Introduction

This section has two aims. The first is to show that, in his critical reflections on ethnology and history, Wittgenstein held views that, by the standards of his time, would have been classified as forms of “cultural relativism” and “historicism.” My second, related, aim is to determine what Wittgenstein meant when he claimed to be using “the ethnological approach.” I shall try to show that this approach inherits the relativism of the critical reflections on ethnology and history.

My section is indebted to Peter Hacker’s paper “Wittgenstein’s Anthropological and Ethnological Approach.”Footnote ³ I agree with Hacker that Wittgenstein’s turn toward this approach occurred in the early 1930s, and that two important thoughts constitutive of this perspective were the “primacy of action and practice” and “a powerful historicist point of view.”Footnote ⁴ I concur with Hacker’s thesis that Oswald Spengler was important for Wittgenstein’s historicism. And I recognize the value of Hacker’s term “historicism without history” for characterizing Wittgenstein’s frequently used technique of inventing ways in which history might have developed.

Still, there are some respects in which I go beyond Hacker. I believe that in order to fully understand Wittgenstein’s anthropological and ethnological approach, we need to focus not just on human universals but also on deep-seated cultural differences. The importance of such differences can best be appreciated by attending to Wittgenstein’s overt and implicit criticisms of influential forms of ethnology. Moreover, I find it surprising that Hacker does not investigate the relationship between the anthropological and ethnological approach and forms of relativism central in anthropology and ethnology. This is all the more unexpected since Hacker elsewhere defends a Wittgensteinian version of conceptual relativism.Footnote ⁵ Nor does Hacker consider relevant that in late nineteenth- and early twentieth-century European culture, historicism was widely perceived as a relativist view. And finally, it is easy to overstate the significance of “historicism without history.” There is more history in Wittgenstein’s historicism than Hacker acknowledges.

Here is the plan for the rest of the section. I start by outlining Wittgenstein’s responses to some facets of the evolutionist and ethnocentric anthropology of color prevalent at Cambridge at the beginning of the twentieth century. I shall argue that these responses were in line with those of influential relativist anthropologists of the time. Subsequently, I shall pursue a similar project with respect to Wittgenstein’s comments on James George Frazer’s classic The Golden Bough.Footnote ⁶ This will be followed by some observations on Wittgenstein’s reflections on Spengler and history. I will conclude by summarizing my results and by characterizing Wittgenstein’s anthropological and ethnological approach in their light.

The Anthropology and Grammar of ColorFootnote ⁷

During 1898–1999, a group of British anthropologists and psychologists visited islands in the Torres Strait, a waterway between Papua New Guinea and Australia. One of the central figures was the Cantabrigian W. H. R. Rivers, who studied the islanders’ color classification and perception. In the report of his results, Rivers later claimed to have shown that the islanders were on a lower level of evolution compared with Europeans since they (allegedly) were unable to distinguish blue from green, gray or black.

In one of Rivers’ experiments, dozens of skeins of wool of various color shades were placed in front of the islanders, and they were asked to sort the skeins into seven piles. Each such pile had a specific color shade as a starting point: red, green, pink, pale green, yellow, blue, and violet. Rivers found that “confusion between green and blue was very common and also between blue and violet.”Footnote ⁸ In another experiment, the islanders had to name, in their own languages, the colors of standardized color papers or items in their environment. Rivers sums up as follows: “there was great definiteness and unanimity in the nomenclature for red […] and very great indefiniteness for blue.”Footnote ⁹

Rivers sought to explain these observations in various ways. Some were physiological (invoking differences in the pigmentation of the retina), some environmental (claiming that blue pigments were rare on the islands), some social-evolutionary: “the savages” lack an “aesthetic interest in nature … The blue of the sky, the […] blue of the sea […] do not appear to interest the savage.”Footnote ¹⁰

Rivers’ research was both influential and controversial. The psychologist E. B. Titchener contradicted Rivers at every point. For instance, Titchener rejected Rivers’ comparisons between the color discrimination thresholds of the islanders and Englishmen. The differences Rivers claimed to find were likely to be artefacts of the unlike lighting conditions in a Torres Strait islander’s hut and the psychological laboratory in Cambridge. Titchener also challenged Rivers’ English translations of some the natives’ terms.Footnote ¹¹

The anthropologist Arthur Maurice Hocart insisted that to understand color vocabularies one needed to consider the details of the relevant cultures. The “analytic strength” or weakness of a given color terminology could only be assessed properly again this backdrop. For instance, in Central Asia there existed no general term for “horse,” and different types of horses were distinguished on the basis of their color. As Hocart had it, these facts allowed for no conclusions concerning the native speakers’ analytic skills and intelligence. In fact, Hocart suggested that the Asian horsemen’s conceptual abilities were no worse than those of English “horsey persons”: “I think I am right in saying that a horsey person never speaks of a stallion or a mare as a horse.”Footnote ¹²

Franz Boas – the leading figure of American “relativist anthropology” – pointed out that in “many primitive languages the groupings of yellow, green, and blue do not agree with ours. Often yellowish-greens are combined in one group; green and blue, in another.”Footnote ¹³ Boas immediately added that this difference in classification was “without any accompanying difference in the ability to differentiate shades of colour.” The speakers of the respective languages simply found it important to pick out “gall-like colour” (in the first case), and “young-leaves colour” (in the second case).Footnote ¹⁴ In other words, semantic decisions reflect “the chief interests of a people.”Footnote ¹⁵ Such decisions differ from language to language and thus “it happens that each language, from the point of view of another language, may be arbitrary in its classifications; that what appears as a single idea in one language may be characterized by a series of distinct words in another.”Footnote ¹⁶

Enter Wittgenstein. Although he never mentions these anthropologists, there can be little doubt that he was familiar with their work. After all, some of the psychologists who went to the Torres Strait Islands were his teachers in Cambridge before World War I. Rivers was a fellow in anthropology at St. John’s College, a stone-throw from Wittgenstein at Trinity College. Be this as it may, a good number of Wittgenstein’s remarks on color from the 1930s onward can be read as critical remarks on Rivers’ methods and conclusions, remarks that echo the relativist perspectives of Hocart and Boas. Here are a few examples.

The first example vaguely recalls Rivers’ experiment with the matching of color skeins. Imagine the experimental subjects or tribesmen match a green skein to a red skein. Does this prove that they have a deficit in color discrimination?

The similarity to Boas’ comment about leaf-colors is striking. Wittgenstein also suggested that the tribesmen’s color taxonomy might be related to an artistic practice: “A type of painting, in which the illuminated side of figures is always painted green, the shadows always red.”Footnote ¹⁸ Perhaps this was an ironic reference to Rivers’ claim that tribes without an equivalent of our term “blue” lack our sophisticated aesthetic appreciation of nature. For Wittgenstein, the explanation lay with salience and practical goals.Footnote ¹⁹

The following case highlights possible connections between color and social organization:

If Rivers had encountered these tribesmen, he might have concluded that they were situated on a lower evolutionary stage of development (compared with us). Wittgenstein disagrees. He again formulates a vignette that renders the tribesmen’s practice with studied neutrality. The tribesmen act in ways in which – under moderately different circumstances – we might well have acted ourselves. They are not deficient in their thinking. Boas was right.

The next thought-experiment brings the exotic back to Cambridge:

Wittgenstein is not making up these color terms: Light greenish is standardly referred to as “Cambridge blue” and darkish blue as “Oxford blue.” To Wittgenstein’s audience in Cambridge, the terms were of course well familiar. Wittgenstein’s vignette is a joke on Rivers’ expense: Given his general framework, Rivers would have had to hypothesize that the English were on a lower evolutionary scale than many other cultures, since there were contexts in which the English ignored the common features of light blue and dark blue.

Wittgenstein also constructs cases that tempt us to speak of “differences in seeing.” In this vein, he introduces a tribe that identifies colors numerically. Wittgenstein comments: “We cannot imagine what it would be like to associate numerals with colours.”Footnote ²² “One wants to say that one cannot imagine their experience.”Footnote ²³ Elsewhere Wittgenstein reflects on a tribe with “different colour poles.”Footnote ²⁴ By “colour poles,” Wittgenstein means color terms (and colors) around which linguistic communities structure their respective “colour geometry.” For “us,” color poles are red, blue, yellow (possibly also green). We define other colors in terms of these poles. Accordingly, we hold for instance that purple is a bluish red. Wittgenstein encourages us to imagine a tribe with different color poles, say, purple, orange, blue-green, and yellow-green. The tribesmen render red as a purplish orange. Wittgenstein denies that the color-terms of the tribe in question would be translatable into our terms: “This tribe and we couldn’t learn one another’s language.” He parallels such divergence in color poles to the difference between a color-blind and a normally sighted person.Footnote ²⁵ This makes it clear that for Wittgenstein the humans we label “color-blind” need not be rendered as deficient; theirs is just a different manner of organizing color-space. The idea is expressed clearly in another passage: “We speak of ‘colour-blindness’ and call it a defect. But there could easily be several different abilities, none of which is clearly inferior to others.”Footnote ²⁶ In other words, even if we found a tribe without our ability to distinguish between shades of color, we would not be entitled to draw, as Rivers does, conclusions about positions in evolution.

To conclude this briefest of selections of Wittgenstein’s vignettes on color, it remains to mention that he also considers languages without any color terms,Footnote ²⁷ as well as languages in which color is encoded together with shape. Again, the upshot is in line with Boas and Hocart:

Lest I am misunderstood, I hasten to stress that Wittgenstein’s interest in amassing and discussing vignettes of this kind is not purely ethnological. Still, Rivers’ ethnology provides him with a target or foil against which he can work out aspects of uses of color terms in his own culture. Moreover, in Hocart or Boas, ethnology furnishes Wittgenstein him with intellectual allies. Ethnology additionally supplies Wittgenstein with a method of analysis. Wittgenstein uses real and imagined uses of color categories for understanding the variability and arbitrariness of color classifications. Central to this method is a methodological relativism that abhors evaluations and seeks to understand each culture on its own terms, each categorization as relative to its respective cultural background. And finally, the resulting relativism is ultimately more than methodological: It is a relativism about concepts, and – insofar as concepts and cultures were inseparable – a form of “cultural relativism.”

Where then did Wittgenstein deviate from the ethnologists? A first important deviation is that for him ethnology – or cultural anthropology – is not all of anthropology. There is also physical or biological anthropology. That is to say, Wittgenstein takes a keen interest in “general natural facts” that are easily overlooked and yet key for understanding why we have our specific color terms:

The second respect in which Wittgenstein goes beyond ethnology is in his collecting “grammatical” or “logical” observations about color terms. Studying the grammar of judgments concerning color allows the philosopher to recognize implicit and explicit commitments concerning colors in their own linguistic community and beyond. A grammar of color studies which types of color attributions are appropriate and acceptable in which contexts. In order to characterize these types of attributions adequately, several more specific questions and issues needed to be addressed: What are the roles of specific color attributions in various practices? How do specific color attributions relate to other color attributions? And how to specific color attributions relate to attributions of qualities in other dimensions, like texture for example?Footnote ³⁴

The ultimate aim, however, is not to formulate a comprehensive philosophical grammar of color terms. The goal is rather to identify and undo conceptual errors in philosophical and scientific theories about color. A case in point is Rivers’ way of confusing interest-relative color categorization with physiological color-discrimination. One cannot draw inferences about the evolutionary stage of a given culture based on its color taxonomy. What Rivers classifies as a case of color blindness is better taken as simply another organization of colors. Nor is it legitimate to render one’s own color taxonomy or color vision as the standard or framework of analysis for all others. Rivers also underestimates the difficulties that might arise when one tries to translate the color terms of another culture. And finally, Rivers’ methodology runs into difficulties when the investigated languages contain form-color concepts. These insights are part and parcel of a cultural-relativist view of color and obtained by relying on a methodological relativism.

From Frazer to Cultural Relativism

I now turn to Wittgenstein’s most overt engagement with ethnology: his discussion of Frazer’s classic The Golden Bough.Footnote ³⁵ I shall relate Wittgenstein’s critique to that of influential contemporaneous anthropologists, especially those that were classified as “cultural relativists.” My aim is again to highlight Wittgenstein’s proximity to these authors.

Frazer’s starting point is an Ancient Roman tale concerning a gruesome custom at the grove of the goddess “Diana of the Wood,” at the village of Nemi, about thirty kilometers south-east of Rome. At any one time, the grove had one, and only one, priest. Any man could rise to that position, provided only that he killed the incumbent. Frazer claims that this practice had no parallel anywhere else in Classical Antiquity, and that to explain it one has look elsewhere. Frazer approaches the task as follows. He starts from the notion that the Nemi custom was “barbarous.” This suggests to Frazer that it was a product of “the early history of man,” that is, prior to civilization. Moreover, Frazer alleges that early-history societies were “essentially similar […].” It follows that crucial elements or “motives” of the barbarous Nemi practice had once to have existed in many other places, too. Frazer also holds that over time the most gruesome features of social practices are generally replaced by more humane arrangements. Thus, the Nemi practice was, by the time of Classical Antiquity, something of a “survival” of earlier times. In any case, Frazer believes that the Nemi custom is explained by the fact that similar customs were widespread in barbarous times. And finally, Frazer alleges that even in his time, we find institutions that are naturally analyzed as successors to early-history practices.Footnote ³⁶

One set of institutions that Frazer discusses in detail are “fire-festivals” as these still existed in Europe in the nineteenth century. The exact script differed from one location and time to others, but in many of these festivals a large doll was crafted out of straw – typically called “the witch” – and eventually thrown into the flames. A case in point is the “Beltane Festival” in Scotland. Among other peculiarities, it included a ritual with a cake especially baked for this occasion. One of its many pieces was marked in some way, for instance, by containing a small hard object. The person who by chance was served this piece was called “the Beltane witch.” Part of the village “made a show of putting him into the fire,” but the “witch” was rescued by the others. For as long as the festival was fresh in people’s memory, they called the “Beltane witch” “dead.”Footnote ³⁷ Frazer suggests that the nineteenth-century forms of fire festivals were successors of earlier versions in which “witches” were burned not “in effigy” but “in the flesh.” Back then the Beltane witches were not rescued: They died in the flames.Footnote ³⁸

I also need to mention Frazer’s “evolutionism” and his psychological explanations of the differences between magic, religion, and science. Magical thought is governed by the “Law of Similarity” and the “Law of Contact or Contagion.” Both are psychological “associations of ideas.” According to the first law, “like produces like”; according to the second, “things which have been in contact with each other continue to act on each other at a distance after the physical contact has been severed.”Footnote ³⁹ One example of the Law of Similarity is a practice of adoption in which the adopting mother pulls the child through her clothes as if she had given birth to the child.Footnote ⁴⁰ An example of the Law of Contact is the belief that by destroying a possession of an enemy, one could destroy the enemy themselves.Footnote ⁴¹ Overall, Frazer calls magic “a spurious system of natural law as well as a fallacious guide to conduct; it is a false science as well as an abortive art.”Footnote ⁴²

Frazer assumes that magic is older than religion. Religion arose when people came to realize that the tools of the magician were ineffective. This led people to think that natural processes are not governed by magicians but by invisible gods whose favors had to be won. It follows that the course of nature “is to some extent elastic or variable”; it can be changed if the believers managed to persuade the gods to intervene. Religion was eventually and slowly replaced by science. Science shares with magic the belief in unalterable natural laws, but in the case of science these laws were discovered by careful observation and experimentation.Footnote ⁴³

This sketch of Frazer’s theorizing suffices for our purposes. We can now turn briefly to its reception by other anthropologists. This will also allow for glimpses at their alternative methodologies and theoretical commitments. Pretty much all elements of Frazer’s work met with criticism: his evolutionism; his assumption that certain ethnic groups today are “primitive” and thus a window into humanity’s early history; his ethnocentrism; his external perspective on societies (a perspective that ignores how the societies see themselves); his atomism (practices are discussed in isolation from their broader cultural context); his psychological mode of analysis; his reliance on the “comparative method” (comparing practices across Europe and historical periods); his insistence that modern institutions are (what came to be called) “survivals” of earlier, barbarous practices; and his rendering of magic as false science.

The American anthropologist John Reed Swanton faulted Frazer for calling elements “primitive” for no other reason than that they “do not occur or have been largely suppressed in our present so-called ‘higher’ culture. We thus assume our own culture as an infallible standard of comparison.”Footnote ⁴⁴ The “functionalist” Bronisław MalinowskiFootnote ⁴⁵ slammed Frazer for neglecting how particular customs fitted with other elements of one and the same culture.Footnote ⁴⁶ Malinowski also blames Frazer for being “addicted to psychological interpretations of human belief and practice.”Footnote ⁴⁷ What is more, Malinowski insists that “the scientific attitude is as old as culture”Footnote ⁴⁸ and laments that Frazer put too much emphasis on magical beliefs and neglected “ritual performance.”Footnote ⁴⁹

Franz Boas likewise objected to Frazer’s reliance on psychology. For Boas, cultural influences can never be filtered out when humans respond in tests or experiments. He concludes that “for most mental phenomena we know only European psychology and no other.”Footnote ⁵⁰ Furthermore, Boas doubts that anthropology can ever identify useful general laws; it thus has to be “a historical science.”Footnote ⁵¹ Boas does not deny “universally human” properties but insists that they invariably take “specific forms” in different cultures.Footnote ⁵² Concerning Frazer’s evolutionism, Boas invokes the “modern tendency to deny the existence of a general evolutionary scheme which would represent the history of cultural development the world over.”Footnote ⁵³ Research has shown, to Boas’ satisfaction, that there exist “different ultimate and co-existing types of civilization.”Footnote ⁵⁴ He contends furthermore that the term “primitives” should not lead one to treat preliterate cultures as somehow closer to the “early history of man” (as Frazer put it).Footnote ⁵⁵

Already in the early twentieth century, it was common to call Boas a “cultural relativist.”Footnote ⁵⁶ Elements of this position included the following. First, Boas recognizes fundamental differences in value-judgments, writing that “the social ideals of the Central African Negroes, of the Australians, Eskimo, and Chinese” are so different from those of American or European populations that the respective “valuations … are not comparable. What is considered good by one is considered bad by another.”Footnote ⁵⁷ Second, the anthropologist therefore has to beware of misconstruing their own, parochial “cultural standards … as generally valid human standards.”Footnote ⁵⁸ And third, the anthropologist should “enter […] into each culture on its own basis [and] elaborate the ideals of each people.”Footnote ⁵⁹

Boas’ student Alexander GoldenweiserFootnote ⁶⁰ accuses Frazer of misconstruing the perspective of “participators in magical processes.” These “participators” do not render nature as a “series of [lawlike] uniform successions of actions and effects.” Much more fundamental is their “belief in [the] power” of “the magical act or ceremony.”Footnote ⁶¹ Moreover, Frazer has not established that magic is “prior to or more primitive than religion.”Footnote ⁶² And finally, Goldenweiser maintains that magic is still very much present in the US American culture of his day: We don’t open umbrella indoors and knock on wood for good luck; we hang horseshoes on the wall to “insure the well-being of many a household”; we use “charms and amulets,” and beware of black cats or thirteen guests at a dinner party.Footnote ⁶³

Boas never himself uses the term “relativism” to characterize his approach to anthropology, but he speaks approvingly of Ruth Benedict’s attempt, in her Patterns of CultureFootnote ⁶⁴ to display “the relativity of what is considered social or asocial, normal or abnormal.”Footnote ⁶⁵ Benedict did her PhD with Boas. She too has frequently been seen as an advocate of “cultural relativism.”Footnote ⁶⁶ Already on the first page of Patterns of Culture, Benedict maintains that a Western anthropologist studying a New Guinea tribe has to “avoid any weighting of one [of the two cultures] in favor of the other.”Footnote ⁶⁷ She goes on to suggest that “no man ever looks at the world with pristine eyes” and that even his “every concepts of the true and the false have reference to his particular traditional customs.”Footnote ⁶⁸ She concludes her book with the remarks that “social thinking at the present time has no more important task before it than that of taking adequate account of cultural relativity,” and that we need to accept “the coexisting and equally valid patterns of life which mankind has created for itself from the raw materials of existence.”Footnote ⁶⁹

Central in Benedict’s relativistic position is the idea of a “great arc” of possible values, institutions, and interests from which particular cultures pick and choose.Footnote ⁷⁰ Benedict follows functionalist anthropologists in highlighting the holistic character of such selections. The criticism of Frazer is explicit on this score: “Studies of culture like The Golden Bough … ignore all the aspects of cultural integration.”Footnote ⁷¹ Benedict invokes Oswald Spengler and Friedrich Nietzsche to capture some of this integration. She uses Nietzsche’s distinction between the “Apollonian” and the “Dionysian” ethe in order to characterize two types of cultures in the Southwest of the United States, the “Zuñi,” and “various surrounding peoples.” The Apollonian ethos of the Zuñi seeks “sobriety” and “measure,” the Dionysian ethos of their neighbors emotional and psychic excesses.Footnote ⁷² Benedict recognizes the similarity with Spengler’s attempts to identify “contrasting configurations in Western civilization.”Footnote ⁷³ Still, while the identification of different ethe strikes her as possible in small and close-knit cultures, she denies that the same holds for “Western civilizations”: Spengler is wrong to “treat […] modern stratified society as if it had the essential homogeneity of a folk culture.”Footnote ⁷⁴

It fell to another one of Boas’ students, Melville Jean Herskovits, to spell out “cultural relativism” in its most uncompromising form.Footnote ⁷⁵ He states the “principle of cultural relativism” as follows: “Judgements are based on experience, and experience is interpreted by each individual in terms of his own enculturation.”Footnote ⁷⁶ Herskovits later adds that the “very definition of what is normal or abnormal” is relative to cultureFootnote ⁷⁷ or that “the relativistic point of view brings into relief the validity of every set of norms for the people whose lives are guided by them, and the values they represent.”Footnote ⁷⁸ Herskovits distinguishes between “absolutes” and “universals” as follows. A universal is a cultural feature shared by all cultures; ‘having a system of morality’ is an example. An absolute is a particular value that had to be, and was, accepted in all cultures. Herskovits accepts universals but denies the existence of absolutes.Footnote ⁷⁹

We can now turn to Wittgenstein’s comments on Frazer’s The Golden Bough. Some of these comments are naturally related to other remarks in Wittgenstein’s notebooks: in particular to reflections on causation and description, on understanding other cultures, and on the nature of religious belief. I shall suggest in each case that Wittgenstein reasoned in ways strongly reminiscent of the cultural-relativist wing of anthropology in the Anglophone world during the first half of the twentieth century. Four main topics can be distinguished: Magical and religious ideas and practices need not involve “errors”; causal hypotheses about the origins of rituals are best replaced by descriptions and “well-ordered synopses”; magical thinking is present everywhere, including in the European culture of today; and magic is best understood in functional and cultural-relativist ways.

Historicism without History?

Above, I applauded Hacker’s use of the term “historicism” for characterizing Wittgenstein’s anthropological and ethnological approach. I also agreed with him his observation that, in order to get clear on Wittgenstein’s historicism, it is necessary to consider Wittgenstein’s response to Spengler’s Decline of the West. Still, there are some respects in which I wish to go beyond Hacker’s valuable observations.

To begin with, we need to remember what “historicism” meant around 1900. The term was used in a variety of ways. Sometimes its opposite was “naturalism,” sometimes “universalism,” sometimes “absolutism.” When interpreted in the last-mentioned way, “historicism” was, or included, “historical relativism.”Footnote ¹⁰⁹ Edmund Husserl criticized historicism in this sense, defining it as the view that an increasing awareness of fundamental historical changes in philosophy, religion, and science has destroyed the belief in “absolute validity” and has prepared the ground for the “recognition of the relativity of the historical form of life [Lebensform].”Footnote ¹¹⁰ The most obvious culprit was Oswald Spengler. Max Scheler called The Decline of the West “the final resonance of romantic historicism.”Footnote ¹¹¹

For present purposes, we need to register only a few central ideas of Spengler’s chef d’oeuvre. Spengler holds that every culture has its own unique “Ursymbol” (original symbol) that underlies all of its artistic, scientific, and philosophical activities. For instance, in Egyptian art the original symbol was the “path,” and for the Occidental or “Faustian” culture, it was and is infinite space. Egyptian painting therefore culminated in drawing and Occidental art in the music of Bach, Haydn, and Mozart. Furthermore, Spengler declares that the fundamentally different original symbols precluded cross-cultural understanding. Thus, modern Europeans with their “Faustian” “Ursymbol” are unable to grasp what the Egyptians felt or thought.Footnote ¹¹² This incommensurability is tied to Spengler’s historical relativism. He asserts that there are “as many moral systems as there are cultures … There is no universal human ethics”;Footnote ¹¹³ “[t]here are several number-worlds as there are several cultures”;Footnote ¹¹⁴ the history of mathematics consists of “a plurality of self-contained, independent developments”;Footnote ¹¹⁵ and Occidentals are unable to fully understand the mathematics of other cultures.Footnote ¹¹⁶

Western culture had reached its final stage of decline, that is, “civilization.” This phase is characterized by rationalism, causal analysis, technology, big cities, democracy, cosmopolitanism, humanism, pacifism, emphasis on human rights, and skepticism. Spengler predicts that this phase will usher into total dissolution. His prediction is based on a comparison between the patterns of development of Egyptian, ancient Greek, Arabic, and Western culture.

Spengler defends his comparative method by referring his readers to Goethe’s work on the morphology of plants.Footnote ¹¹⁷ Spengler’s central move is to interpret cultures as organisms:

Throughout his book, Spengler contrasts morphological investigations to the study of causal relations in nature. For instance,

Systematics is dominant within the natural sciences, whereas physiognomics can be found in historical studies. And the natural sciences are based on “understanding,” whereas history relies on “intuition.”Footnote ¹²⁰ According to Spengler, Kant and Aristotle were the philosophers of understanding, whereas Plato and Goethe were the philosophers of intuition. Or, in Spengler’s own words: “Plato and Goethe accept the secret, Aristotle and Kant wish to destroy it.”Footnote ¹²¹

We know from Georg Henrik von Wright that Wittgenstein “lived the Untergang des Abendlandes.”Footnote ¹²² This is confirmed in part by Wittgenstein’s one-time exclamation: “I am so influenced in my thinking by Spengler!” (Wie sehr ich doch bei meinem Denken von Spengler beeinflusst bin!).Footnote ¹²³ In another place he remarks:

I have numbered the sentences for ease of reference. The core claims [3] and [4] are not necessarily a commitment to Spengler’s radical incommensurability between cultures. Indeed, Wittgenstein elsewhere rejects as “dogmatism” the idea that the Ursymbol is some kind of “prejudice to which everything [in a given culture] has to conform.”Footnote ¹²⁵ Wittgenstein offers “family resemblance” as a tool for improving Spengler’s position:

There are thus resemblances between cultures and comparisons between cultures are both possible and desirable. In other words, Wittgenstein is not willing to throw cultural anthropology under Spengler’s bus. He also refuses to be realist about Spengler’s original symbols. They might be fruitful for some specific forms of reflection but not for others:

Although Wittgenstein thus distances himself from the strong form of Spenglerian incommensurability, he finds merit in the thought that one’s tradition and position in history limits one’s ability to understand people of other cultures and times. This comes out clearly in his lectures on aesthetics: “What we now call a cultured taste perhaps didn’t exist in the Middle Ages. An entirely different game is played in different ages.”Footnote ¹²⁸ Wittgenstein lists as different “traditions”: fifteenth-century painting,Footnote ¹²⁹ “Negro Art,”Footnote ¹³⁰ tailoring of Coronation robs in thirteenth-century England,Footnote ¹³¹ or dressmaking in eighteenth-century Europe.Footnote ¹³² He denies that one can choose to adopt an alien tradition, or indeed, even learn to understand it: “Tradition is not something a man can learn; not a thread he can pick up when he feels like it; any more than a man can choose his own ancestors.”Footnote ¹³³ This does not rule out appreciating the art of another tradition. But Wittgenstein is adamant that the appreciation of another tradition is completely different from the appreciation of that tradition by its own members.Footnote ¹³⁴

Similar motifs can be found in Wittgenstein’s discussions with Rush Rhees in the 1940s. He declared himself unable to answer the question “whether Brutus’ stabbing Caesar was a noble action (as Plutarch thought) or a particularly evil one (as Dante thought).” He takes a similar view on Kierkegaard’s question “Has a man a right to let himself be put to death for the truth?”Footnote ¹³⁵ The ethical vocabularies in which these questions were posed are no longer his.

Note though that Wittgenstein is not saying, as Spengler and Benedict would, that all ethical systems are equally valid: “If you say there are various systems of ethics you are not saying they are all equally right. That means nothing.”Footnote ¹³⁶ But although Wittgenstein dismisses the equal validity of different ethical systems, he clearly advocates the idea of a methodological relativistic neutrality in our study of different such systems. Here is another passage from Rhees:

Spengler is also in the background when Wittgenstein rebuffs Frazer’s hypotheses about proximate and distant causes. Indeed, Wittgenstein calls his noncausal well-ordered synopsis an expression of “a kind of ‘world-view’, as is apparently typical of our time.” After a full stop, he added “Spengler.”Footnote ¹³⁸ I take this to be a reference to Spengler’s preference for Goethean morphology over Kantian natural laws and causal relations. Again, it is worth recalling that Benedict also finds Spengler a useful inspiration in this regard. To the above-cited criticism of Frazer, we can add similar comments elsewhere in Wittgenstein’s oeuvre:

These passages must be taken aright. Wittgenstein is not dismissing causal reasoning per se. The first quote is best understood in light of the remark: “Who knows the laws according to which society develops? I am sure that even the smartest people have no idea.”Footnote ¹⁴² The laws in question are causal laws. Wittgenstein is making the obvious point that causal historical laws are very hard to come by. The idea was standard lore in German philosophy of history and reflections on historicism around 1900.Footnote ¹⁴³ Boas, trained as he was in Germany, says much the same thing. The second passage does not postulate that it is always and everywhere wrong to investigate the history of buildings. It suggests merely that the focus on causal history must not overwhelm an account of the present. There is something to be said for taking in the Mona Lisa in the here and now, independently of reading up on the history of the painting. And the third comment insists on the value of counterfactual historical thinking. Such thinking is crucial for appreciating the contingency and genealogy of things one takes for granted: “One of the most important methods I use is to imagine a historical development for our ideas different from what actually occurred. If we do this, we see the problem from a completely new angle.”Footnote ¹⁴⁴

I thus agree with Hacker that Wittgenstein was a historicist and influenced by Spengler – albeit that Wittgenstein does not accept Spengler’s realism about original forms or Spengler’s radical ideas on cultural incommensurability. Still, other than Hacker, I have sought to identify the relativistic motifs that Wittgenstein finds noteworthy in Spengler, like the notion of the plurality of self-contained systems, and the idea of limits of understanding across history and cultures.

Finally, the phrase “historicism without history,” catchy though it is, seems to me not to fit Wittgenstein all that well. It downplays the interest Wittgenstein takes in actual historical episodes in the arts and sciences. Wittgenstein had a sophisticated understanding of the development of musicFootnote ¹⁴⁵ and lectured on aesthetics with reference to painting and sculpture.Footnote ¹⁴⁶ He also made extensive use of his knowledge of recent mathematics and logic, from intuitionism to finitism, from logicism to logical pluralism. He commented on Kurt Gödel’s famous theorem. He made interesting metaphorical use of Einstein’s ideas of clock-coordination.Footnote ¹⁴⁷ In his philosophy of psychology, he commented in an informed way on the “thought psychology” of the Würzburg School and on Gestalt psychology.Footnote ¹⁴⁸ And, as I hope to have shown earlier, Wittgenstein was well aware of important trends in anthropology.

Summary

Hacker takes his starting point from the following passage in Wittgenstein:

If we want to understand what Wittgenstein meant by “the ethnological perspective,” we need to investigate his views on physical and cultural anthropology and ethnology. And this is best done by situating these views relative to positions in these emerging sciences that Wittgenstein engaged with or can reasonably be assumed to have known about, at least in broad outline. This is what I have done above. If I am right, then Wittgenstein’s ethnological perspective is cultural-relativist and historicist, in both methodological and substantive ways. To study phenomena ethnologically is to study them in a neutral and context-sensitive way without ready-made yardsticks of progress or reason. It needs this methodology to realize the irreducible variety of cultural practices and worldviews, and to undermine absolutist conceptions of meaning, value, and knowledge.

Introduction

One important strand of Wittgenstein-scholarship addresses the question whether Wittgenstein was a naturalist. The replies can be situated on a spectrum, ranging from a bold scientific-explanatory naturalism at the one end to a therapeutic non-scientific naturalism at the other end. David Bloor is the most outspoken defender of the first option.Footnote ¹⁵⁰ Bloor takes Wittgenstein’s later philosophy to be continuous with the empirical sciences, especially with psychology and sociology.

The other end of the spectrum has too many advocates for me here list them all.Footnote ¹⁵¹ They are united in giving special weight to the following passages from the Philosophical Investigations:

Daniel D. Hutto and Glenda Satne are defenders of the therapeutic-naturalistic interpretation of Wittgenstein, and they render these and some other relevant passages as follows. Wittgenstein’s naturalism is “liberating”: It liberates us “from certain mystifying tendencies of thought,” and it stands in a complementary relation to the sciences.Footnote ¹⁵⁴ Liberating naturalism offers “descriptions” of facts “that lie right before our eyes but which can, due to our own tendencies, become blocked from our view.”Footnote ¹⁵⁵ Liberating naturalism removes conceptual confusions by reminding us of our natural history. But such reminders are not explanations, hypotheses, or speculations. Rather than explaining our practices, “philosophical investigation … comes to an end with an acceptance that it is a fact of our natural history that these practices have the shape that they have.”Footnote ¹⁵⁶ Still, it is the invocation of natural history that justifies calling liberating naturalism a species of the genus naturalism.

Advocates of therapeutic nonscientific naturalism do not generally engage with scientific-explanatory naturalism. Their animus is toward positions that lie halfway between the endpoints: these are views offering philosophical – though not scientific – theories about the nature of logic, mathematics, the mind or language. Prominent representatives include Saul Kripke, John McDowell, and Crispin Wright.Footnote ¹⁵⁷ (I shall not sketch their views here. I have discussed all three in my book-length defense of Kripke’s Wittgenstein.Footnote ¹⁵⁸)

In this section, I want to make a small contribution to the discussions over Wittgenstein’s naturalism. To begin with, I want to move from the genus naturalism to two of its species that were much better known to Wittgenstein’s philosophical contemporaries: to wit, “psychologism” and “sociologism.” I shall show how Wittgenstein’s later philosophy related to these positions, and how some influential authors in his day would likely have situated him concerning these “isms.” The topic is relevant for my overall project in this Element, since psychologism and sociologism were widely regarded as important forms of relativism. In working out these issues, it is important to focus on how Wittgenstein’s relativistic “ethnological perspective” operates in his reflections on metrology, logic, and mathematics.

The Psychologism and Sociologism WarsFootnote ¹⁵⁹

The period between the 1890s and early 1920s witnessed an intense debate concerning the relationship between empirical psychology, on the one hand, and logic and mathematics, on the other hand. The following arguments were paradigmatic cases of what came to be called “psychologism”:

• (Argument 1)

• Psychology identifies and investigates all laws of thought.

• Logic identifies and investigates a subset of the laws of thought.

• Ergo, logic is a subfield of psychology.

• (Argument 2)

• Logic tells us how we ought to reason.

• Since ought implies can, logic needs to be based upon an adequate descriptive account of human reasoning abilities.

• Only psychology can deliver the required descriptive account.

• Ergo, logic must be based upon psychology.

• (Argument 3)

• Logic investigates concepts, judgments, and inferences.

• Concepts, judgments, and inferences are mental entities.

• Mental entities are the subject matter of psychology.

• Ergo, logic is a subfield of psychology.

These arguments were only rarely formulated explicitly. But it was widely thought that authors like Wilhelm Wundt,Footnote ¹⁶⁰ Wilhelm Jerusalem,Footnote ¹⁶¹ Christoph Sigwart,Footnote ¹⁶² and Benno ErdmannFootnote ¹⁶³ were at least implicitly committed to these arguments. In the background of these “psychologistic” views was John Stuart Mill, and especially the following passage:

Mill is important in the present context also because of his empiricist account of mathematics. Mill held that the basic laws of geometry and arithmetic are learned by way of inductive reasoning.Footnote ¹⁶⁵

So much for the psychologistic thinkers. Wittgenstein is likely to have encountered them only through the work of Gottlob Frege.Footnote ¹⁶⁶ Frege insists that neither mathematics nor logic are parts of psychology, and that mathematical or logical objects and laws are not explained or justified by psychological observation. One of Frege’s key considerations is that whereas mathematics is the most exact science, psychology is invariably imprecise.Footnote ¹⁶⁷ This difference makes it hard to see how mathematics could be based on psychology. Against Mill, Frege maintains that when children learn how to calculate, they do not thereby gain knowledge about the empirical world. And there is no general inductive law from which all mathematical propositions follow. More generally, empirical knowledge cannot justify mathematical propositions.Footnote ¹⁶⁸

Frege distinguishes between descriptive and prescriptive laws. All prescriptive laws might be called “laws of thought” since they tell a relevant group of people how they ought to think. The case of descriptive laws is different. Only one kind of descriptive law can properly be rendered as consisting of laws of thought: to wit, the laws of empirical psychology.Footnote ¹⁶⁹ Prescriptive laws are always based on descriptive laws. Logic is no different from other sciences in this regard. Frege maintains that logic thus has two levels: (logical) descriptive laws and (logical) prescriptive laws.Footnote ¹⁷⁰ Prescriptions based on psychological laws can be no more than demands to conform to certain thinking habits. But such prescriptions cannot serve as yardsticks by which these very thinking habits can be assessed regarding their truth. Furthermore, the descriptive psychological laws – that psychologistic logicians take to constitute the bases for logical prescriptions – are laws of “taking-to-be-true”: They state the conditions under which humans recognize the truth of a judgment or the validity of an inference; descriptive psychological laws do not state the conditions under which the judgment is true, or the inference valid. Frege concludes: “If being true is … independent of being acknowledged by somebody or other, then the laws of truth are not psychological laws: they are boundary stones set in an eternal foundation, which our thought can overflow but never displace.”Footnote ¹⁷¹ In the same context, Frege also dismisses the notion according to which logical laws are merely “hypothetically necessary,” that is, necessary only given the (current) psychology of the human species. As Frege has it, if we were to encounter creatures who reject or lack our logical laws, we would take them to be “insane.”Footnote ¹⁷²

Finally, Frege contends that the realm of the “non-real” is not identical with the realm of the psychological and the subjective. Numbers, for instance, are not ideas since they are the same for all subjects.Footnote ¹⁷³ To deny the objectivity and nonreality of numbers and concepts leads to idealism and solipsism. Note that Frege does not accuse the “psychological logicians” of relativism or skepticism. But it is not difficult to see such charges as implications of the diagnosis that psychologism is tantamount to (subjective) idealism and solipsism.

Edmund Husserl’s criticism of psychologism in his Logische Untersuchungen (1900) was, at least initially, much more influential than Frege’s. For present concerns, it is imperative to focus on one important difference; to wit, that Husserl explicitly links psychologism to relativism. Husserl insists that all forms of psychologism are forms of “species relativism,” or more precisely “anthropologism.” Roughly put: If logic varies with psychological make-up, then different species have different logics. Anthropologism is the view that the logic developed by human logicians is tied to their human physiology. Husserl presents a number of arguments against species relativism. The most important such arguments are: (a) that since the Principle of Non-Contradiction is part of the meaning of truth, one and the same thought cannot be true for one species and false for another; or (b) that truths are eternal and thus not temporally and spatially restricted.Footnote ¹⁷⁴ More specifically, the “psychologistic thinker” Sigwart found it incoherent to assume that a judgment could be true even if no intellect had ever thought of it. In reply, Husserl asserts that it is part of the meaning of the law of gravity that it is true for all times (even prior to its discovery) regardless of whether it was ever thought by anyone.Footnote ¹⁷⁵

Husserl’s Prolegomena were widely discussed in German-speaking philosophy at the time. Let me mention just one critic whose work was familiar to Wittgenstein: Moritz Schlick. Husserl (and Frege) had insisted that logical laws could not be psychological laws: psychological laws are invariably vague; logical laws are always exact. Schlick is unconvinced:

Moreover, Frege’s and Husserl’s case against psychologism centrally relies on the claim that truths exist independently of whether or not they are ever grasped by anyone. Schlick calls this assumption “the independence theory of truth” and opposes it as follows: “There is no truth of judgments such that this truth is independent of the judgments’ existence in mental acts. Only the facts upon which true judgments are based, are independent of us.”Footnote ¹⁷⁷ Schlick continues: “[I]n the case of abstract ideas, object and content coincide, i.e. the object of the idea is nowhere to be found, except within that very idea. And thus, logical sentences and acts of judging are absolutely inseparable.”Footnote ¹⁷⁸

The psychologism controversy of the first two decades of the twentieth century was followed – in the German-speaking world of the late 1920s and early 1930s – by another fierce debate: the dispute over the philosophical aspirations of the sociology of knowledge. Given how central the accusation of “sociologism” figured in this debate, it seems adequate to call it the “sociologism wars.” Much of these wars centered on the work of Karl Mannheim, but two other authors, Wilhelm Jerusalem and Max Adler also deserve to be mentioned, at least briefly. Sociologism was typically thought of as involving “two thesis: first, all thinking and knowing is bound to existence; and second, this dependence on existence has a bearing on the validity of a judgement.”Footnote ¹⁷⁹

Jerusalem is highly critical of philosophers like Husserl who

Adler believes that the central question of Kant’s epistemology – “how is science possible?” – is a sociological question and that Kant had intended it as such. This can be seen, Adler thinks, from the fact that Kant did not pose the question as “how can the individual have knowledge?” Indeed, to substitute this second question for the first is to be guilty of psychologism.Footnote ¹⁸¹ Adler intends the sociology of knowledge to show that all forms of knowledge are socially constituted.Footnote ¹⁸² Adler explicitly presents the sociological approach as a way to overcome the shortcomings of psychologism. At the same time, Adler is not altogether dismissive of psychologism. Indeed, he regrets the contemporary “fear of psychologism,” which in his view has led to an unfortunate separation of “the problem of validity from all links to the fact of experience.”Footnote ¹⁸³

Mannheim’s central thought is that at least certain types of knowledge – the knowledge of the social sciences and humanities – are “seinsgebunden” or “seinsrelativ”: that is, they were relative to specific natural and social conditions. These conditions are at least partial causes of knowledge, and the latter cannot be fully grasped unless the former are also understood.Footnote ¹⁸⁴ Moreover, Mannheim emphasizes that epistemology needs to be radically reformed and that the sociology of knowledge is instrumental in this endeavor. Epistemology needs to be rebuilt around the insight that the knower is always the member of a group. Although the psychological genesis of a judgment is often irrelevant to its content, the sociopolitical context of a judgment is typically part of its meaning. Finally, even in the case of 2 × 2 = 4 it is misleading to say that there is such a thing as “truth in itself.” Speaking in this way is of heuristic value only.Footnote ¹⁸⁵

Mannheim’s views were criticized from many directions. The standard criticism was that sociologism was a form of relativism and close kin to anthropologism and psychologism. Eduard Spranger remarks that “this new sociologism … seems to take over the part of … psychologism.”Footnote ¹⁸⁶ Siegfried Marck sees psychologism and sociologism as the two main species of what he called “existentialism.”Footnote ¹⁸⁷ Mannheim was not happy with either charge and deplores the tendency in German philosophy “to shout ‘sociologism’ or ‘psychologism’ as soon as a special science starts to tackle general problems of knowledge.” To take such invectives seriously, Mannheim writes, would be “intellectual suicide.”Footnote ¹⁸⁸

The Later Wittgenstein on Metrology, Logic, and Mathematics

I now turn to Wittgenstein’s reactions to some of the issues raised in the psychologism and sociologism wars. Limitations of space do not allow me to discuss the Tractatus or to distinguish between the “middle” and “later” Wittgenstein. A natural starting point is Wittgenstein’s reflections on metrology. This is because, for Wittgenstein, measurement practices, mathematics, logic, and language have important properties in common: All four are, in an important sense, arbitrary; all four exist in the plural; and all four depend upon the same kind of general conditions: human physiology and psychology, social institutions, and general facts of nature.

Introduction

Wittgenstein’s last notebooks were published under the title “On Certainty.”Footnote ³⁰¹ The work has attracted considerable attention over the past three decades. Indeed, On Certainty has become a key text in a new subfield of philosophy called “hinge epistemology.”Footnote ³⁰² “Hinge” is one of the metaphors Wittgenstein uses for certainties: [T]he questions that we raise and our doubts depend on the fact that some propositions are exempt from doubt, are as it were like hinges on which those turn.”Footnote ³⁰³ I agree with “hinge epistemologists” that On Certainty is of considerable philosophical interest. And yet, it seems to me that something important is lost about the book when it is straightforwardly inserted into contemporary discussions of epistemic skepticism, epistemic justification, or knowledge. What is lost is the fact that Wittgenstein’s last notebooks are in many ways continuous with his writings of the 1930s and 1940s, writings that focus on logic and mathematics as “ethnological phenomena.”Footnote ³⁰⁴ This fact is important given the overall aim of this Element: to identify ideas and motifs in Wittgenstein that his contemporaries would have classified as relativistic.

There are many philosophers of the early twentieth century who very likely would have categorized the views of On Certainty as leaning toward relativism. Not all such authors work well as foils for Wittgenstein’s ethnological approach to certainty. We need authors who worked on themes related to On Certainty but whose approaches differ in interesting ways from Wittgenstein. I believe that two philosophers of the “Vienna Circle,” Moritz Schlick and Otto Neurath, are just the authors we need. I am thinking here primarily of their debate over “protocol statements,” as well as some of Neurath’s later writings. In the Protokollsatzdebatte, Schlick is centrally preoccupied with trying to refute epistemic skepticism and relativism. Neurath rejects Schlick’s “absolutism” and develops a form of relativism that resonates with some themes in Wittgenstein. Comparing Wittgenstein with both Schlick and Neurath brings out both commonalities and differences. Most importantly, it helps once more to appreciate the role of relativism in Wittgenstein’s philosophy.

Here is the plan for the rest of the section. I start by outlining the central theme of the Protokollsatzdebatte, especially the disagreement between Schlick and Neurath. Subsequently, I summarize the key ideas of On Certainty. I conclude by highlighting the main differences between Schlick, Neurath, and Wittgenstein, especially as far as their methodology and positions vis-à-vis relativism are concerned.

Some Central Threads of the Protokollsatzdebatte

The debate over protocol statements was triggered by Schlick’s 1934 paper “The Foundation of Knowledge.”Footnote ³⁰⁵ Schlick begins by claiming that epistemologists – including “relativists” and “sceptics” – are motivated by the search for “absolute certainty,” and that early work on protocol statements interpreted them as expressing facts directly and without distortion, and thus as “absolutely indubitable starting points of all knowledge.”Footnote ³⁰⁶ Unfortunately, Schlick laments, discussions of protocol statements have started to drift towards “a peculiar relativism.”Footnote ³⁰⁷ This is because protocol statements have increasingly been interpreted as (mere) hypotheses. Neurath is the chief culprit.

Assume we formulate a protocol statement as follows:

(*) does not state an absolute certainty. John might have misdescribed what he observed. Or X might have malfunctioned without John noticing.Footnote ³⁰⁸ It follows that (*) has the character of a mere hypothesis. Thus, if (*) is a protocol statement, then we have to find a new way of telling protocol statements apart from other statements. Some philosophers invoke “convenience,”Footnote ³⁰⁹ others, like Neurath, coherence considerations. Schlick urges his readers to focus on the shortcomings of the coherence theory.Footnote ³¹⁰

Coherence, Schlick alleges, boils down to consistency. And if the mere absence of contradictions in a system were to guarantee its truth, then consistent fairy tales would be true. If consistency were to guarantee truth, then there could be many true systems of statements, systems that are incompatible with one another. Under the conditions of a coherence theory, the only way to maintain a special status for protocol statements is to define them as those statements, the “retention” of which “requires a minimum of alternation in the whole system … in order to rid it of all contradictions.”Footnote ³¹¹ Schlick objects to this “economy principle” on the ground that it allows, over time, for downgrading a statement from “protocol statement” to mere hypothesis.Footnote ³¹² Nor does it help to render protocol statements as statements accepted by most scientists. We give value to the majority view only because it is based on something which is “not corrigible or annulable.”Footnote ³¹³

At this point, Schlick moves to outlining his own proposal. The core idea is the distinction between “protocol statements” and “basic statements.” The former are statements reporting someone’s sensory observations; the latter are statements that can function as “the ultimate ground” of all empirical knowledge.Footnote ³¹⁴ The latter express what is “immediately observed” by the observer, for the observer, during the very act of observing.Footnote ³¹⁵ Such basic statements are absolutely certain; indeed, I might even dismiss a whole field of science on the grounds that it conflicts with my basic statements: “What I see, I see!” can be my credo.Footnote ³¹⁶ Nevertheless, there is something peculiarly fleeting and vanishing about such basic statements. The moment they are written down, or memorized, they lose their absolute certainty and turn into hypotheses. This is because the passing of time allows for error to creep in.Footnote ³¹⁷

To further sharpen his account, Schlick introduces the category of “Konstatierungen” (affirmations). Affirmations capture the contents of basic statements before the latter are formulated and thereby turned into hypotheses. Consider how we test scientific predictions in the light of sensory experience. Think of the prediction that a given piece of blue litmus paper will turn red when exposed to (acidic) material in a test-tube. To verify this hypothesis, we ultimately need an affirmation of the form “Here red now.” This affirmation triggers in us a “feeling of fulfilment, a quite characteristic satisfaction.” Put differently, affirmations “bring verification (or falsification) to completion … They constitute an absolute end.”Footnote ³¹⁸ And thus “it gives us joy to reach them.”Footnote ³¹⁹

Schlick concludes by comparing the truth of affirmations with the truth of analytic statements. What makes an analytic statement true is the fact that it has been correctly formulated in light of the definitions of its terms. And to understand an analytic sentence is to understand that it is true.Footnote ³²⁰ The situation of an affirmation is both similar and different. The difference is that an affirmation is synthetic and expresses an experience. The similarity is that to understand an affirmation is to understand its truth. An affirmation necessarily contains “demonstrative terms” like “here” or “now.” Their referents, in given situations, depend upon the pointing gestures of the speaker. When I say to myself “here red now,” I know what “here” and “now” refer to since it is me who is doing the (implicit or explicit) pointing. I understand what my words mean in the very act of referring to actual objects. And therefore, Schlick insists, affirmations behave relevantly like analytic statements.Footnote ³²¹

We can now turn to Neurath. For present purposes, his 1934 reply to Schlick, entitled “Radical Physicalism and the ‘Real World’,”Footnote ³²² is of special interest. Neurath first recalls that, not so long ago, Schlick held that there are “no special philosophical questions at all.”Footnote ³²³ This does not fit with Schlick’s hunting for “absolute certainty.”Footnote ³²⁴ Moreover, Neurath is annoyed about being classified as a coherence theorist. The accusation stings since Schlick, in earlier work, has tied the coherence theory to “idealistic metaphysics.” Neurath pleads “not guilty”: Idealistic metaphysics is incompatible with Neurath’s commitments to physicalism and empiricism.Footnote ³²⁵ In line with these principles, Neurath avoids speaking of “personal experiences” and uses the term “experience statement” instead. Accordingly, protocol statements are to be formulated in “physicalist language”:

Neurath highlights the fact that his protocol statements do not have indexicals. He goes on to list four further key tenets of “radical physicalism.” First, all scientific statements and all protocol statements are adopted by way of “decisions” that can subsequently be overturned. Second, scientific statements or protocol statements are false if they fail to fit with “the whole structure of science.” In the case of protocol statements, we have the further option of changing the structure of science so that the protocol statement conforms to the (new) structure. Third, instead of “reality,” Neurath speaks of a multitude of internally consistent “totalities of statements”; these totalities cannot be combined into a consistent whole. Fourth, for the radical physicalist, claims about what is “unsayable” or “unwritable” are ruled out.Footnote ³²⁷

Neurath subsequently attacks central tenets of Schlick’s paper. To begin with, Neurath takes exception to the notion of “absolute certainty.” One questionable element here is Schlick’s imagery of horizontal layers of more or less certain entities. This imagery is a residue of metaphysics. Physics teaches us that “everything is in connection with everything else.” Moreover, Schlick seeks to dismiss positions by labeling them “relativistic” – this is too quick.Footnote ³²⁸ Schlick’s most important attempt to defend the absolute certainty of “affirmations” centers on the comparison with analytic statements. Neurath denies that logical or mathematical statements are absolutely certain; after all, we occasionally correct calculations or inferences. And when we do so, we typically refer back to protocol statements.Footnote ³²⁹ Here Neurath approvingly cites Karl Menger’s criticism of Henri Poincaré. Menger chastises the latter for demanding “from mathematics a certainty surpassing that of all other human activities not only in degree but in essence.”Footnote ³³⁰

Neurath’s second major objection concerns Schlick’s failure to appreciate “ambiguity.” “Ambiguity” is the opposite of “absolute certainty.”Footnote ³³¹ Neurath reminds his readers of Pierre Duhem’s notion that available protocol sentences always underdetermine theory choice. Neurath goes one step further by insisting on the underdetermination of protocol sentences as well. This invariably leaves us with a plurality of “totalities” of protocol statements and scientific theories. It is the “practice of living” that reduces this plurality, and this process involves many individuals and groups.Footnote ³³² In passing, Neurath suggests that it is the “metaphysics” of Wittgenstein’s Tractatus that likely underpins Schlick’s opposition to ambiguity and plurality: Schlick’s affirmations seem to function similarly to Wittgenstein’s “elementary sentences.”Footnote ³³³

Neurath’s longest critical section concerns Schlick’s correspondence theory of truth. For Neurath, statements can only be compared with other statements, not, as Schlick has it, with “reality,” the “true world” or “things.”Footnote ³³⁴ In this context, Neurath rejects Schlick’s contention that even “relativists” and “sceptics” strive to grasp “the true world.” Philip Frank has shown, to Neurath’s satisfaction, that “‘true world’ is a meaningless arrangement of script signs or sounds.”Footnote ³³⁵ Furthermore, Schlick thinks that one’s own protocol statements have special weight. For Neurath such “obstinacy” is of merely psychological or historical interest. We always have the option of preferring another person’s protocol sentences to our own.Footnote ³³⁶

This leads naturally to Neurath’s fourth and final objection. It targets Schlick’s claim that “a genuine affirmation cannot be written down.” For Neurath this is tantamount to saying: “there are statements that are not statements.”Footnote ³³⁷ Schlick also maintains that “feelings of satisfaction” accompany affirmations. Neurath dryly remarks that “a consistent empiricist would suggest experiments to test this hypothesis.”Footnote ³³⁸ Nor does it make sense to speak of a “mission” of affirmations: “It looks as if here we come across the last residues of a network of metaphysics.”Footnote ³³⁹ And when Schlick writes that the “moments of fulfilment” in affirmations are the “the light for whose source the philosopher is actually asking,” Neurath sarcastically replies: “the advocate of a radical physicalism in the service of science will not claim to be a philosopher in this sense.”Footnote ³⁴⁰

Up to this point, I have given a rough and selective sketch of Schlick’s and Neurath’s positions in the Protokollsatzdebatte. I shall now go further and summarize briefly Neurath’s later pronouncements on relativism and pluralism.

The natural starting point is Neurath’s Anti-Spengler (of 1921).Footnote ³⁴¹ Neurath submits that Spengler has not made the best possible case in favor of cultural relativism. Spengler could have relied on Duhem, Poincaré, and Wilhelm Dilthey, and moved from the underdetermination of theories by observations to the underdetermination of worldviews by science.Footnote ³⁴² Some of the objections first used against Spengler are later marshalled – mutatis mutandis – against others. For instance, Neurath argues against Lucien Lévy-Bruhl that tribes who believe in magic do not reason according to a different logic from that of the Europeans: It is the content not the form of thought that differs in the two cases.Footnote ³⁴³ And in texts of the 1940s, Neurath insists on a common core of all languages. And yet, even here certain relativistic elements are visible. For example, Neurath contends that there are social and intellectual discourses that are incommensurable between traditions: “speculative metaphysics,”Footnote ³⁴⁴ ethics, or politics.Footnote ³⁴⁵

Neurath’s occasional remarks on the sociology of knowledge also have a relativistic flavor, for instance where he stresses that even natural-scientific investigations can result in fundamental and undecidable disagreements, disagreements that cannot be resolved by collecting more data.Footnote ³⁴⁶ However, Neurath’s chief goal concerning the social sciences is to purge them of the metaphysical commitments and “absolutist formulations” of Max Weber, Mannheim, or Max Scheler.Footnote ³⁴⁷ Sociology, correctly understood, displays an intellectual “modesty” that combats “all absolutist” views.Footnote ³⁴⁸ Indeed, Neurath alleges that the social sciences are already essentially “pluralist.”Footnote ³⁴⁹ “Pluralism” is the notion that in politics as well as in science “we start with many divergent statements and remain with divergent statements forever, as it were.”Footnote ³⁵⁰

Relativism comes to the fore most clearly where Neurath discusses epistemology and “pseudorationalism.” For Neurath, epistemology is an absolutist project and a form of pseudorationalism. In 1931 Neurath writes: “[T]here can be no ‘theory of knowledge,’ at least not in the traditional form. It could only consist of defense actions against metaphysics.”Footnote ³⁵¹ Pseudorationalism rejects the role of “judgement” in practical and theoretical contexts.Footnote ³⁵² One important enemy here is the Tractatus’ “absolutist” notion that our knowledge can be analyzed into elementary sentences that in turn can be immediately verified.Footnote ³⁵³ Karl Popper’s epistemology also belongs in this context. It involves an “absolutism of certainty,” an “absolutism of precision,”Footnote ³⁵⁴ and an “absolutism of falsification.”Footnote ³⁵⁵

Neurath’s “encyclopedism” is of a piece with his relativism. From the late 1930s onward, Neurath is increasing concerned that the project of “unified science” counteracts his relativism: He worries that unified science might take on the form of a single “hierarchical,” “absolute,” and “pyramidistic” system.Footnote ³⁵⁶ Speaking of “encyclopedias” or the “mosaic” of the sciences is intended to rectify these distracting models. An encyclopedia is “a preliminary assemblage of knowledge,” or a possible structure linking the sciences.Footnote ³⁵⁷ In any case, an encyclopedia is a product of history.Footnote ³⁵⁸ In words most relevant for the present context, the theory of encyclopedias has to reject “the traditional absolutistic terminology that allows reference to ‘the real world’, the ‘ideal totality of statements’ and other similar things.”Footnote ³⁵⁹

To fully appreciate Neurath’s take on “the real world,” it is imperative to turn briefly to Philipp Frank’s Das Kausalgesetz und seine Grenzen (1932). In the chapter entitled “Of the So-called True World,”Footnote ³⁶⁰ Frank attacks the scientific-realist assumption that our theories become ever better representations of the way the world is.Footnote ³⁶¹ As Frank has it, this assumption implicitly commits one to the metaphysical-theological idea of a “higher intelligence” whose theory of the world we increasingly come to approximate.Footnote ³⁶² Frank’s counterclaim is that scientific progress consists in ever “better … orderings of experiences,” where the choice between different senses of “better” is “arbitrary.”Footnote ³⁶³

Neurath wholeheartedly agrees with Frank’s antirealism.Footnote ³⁶⁴ Neurath draws several conclusions from the antirealist position. First, he dismisses the correspondence theory of truth. The latter is an “absolutist conception of truth” and undermined by Frank’s arguments.Footnote ³⁶⁵ Moreover, Neurath puts forward a number of suggestions on how to approach truth and related notions: Stop using “true,” “verification,” “progress,” “reality” altogether.Footnote ³⁶⁶ Develop a novel rendering of truth along roughly coherentist lines.Footnote ³⁶⁷ Or supplant truth-discourse with talk of what is ‘accepted by us at a given time and place’. This avoids all “absolute uses” of “true.”Footnote ³⁶⁸

As we saw earlier, Schlick and Neurath disagree over the question whether Neurath’s position in the Protokollsatzdebatte is a coherence theory of truth. A passage like the following sure fits most bills:

That said, Neurath’s ‘coherence theory’ differs from standard formulations in that he gives special weight to protocol sentences.Footnote ³⁷⁰

To sum up, Neurath dismisses a wide range of positions as “absolutist”: talk of ‘reality as such’, or of ‘true reality’; truth as correspondence; Schlick’s and Wittgenstein’s foundationalism; scientific realism; fixed standards or norms in epistemology or politics; pseudorationalism; and pyramidism, monism, and unrestricted commensurability. And finally, there are numerous relativist conceptions that Neurath supports in no uncertain terms: the idea of deep cultural differences; various forms of underdetermination: of worldviews, of theories, of the priority between theory and protocol statements; social values and interests are elements needed to overcome underdetermination; pluralism in the sciences and politics; irresolvable disagreements; truth as coherence or consensus; truth as a concept to be re-engineered; truth as an obsolete concept; incommensurability in certain areas: for instance, in metaphysics, ethics, and politics; or the historicity of encyclopedias.

On Certainty and the Ethnological Perspective

We can now turn from Schlick and Neurath to Wittgenstein’s On Certainty. The following summary of On Certainty is based upon a controversial reading. Since I have defended it elsewhere, I shall not do so again here.Footnote ³⁷¹ (I list important alternatives to my interpretation in this footnote.)Footnote ³⁷² On my reading, the view On Certainty is at pains to undermine carves up our system of beliefs into two strata: beliefs about the meaning of words and mathematical axioms or theorems are in the lower, more fundamental stratum and they cannot be doubted; beliefs about empirical matters are in the upper, less fundamental stratum and they are invariably in principle doubtful:

These quotes hint at the following idea. Some empirical beliefs or empirical judgments can, on some occasions, be as fundamental as are meanings of linguistic expressions or mathematical propositions. Moreover, not all beliefs about linguistic expressions or mathematical propositions are as certain as at least some beliefs about the empirical world.

I read Wittgenstein as committed to the views that, first, most types of certainties are propositional, and based upon epistemic evidence; and, second, that all certainties often function as standards for the other beliefs. All certainties are: “Cases in which I rightly say I cannot be making a mistake.” What is more: “I can … enumerate various typical cases, but not give any common characteristic.”Footnote ³⁷⁵ This comment makes clear that for Wittgenstein “certainty” is a “family-resemblance concept.” Different members of that family share different properties with some other members, but there is no overall common characteristic over and above the “cases in which I rightly say I cannot be making a mistake.” On my reading, when the principle of classification is the subject matter of certainties, Wittgenstein mentions and discusses eleven (content) types of certainties. These fall into five different epistemic kinds.

Epistemic Kind I consists of beliefs based on an overwhelming amount of evidence. For this very reason, it is impossible to share the evidence all at once. Epistemic Kind I certainties fall into seven content types:

• Beliefs about perceivable parts, or the whole, of one’s own bodyFootnote ³⁷⁶

• Perceptual beliefs about medium-size familiar objects in our proximityFootnote ³⁷⁷

• Beliefs based on introspectionFootnote ³⁷⁸

• Autobiographical beliefsFootnote ³⁷⁹

• Simple beliefs based upon inductive inferenceFootnote ³⁸⁰

• Beliefs informed by the testimony of textbooksFootnote ³⁸¹

• Simple semantic beliefsFootnote ³⁸²

Cases of Epistemic Kind II are beliefs that constitute domains of knowable things. An example is the certainty: “[T]he earth exists.”Footnote ³⁸³ Such certainty cannot be directly supported by empirical evidence since such evidence would have to come from within the world or the earth, rendering the support circular. Epistemic Kind III features certainties the evidence for which is so comprehensive that it is reasonable to immunize them, “for the time being,” against possible refutation. We are here dealing with foundational empirical-scientific beliefs.Footnote ³⁸⁴ Epistemic Kind IV goes beyond Kind III in that here the immunization against refutation is “once and for all.” Kind IV comprises mathematical propositions that have “officially, been given the stamp of incontestability.”Footnote ³⁸⁵ Finally, Epistemic Kind V cases are religious beliefs.Footnote ³⁸⁶ In this case, the evidence is accessible only to the person who has already adopted the religion in question.

There is a range of sketches of responses to radical epistemological skepticism in On Certainty. Particularly important for our concerns are those that involve the idea of an “epistemic system.” Our beliefs “form a system.”Footnote ³⁸⁷ Such system forms gradually, and while some things “stand unshakably fast” within the system, others “are more or less liable to shift.”Footnote ³⁸⁸ Wittgenstein applies system-talk also to knowledge,Footnote ³⁸⁹ language-games,Footnote ³⁹⁰ sciences,Footnote ³⁹¹ hypotheses,Footnote ³⁹² evidence,Footnote ³⁹³ verification,Footnote ³⁹⁴ and doubts.Footnote ³⁹⁵ As far as the items within epistemic systems are concerned, Wittgenstein toys with the idea of epistemic positions of different strengths: “I make assertions about reality, assertions which have different degrees of safety (Sicherheit).”Footnote ³⁹⁶ Which strength safety is required depends upon our contexts of action, our purposes and needs. This interpretation seems natural in light of the fact that Wittgenstein notes his proximity to pragmatismFootnote ³⁹⁷ and insists that the end of justification is “not an ungrounded presupposition [but] an un-grounded way of acting.”Footnote ³⁹⁸

All beliefs, including certainties, can be measured in terms of safety. A belief scores high in terms of safety if we can doubt it only by considering very remote error possibilities. Certainties have a very high safety score. It seems plausible to suggest that the different epistemic kinds do not become doubtful at one and the same level of safety. Put differently, within each domain of subject matter – beliefs about my body, beliefs about medium-size objects etc. – certainties mark the point at which doubts no longer make sense; in this sense certainties are equal. But they may well be unequal when measured in terms of error-possibilities across domains.

I can now explain one of Wittgenstein’s anti-skeptical arguments involving the idea of an epistemic system. The skeptic needs to be able to show that what she calls her skeptical or “philosophical doubt” is a sharpening or tightening of “ordinary doubt.” Ordinary doubt needs grounds,Footnote ³⁹⁹ it must be located within our systems of doubts or beliefs,Footnote ⁴⁰⁰ and it must have characteristic manifestations.Footnote ⁴⁰¹ Ordinary doubt proceeds against a background that is exempted from doubt.Footnote ⁴⁰² That is to say, ordinary doubt proceeds within an epistemic system, using it as a background, basis, or filter. The system itself – and its fundamental components – is not doubted. The level of ordinary doubt is set by our practical concerns, as already suggested. And our ordinary doubt has characteristic practical manifestations.

All this does not mean, however, that one-time individual certainties might not lose their status, become doubtful and negated. And they can become doubtful and negated even within the realm of ordinary doubt. This happens when new evidence becomes available at a given level of safety, that is, without that the level of safety changes. When the evidence of the moon landing became available, we gave up the earlier certainty that no-one is able to travel there. But this process did not necessarily involve a change in our system of doubt, in our rules for fixing levels of safety. To sum up, ordinary doubt is local within an epistemic system; based on domain-specific grounds that are less doubtful than what is doubted, it has it has typical manifestations in action; and its level is set by practical concerns. In all these dimensions, radical philosophical doubt differs from ordinary doubt: it is not local,Footnote ⁴⁰³ it is based on very broad general grounds,Footnote ⁴⁰⁴ it is unclear whether its grounds are less doubtful than what is doubted,Footnote ⁴⁰⁵ it has no typical manifestations,Footnote ⁴⁰⁶ and its level is not set by practical concerns.

As far as the motivation for skepticism is concerned, Wittgenstein targets primarily Cartesian themes. A key passage is the following:

To correctly interpret this passage, we need to recall Wittgenstein’s criticism of blanket demands for consistency proofs in mathematics. He questions whether such proofs can increase our trust in a given formalism when the alleged re-assurance is given in terms of a second formalism.Footnote ⁴⁰⁸ Mutatis mutandis in the case of epistemology: Once we raise the degree of safety to the level that the Cartesian skeptic demands, then the meta-investigation will be as doubtful as any of the first-level investigations for which we are seeking reassurance.Footnote ⁴⁰⁹

Finally, what should be our relationship to our epistemic system as a whole? The answer lies with paragraphs like the following: “I did not get my Weltbild by satisfying myself of its correctness; … No: it is the inherited background against which I distinguish between true or false.”Footnote ⁴¹⁰ What holds for the “Weltbild” also holds for the epistemic system. There is no position from which we can evaluate and assess our epistemic system as a whole.Footnote ⁴¹¹ What goes for language-games goes for our epistemic system, too. We cannot justify our epistemic systems prior to using it, nor can we ever hope to justify it in anything but a circular fashion.

So far I have focused on those elements in On Certainty that lend themselves more easily to an epistemological reconstruction. I now turn to elements in which the ethnological dimension is more prominent. Wittgenstein reflects on thirty-odd scenarios in which someone denies, or seems to deny, one of our certainties. These cases can be ordered in terms of the distance between Wittgenstein and the speaker:

1. (i) Speaker and audience are identical; Wittgenstein occupies both roles.Footnote ⁴¹²

2. (ii) The speaker is a “friend.”Footnote ⁴¹³

3. (iii) The speaker is G. E. Moore.Footnote ⁴¹⁴

4. (iv) The speaker is “someone,” presumably an adult in Wittgenstein’s culture.Footnote ⁴¹⁵

5. (v) The speaker is a “child” or “pupil.”Footnote ⁴¹⁶

6. (vi) The speakers are Christian believers.Footnote ⁴¹⁷

7. (vii) The speakers are from a “wild tribe” or people of the past.Footnote ⁴¹⁸

8. (viii) The speakers are Martians.Footnote ⁴¹⁹

The most obvious pattern in our responses to people who (seem to) deny one of our certainties, Wittgenstein suggests, is that these responses vary with social, temporal, or cultural distance. Simplifying a little, we can capture the degrees of distance from the “I” (i.e. Wittgenstein) and others in the following figure (Figure 1).Footnote ⁴²⁰

Figure 1

Representing cultural distance.

I shall work from the inside out. Consider first the scenario where someone, whom we deem an intellectual peer, acts or speaks in a way that suggests to us a rejection of one of our certainties (cases ii and iii). Wittgenstein reminds us that we would have little time for such behavior. We would not be able, or willing, to treat it as a blunder, error or mistake. Instead, we would feel compelled to invoke categories of intellectual disability. For Wittgenstein the most dramatic such case would be one in which Moore himself denies his celebrated “certainties.”Footnote ⁴²¹

“Someone” is not as intellectually close as a friend or colleague (case iv). But “someone” is held strictly to our standards. On Certainty includes cases where adult members of our own culture doubt what we take to be well-established historical truths; who deem all our calculations uncertain; who believe that cars grow out of the Earth, or cats on trees; who deny that water boils at around 100 C°. The natural verdicts are again couched in psychopathological terms.Footnote ⁴²²

Some of Wittgenstein’s comments suggest why our response to adult deviants of our culture – that is, adults denying one of our common-sense certainties – is strict and unforgiving. It is not just that such people are often impossible to understand and interact with. It is also important that they constitute challenges to our social order. They opt out of our “community which is bound together by science and education,”Footnote ⁴²³ and they refuse to accept the “certain authorities” that one “must recognize … in order to make judgements at all.”Footnote ⁴²⁴

This brings us to the children in our culture (case v). Wittgenstein believes that, even when they deny one of our certainties, children are not (normally) dismissed as having an intellectual disability. Instead, we take them to be joking or misled; we reproach them for talking about things they do not yet understand, and we educate them further.Footnote ⁴²⁵

Finally, we come to the three outmost rings in Picture 1, entitled “Other Cultures, Religions” “a,” “b,” and “c.” I subdivide the broader class into three subcategories since, on Wittgenstein’s account, we do not have one uniform response to the various encounters with people in other cultures who do not share all of our common-sense certainties. Roughly put, in a-cases, we treat the members of other cultures and epistemic practices as virtual members of our own and evaluate their actions and beliefs by the standards of our community. In c-cases, we are inclined not to engage in any form of evaluation. Finally, b-cases are ambiguous: we are torn between assimilating them to either a- or c-cases.

In a-cases, members of other cultures make claims that flatly contradict, or at least present as doubtful, one of our certainties. For instance, they claim that Moore comes from somewhere between Earth and Moon. They thus challenge some of our most cherished scientific and technological knowledge. Wittgenstein’s response to these imagined “people” (Völker) or “tribes” is uncompromising and critical: “[T]hey are wrong and we know it,” and their “system of knowledge … is the poorer one by far.”Footnote ⁴²⁶

An interesting transitional or ambiguous (type b) situation is this: “People have judged that a king can make rain; we say this contradicts all experience. Today one judges that airplanes, radios etc. are instrumental in bringing nations closer together and instrumental in spreading culture.”Footnote ⁴²⁷ On the face of it, the first sentence condemns the past people’s judgments as contradicting all experience. But the italicized “we” might make one wonder whether Wittgenstein wishes himself to be included. The same distancing strategy is apparent in the second sentence where Wittgenstein recalls what “man” judges. “Man” is a German indefinite pronoun standing for an unidentified group of people. It is likely that Wittgenstein’s attitude towards this man is critical and that he deems man’s judgment concerning the beneficial effects of airplanes and radios naïve. Read in this light, the remark does not suggest that the people who believe in rain-making kings are as wrong as we are when we believe in the positive power of air-travel and the radio. Instead, the paragraph tells us not to be too quick in accusing other cultures of contradicting experience. Maybe we are as wrong when making such accusation as we are – in Wittgenstein’s view – when we praise airplanes and radios.

Finally, we can turn to the c-cases in the outermost ring of Picture 1, where Wittgenstein favors a nonjudgmental attitude. It is important to understand what features the positions in this category share. First, the people holding these views may be of our own or of other cultures, past and present: “a king,”Footnote ⁴²⁸ “someone,”Footnote ⁴²⁹ “a human being,”Footnote ⁴³⁰ “Catholics,”Footnote ⁴³¹ and “human beings” in other historical “periods.”Footnote ⁴³² Second, no reason is mentioned that would imply that members of this diverse group have an intellectual disability.Footnote ⁴³³ Third, despite the fact that On Certainty emphasizes the possible intellectual strengths of the humans we encounter in this category, Wittgenstein also notes that we might feel “intellectually very distant” from them. We struggle to understand their reasoning and we “might therefore interrogate [them].”Footnote ⁴³⁴ There is no guarantee that such interrogation will be successful. Fourth, there is little that can be done in c-cases to rationally convince the other person, starting from assumptions that one shares with them. It is possible that we and they are so far apart that we simply “have to put up with it.”Footnote ⁴³⁵

Central to almost all of the cases where Wittgenstein favors a nonjudgmental attitude is an encounter with a religious system that one does not share. Either the case in question is straightforwardly one of this kind, or else the terminology used, or the way examples are juxtaposed, suggests that this type of encounter acts as the central model or analogue. If our interlocutors flatly deny a scientific result that has acquired the status of a certainty of (scientific) common sense, then we cannot but treat them as wrong. But the same attitude is not compelling when the encountered group expresses religious (magical, mystical) beliefs that do not clash with any particular such scientific certainty. In such cases, we have at least the plausible and permissible option of regarding the other system or practice as too intellectually distant for our vocabulary of appraisal to properly apply to it.

I now turn to a different string of remarks that focus on certainties as ethnological phenomena. For Wittgenstein, many certainties are best thought of as crystallizations of practices that subsequently are refined, purified and domesticated by epistemology. Here the ethnological perspective involved a criticism of epistemology. Wittgenstein’s analysis involves the use of sociological and metrological analogies. His main suggestions seem to offer not just a helpful and fine-grained vocabulary for the route from practices, through certainties, to epistemological theories; his suggestions can also be taken as offering a grid of analysis, or a spectrum of possibilities, that can usefully guide both our philosophical curiosity and our historical investigation. I begin by listing four transformations from elements of practice into certainties.

Transformation from result to unit of measurement. The proposition “Water boils at 100°C” no longer functions as the report of a measurement but as standard against which other claims can be measured.Footnote ⁴³⁶ The same is true for the mathematical proposition “2 × 2 = 4”: for adults this is no longer a genuine calculation but a measure for calculating correctly. According to Wittgenstein, such propositions are no longer true or false, known or dubitable; they are instead measures of truth and measures of knowledge.

Transformation from implicit assumption to criterion of group membership. Wittgenstein holds that we only ever formulate certainties explicitly once we encounter cognitive anomalies. Such cognitive anomalies arise with different groups of people who breach our routines and practices: children, the mad, philosophers, and the members of other cultures. In such cases our vocabulary of appraisal fails to get a proper grip. All we can do is acknowledge that they just are not “people like us.”Footnote ⁴³⁷

Transformations between the propositional and the non-propositional. Explicitly formulated common-sense certainties are verbal glosses on taken-for-granted routines and practices. These routines and practices are primary with respect to the glosses. The main function of using the verbal glosses is to re-establish a situation in which the verbal gloss is not needed.

Transformation of empirical observations to platitudes or proverbs. This is a transformation that Wittgenstein missed but that is explored by two ethnologists – one a card-carrying ethnologist, the other a social historian of science, both of whom take their starting point from Wittgenstein. I am referring to Clifford Geertz’ “Common Sense and Cultural System”Footnote ⁴³⁸ and Steven Shapin’s “Proverbial Economies.”Footnote ⁴³⁹ Shapin finds proverbs all over scientific practices of training and justification. The following list is but a small sample: “Ontogeny recapitulates phylogeny. For every action there is an equal and opposite reaction. Opposites attract. Nature abhors a vacuum. Nature doesn’t make leaps. Don’t waste clean thinking on dirty enzymes. Statistics is a way of making bad data look good. Absence of evidence is not evidence of absence.”Footnote ⁴⁴⁰ I am sure Wittgenstein would have gratefully accepted this proposal.

Wittgenstein also provides an interesting grid for capturing ways in which epistemological theorizing may go wrong in its treatment of common-sense certainties. His work on this score is part and parcel of his general critical attitude towards the epistemologists of his day. It comes out clearly in this passage from 1948:

Wittgenstein distinguishes the following forms of sublimation.

Sublimation qua reification. Certainties are reified into fixed and most certain items of knowledge. The key example here is G. E. Moore and his list of certain knowledge in “A Defense of Common Sense”Footnote ⁴⁴² and “Proof of an External World.”Footnote ⁴⁴³

Sublimation into universal intuition. Moore also thinks of certainties as universally valid insights or intuitions and ignores the fact that certainties might differ from one context to another. (”Catholics believe … that in certain circumstances a wafer completely changes its nature … And so if Moore said ‘I know that this is wine and not blood’, Catholics would contradict him.” OC 239)

Sublimation into mental state. Philosophers like Moore treat common sense not so much as a set of socially shared assumptions but as a quality of the state of mind of an individual. The criticism of this form of sublimation is of course a central theme in Wittgenstein throughout. He often attacks the tendency of philosophers to psychologize social phenomena, the tendency to misunderstand social states and processes as psychological states and processes.Footnote ⁴⁴⁴

Sublimation as rejection of platitude and proverb. Even though Wittgenstein overlooks the phenomena of proverb and platitude – the word “proverb” (Sprichwort) appears nowhere in his notebooks – attention for these phenomena would have fitted well with his emphasis on the ordinary. If Wittgenstein had paid attention to this topic he would undoubtedly treated philosophers’ inability to appreciate proverbs and platitudes as part and parcel of their general tendency to disregard and miss the ordinary and the everyday.

Sublimation as denying contingency. A fifth and final form of sublimation is philosophers’ lack of interest in the question whether and how historically contingent common sense shapes and structures their reflections. In order to combat this loss of contingency, Wittgenstein urges: “[I]t is important to imagine a language in which our concept ‘knowledge’ does not exist.”Footnote ⁴⁴⁵

Discussion

We now have enough material to highlight the ethnological – and relativist – approach of On Certainty, and to compare and contrast it with Schlick’s and Neurath’s positions. First, I hope to show that Schlick and Neurath relied on two different definitions of relativism, and that On Certainty was relativistic in Neurath’s, but not obviously in Schlick’s sense. I take this to be a result that is of both historical and systematic interest. After all, both definitions are still influential today, and the fact that their difference is rarely recognized accounts for much unfruitful controversy. Second, I shall highlight some intriguing parallels and tensions between Neurath’s and Wittgenstein’s versions of relativism. This topic has been largely invisible since both Neurath- and Wittgenstein-scholars have gone to great lengths to defend Neurath and Wittgenstein against the “charge” of relativism.Footnote ⁴⁴⁶ I would like to think that this section undermines these efforts in both cases.Footnote ⁴⁴⁷

In the Protokollsatzdebatte, Schlick and Neurath advocate two rather different forms of empiricism. The main goal of Schlick’s “consistent empiricism” is to identify the ultimate foundations of empirical knowledge, and to refute skepticism and relativism in the process. Relativism here centrally involves the idea that different systems of belief are equally valid (every consistent story is equally justified). As Neurath rightly points out, the battle against relativism and skepticism becomes so important to Schlick that he is willing to set aside the fight against metaphysics and idealism. And indeed, it is striking that Schlick’s discussion is not continuous with empirical science: he nowhere draws on, say, the psychology of perception, or highlights the very different ways in which the empirical sciences generate and process their empirical data. The account is domain-general and not specific to scientific observation. It is also foundationalist, individualist, atomistic, and solipsistic.

The overriding aims of Neurath’s “radical empiricism” throughout his life are to combat metaphysics within and outside of the sciences, and to give an account of scientific knowledge that highlights the commonalities and differences (including irresolvable disagreements) between and within particular sciences. This is not a project of pure philosophy; on the contrary, relying on empirical, historical, and sociological, studies of the sciences, Neurath aims for something like a naturalistic perspective on the sciences. To be sure, Neurath is not uncritical vis-à-vis the sciences; after all, he goes to great lengths to identify metaphysics in some of the social sciences. Neurath’s naturalism also explains why he regards skepticism as a pseudo-problem: For him the success of empirical science is a given. Relativism simply falls out of the empirical study of knowledge (in history and sociology). But note that relativism for Neurath is not the general claim that all coherent systems of beliefs or sentences are equally valid. Here Neurath deviates from Schlick. For Neurath, relativism is the rejection of absolutism in all its forms. Pseudorationalism is a central form of absolutism. Important ingredients of Neurath’s relativism are an emphasis on instability – “science with all its parts is always up for discussion” – and a stress on decision or judgment. And finally, the foundations of empirical knowledge can be studied only “top-down” rather than “bottom-up”: they must be studied holistically and dependent upon theories, scientific practices and communities.

Enter Wittgenstein. It is likely that Schlick would have classified On Certainty as relativist in his sense of the term. On the one hand, Schlick would have had little appreciation for Wittgenstein’s ethnological approach to knowledge and certainty. Schlick would have been quick to point out that such approach is at least methodologically relativist. On the other hand, Wittgenstein allows that different communities or cultures can have incompatible certainties, and he leans toward a coherence conception of certainties within the same system. To be sure, Wittgenstein does not clearly commit to a coherentist conception of truth or justification. But there is no denying that his conception of certainties and knowledge has much more in common with coherence theories than with the foundationalism of Schlick’s ilk. Wittgenstein is ready to speak of certainties as “foundation-walls.” But he immediately adds that these foundation-walls are of an unusual kind: they are “almost … carried by the whole house.”Footnote ⁴⁴⁸

Still, Wittgenstein does not think of different systems (of certainties, judgments, or beliefs) as equally valid. And in contemporary philosophy, the jury is very much out on the question whether a coherentist account of epistemic justification is invariably relativist.Footnote ⁴⁴⁹ In any case, it seems much more in line with Wittgenstein’s general commitments, to attribute to him the rejection of absolutes rather an adoption of the equal-validity thesis. The point is of more than historical interest: some authors today present Wittgenstein as an antirelativist on the grounds that he is not committed to equal validity.Footnote ⁴⁵⁰ Others call him a relativist insofar as he rejects absolutes.Footnote ⁴⁵¹ Both sides may be right.

Presumably Schlick would have been shocked by Wittgenstein’s relativist turn in his later philosophy. After all, Schlick’s foundationalism is reminiscent of the architectonics of the Tractatus, at least if we draw a parallel between elementary propositions and affirmations. No doubt, Schlick felt vindicated by this parallel. And thus he might have taken the author of the Tractatus to be firmly in the absolutist camp. Schlick was not the only Viennese philosopher who thought of Wittgenstein as an absolutist in the early 1930s. Neurath agreed with this assessment. But Neurath knew little about how Wittgenstein’s views developed after the Tractatus. Only Schlick and Friedrich Waismann heard from Wittgenstein directly, and they shared the information with but a select few. Still, not all of these “select few” kept quiet. In April of 1935, Neurath learned from Schlick’s former PhD student, Heinrich Neider, that Wittgenstein “continuously develops in a relativistic sense”: “I see his relativistic development in that he has dropped the ideas of a primary ideal language and the doctrine of atomic sentences, and in that he emphasizes the conventional character of all forms of language.”Footnote ⁴⁵² Neurath was greatly surprised.Footnote ⁴⁵³ Unfortunately, he was not kept up-to-date subsequently, and it is unlikely that he received detailed news about Wittgenstein’s lectures and manuscripts of the 1940s, when Wittgenstein was in Cambridge and Neurath in Oxford.

Why does Wittgenstein never mention relativism anywhere in his oeuvre? He can hardly have been ignorant of the many German- and English-language debates surrounding relativism at the time. Perhaps he avoided the term since the debate between Schlick and Neurath had alerted him both to the radically different available definitions of relativism and to its many negative connotations.

It is interesting to note that Schlick attacks both skepticism and relativism in his 1934 paper. Other philosophers too saw skepticism and relativism as one problematic package. For instance, Husserl attacked psychologism as a “sceptical relativism.”Footnote ⁴⁵⁴ This should not occasion surprise: From the perspective of the absolutist, relative knowledge or relative morality do not qualify as knowledge or morality. Even the skeptic might see things in this way. Relativists disagree. They wish to distinguish themselves from both absolutists and skeptics. In this vein, Neurath is making an effort to distance himself from skepticism. He takes relativism to be an obvious implication of the plurality of the sciences and of the rejection of metaphysics and pseudorationalism. But he nowhere shows any sympathy for skepticism. Indeed, he largely ignores skepticism. Wittgenstein attacks skepticism, agreeing on its relevance with Schlick. But Wittgenstein leaves his own relativistic leanings unlabeled. Almost as if he took these leanings to be obvious and unremarkable.

Needless to say, Schlick’s and Wittgenstein’s attempts to answer skepticism are quite different. Schlick attacks skepticism at the very starting point of empirical knowledge. Skepticism is taken by Schlick as a threat to the individual’s perceptual knowledge. Wittgenstein attacks skepticism much more broadly: to wit, as a challenge to individual and communal knowledge, scientific and everyday. On Certainty focuses on the role of doubt in systems of knowledge and tries to refute the skeptical idea according to which doubt can infect our empirical knowledge all at once. We might call this a “system response” to skepticism. It is noteworthy that the central role played by our system of empirical beliefs is much more in line with Neurath’s than with Schlick’s empiricism. Wittgenstein here stands roughly halfway between Schlick and Neurath: Schlick is right to combat skepticism, but a Neurathian view of knowledge is what is needed to combat skepticism successfully.

I now turn to highlighting some intriguing parallels and tensions between Neurath’s and Wittgenstein’s versions of relativism. One difference is in the scope of their respective pluralism. In Neurath, we get a naturalist pluralism about the sciences that develops into a more relativistic direction over time. A pyramidistic conception of the unity of science gives way to the idea of an encyclopedia or mosaic of the sciences. The latter conceptions go together with the acceptance of fundamental and irresolvable disagreements between and within individual sciences. In Wittgenstein’s On Certainty we also find pluralistic motifs. One such motif is that there are many different kinds of certainties. More important for present concerns is the notion that different disciplines or cultures might have different certainties of the same kind – different certainties, say, about religious matters. That is to say, Wittgenstein’s pluralism is not confined to the sciences. He is also interested in religions, in everyday or common knowledge, and in the knowledge that an individual has of themselves. This is of course a central ingredient of his ethnological approach.

Neurath’s relativism centrally involves a stress on instability, ambiguity, judgment, and “decision,” on “underdetermination,” and on the rejection of “pseudorationalism.” How does On Certainty stand relative to these ideas? To begin with instability, Neurath writes that “[s]cience with all its parts is always up for discussion.”Footnote ⁴⁵⁵ Clearly, Wittgenstein does not accept this view. On the contrary, the emphasis on certainties is precisely meant to highlight that in all types of communities, certain beliefs are definitely not up for discussion. Certain beliefs are like “hinges” that have to stay in place so that other beliefs can shift and change: “[S]ome propositions are exempt from doubt, are as it were like hinges on which those turn.”Footnote ⁴⁵⁶ This does not of course mean that there can be no changes in certainties. As Wittgenstein’s “river-bed” analogy indicates, even certainties get pulled into processes of change and replacement: “[T]he river-bed of thoughts may shift.”Footnote ⁴⁵⁷ We also get a more straightforward account without the metaphor: “what men consider reasonable or unreasonable alters. At certain periods men find reasonable what at other periods they found unreasonable. And vice versa.”Footnote ⁴⁵⁸ Still, the gap between Neurath and Wittgenstein remains. It foreshadows debates that have been central in the philosophy of science ever since the Kuhn-Popper debates of the second half of the twentieth century.Footnote ⁴⁵⁹

The difference regarding instability is mirrored in Neurath’s frequent emphasis on decision, judgment and underdetermination concerning protocol sentences. Interestingly enough, On Certainty contains passages that go in a similar direction. At one stage, Wittgenstein wonders “whether knowledge is akin to a decision.”Footnote ⁴⁶⁰ And he writes that “experience is not the ground of our game of judging.”Footnote ⁴⁶¹ A slack between our epistemic system or practice and the world is also suggested elsewhere in On Certainty, for instance by: “It is always by favor of Nature that one knows something.”Footnote ⁴⁶² If there were no slack, no favors would be needed. Or consider Wittgenstein’s answer to his own question: “What if you had to change your opinion even on these most fundamental things?” The answer is: “You don’t have to change it. That is just what their being ‘fundamental’ is.”Footnote ⁴⁶³ It also seems plausible that Neurath and Wittgenstein share the animus against “pseudorationalism.” Although Wittgenstein never mentions the term, it seems to fit quite well his criticism of philosophers meddling in sublimation. Such sublimation is challenged by an ethnological perspective that recognizes certainties as hard to reconstruct and ultimately based in our biological and social history.

Finally, it is intriguing that both Neurath and Wittgenstein – each in their own ways – take an interest in the budding social sciences of their day and both Neurath and Wittgenstein use social-scientific perspectives on philosophical issues. Neurath hints in the direction of a sociology of scientific knowledge and emphasizes the political nature of scientific knowledge. Moreover, Neurath takes a genuine interest in the history of science, and many of his observations of the state of the sciences of his day are well informed. But Neurath’s philosophy and criticism of Schlick does not rest on an ethnological or sociological approach to certainty and knowledge. When arguing against Schlick, and when developing central elements of his “physicalism,” Neurath is not interested in our attitudes toward people at different cultural distances indicating commitments to different certainties, not concerned with how communities are constituted by shared certainties, and not preoccupied with transformations of practices into certainties, or with philosophical sublimations. (To be fair, he does pay some attention to sociologists’ sublimations in his philosophy of the social sciences.)