expressive properties. But for the logical positivists, philosophy properly belonged with science, not art, and had to be conducted transparently.
The logical positivists’ defence of science was uncompromising at a time (once thought well past) when quackery and scepticism were flourishing. But when they abandoned strict or complete verificationism and allowed for mere testability (fallible testing), they discovered that it was difficult to formulate a logically precise criterion (either plainly metaphysical statements slipped through the net or scientific statements were excluded). Verificationism was long considered a philosophical failure, though the idea that Carnap’s own reformulation is amendable as required has recently gained traction among some philosophers. In any case, one common charge against verificationism is that the principle is self-refuting: being neither tautologous nor supported by empirical evidence, it appeared to rule itself out from being meaningful. But for Carnap, the criterion was not meant to apply to itself, for it made no assertion but represented a proposal for a language form. To understand this crucial move and its significance we need some background.
By his own account, Carnap was born to deeply religious parents, but it was his mother’s stress on the importance of tolerant humanism that left the more lasting impression. As a student in Jena and Freiburg he was active on the progressive wing of the German youth movement, served in the First World War, welcomed the German Revolution, found himself disappointed by the conservative restauration in the Weimar Republic and immersed himself in his studies.
Carnap’s first major work, The Logical Structure of the World (1928), constituted an attempt to capture all concepts (including those for sense-data, those for middle-sized objects and their properties and relations, and those for the theoretical objects of science) in one hierarchically organized logical system. The work is best understood as Carnap’s attempt to provide an account of the objectivity of physics (and empirical science generally), given its origin in individual sense experiences. Objectivity was located in structures Carnap took all thinkers to share, namely in the logical form of the sentences reporting experiences and scientific findings. Contrary to philosophical folklore, Carnap did not seek to provide empiricism with foundationalist underpinnings – an epistemology that built science on the certainty of sense experience – but rather to safeguard its intersubjective accountability against relativists and sceptics.
Next Carnap turned to the highly technical dispute between logicists, formalists and intuitionists over the foundations of arithmetic in Logical Syntax of Language (1934). In approaching the subject, Carnap had to confront not only the consequences for logicism of Kurt Gödel’s incompleteness theorem (every system strong enough to express classical arithmetic contains sentences that can only be proved in a stronger system which in turn contains undecidable sentences of its own, so arithmetic does not simply reduce to logic) but also the challenge of mathematical intuitionism (which rejected the logical principle of excluded middle: either p or not-p). In response Carnap abandoned the search for one unique epistemological base to cover all of mathematics. This allowed him to finesse Gödel’s incompleteness result by declaring arithmetic to consist of a hierarchy of ever stronger logical languages (languages of greater expressive power) and also to accommodate intuitionism’s highly restrictive logic by assigning it to languages too weak to reconstruct classical mathematics. By abandoning the ideal of one true universal logic that his predecessors still sought to delineate, Carnap introduced logical pluralism.
Carnap’s pluralism was encapsulated in his “Principle of Tolerance”: “Everyone is at liberty to build up his own logic, i.e. his own form of language, as he wishes. All that is required of him is that, if he wishes to discuss it, he must state his methods clearly, and give rules instead of philosophical arguments”. This principle reshaped Carnap’s philosophies of science as much as his conception of philosophy itself. With plural logics came a multiplicity of languages and, confronted with both, philosophy became the “logic of science”: it investigated the languages of science, explored representational possibilities for them and offered recommendations or proposals for revised or new language forms – always in terms of their knowledge function.
Consider how Carnap now spoke of verificationism and empiricism. In “Testability and Meaning” (1936-37) he wrote “It seems to me that it is preferable to formulate the principle of empiricism not in the form of an assertion – ‘all knowledge is empirical’ or ‘all synthetic sentences that we can know are based on (or connected with) experiences’ or the like – but rather in the form of a proposal or requirement”. Empiricism was no longer dogma, but became a logico-linguistic proposal. Carnap continued: “As empiricists, we require the language of science to be restricted in a certain way and hence synthetic sentences are not to be admitted unless they have some connection with possible observations, a connection which has to be characterized in a suitable way”.
The importance of verificationism or empiricism was not diminished by Carnap’s conception of them as “language choices”; they remained at the centre of his philosophy. This was not to pre-suppose empiricism uncritically, but to discard the question of its viability as unfruitful. Just as Carnap foreswore comprehensive foundations for mathematics, so he abandoned the idea that philosophy can ground all empirical knowledge in sense experience. (To be sure, sense experience is causally indispensable, but the logic of science concentrated on the propositional output of scientific activity, not its psychology, sociology or history. Such questions were instead dealt with by the complementary “pragmatics of science”, pursued by his colleagues Philipp Frank and Otto Neurath.)
With no tools at its disposal beyond those also employed by the formal and empirical sciences, Carnap’s philosophy became science in the reflective mode. As one of the sciences of science, the logic of science investigated the formal characteristics of scientific concept formation and methodologies of hypothesis-testing and confirmation; it analysed historical conceptual frameworks and developed and explored alternative ones. While it was not meant to double-guess science as to the nature of reality, philosophy was not intended to genuflect before science either, but to serve as its critical conscience. Philosophy clarified concepts and frameworks. It did not solve problems in a direct way, but it contributed to the stock of human knowledge and, importantly, to our self-understanding as cognitive agents.
By 1950, Carnap had a name for what philosophy could achieve: “explication”. Explication was what contested concepts required – concepts of which the precise meaning was in dispute, such as, in Carnap’s day, the concept of probability. (Did it designate the relative frequency of events occurring, or the state of epistemic ignorance concerning them?) Carnap’s answer was that there were two concepts of equal validity with different domains of application. For this process of explication, a contested concept was subjected to a preliminary clarification of its intended meaning(s), and was then reconstructed in fragments of logically clear formal languages. But successful outcomes were not guaranteed. Sometimes intended meanings turned out to be incoherent; and typically, the logically sound version of the concept under explication (the explicans) had a narrower field of explication compared to the original (the explicandum).
Significantly, the explicandum does not govern, but only guides, the explication. Success does not require that the explicandum and explicans be synonymous, only that the latter be non-contradictory and practically useful in at least one of the senses the explicandum carried. This avoids paradox and allows progress.
Carnap was criticized not only by philosophers upset by this deflationist attitude towards philosophy, but also by fellow-travellers who thought his methodology still too concessive. Famously, in 1951, W. V. O. Quine attacked the analytic/synthetic distinction (which was essential both to verificationism and explicationism) as groundless dogma. What is worth noting here is that Carnap’s concept of analytic sentences – sentences whose truth is fixed by a semantical rule – makes no claims to express eternal truths of reason. Analytic truths, as he conceived them, are language-relative and can be discarded by speakers by changing their language. (That language change can be revolutionary, Carnap noted long before Thomas Kuhn wrote The Structure of Scientific Revolutions, 1962.)
Carnap’s explicationism has applications that go far beyond the philosophy of science: the potential reach of his work extends into various spheres of human affairs. The language constructor’s approach can be applied to areas of social life, to the means of public communication and political debate, where contested concepts abound.
Some thirty years ago, Richard Creath designated Carnap’s vision of his new role for philosophy “conceptual engineering”. Independently, “conceptual engineering” has since become a popular catchphrase in contemporary philosophy. Carnap’s historical role is not always recognized, but that should not detract from the centrality or the radically modernist nature of his work. For him, a clear – but partial – replacement was much to be preferred to a more comprehensive – but vague – original. His explications aimed to increase accountability by increasing transparency: the conditions of correct applicability of an explicans had to be intersubjectively available and decidable. One could do worse than recommend Rudolf Carnap’s take on conceptual engineering to our multiply challenged postmodern times.
Thomas Uebel is Professor Emeritus at the University of Manchester