Chapter 1 - Carnap’s Noncognitivism

Summary

With a view to highlighting the importance of archival and less well-known primary sources for our understanding of Rudolf Carnap’s philosophy, this chapter investigates several examples of concrete influences on his thinking, from nineteenth-century Herbartianism and empiriocriticism, the German Youth Movement, Bauhaus modernism and the revolution from the right, toward the Vienna Circle and post-WWII analytic philosophy. These examples demonstrate that Carnap’s philosophy had always been shaped by practical motives; he developed a philosophical stance that is directed at the reality of life and integrates cognitive as well as non-cognitive elements. This meta-philosophical view that carefully investigates the borders between the scientifically comprehensible (viz., the cognitive) and those aspects of reasoning that merely comprise personal attitudes (viz. the noncognitive) developed through various stages, from the "scientific world-conception" and antimetaphysics of the Vienna Circle toward Carnap’s mature views on inductive logic and human decision-making. The upshot is that noncognitivism as being understood by Carnap and his philosophical allies rather than denouncing value statements as arbitrary and irrational embeds them into a rational scientific discourse, to maximize rationality in connection with moral and political decision making.

1.1 Introduction

With a view to highlighting the importance of archival and less well-known primary sources for our understanding of Rudolf Carnap’s philosophy,¹ this chapter investigates several examples of concrete influences on his thinking. These influences, which enable us to better understand Carnap’s philosophical views, include a broad range of heterogeneous movements: nineteenth-century Herbartianism and empiriocriticism, the German Youth Movement, Bauhaus modernism and the revolution from the right, as well as the Vienna Circle and post–World War II (WWII) analytic philosophy.

While Carnap is best known for his work on logic and the philosophies of science and language – that is, as a theoretical philosopher – it is fair to say that his philosophy had always been shaped by practical motives as well (Richardson 2007). This becomes evident not only in Carnap’s involvement in a variety of political activities (see the contribution by Audrey Yap to this volume [Chapter 4]) but also in his development of a philosophical stance that is directed at the reality of life and integrates cognitive as well as noncognitive elements – as such, this stance is as theoretical as it is practical/political.

At the heart of Carnap’s philosophy lies his noncognitivism. The origin of this philosophical agenda is the Vienna Circle’s so-called criterion of meaning (Sinnkriterium), which argues that statements being neither completely verifiable/falsifiable nor logically true/false are meaningless.² Put in the more careful language that Carnap adopted in his later years, such statements can be “interpreted neither as factual nor as analytic (or contradictory)[; they are] devoid of cognitive meaning, and therefore the distinction between truth and falsity is not applicable to [them]” (RSE, 999). Noncognitive statements represent attitudes rather than epistemically justifiable claims. They comprise a very broad spectrum of statements: moral, political, and aesthetic value statements (RSE, 999–1013), as well as all kinds of scientific attitudes. The latter include both theoretical statements (Carnap 1962) and “linguistic frameworks” (ESO).

There are important differences between value statements and scientific attitudes. Even though a scientific attitude, not unlike a value statement, can neither be absolutely true nor false, scientists can still epistemically justify it. If the attitude is a scientific theory, then its empirical predictions can be tested; if the attitude is a linguistic framework, then the experts are able to identify it as more or less “expedient” or “fruitful” (ESO, 31). The application of tests (theories) and practical trials (frameworks) is crucial for every scientific attitude and the reasons for scientific attitudes being either accepted or rejected by the scientific community tend to be clear and consistent. However, even protocol statements and proven mathematical theorems are never absolutely cognitive,³ seeing as our senses and reasoning abilities might fail. Yet there does exist a strong cognitive backup: A statement, theory, or framework becomes justified via confrontation with the available facts.

In the case of values, we cannot match the cognitive backup that justifies scientific claims and attitudes, for where values are concerned, our final word tends to be a matter of sentiment rather than cognitive evidence. Cognitive backups nonetheless do play a crucial role here, as there are various factors that might influence people’s personal values. For instance, do you still stick to a certain attitude once you understand the consequences it will bring about, once you know about its logical relationships (the statements it logically implies and the statements from which it can be logically derived), or once you recognize its links to the values of other individuals in your social group? We may certainly criticize people who seem immune to scientific and rational advice of that kind as being illogical and anti-scientific (see Section 1.3); at the same time, we have to acknowledge the importance of sentiment when it comes to values. A value statement cannot reasonably be called “true,” “corroborated,” or “fruitful” – we can only say that we adopt it in a most reasonable and rational way: After considering all relevant cognitive information we identify this value statement as “our pure optative” (RSE, 999–1013). Expert knowledge is indispensable to ensure value statements being reasonable. However, a value statement is in itself neither knowledge nor a matter for experts at all. Every moral question is relevant for everybody and therefore everybody is a lay person when moral rather than scientific questions are due. Values are positioned within an epistemic spectrum that gradually moves from the almost entirely cognitive toward the almost entirely noncognitive (Table 1.1).

Table 1.1 The epistemic spectrum of values

Protocols and proven mathematical theorems	Scientific attitudes:		Value statements
	Scientific theories	Linguistic frameworks
… are almost entirely cognitive due to reproducing empirical facts and logical truths.	… enable predictions, procedures of testing, and corroboration.	… can be identified as more or less expedient or fruitful.	… are a matter of noncognitive choice of a lay person, although there are cognitive factors involved.
… are considered true by the relevant experts on the basis of observation or proof.	… are considered corroborated by the relevant experts on the basis of their inductive intuition and empirical evidence.	… are considered expedient or fruitful by the relevant experts based on their scientific experience.	… are adopted by a lay person or a group of lay persons after carrying out a cognitive backup that may include expert knowledge.

Carnap and other logical empiricists discussed two caveats to this relativist view on value statements. First, one might think that there is a third form of epistemic justification, in addition to empirical and logical justification; namely, synthetic a priori truth: Though failing to be empirically or logically justifiable, value statements are justified at the level of synthetic a priori reasoning and therefore can be seen as cognitive. As anti-metaphysical thinkers, all logical empiricists reject this idea. Second, one might conceive a situation where value statements are perceived as objective once they – for reasons such as cultural development or divine mercy – are shared by everybody (or every reasonable person at least). This quasi-realism is nothing but a fallible empirical claim: “There are value statements that every reasonable person actually shares.” As an empirical claim, this idea is an option for the logical empiricist. Indeed, in his later years Carnap sympathized with the “optimistic opinion” of his friend Wilhelm Flitner who claimed “that through the development of culture first in smaller groups, then in nations etc., and finally in the whole of humanity a shared system of values emerges” (Carnap in preparation, entry on September 7, 1964).

This is exactly the point where noncognitivism and naturalism overlap. Moral objectivity becomes an empirical claim stating that all reasonable individuals tend to share certain value statements. If we accept this to be empirical (rather than logical or metaphysical), then the following philosophers appear to be noncognitivists who disagree with Carnap only insofar as they put a different weight on the option of universal consensus: theologians in the Herbart tradition such as Carnap’s grandfather Friedrich Wilhelm Dörpfeld (see Section 1.2); naturalist theologians à la Henry Nelson Wieman, who enthusiastically subscribed to Carnap’s noncognitivism;⁴ noncognitivists in the style of Franz Brentano and Oskar Kraus, who thought that all reasonable individuals must share the very same attitudes (Damböck 2022a, section 2.1); Gibbard-style quasi-realists (Gibbard 1990); and pragmatists such as Dewey (1939). Dörpfeld, Wieman, Brentano, Kraus, Gibbard, Dewey, and Carnap share the idea that consensus among all reasonable individuals as regards certain fundamental value statements is an empirical fact rather than a consequence of certain synthetic a priori judgments or logical truths.

Carnap’s noncognitivism is a philosophical theory about the relationship between science and values. This theory is political in a specific and far-reaching (as well as timely) manner, in that it proposes the following value as being relevant to absolutely everybody: You must adopt all possible strategies to maximize rationality, rather than act illogically (see Sections 1.4 and 1.5). Though scientifically sound, this is no longer mere internal scientific advice but rather a value whose adoption the “scientific humanist” (IA, 83) recommends to both the scientist and the lay person. As John Dewey put it and as it was later echoed by Carnap:

[The] science that is put to distinctively human use is that in which warranted ideas about the nonhuman world are integrated with emotion as human traits. In this integration not only is science itself a value (since it is the expression and the fulfilment of a special human desire and interest) but it is the supreme means of the valid determination of all valuations in all aspects of human and social life.

(Dewey 1971, 446, original emphasis)

Happiness, moral responsibility, and aesthetic pleasure are purely noncognitive matters insofar as everybody must evaluate these matters for themselves. A scientist is no more competent here than a layperson. However, to study the cognitive aspects of values before we accept or reject them is something we can very well learn from science. And exactly for that reason is science (or “scientific humanism“) both a value in its own right and a political enterprise that is relevant for the whole of society.

To think that there exists an aporia between noncognitivism and political engagement/moral responsibility was a trivial though widespread misunderstanding among those who erroneously assumed that noncognitivism implies the abstinence from taking a stance (cf. Hegselmann 1979, 59–66). Being free to take a stance does not mean, however, that scientists in their role as experts are asked to take a stance always and everywhere. In their role as experts, scientists are certainly wise to only propagate those scientific attitudes that directly belong to their field of competence. The scientist’s genuine moral sentiments may exclusively refer to science:

We too, have “emotional needs” in philosophy [and science], but they are filled by clarity of concepts, precision of methods, responsible theses, achievement through cooperation in which each individual plays his [sic] part.

(Aufbau, xvii)⁵

Once scientists go beyond recommending well-established scientific findings of their field, they stop acting as scientists and become laypersons. Scientists who fail to reveal this difference were criticized for good reason (see Weber 1919). Carnap reformulated Max Weber in an anti-metaphysical fashion: People who sell their layperson opinions in scientific disguise are not to be labeled as great philosophers but rather as bad scientists and philosophical charlatans who produce “meaningless metaphysics” (Carnap 1932/2004) and “opium for the educated”;⁶ the only reasonable moral sentiment that is genuinely scientific is science itself. The “scientific world-conception“ is the ridge walk of the politically engaged scientist who seeks to negotiate their position in between the abyss of philosophical charlatanry and apolitical technocracy.

1.2 Friedrich Wilhelm Dörpfeld: Herbartian Roots of Noncognitivism

Carnap’s intellectual development started surprisingly slowly. Before World War I (WWI), he was more interested in the social activities of the German Youth Movement than intellectual classroom activities (Werner 2003, 231–307). Since the war later interrupted his growing intellectual ambitions, he only arrived at the decision to aim for a career in science in 1920 at the age of twenty-nine (Carnap 2022a, Einleitung). Nevertheless, during his studies in Jena and Freiburg before and immediately after WWI, he not only gained a solid education in experimental physics and took classes in philosophy (with Herman Nohl, Bruno Bauch, Heinrich Rickert, and Jonas Cohn) and logic (with Gottlob Frege) but also studied some of the works of leading philosophers and scientist such as Ernst Haeckel, Ernst Mach, Wilhelm Ostwald, Gustav Theodor Fechner, Henri Poincaré, Wilhelm Wundt, Moritz Schlick, and Paul Natorp (Carnap 2022a, Leselisten). Some key influences on Carnap’s early intellectual development also came from his own family. In the extended version of his autobiography, he mentions two relatives of his mother’s family whom he “regarded from childhood as models of men.”⁷ These were his uncle Wilhelm Dörpfeld, the famous archaeologist, and his grandfather Friedrich Wilhelm Dörpfeld (Siegetsleitner 2014, 92; Damböck 2022c; Heidelberger 2024), a pedagogue firmly rooted in the Herbartian tradition, who was quite famous in his time as well. Grandfather Dörpfeld accumulated a remarkable corpus of philosophical and pedagogical writings that were collected in an eleven-volume edition at the prestigious publishing house of C. Bertelsmann. Carnap’s mother Anna Carnap wrote a biography of her father when Carnap was a child (A. Carnap 1897) and taught him and his sister Agnes at home. This she did in the spirit of her father,

[who] had always strongly emphasized that in the education of a child’s character, the moral principles should be based only on the child’s own conscience and not on God’s will. He criticized the church severely for making ethics dependent upon theology because once young people would begin to doubt the dogmas, they would also be in danger of losing their moral ground.

(UCLA CM3 M-A5, A8–A9)

Anna Carnap read to her children from her father’s main philosophical work On Ethics (Dörpfeld 1895), a book that Carnap himself also read several times as an adult, in 1920 and 1964.⁸ The key idea of Dörpfeld’s book is that it is impossible to arrive at moral judgments at an epistemic level, either by way of induction and deduction or through metaphysics. Rather, moral judgments are based on “a certain feeling” that the “purely objective observer,” whose view “may not be clouded and affected by any personal interest, partisan sympathy, or antipathy etc.” (Dörpfeld 1895, 14), develops. We ought to trust these feelings alone. Objectivity is not meant here in the semantic sense of “correspondence with an external objective reality” but in the sense of an “objective” attitude of impartiality or disinterestedness. In referring to his mother, Carnap said: “What convictions, including religious beliefs, anybody had, was for her a morally neutral matter, as long as he would seriously search for the truth and in the forming of his convictions follow his best insights.”⁹

Although Dörpfeld embedded his ideas in a deeply religious world view, he clearly was a noncognitivist, pretty much in the same way that Carnap later adopted himself. Two crucial theses of Carnap’s later noncognitivism can already be found in Dörpfeld. His first thesis was:

(NCOG)

Moral statements are not based on any cognitive insight, they are neither (1) empirically nor (2) logically nor even (3) metaphysically justifiable; because there is no objective reality to which they correspond, they are not truth-apt.

Option (3) – that is, the metaphysical insights – falls away for Carnap and Dörpfeld, because they both adopt an anti-metaphysical stance and reject the idea that there could be any cognitive knowledge based on either logical or empirical insights. Options (1) and (2), in turn, form what he describes as cognitive statements, namely, statements that are either factual or logical (RSE, 999). Both Carnap and Dörpfeld understand moral statements as being noncognitive, as they can be justified on neither factual nor logical grounds.

Carnap’s second thesis was:

(IMPART)

We should trust only impartial and genuine attitudes.

“[V]alue statement[s] express more than merely a momentary feeling of desire, liking, being satisfied or the like” (RSE, 1009), as he points out with reference to John Dewey; rather, value statements ought to express “satisfaction in the long run” (RSE, 1009). In matters of moral discourse, we may only trust our own feelings, which is why we should ensure that what we feel now is a reflection of our genuine attitude. As such, we are not in danger of entirely changing our views tomorrow, after today’s feelings have turned out to be nothing more than momentary emotions or even something that was imposed upon us by others (Damböck 2022a, section 3.1). Whether impartiality in this sense also means that everybody who only trusts their impartial feelings also shares all feelings with every other impartial individual is ultimately an empirical question. People like Dörpfeld, Brentano, Kraus, and Gibbard seem to think that the answer to this question is yes; Carnap and Flitner are more careful, even though they tend to sympathize with such a view. (Cf. the remarks at the end of the Introduction as well as at the end of Section 1.5).

While sharing the above-mentioned two aspects of Dörpfeld’s noncognitivism, Carnap’s metaethics diverges from Dörpfeld in that Carnap rejects the idea that people who act impartially and in accordance with their very own feelings are necessarily bound to arrive at the same moral judgments. If we all arrive at similar moral judgments then this is just a matter of empirical coincidence and not a matter of divine mercy or logical or transcendental reasoning. For Dörpfeld, the matter was different: if we manage to follow our own authentic feelings and attitudes, this means nothing less than following the voice of God, which must be the same for every individual. In Dörpfeld’s view, therefore, moral noncognitivism is inevitably linked with a religious world view and moral objectivism that follows from the unequivocalness of God’s voice (Damböck 2022c, section 2). Carnap, by contrast, defines as central to his view the understanding that (even if there is an empirical convergence of some kind that unites the humanists in some widely shared moral ideals) we must at least consider the logical possibility that even those who share the very same knowledge may adopt diverging moral views:

It is logically possible that two persons A and B at a certain time agree in all beliefs, that their reasoning is in perfect accord with deductive and inductive standards, and that they nevertheless differ in an optative attitude component.

(RSE, 1008)

In contrast to Dörpfeld, Carnap claims:

(DISAG)

It is logically possible that two persons that agree on all matters of rational belief might still diverge in questions of moral attitude.

There are numerous other currents of thought that might have influenced Carnap’s noncognitivism (Carus 2021, 2022). Yet Carnap himself highlights the importance of Dörpfeld’s views in his autobiography. He read Dörpfeld again in 1920 and referred back to him in 1964. It is therefore very likely that the three cornerstones of his moral noncognitivism as highlighted above directly stem from Dörpfeld: (NCOG) and (IMPART) appear as results of Dörpfeld’s positive influence, while (DISAG) presents as a statement that Carnap formulated in negation of the religious views of his grandfather.

1.3 Hans Freyer versus Bauhaus: Revolution from the Right Negated

The influence of Hans Freyer – who was a friend of Carnap in the early 1920s – on the views in the Aufbau on mental objects is already well investigated (Dahms 2016, 172–179; Tuboly 2022). In this section, I will therefore focus on another aspect of the philosophy of Freyer – the well-known sociologist and major representative of late-1920s revolution from the right (Muller 1987, ch. 4) – namely, on his metaethical stance (Damböck 2022b).¹⁰ Like Carnap, Freyer was a noncognitivist who shared (NCOG) and (DISAG). At the same time, Freyer’s views were also connected to a strong anti-rational world view, here combining noncognitivism with fascism. This, in turn, allows us to highlight the importance and nontrivial nature of the fact that the rational attitude is so crucial for Carnap’s noncognitivism. Freyer can be seen as the typical defender of a view that Carnap rejects.¹¹

Freyer’s views that combine moral noncognitivism with a fascist and völkisch ideology are outlined in (1926, 108–120; 1930). On the one hand, Freyer (1930, 112) sees the “moral subject” as “the ultimate authority, a solitary judge, an organ that perceives the demands of the world within itself.” This is basically a rephrasing of (NCOG): there is no other reason for or against a moral statement than our feelings, our “inner voice.” Freyer (1930, 112) also adopts (DISAG) while instrumentalizing it in a way that is not intended by Carnap: “The political powers are put into the world to historically realize a closed value gestalt that is contained within a people.” There must be a strong dictator or “Führer” (Freyer 1926) who follows his instincts and puts into practice a new closed value gestalt for his state. Other individuals may also have their inner voices, but Freyer claims that the people/nation – that is, the mass of ordinary non-Führer individuals – must be brought into line with the Führer by way of strong propaganda institutions. “The plain secret of all Führung is this: to take the others as they ought to be and to commit this noble fraud in such a way that they actually become it.” “The people is the Führer’s ever-growing, always renewed work.” (Freyer 1926, 110, 114) In reality, it is only the Führer who follows his inner voice, creating new values that shape his people. For that purpose, a propaganda machinery is needed that manipulates the inner voice of each member of the people in a Führer-conducive way.

The crucial feature of this aspect of fascist noncognitivism is not the emotional side of propaganda. It is not the framing of an announcement with bombastic music, strong words, nice colors, and props. To propagate one’s views and to try “to influence other people by a suitable choice of emotive language” (Ayer 1946, 22) is not necessarily a bad thing. Not only fascist ministries of propaganda use music, words, colors, props, and other emotive sources to make their views appear favorable – everybody has the right to do so, and everybody becomes emotional as soon as there arises a disagreement on questions of utmost moral importance. What makes Freyer’s idea of fascist propaganda special is that it recommends to systematically hide the truth and produce misinformation in order to influence people. The key idea of fascist propaganda, in the sense of Freyer, is the “noble fraud” of making people share our values although we know perfectly well that they only accept them because we manage to hide the consequences of what we propagate. Thus, Freyer’s main principle of noncognitivism as a means of fascist propaganda is this:

(OBSC)

You ought to obscure facts, spread misinformation, and make people act irrationally whenever it becomes necessary to make them share the closed value gestalt of their Führer state.

What Carnap could learn from fascist noncognitivists like Freyer was that the key principle for them, while clearly rejecting (IMPART), was (OBSC): to make people act irrationally, without them being aware of the causal consequences of an action. Therefore, as Carnap realized, the entire way of describing the noncognitivist stance had to change fundamentally. It was no longer sufficient to highlight (NCOG) – value statements are neither factual nor logical – but it needed to be highlighted that, for a scientifically minded person, noncognitive values must relate to cognitive facts. This led to a new way of framing noncognitivism that was entirely different from older varieties, such as those found in the German Youth Movement before WWI, when Gustav Wyneken, Hans Reichenbach, and Carnap all exclusively focused on (NCOG), varieties of (IMPART), and (DISAG) without any explicit appeal to rationality (see Damböck 2022b, sections 2–4).

It was only in 1929 that the picture changed, when Carnap provided the first account of his mature noncognitivism in the context of a lecture entitled “Science and Life” that he delivered at the Bauhaus in Dessau (Dahms 2004, 368–369). Though values still are considered noncognitive in the sense of (NCOG) – “the direction of our acts” is determined by “irrational instincts” (RC 110-07-49, 4) – the main objective of value philosophy now is to highlight the various ways in which science and rationality are important for value statements – not as a “leader [Führer]” (who imposes his intentions upon us) but rather a “signpost [Wegweiser]” (that allows us to identify and follow our own inner voices and aims; RC 110-07-49, 4).

“The valuation itself cannot be found via theoretical knowledge, because it is not the grasping of a fact, but a personal attitude.”¹² Science still plays a role for our determination of “the direction of the will, of practical action” (RC 110-07-49, 2), namely, in a twofold way: “Through reasoning, theory, knowledge, [and] science can and must be examined (1) what are the inner consequences of a valuative attitude … (2) what are the means that we have to achieve an intended aim.”¹³ Whereas a fascist will hardly reject point (2) – seeing as technical-scientific progress is not something that fascism typically wants to avoid – point (1) is entirely ruled out by fascist ideologues like Freyer. Therefore, (1) becomes one of the most crucial features of Carnap’s and other logical empiricists’ branch of noncognitivism. Science becomes a normative demand:

(CAUS)

One ought to investigate the causal and logical consequences of a possible aim and accept that aim only if one is also willing to accept its consequences.

Is my value system consistent (logical consequence)? Am I willing to embrace/accept every probable outcome in a world that fits my value system (causal consequence)? All varieties of Carnap’s practical philosophy – as exemplified in the 1929 lecture, (Carnap 1934), as well as (RSE, 999) – are centered around this fundamental moral commitment: one’s practical decisions must take into account all relevant logical and empirical questions rather than hide them. This message was clearly directed against those who, like Freyer, wanted to set up a system of noncognitivism that explicitly negates rationality. The core of Carnap’s noncognitivism, in other words, is the normative claim that for all practical purposes we ought to maximize rationality and minimize irrationality. This is also the key point of his Bauhaus lecture:

It is wrong to grant the irrational an influence beyond its territory, namely, in the rational: If we do not want to commit fraud ourselves, we must be extra careful in our judgments whenever feelings and our will attempt to seduce us.

(RC 110-07-29, 4)

Looking at Freyer’s take on noncognitivism is an important lesson for a noncognitivist. For, without facing the reality that there exists an understanding of noncognitivism that forcefully and systematically rejects rationality, it may not be possible to fully understand the normative (and fundamentally moral) nature of Carnap’s defense of rationality in connection with moral questions. Carnap’s noncognitivism was clearly developed as an antidote to irrational and fascist modes of reasoning.¹⁴

1.4 The Neurath Circle: Toward a Political Philosophy of Science

In 1929, Carnap delivered the Bauhaus lecture. The Verein Ernst Mach also published the famous Vienna Circle Manifesto (Carnap, Hahn, and Neurah 1929/2012), a text that was mainly written by Carnap and Otto Neurath (Uebel 2012b). This manifesto is very important for the practical and political aspects of the Vienna Circle’s scientific world-conception. Its key passage connects the scientific world-conception with socialism, stating that “in many countries the masses” tend to see a convergence between “their socialist attitudes” and “a down-to-earth empiricist view” (Carnap, Hahn, and Neurah 1929/2012, 90). This development, as the authors of the manifesto suggest, may lead to an entirely new foundation for Marxism:

In the past, materialism was the expression of this view; meanwhile, however, modern empiricism has left behind a number of inadequate forms in its development and has found a defensible form in the scientific world-conception.

(Carnap, Hahn, and Neurah 1929/2012, 90)

This statement was formulated against the backdrop of Lenin (1927), who intended to remove all empiricist interpretations from Marxism and replace them with “dialectic materialism.” Though Lenin was quite successful in his attempt – after 1918, the official communist ideology had nothing left in common with empiricism – his approach was criticized by Austrian social democrats such as Otto Bauer, Friedrich Adler, Max Adler, Rudolf Goldscheid, Otto Neurath, and other unorthodox Marxists (such as Karl Korsch) who were affiliated with the logical empiricist movement.¹⁵ Though Carnap hardly engaged with Lenin’s discussion, the principal challenge was clear: Carnap, Neurath, and the Austro-Marxists were convinced that (a) dialectic materialism is ultimately meaningless metaphysics and unsuited as a proper basis for rational politics; therefore, (b) the only possible road for the scientific world-conception is a strictly empiricist one, arguing along the lines of those currents of thought that Lenin rejects; hence, the new scientific world-conception has to offer a nonmetaphysical approach, which, at the same time, can form the basis for a Marxist (or social democratic) world-conception.

The immediate result of the manifesto’s vision was the founding of a new discussion group, the so-called Neurath Circle, which met on Mondays and stayed active for several months in the first half of the year 1930 (Damböck 2022a, section 2). Neurath and Carnap utilized the Neurath Circle to start rephrasing their anti-metaphysical and noncognitivist views, now calling “Marxist” everything that they had previously called “scientific.” This meant that this new version of Marxism was indeed foundationless insofar as it was based on noncognitivism, which “only restricts the way in which the Marxist individual may justify their actions in a Marxist manner.”¹⁶ Marxism is reduced to a fallible empirical prediction: if the working class is given access to higher education and freedom of opinion, then it will choose in free elections those parties that propose socialist values. In dialectical materialism, by contrast, the socialist values become identified as absolute truths being justified by synthetic a priori judgments. Ideology thus becomes deflated: “Marxist reasoning makes it impossible to derive demands from reasoning and to create a ‘Marxist ethics’ in this way.”¹⁷

The great question that is now left to ask is in what sense this deflationary form of Marxism – which is only seen as the empirical prediction of success of a value system, no epistemic justification of this system is available at all – is still Marxist and, indeed, in what sense it is political at all. My interpretation is similar to the one articulated by Romizi (2012) and Uebel (2012a). I share with Uebel and Romizi the rejection of S. Richardson (2009a), who attributes the following view to the “Left Vienna Circle”:

(NEUT)

A politically neutral philosophy of science (PNPS) is defended. (PNPS) is compatible with both the political program and practical rationality (PCAUS) propagating (CAUS) being embraced by Carnap and Neurath. However, (PNPS) would also be compatible with a political agenda (POBSC) propagating (OBSC) and negating (PCAUS), as well as with an entirely apolitical world view.

(NEUT) states that the theoretical views of the (left wing of the) Vienna Circle are political only insofar as they do not explicitly rule out (PCAUS). Defenders of (PNPS) can also be defenders of (PCAUS), but – as noncognitivists – they could very well also adopt an entirely apolitical stance or even defend a different political agenda (POBSC) that is totally at odds with (PCAUS).

The picture I draw here and Section 1.3 suggests a different view. The approach Carnap (together with Neurath) recommends is political not despite but rather because of the specific form of noncognitivism it promotes. Carnap’s theoretical and political doctrine is twofold: (a) one must take into account all available scientific knowledge and base one’s decisions on it; (b) one must reject every pseudo-justification that cannot be carried out on mere scientific grounds. The political implications of this doctrine are massive, because (a) and (b) force us to reject the cornerstones of totalitarian politics that I outlined in Section 1.3. In other words, there is a political agenda (PCAUS), which is logically implied by the Vienna Circle’s non-neutral philosophy of science (PNNPS). The result is an alternative to (NEUT) that now ascribes to the Vienna Circle a political agenda in a much more specific sense:

(NNEUT)

A politically non-neutral philosophy of science (PNNPS) is defended, which forces us to accept the political agenda (PCAUS) and reject (totalitarian) alternatives (POBSC) even when the defender of (PNNPS) remains passive in the political arena (in the sense of refraining from openly propagating measures in the field of [PCAUS]).

When Carnap states in his autobiography that his political convictions are independent from his philosophical views – and that it was a key stance of the Vienna Circle to keep these things separate (IA, 23, 82–83; cf. Uebel 2012a, 133) – he certainly means this in the way I just specified. An advocate of (NNEUT) does not necessarily have to be a social democrat and an atheist like Carnap – there are other options. But an advocate of (NNEUT) cannot defend any political view that is incompatible with (PCAUS).¹⁸

Those instances of (CAUS) that have a more publicly oriented nature are here labeled (PCAUS), which includes any practical attempt to spread the scientific world-conception throughout society and culture. Whereas Neurath was quite active at the level of (PCAUS) throughout his life, Carnap hardly ever engaged at this level. Nonetheless, Carnap was certainly more prolific at the level of (PNNPS) than Neurath: Carnap’s agenda was the development of the theoretical foundations of a politically non-neutral philosophy of science, without much engagement at the public level of (PCAUS).

1.5 The Capstone of Carnap’s Philosophy

How did Carnap proceed? As it became obvious already at the end of the 1920s, the question of confirmation or, as Carnap then called it, “constitution of the non-given”¹⁹ is crucial for both theoretical and practical matters. This is the case because most of the statements we are making in real-world conditions are neither purely logical nor purely empirical. Therefore, the following question arises: to which degree might any such statement $h$ be supported or confirmed by our present empirical knowledge $e$? Carnap spent several decades of his philosophical career investigating this question, specifying functions of the type $𝔠\left(h,e\right)$ that express the degree of confirmation or “logical probability” of $h$, when faced with evidence $e$ (Sznajder 2018).

There is a crucial aspect of Carnap’s theory that I can only touch in passing, namely, the question of what exactly is empirical and how we constitute the set of empirical knowledge $e$ of a person or group. In short: The Vienna Circle’s protocol-sentence debate led Carnap to overcome his initial prioritization of the autopsychological and to finally arrive at a thing language empiricism that no longer provided room for traditional Erkenntnistheorie (Uebel 2021). The goal of this process of overcoming Erkenntnistheorie was to develop a language that enables us to get a grasp of the empirical, in an intersubjectively determinable way. The Aufbau failed to serve that purpose because of its autopsychological basis. Only the subsequent choice of a heteropsychological basis allowed for a consistently practicable empiricist strategy; and it is precisely here where Carnap’s philosophy finally converges with empiriocriticism. That this is the case was also acknowledged by Carnap in his autobiography (IA, 16, 18, 20, 45, 50); and, most importantly, in the preface to the second edition of the Aufbau, where he indicated that he would no longer use elementary experiences as basic elements “but something similar to Mach‘s elements, e.g. concrete sense data, as, for example ‘a red of a certain type at a certain visual field place at a given time.’” (Aufbau, vii) The key aspect of this shift is certainly the transition from the autopsychological to the heteropsychological. In 1967, “sense-data” no longer refers to empirically unfounded “phenomena” but rather to the empirically detectible elements of a heteropsychological thing language in the spirit of Avenarius and Mach.²⁰

In Carnap (2017), a sketch from 1958 that was only recently published by A. W. Carus, we are shown how the framework for human decision-making that Carnap developed in the final four decades of his life is directly related to the initial metaphilosophical proposal from the Bauhaus lecture and Carnap (1934). Carnap addresses the notion of a “possible history of the world” $W$, here represented as a set of atomic propositions of the thing language that specify a causally possible world history (2017, 191). If $W_T$ is “the true history,” that is, everything that empirically characterizes our world up to the present day, then we intuitively obtain a bundle of possible histories of the world that characterize different possibilities of how $W_T$ might develop in the future.²¹ Carnap evaluates these possibilities by means of the function $𝔠\left(W,e\right)$ where $W$ is a possible history of the world, $e$ is the present empirical knowledge of an agent, and $𝔠\left(W,e\right)$ is the probability that the agent ascribes to $W$ on the basis of the empirical evidence $e$. This function allows us to evaluate all causal possibilities in light of our present knowledge.

The second building block of Carnap’s proposal is what he calls a “value function“ $V$. This function assigns to each possible history of the world a value $V\left(W\right)$: if $W$ is more favorable to an agent than $W^{\prime }$ then it must hold that $V\left(W\right)>V\left(W^{\prime }\right)$ for the agent’s value function. What distinguishes value functions representing moral attitudes from any other value or utility function is this: “A person X at a given time has not just a single value function, but a great many of them,” all representing local preferences regarding matters of politics, health, economics, and the like. At the same time, Carnap thinks that there is one “comprehensive value function of X that comprises all aspects, and in which the relative weight of each aspect in any possible overall situation finds expression.” It is this comprehensive value function – where all conflicting moral matters must be resolved and weighted – that Carnap (2017, 192) recommends to use as a basis for “moral value judgments.” Comprehensiveness is crucial here, because only a function that weights all conflicting interests and value systems is a proper basis for practical decisions that reflect the personality of an actor. Which relative weight do I, personally, put on issues such as dietary measures, health, economic growth, environmental sustainability, employment, individual pleasure, and social life? It is everyone’s very own moral duty (as it is the moral duty of politics and society as a whole) to arrive at decisions here, in every single case. Only comprehensive value functions will serve that purpose.

Besides comprehensiveness, there are some additional features of moral value functions that Carnap (2017, 193) finds desirable, namely, (1) that value functions are “derivable from general principles” – there must be certain main values from which others are logically derived – and (2) that they are “expressible by mathematical functions … that are continuous and relatively smooth, rather than jumping up and down.” The latter feature seems to somewhat converge with (IMPART) – since only impartial and genuine attitudes are pure and long-term – but Carnap’s proposal remains quite sketchy here.

What makes the 1958 manuscript so special is that it demonstrates something that other writings by Carnap on the topic of human decision-making hardly touch: namely, that he had intended the famous function for calculating the preference $𝔓$ for an action $a$ as a framework for human decision-making. It is a framework that brings together the moral/noncognitive and the rational/cognitive side of preference formation. We calculate

$𝔓\left(a\right)=\sum V\left(W\right)𝔠\left(W,a\cdot e\right)$

for each possible action and then choose the action $a$ that receives a maximum value for $𝔓\left(a\right)$.²² Here, $a\cdot e$ is the evidence that results from $e$ when the action $a$ is adopted. An action is “perfectly rational” if it results from a calculation like this that uses the best possible normative standards at all levels: regarding inductive logic, deductive logic, and value functions (Carnap 2017, 192–194).

What we learn here is that in his later years, Carnap identifies the two main philosophical problems that remain important to him as inductive logic and the analysis of value functions (Carnap and Hochkeppel 1993, 143–147). We also learn that these two realms are combined in a framework of decision theory that covers the entire spectrum of cognitive theoretical questions and noncognitive practical decisions.

For the purpose of this chapter, I will define the crucial point of Carnap’s considerations on value functions as follows: value functions show that and how the political stance of any individual is both (a) an essentially private matter and (b) affected by the scientific world-conception. That value functions need to be comprehensive and “smooth” is required by the scientific world-conception, in the same way that we always need to take seriously the available evidence and strive to act rationally instead of anti-rationally. This directly follows from the scientific doctrines of (PNNPS). Yet even while accepting all of these rules, we are still free to choose a value function that meets the criteria of (PNNPS): we obtain (DISAG) as the noncognitive limes of a scientific world-conception.

This also allows us to rethink the seemingly neutral and meager considerations on “scientific humanism“ that Carnap formulated at the end of his autobiography (IA, 83). His attitude can be rephrased as the claim that everybody should always try to make their decisions in a maximally rational way. True, there remains the possibility of diverging value functions (DISAG). On the other hand, how likely is it that people knowingly adopt value functions that are entirely misguided from the standpoint of democracy and human rights? The real fascists arrive at catastrophic decisions not because they knowingly adopt catastrophic value functions but because they unknowingly act illogically and irrationally (Carnap 1937). It seems very likely that the only thing that we need to do to change the world for the better is to make people act in accordance with “scientific humanism.” The pragmatist and naturalist optimism of noncognitivist reasoning implies that only a demographically irrelevant minority (which can be kept under control in a democratic society where everybody is willing and able to carry out cognitive backups) would still mess things up by using anti-humanist value functions. Scientific humanism is all we need to insist on, because the real threat to political progress is not the existence of deviant moral systems but people’s unwillingness or inability to act rationally.