Empiricism at the Crossroads: The Vienna Circle's Protocol-Sentence Debate Revisited

By Thomas Uebel

Rather than a monolithic movement of naïve empiricists, the Vienna Circle represented a discussion forum for what were sometimes compatible, sometimes conflicting philosophical approaches to empirical evidence. The Circle’s protocol-sentence debate here reconstructed and analyzed provides an exceptional vantage point from which to survey the various options and choices of the participants. Author Thomas Uebel mines the diaries, letters, and notes of the group’s leading philosophers to show how their ideas emerged from real-world arguments, personal relationships, and historical settings.

• Language: English
• Publisher: Open Court
• Release date: Nov 2, 2015
• ISBN: 9780812699296

Book preview

Preface

Empiricism, like many other isms, has been in an on-going process of development. To speak of empiricism at the crossroads is not to single out a specific episode or period. The point of the title of this book, therefore, is not to draw attention to a type of crisis which empiricists never sensed before or since the episode designated by the this book’s subtitle. Rather, it is to draw overdue attention to the fact that in this context and period, too—at the high-point of Viennese neopositivism, contrary to its common caricature as naive scientism—empiricism was in question. And it is to claim renewed attention for the way in which empiricism was questioned there.

What justifies the excavation of the trenches of the Vienna Circle’s protocol-sentence debate undertaken here is not just that it allows us to see in close interaction and at a critical stage the development of the views of three leading theorists of a group of self-styled philosophical revolutionaries. What justifies it is also the suspicion that certain forms of Vienna Circle philosophy, especially in judicious combination, possess far more resources for post-metaphysical philosophy of science than logical positivism is usually credited with. Just what these resources are remains controversial, but certain discussions of Vienna Circle philosophy easily double as discussions about empiricism today. For instance, consider what have become standard—though by no means universally accepted—interpretations within revisionist Vienna Circle scholarship: Carnap as a neo-Kantian thinker of highly original cast; Neurath as a naturalistic epistemologist avant-la-lettre. In this respect, finding empiricism at the crossroads in the Vienna Circle has to be understood widely: one of the alternatives was to abandon empiricism. Empiricism itself was in question. Whatever their philosophical prehistories, however, the fact remains that all three of Carnap, Schlick and Neurath insisted upon thinking of themselves as empiricists, despite the fact that they disagreed over precisely what that meant. So even though we will not here pursue Carnap’s neo-Kantian proclivities too far, it must be noted that it cannot be discounted from the start that the contents of some forms of early logical empiricism do not conform to the label under which they are catalogued.

Some sixteen years ago I ploughed these fields for Overcoming Logical Positivism From Within. While I still agree with most of what I wrote then, there is also much that I now would prefer to have put differently. More importantly, there are some claims in that book that I no longer agree with and some respects in which the story I told, I now recognize, is incomplete. I am grateful therefore to the publisher and the editor of the series Full Circle: Publications of the Archive of Scientific Philosophy, Hillman Library, University of Pittsburgh for their invitation to revisit the topic for their series. This has given me the opportunity to put things right as best I can. While the overall structure of this book bears some resemblance to the older one, the text has been mostly rewritten and a considerable amount of additional source material has been worked in, with new chapters added and old ones dropped or radically altered. Still I must resist claiming completeness, if only because numerous topics related to the protocol-sentence debate have not been pursued as far as they might have been. The most important change, however, is due to the changed dialectical situation in this part of the history of philosophy of science and to a certain change in the general philosophical climate.

Sixteen years ago my main aim was to reintroduce Neurath as a serious thinker to the analytical tradition which remembered him only vaguely as an organizer and polemicist, if at all. This required above all bringing out the differences between his position and that of his fellow-disputants in the protocol-sentence debate and I chose to do so by characterizing him as an early proponent of naturalistic epistemology. That Neurath was an independent and, indeed, naturalistic voice in the Vienna Circle is not less worthy of investigation today than it was then, but it is less the news it once was. Fortunately, the last sixteen years have also seen a veritable explosion of Vienna Circle scholarship and, while my thesis held up, this brought advances in the understanding of many other aspects of Vienna Circle and early continental logical empiricist philosophy. Writing a new book on the protocol-sentence debate thus gives me the opportunity not only to update the old one in relevant respects, but also to give the investigation a somewhat different setting.

If it is fair to say that sixteen years ago the rediscovery of the actual work of the early continental logical empiricists was still gathering speed, then it may be said that present-day emphasis is moving again towards the re-evaluation of their contribution to philosophy as a whole. In step with this, my concern now is less exclusively the differentiation of Neurath within the Vienna Circle and more the delineation of an emerging consensus on metaphilosophical matters on the so-called left wing of the Vienna Circle (which included, besides Carnap and Neurath also Philipp Frank and, until his early death, Hans Hahn). This consensus was always only partial and in fact more potential than real, yet it is, I believe, of great doctrinal interest. It aimed precisely at an improved understanding of what empiricism could be as a post-metaphysical philosophy of science—and so speaks directly to contemporary debates. The analysis of the Vienna Circle’s protocol-sentence debate given here accordingly seeks to provide not only an account of what happened that is as accurate as possible and as detailed as needed, but also to contribute some suggestive insights into what could have been but was not and yet remains of interest. To explore this potential consensus more fully, however, must be left for another book.

There are three presentational conventions used in this book. Cross-references to other parts of the book are given in the text by citation of section or subsection in parentheses. In line with the historiographical nature of much of this work, references in the text and footnotes are given by year of first publication, followed by year of publication of the edition or translation employed in square brackets (both decoded in the Bibliography). Translations from German of previously untranslated texts (and insertions in square brackets in quotations) are by the present author. The German text is provided in cases where the original has remained unpublished; otherwise the reference is to the German edition used.

For their generous help in locating and making available to me unpublished materials I wish to thank Brigitte Parakenings of the Philosophisches Archiv, Universität Konstanz (especially for her transcriptions of shorthand jottings on Carnap’s manuscripts), Brigitta Arden of the Archive of Scientific Philosophy at the University of Pittsburgh (especially for her transcriptions of shorthand from Carnap’s diary), Reinhard Fabian of the Dokumentationszentrum für östereichische Philosophie, Graz, and Godlieve Bolten of the Noord Hollands Archief, Haarlem. For permission to quote form their holdings I wish to thank the Trustees of the Archives of Scientific Philosophy, Hillman Library, University of Pittsburgh, and the trustees of the Wiener Kreis Stichting, Amsterdam. I also wish to thank Fred van der Zee of Editions Rodopi for permission to use material from Overcoming Logical Positivism from Within, Steven Gerrard of Acumen for permission to use material from The Central Works of Philosophy, Vol. 4, edited by John Shand, and the Editors of Studies in History and Philosophy of Science, Philosophy of Science, Synthese, History of Philosophy Quarterly, British Journal for the Philosophy of Science, International Studies in the Philosophy of Science, and Proceedings of the Aristotelian Society for permission use material that appeared in their pages.

Unfortunately I cannot here name all the people who have helped me form the views expressed on this book, though I am grateful to all of them—like Sylvain Bromberger, Jim Higgenbotham, and Paul Horwich who helped at the very start. But I must thank those with whom I have had the privilege of discussing matters raised in this book or pertaining to them over some part or even all of the last sixteen years: Steve Awodey, Michael Beaney, Jacques Bouveresse, Nancy Cartwright, André Carus, Jordi Cat, Delphine Chapuis-Schmitz, Robert S. Cohen, Richard Creath, Hans-Joachim Dahms, Arthur Fine, Michael Friedman, Warren Goldfarb, Rudolf Haller, Michael Heidelberger, Don Howard, Eckhart Köhler, Juha Manninen, Thomas Mormann, Elisabeth Nemeth, Paul Neurath (d.2001), Camilla Nielsen, Thomas Oberdan, John O’Neill, Paolo Parrini, Christopher Pincock, George Reisch, Alan Richardson, Thomas Ricketts, Thomas Ryckman, Friedrich Stadler, Graham Stevens, Michael Stöltzner, Rowland Stout and Pierre Wagner. In addition I’d like to thank Christopher Pincock for comments on an earlier version of Chapter 2, Juha Manninen for comments on versions of Chapters 4 and 5, and John O’Neill for comments on the penultimate versions of Chapters 12 and 13. Special thanks go to George Reisch for copy editing not only with professional care but a specialist’s eye for detail and nuance. Needless to say, all remaining mistakes are my own.

Overcoming was dedicated to my mother, Doris Uebel. This book I wish to dedicate to her memory and to my wife Susan Watt and our son Felix with love and gratitude.

Chapter 1

Introduction: Naturalism and Deflationism, the Vienna Circle and the Protocol-Sentence Debate

The Vienna Circle’s so-called protocol-sentence debate concerned the form, content and status of scientific evidence statements and extended throughout most of the public phase of the Circle from 1929 to 1936, with Rudolf Carnap, Moritz Schlick and Otto Neurath as its main protagonists. Notwithstanding its being remembered mainly as an accumulation of errors—or perhaps because of it— this debate has never been far from the horizon of analytical epistemology. W.V.O. Quine’s proposal to naturalize epistemology was clearly marked by his reading of Carnap’s early contributions to this debate. Wilfred Sellars’s critical remarks on Schlick’s Konstatierungen form a major premise of his argument against the myth of the given. And, more recently, Ernest Sosa sought to gain neo-foundationalist ground by castigating the supposed errors of Neurath’s coherentism in a Schlickean vein.¹ My purpose in revisiting the debate is to investigate its own course in the first place and only then to pronounce, in a preliminary way, on what might still be learned from it.

See Quine (1969a), Sellars (1956 [1997, 68–73]), Sosa (1997).

Like the philosophy of the Vienna Circle generally, the protocol-sentence debate is much richer than common portrayals of it suggest, both in terms of the variety of positions taken and in terms of their fruitfulness. For instance, when we do consider the debate in due detail, we find that Quine was not the first naturalizer of epistemology in the history of logical empiricism, that Schlick was not quite the foundationalist he was portrayed as, and that Neurath was not committed to internalist justifications of an individual’s beliefs. In other words, many of the types of meta-epistemological reflections that philosophers became used to in response to the failure of the logical empiricist project as commonly conceived—putting science on secure foundations—were already underway within the Vienna Circle itself. Far from being naive empiricists, its protagonists saw empiricism itself under threat and in trying to meet this threat accepted that their efforts would not leave empiricism unchanged. With the protocol-sentence debate serving as the public forum for these reflections, it is not only their results but also the discussions themselves that are of interest for a philosophy of science that claims awareness of its own past (if only not to repeat its errors).

This introduction will set the stage for the lengthy inquiry needed to make good these claims. First, to provide some further motivation for considering the protocol-sentence debate as a site for discussions about the renewal of empiricism, the simile of Neurath’s Boat is introduced as a motto for naturalism in the context of systematic epistemology. Its promise is discussed in the wider context of the Vienna Circle’s philosophy of science, especially Carnap’s deflationism (1.1). Then we backtrack a bit to give a first overview of the Vienna Circle as a whole and its philosophy (1.2) and the different interpretative strategies that have been employed to come to terms with it (1.3). (These intermediate two sections can be skipped by readers familiar with the terrain.) Then the setting for the detailed reconstruction of the protocol-sentence debate that follows is provided (1.4), ending with an overview of the course of investigation (1.5).

1.1 Neurath’s Boat as a Motto for Naturalism

Everybody familiar with contemporary analytical philosophy is likely to have come across the simile known as Neurath’s Boat: We are like sailors who must rebuild their ship on the open sea, never able to dismantle it in dry-dock and to reconstruct it there out of the best materials. (1932d [1983, 92]) For Quine, to whose adoption its wide coinage is owed, this simile eloquently expresses that it is impossible for philosophers to succeed in basing scientific knowledge on indubitable foundations, and the moral that flows from this: epistemology should be naturalized. Very broadly, this means that the theorist of knowledge should forego the a priori standards of traditional epistemology and presuppose the available results of scientific investigations in order to understand scientific knowledge. Science must be made to ‘explain itself’. The boat simile has come to encapsulate this naturalistic approach to epistemology. Despite Quine’s frequent remarks on the Neurathian provenance of the parable of the mariners, however, it is by no means obvious whether it meant for Neurath what it meant for Quine.²

See, e.g., Quine (1950 [1953b, 79]), (1953a [1966, 223, 225]), (1958 [1969b, 16]), (1960, 3f., 124, 210), (1968a [1969b, 127]), (1968b [1981b, 178]), (1969a, 84f.), (1981a, 72).

The answer that will be developed here is that Neurath did indeed develop a conception of naturalistic philosophy of science, albeit a rather unusual one that must not be mistaken with Quine’s. In different respects, Neurath’s naturalism is both more radical and less radical than Quine’s and, as a result, holds out different promises to epistemologists of science. In fact, it will be argued that Neurath’s naturalism represents but one part of the Vienna Circle’s enduring legacy for the philosophy of science. Another equally valuable part consists in Carnap’s contribution; the complementarity of their approaches is the topic of Chapter 12. The initial focus on Neurath’s naturalism recommends itself for good reasons, however. It confounds in the most spectacular manner what is still the received view of Vienna Circle philosophy and it allows the most direct interpretive route to what this has long obscured. For to ask what Neurath’s naturalism was and how it manifested itself is to ask just what went on in the Vienna Circle’s notorious protocol-sentence debate. Reconstructed in situ in this debate, a rather different conception of the philosophy of science emerges than what standard histories of the logical positivism provide. Indeed, the recognition of the plurality of conceptions of philosophy of science entertained in the Vienna Circle is the very first result of revisiting its protocol-sentence debate.

1.1.1 Neurath’s Boat and the Theory of Scientific Knowledge

The systematic questions in the theory of knowledge are generally organized most easily around the question of scepticism. To be sure, Neurath had no sympathy for radical scepticism as a positive theory, nor did he think it worth his while to try to refute it. Yet sceptical questions provide a fitting heuristic for his program, if only partially so. Scepticism can frame questions that Neurath did respond to, albeit at a level skeptics would find insufficient to allay their concerns. Neurath sought to keep in check the individualist excesses of traditional philosophy, but he saw no need for systematic sceptical challenges once reasoned and evidence-based intersubjective agreement underlay judgments. (Like idealism, radical scepticism fell by the wayside of his naturalism on account of its unsustainable absolutization of the individual knower.)

Systematically speaking then, the central problems of the theory of knowledge concern the justification of knowledge claims and are most directly raised by sceptical questions. Sceptical questions challenge the claim that we do possess knowledge at all. (They can be raised about all kinds of knowledge claims, and within each kind, about all areas; in this study, we are concerned with empirical scientific knowledge claims.) Over the course of history, sceptical challenges became more radical. Whereas the sceptic of antiquity merely doubted our knowledge of the true nature of the natural and moral world, the modern sceptic doubts our knowledge of the very existence of the purported object of knowledge. Cartesian scepticism doubts the existence of the physical world; since Nietzsche, scepticism even doubts the existence of an inner world, in particular, the correct grasp of the content of one’s own mental states.³ The interest of sceptical arguments, however, lies not only in raising a question of fact, namely, whether we have knowledge, but also in raising a more conceptual question, namely, what knowledge is. They do so because they raise the question of right—Are we justified in our knowledge claims?—and so raise the question of what kind of justification is appropriate for knowledge claims.

On the absence of idealism in Greek philosophy, see Burnyeat (1982); for the anticipation of contemporary radical scepticism in Nietzsche, see his (1883–88 [1968, #476]): Our inner world, too, ‘appearance’! For a Vienna Circle appreciation of Nietzsche see Frank (1917); for observations on the Vienna Circle’s reception of Nietzsche, see Fischer (1982a); especially on Neurath’s, see Nemeth (1992).

Against the sceptical challenge, the defense may either accept the sceptic’s premises and show the mistakes in the reasoning, or accept the sceptic’s reasoning and challenge the premises in more or less radical ways. In common terminology, the defense against the sceptic may either be foundationalist or antifoundationalist. Foundationalists hold that all empirical knowledge consists of basic and non-basic beliefs, and that basic, non-inferentially justified beliefs can be found.⁴ Antifoundationalists, by contrast, hold that there are no non-inferentially justified beliefs that could help the justification of empirical knowledge claims; consequently, they must provide an alternative account of the structure of justification.

As is apparent from this characterization of foundationalism, justification is here understood to concern the grounds or reasons for a person’s believing a proposition. The focus is on doxastic justification, as opposed to propositional justification, which justifies the proposition a person believes. In so far as traditional epistemological terminology applies to the philosophy of science of the Vienna Circle at all, it is doxastic justification that was at issue; when Carnap turned his back on epistemology, however, it was propositional justification that became the issue.

Different ways in which theorists may understand the sceptical challenge must also be distinguished. Concerned with the knowledge claim of science, for instance, they may understand their task to provide its defense ‘from without’ or ‘from within’. That is to say, they may seek to defend it starting from premises held uncontroversially true or by assuming the truth of some of the claims of science to be defended. The naturalistic approach engages with scepticism only as a heuristic strategy and so attempts a defense ‘from within’. Accordingly, naturalists may entertain conceptions of the nature of justification at odds with attempts to answer the radical sceptic. For instance, externalist conceptions have been developed that deny what traditional internalist ones insist upon, roughly, that knowers themselves must have access to or be in possession of the justification for their beliefs.⁵

This is very rough; see Kornblith (1999) for survey of recent externalisms and internalisms.

Whereas the foundationalist owes an account of the nature of basic beliefs and of the relation of epistemic ascent from the basic to the non-basic beliefs, the antifoundationalist owes an argument establishing that no non-inferentially justified propositions can be found and an account of the nature of justification without foundations. Neurath’s simile of the Boat suggests that inferential justification was all the justification that could be provided for scientific knowledge claims and that any hope for a secure defense of science ‘from without’ had to be abandoned. Thus Neurath requires, first, an argument against foundations, and, second, an account of how that non-foundationalist, naturalistic justification would go. Neurath’s argument against foundations and his positive account long remained unknown. In fact, there continues to be considerable puzzlement over how Neurath could have anticipated a naturalistic approach to scientific knowledge, for the Vienna Circle to which he belonged is commonly held to have pursued the radically different strategy of opposing the sceptic head-on with, more or less, traditional foundationalism.

Among others, it is this matter that this study seeks to address in its full complexity. In order to do so, care must be taken not to let our expectations of what we might find frame our perception of the protocol-sentence debate too rigidly. For instance, as introduced above, Neurath’s Boat speaks to the phenomenon of knowledge and the fallibility of the justifications we can provide. We should be wary, however, of translating this concern without much ado into that of traditional epistemology. What the protocol-sentence debate was about was the basis of scientific knowledge claims. Whether this easily translates into concern with the justificatory basis for the beliefs of individual knowers is debatable. Neurath and, in a different way, also Carnap were clearly concerned with epistemological matters broadly speaking. But we would bar access to their most distinctive contributions were we to read them as theses pertaining to individualist epistemology.

1.1.2 Neurath and Quine on Naturalism

Much contemporary philosophical interest in the Vienna Circle concerns its role in the formation of the views of Carnap and Quine. When such interest is focussed upon the Circle’s philosophies of empirical science, as opposed to its philosophies of logic and mathematics, contemporary discussion typically sees the Vienna Circle opposing Quine on the topic of naturalism, for it rightly views Quine to be in opposition to Carnap. Typically, one wants to know whether Quine was justified in his particularly radical conception of naturalism: need naturalistic epistemology renounce appeal to meanings and propositional attitudes; indeed, must it shun all social science? Moreover, was Quine justified in rejecting any separation of logical and empirical statements such that his epistemological holism was unable to draw any distinction between the two and had to reject any conception of the a priori whatsoever?

But Quine’s radical opposition was not the first naturalistic response which Carnap had to face.⁶ In the Vienna Circle, Neurath urged that the philosophy of science be naturalized.⁷ What is more, Neurath’s challenge to Carnap was not only, as a matter of historical fact, a precursor of, but also, as a matter of doctrine, an alternative to Quine’s. In particular, and quite apart from other doctrinal differences that obtained between them, Neurath sought to develop something quite simple in conception yet very hard to realize: a conception of scientific knowledge as a tool for empowerment, an instrument of emancipation. With Neurath, naturalism shouldered the concerns that motivated the early Marx and Engels in their Theses on Feuerbach and German Ideology.⁸ What makes Neurath interesting therefore is not just his de facto anticipation of a naturalistic philosophy of science, but the different, wider objectives behind his naturalism. Yet still before this ideological dimension of the philosophy of science is touched upon, there are significant differences that must be noted lest the real Neurath once again slip through the interpretive net.⁹

For general discussions of contemporary epistemological naturalism, see e.g. Kornblith (1985), Maffie (1990), Laudan (1990), Kitcher (1992), Goldman (1994).

That, as naturalists, both Quine and Neurath reject the challenge to disprove radical scepticism, is argued in Uebel (2004a).

For further discussion of this see sections 12.1.3 and 13.6.

On the view of the weltanschauliche background to much of the Vienna Circle’s philosophy, especially that of its so-called left wing that is presupposed here, see Uebel (2004b), (2005a). For more literature in this topic, see fn. 39 below.

To fix ideas, let’s ask what makes for a naturalized or naturalistic theory of knowledge. Here recall the warning that all too often no distinction is drawn between general epistemology and philosophy of science, with the result that, as in the case of Quine, it is left unclear whether what he called epistemology naturalized was his model for philosophy of science or for epistemology in general. Yet it is not clear that the two are indeed coincident: their units of analysis can reasonably be claimed to be different. In epistemology it is the beliefs of an individual for which justification is to be provided; by contrast, in philosophy of science it is public knowledge claims that are to be substantiated. This difference may not seem to amount to much at first glance, but the pursuit of justification for these different types of knowledge claims can lead to the adoption of quite different philosophical strategies. In Chapter 11, we shall discuss in detail the striking results that emerge when knowledge claims are discussed as public ones of a particular type rather than as expressions of an individual’s beliefs. Yet even before that we shall see how related differences over how to legitimate scientific knowledge claims emerged in the course of the protocol-sentence debate and led to different positions by different theorists. While Neurath must be read as a philosopher of science, Schlick, by contrast, drew no such distinction between epistemology and philosophy of science; Carnap displayed a different attitude again, arguing initially from the perspective of individualistic epistemology but switching to the explication of public knowledge claims roughly half-way through the protocol-sentence debate.

So what does it mean for naturalistic philosophy of science to ‘explain science by science’? Given the ongoing developments in naturalism, it is best to characterize this program not in terms of any particular doctrines which would prejudge its outcome, but rather in terms of its overall methodology. One characterization that suggests itself is that what makes inquiries naturalistic is the decision to employ only such concepts as can be explicated, in turn, by concepts which have proven themselves in the explanation of processes antecedently understood to be natural ones.¹⁰ Only ultimately scientifically legitimate concepts may thus figure in the justification of knowledge claims. Clearly, everything depends on what’s considered natural. Note here that Quine tended to take a very strict view of the matter. Natural processes were processes happening in space and time and open to empirical inquiry. Quine thus replaced talk of epistemic priority by talk of causal proximity to sense receptors (1969a, 84–5), talk of subjective similarity by talk of innate quality spaces (1968a [1969b, 123], 1975, 69–71), and talk of the given by talk of observation sentences (1975, 73). Observation sentences in turn were defined as occasion sentences—as opposed to standing or eternal sentences—whose occasion is not only intersubjectively observable, but generally adequate, moreover, to elicit assent to the sentence from any present witness conversant with the language. For Quine then, neurophysiology, evolutionary biology and behavioristic psychology furnished the explication of the epistemological concepts required by naturalism. Both Quine and Neurath required a scientific explication of the epistemological concepts of science. They differed on what kind of sciences could be employed in this explication.

Paul Pietroski helped with this formulation long ago. Since the formal sciences are typically involved in explanations of natural processes, Ernest Nagel and Richard Brandt’s classic definition of the meta-epistemological doctrine of naturalism is upheld, in particular its second part. According to them, naturalism holds that epistemological statements are truth-valuable, that epistemological terms can be explained by way of empirical and logical concepts exclusively, without using other epistemological terms in the process and that epistemological principles have the same cognitive status as statements in the empirical sciences, and that like the latter they must be judged in the light of relevant data obtained by sensory observation. (1965, xiii).

To get a sense of what is involved in Neurath’s non-standard naturalism, consider two different ways of reading Neurath’s Boat.¹¹ There are what can be called the austere and the rich readings of what is involved in re-building the metaphorical ship, understood to represent scientific knowledge. Both readings agree that it is impossible to repair any deficiencies in the body of scientific knowledge by stepping beyond it and appealing to insights that could not be challenged. The two readings disagree on the standards with which such an repair of science ‘from within’ has to meet. According to the austere reading, whatever cannot meet the individuation conditions of physical kinds has to be consigned to the scrap-heap of mythology and cannot be used in reconstruction. Given the austere reading, Neurath’s Boat means that natural science explains (knowledge of) natural science. According to the rich reading, all human cognitive endeavors could serve as replacement planks and beams to keep the ship of science afloat, provided one condition is met. This condition is that the assertions characteristic of these various enterprises be empirically testable (and thus, presumably, something less than a reduction of all explanatory types to physical types is required). Given the rich reading, Neurath’s Boat means that natural and social science explains (knowledge of) natural and social science.

For two different readings of the Boat by way of comparison of the theories of Quine and Davidson see Hookway (1988).

It is clear that the austere reading is Quine’s, but which one was Neurath’s? To claim that the rich reading was what Neurath intended would support the earlier claim that Neurath’s naturalized philosophy of science not only constituted a precursor, but also an alternative to Quine’s. It is also pretty clear that only the rich reading supports a conception of scientific knowledge compatible with the idea of knowledge as an instrument of emancipation. However, what I called the rich reading of Neurath’s Boat is more liberal also in still another sense. It allows not only the empirical sciences to aid in the explanation of scientific knowledge claims. This is important. To say that the assertions of the sciences that help in explaining science must be empirically testable suggests that Neurath’s naturalism shares Quine’s rejection of the distinction between the logical and the empirical and his semantically undifferentiated epistemological holism. This suggestion is problematical, however.

To begin with, this suggestion is misleading for one simple reason: as a logical empiricist Neurath shared in the overall strategy of considering mathematical knowledge as broadly speaking logical knowledge and its propositions as analytical ones, in contrast to the synthetic propositions of the empirical sciences. Nowhere does Neurath indicate dissent from this strategy that was basic to the Vienna Circle’s rejection of the a priori philosophies of old and of the synthetic a priori of Kant. But this suggestion is also problematical in another respect. It presupposes that naturalism in the theory of knowledge must reject the epistemological distinction between the logical and the empirical in the manner of Quine. Yet it is not clear that naturalism must do so, unless it were only another name for radical empiricism.¹²

Note, for instance, recent attempts by self-styled naturalistic epistemologists to reserve a place for the a priori under some description: Rey (1998), Goldman (1999).

1.1.3 Naturalism and Deflationist Philosophy of Science

An alternative emerges once the motto ‘explaining science by science’ is reconsidered. What according to this motto is considered unnatural, is only the attempt to explain science by supra-scientific means: to explain it by appeal to the kind of insight that cannot be sustained by any science. (Just that, of course, the Vienna Circle opposed and combatted as metaphysics.) Under such a construal, naturalism does not rule out appeal to the sciences of logic and mathematics as unnatural only if they were assimilated, via their undeniable indispensability, to the empirical sciences; rather, logic and mathematics are accepted on their own terms. Of course, to sustain the Vienna Circle’s antipathy to metaphysics, naturalism must still delimit sharply the kind of insights that logic and mathematics build upon. So if these sciences can be considered purely formal (as they were in the Vienna Circle), then appeal to logic and mathematics will not contravene this non-Quinean version of naturalism. (Quine’s naturalism, it will be remembered, built upon the perceived failure of Carnap’s own, nonstandard version of logicism.) Neurath’s naturalism, according to the claim defended here, was of this non-Quinean sort.

At this juncture it may be asked, of course, what sense there is in designating this reading of Neurath’s philosophy of science as a naturalistic one. Given the enormous weight that Quine’s version possesses as the paradigmatic version of naturalism, it must be conceded that the puzzlement is justified. Yet there are two good reasons to retain this label for Neurath. First, in considering the motto of explaining science by science, it should be readily intelligible that Quine’s is just one way of trying to realize the program for which that motto stands: to make scientific reason self-sufficient, to lead it out of its supposed dependency on philosophy. If that is what is really meant by naturalism, then Quine’s radical rejection of the distinction between the logical and the empirical (and all that it entails) is optional but not obligatory. Second, we may consider Neurath’s own most distinctive contribution to philosophy of science. Typically, it did not consist in logical or meta-mathematical investigations. Rather, it consisted in investigating aspects of science that can be called behavioral in the broadest sense. Neurath’s contribution to the philosophy of science consisted in deriving lessons from the empirical sciences of science.

While it remains justified to call naturalistic the entire philosophical approach to science of which Neurath developed one part, it may be desirable to avoid the misleading Quinean associations of the term.¹³ Bearing in mind that aiming for the self-sufficiency of scientific reason is the ultimate point of naturalism, one can use another term to the same effect. Recall that in the perspective to which Neurath contributed, philosophy of science belongs to science as its metatheory. The central notion is that of metatheory being part of science and employing empirical and formal subdisciplines. This not a merely terminological issue. Scientific metatheory belonging to science means that the metatheory does not use resources that are not used in first-order scientific inquiries (or that could be used there). What sets meta-theory apart from the first-order sciences, therefore, is that it uses these concepts reflexively, that it applies them to science and employs them, as it were, in the science of science. The philosophy of science that seeks to make science self-sufficient in this sense is best called philosophically deflationist.

Note, however, that such use is consistent with the definition naturalism given by Nagel and Brandt; see fn. 8 above.

Philosophical deflationism denies as mere pretense the claim that philosophy can give answers where science cannot, that it there can give answers that nevertheless have a cognitive standing commensurate with those given by science. There are different types of such purely philosophical answers that deflationism precludes. Most broadly, it precludes answers to the issue of realism (as well as idealism and solipsism), answers that turn on apodictic knowledge claims of matters of fact, and answers that involve unconditional value statements. The reason why philosophical deflationism is not philosophical nihilism is that a great many philosophical questions remain. Seeking to uncover the hidden nature of things is not needed for philosophy to pursue its meliorative office of old: to engage reflection that helps improve the way we interact with the world. How best to conceive of certain matters if certain aims are to be pursued does not require apodictic insight into how things are or should be.

Such an engineering conception of philosophy is closely associated with Carnap’s philosophy of logical tolerance, of course, but it equally fits with the attitude expressed in the simile of Neurath’s Boat. As will be argued, Neurath’s actual naturalistic contribution to the Vienna Circle’s philosophy of science requires complementation by what Carnap’s formal philosophy of science can offer (and vice versa). But this conception of a deflationist philosophy of science, encompassing the contributions of both Carnap and Neurath, was not always realized as such by them. To help us notice this potential combination of philosophical voices is yet another reason to revisit the debates reconstructed here. By contrast, Schlick’s attempts to stay true to a more traditional conception of empiricism by adopting salient elements of Wittgenstein’s fast-changing ideas about representations may serve more as a cautionary tale.

Clearly, the Vienna Circle’s protocol-sentence debate represents a particularly multi-dimensional object of inquiry. Not only are we dealing with a plurality of views, but each of their proponents also has its own characteristic history of changing past and future allegiances. For instance, telling the story of Neurath’s Boat would be a matter not just of getting to understand one voice in the Vienna Circle. To start with, one must also understand the early Neurath and the late Neurath: they also have their own Boat similes and each requires careful consideration.¹⁴ (Neurath was a very contrapuntal writer: each use of the simile comments on specific writers and positions in ways which are not immediately obvious.) Here, however, attention is nearly exclusively limited to the middle Neurath’s engagement in the Vienna Circle’s protocol-sentence debate.¹⁵ Yet understanding the potential of Neurath’s conception of scientific knowledge also requires, as just noted, understanding Neurath’s interlocutors in this debate. His main interlocutors there, Carnap and Schlick, of course also possessed a philosophical past that informed their Vienna Circle present, but in their cases too, only fairly brief indications of earlier positions can be given when so required.¹⁶

See Ch.4, fn.1.

On the story of Neurath’s Boat in the different contexts from 1913 to 1944, see my contribution in Cartwright, Cat, Fleck and Uebel (1996, 89–166). An additional significant but unpublished use of the simile is discussed below in 6.4.2.

For monograph-length studies which include the early Carnap, see Rungaldier (1984) and Richardson (1998), for shorter studies of the pre-Vienna Circle Schlick see Daum (1982), Quinton (1985), Shelton (1989), Gower (2000).

1.2 The Vienna Circle: Basic Doctrines and Aims

There are numerous introductions to the Vienna Circle to which readers may be referred who wish a more comprehensive perspective.¹⁷ Here it must suffice to note briefly its personnel and basic doctrines. The Vienna Circle met from 1923/4 to 1936/7 during the academic terms for often weekly discussions of questions in the philosophy of formal and natural science. Its nominal leader was Moritz Schlick and its principal members were Hans Hahn, Otto Neurath, Philip Frank (since 1912 in Prague) and Rudolf Carnap (since 1931 in Prague); other members included Herbert Feigl (since 1930 in the USA), Friedrich Waismann, Kurt Gödel, Viktor Kraft, Gustav Bergmann, Bela von Juhos and Heinrich Neider. Viennese associates of the Circle included Karl Menger and Edgar Zilsel.

For more or less traditional overviews of the Vienna Circle and its philosophy see, e.g., Kraft (1950), Jørgensen (1951), Ayer (1959b), Passmore (1967), Hanfling (1981a). For succinct introductions from more contemporary perspectives, see Rutte (1977), Dahms (1985b), Hegselmann (1987), Stadler (1998), Richardson (2003), Uebel (2006). For a reliable monograph-length introduction, see Haller (1993), for comprehensive documentation, Stadler (1997 [2001]).

The central concern of the Vienna Circle was the understanding of scientific knowledge.¹⁸ Somewhat unusually so for philosophers of their time, all of the Vienna Circle’s members and most of its associates were trained in the formal and/or empirical sciences.¹⁹ (Appropriately enough, the Circle’s heroes were not only philosophers like Frege, Russell, Wittgenstein and Mach, Poincaré, Duhem, but also scientists like Helmholtz, Hertz, Boltzmann and, supreme among contemporaries, Einstein.) Apart from the pre-history of its older members (Schlick, Hahn, Neurath, Frank and, much shorter, Carnap), the Circle’s existence fell into two periods. During the formative period from 1923 to 1928, its activity was mostly confined to weekly discussion meetings. In 1929, the Circle entered its public period with the formation of the Verein Ernst Mach, the publication of its manifesto Wissenschaftliche Weltauffassung: Der Wiener Kreis by some of its leading members, the first public conference for Epistemology of the Exact Sciences in Prague, and the first publications in the monograph series Schriften zur wissenschaftlichen Weltauffassung, edited by Frank and Schlick.²⁰

Neurath rejected the term philosophy as a label for the object of his efforts and preferred the more neutral term scientific world conception. See Rutte (1982a) for a discussion of the justification of calling Neurath a philosopher despite his own disclaimers.

Though a philosopher by profession, Schlick had gained his doctorate in physics under Max Planck and Carnap’s doctorate on different conceptions of space was also originally intended to be in physics. Hahn, Menger and Gödel were mathematicians, Frank was a physicist, and Neurath a political economist. Zilsel made a name for himself as a historian of science.

The distinction between the non-public and the public phases of the Vienna Circle was first introduced by Mulder (1968).

The Circle soon attracted international attention and collaboration, first with Hans Reichenbach and the Berlin Society of Empirical Philosophy, in the acquisition of the editorship of a philosophical journal renamed Erkenntnis from 1930 onwards, and soon with further colloquia and international congresses in Königsberg (1930), Prague (1934), Paris (1935, 1937), Copenhagen (1936), Cambridge, England (1938) and Cambridge, Mass. (1939). International visitors and associates included, from Berlin: Carl Gustav Hempel; from Poland: Alfred Tarski; from England: A.J. Ayer; from the USA: Albert Blumberg, Ernest Nagel and W.V.O. Quine; from Scandinavia: Eino Kaila, Ake Petzäll, Arne Naess and Jørgen Jørgensen. Partly due to these wide contacts, the death and dispersion of key members from 1934 onwards (Hahn d.1934, Neurath exiled to Holland 1934, Carnap to USA 1936, Schlick d.1936) did not mean the extinction of the Vienna Circle’s philosophy.²¹ Through the subsequent work of earlier visitors (Ayer, Naess, Nagel, Quine) and particularly through the work of their members and collaborators who had emigrated to the USA (Carnap, Feigl, Frank, Hempel, Reichenbach), their philosophical explorations proved extremely influential on the development of Anglo-American analytic philosophy.²²

On the dispersion of the Vienna Circle see the short remarks in Neurath (1946a), the recollection in Feigl (1969a) and the detailed studies by Dahms (1985c), (1987), (1995) and Stadler (2003a), (2003b).

The received view of scientific theories and orthodox logical empiricist philosophy of science were not overthrown until quasi-internal and external critiques had sunk in. The former is represented by Quine’s long campaign against Carnap and papers like Putnam (1962) and (1969); the latter by works like Hanson (1958), Kuhn (1962) and Feyerabend (1970). For a historical overview of this decline see Suppe (1977). That some of these critics still harbored presuppositions afflicting their predecessors was noted in Galison (1988).

1.2.1 The Challenge to Empiricism

Here, very briefly and very roughly, is the historical situation in which the Vienna Circle and logical empiricism generally saw itself located.²³ Kant’s answer to the sceptical claim that objective knowledge of nature was impossible involved the synthetic a priori: regulative principles of reason that are imposed upon the deliverances of sensibility, just as our faculties of sensibility impose upon stimulation the intuitions of space and time. These synthetic a priori principles, according to Kant, made intersubjectively intelligible not only the phenomena of contingent experience, but also the necessities of nature. However, all such objective knowledge was attained only against the background of our confrontation with noumena, unknowable things-in-themselves. As with Fichte, Schelling and Hegel, philosophy became more extravagant in its positive claims, practitioners of science themselves became more modest in what they claimed scientific knowledge to be. The positivistic program of reducing the knowledge claim of science and then providing legitimization for what’s left found favor with more radically empirically minded thinkers like Mach. Starting with Helmholtz, however, there also arose Neo-Kantianism which sought to distinguish the scientific respectability of the Kantian tradition from the speculations of Naturphilosophie. Still, description, not the explanation, of natural phenomena became the professed task of science for positivists who no longer looked to philosophy for metaphysical foundations, but rather found them in the utility of their preferred empirical procedures.²⁴ Hypotheses and all other generalizations and abstractions were viewed as legitimated only by the results of whatever were the preferred empirical methods.²⁵

On the parting of science and philosophy in early 19th-century Germany, see Schnädelbach (1983, Ch. 3), but note also Köhnke (1986) for the history of Neo-Kantianism in Germany. See Neurath (1936a), (1936b) for a characterization of the somewhat different development of Austrian philosophy and Haller (1977), (1979), (1983), (1986), Rutte (1977), for the elaboration and discussion of Neurath’s thesis of a distinct Austrian philosophical tradition. See Smith (1987), (2000) for sideways views on Neurath’s thesis and his (1994) for an account of Austrian philosophy centered on the Brentano school.

Such an empiricist view left non-empirical, philosophical justifications of science out in the cold; at best, as with Helmholtz—whose epistemological writings (1921) were edited by Schlick with Paul Hertz—even Kant’s speculations were viewed as standing in need of empirical support.

In this all agreed, whether they followed Mill’s inductivism (as did Helmholtz) or Whewell’s deductivism (as did Heinrich Hertz).

Along with the positivist Mach, the Vienna Circle thought the Kantian answer to the sceptical challenge uneconomical. Moreover, the Vienna Circle also saw it contradicted by formal science, logic and mathematics. Riemann’s abstract geometries showed that questions about the geometry of physical space, whether it was Euclidean or variously non-Euclidean in nature, were not answerable a priori. It fell to scientists like Einstein to see and the Vienna Circle itself to argue that physical science, namely, relativity theory, showed the untenability of Kant’s conception of space and time as forever fixed synthetic a priori forms of intuition.²⁶ The obvious answer to these problems of Kant’s grounding of science might seem to be a radical empiricism like Mill’s, but Frege’s anti-psychologistic critique had shown empiricism unable to account for the knowledge of arithmetic. Moreover, as Frank once recalled the problem situation of the Vienna Circle, familiarity with the then new theories in physics suggested that the return to pure empiricism was blocked anyway.²⁷

This formulation reflects the Circle’s view of the matter but it clearly negelects to take account of the neo-Kantian answer offered by Ernst Cassirer. On the latter, see Ryckman (1991a), Richardson (1998, Ch. 5). For Schlick’s important role in the refutation of the neo-Kantian defense see section 3.1 below.

Frank talked specifically about the first Vienna Circle but his narrative implied the thematic continuity between it and the Vienna Circle proper.

Our whole group understood and fully agreed that the human mind is partly responsible for the content of scientific propositions and theories. . . . We admitted that the gap between the descriptions of facts and the general principles of science was not fully bridged by Mach, but we could not agree with Kant, who built this bridge by forms or patterns of experience that could not change with the advance of science. (1949a, 7–8)

At this point Henri Poincaré, Pierre Duhem and Abel Rey offered the required assistance with their concepts of conventions in science. For Neo-Kantians like Cassirer, by contrast, the answer lay in stripping the synthetic a priori of its apodictic validity and developing transcendental idealism along the route of the relative a priori. The young Reichenbach too employed this tactic, but Schlick convinced him to change his views.²⁸ Henceforth, it was as either freely chosen determinations of individual scientific principles, as conventions, or as adoptions of entire sets of theoretical propositions, of networks of empirically underdetermined hypotheses, that logical empiricism tried to comprehend the contributions made by the human mind to scientific theories like Einstein’s.²⁹ The rejection of the synthetic a priori became logical empiricism’s cardinal doctrine.

See Coffa (1991, Ch. 10) and Friedman (1994) for discussions of this episode. I return to this in section 3.1.3 below.

Whether the former strategy succeeds in the case of the theory of general relativity—the logical empiricists’ own favorite example—has been questioned by Ryckman (1992) who subsequently set about developing an alternative transcendental idealist perspective on it (2005).

1.2.2 Logical Empiricism

The Vienna Circle’s philosophy of science was informed by the fundamental assumption—shared, before them, by philosophers as different as Occam, Leibniz, Kant, Peirce and Mach—that only those propositions are cognitively meaningful whose truth or falsity makes a difference that is discernible, at least in principle and however fallibly, by scientific means. (Cognitive meaning, unlike non-cognitive meanings, always concerns a factuality of sorts.) What distinguished the Vienna Circle and logical empiricism generally was the sharp division drawn between the empirical (physics, biology, the social sciences, etc.) and the formal sciences (logic, mathematics). This division reflects the logical empiricist strategy of attempting to renew empiricism by freeing it from the impossible task of grounding logical and mathematical knowledge. (Their factuality was evidenced not in empirical but formal reasoning.) This strategy was codified in the basic principle all logical empiricists accepted, whatever their further positions. This was that either propositions were of a synthetic nature and their assertion justifiable only a posteriori, or they were analytic in nature and justifiable by a priori reasoning—tertium non datur.

The Circle’s basic principle is neither without appeal nor without problems. The knowledge claims of logic and mathematics were thought to gain their justification on purely formal grounds, by proof of their derivability by stated rules from stated axioms and premises. Depending on the standing of these axioms and premises, justification was conditional or unconditional; axioms and principles of derivation in turn were considered linguistic rules and determined by convention. According to some such form of logicism, logic and mathematics were thought easily integrated into the empiricist framework. Gödel’s incompleteness results complicated matters, as we will see, but Carnap proposed to accommodate these by separating analyticity from effective provability and by postulating arithmetic to consist of an infinite series of ever richer arithmetical languages.³⁰

See Carnap (1934 [37, §60a–d]) and the discussion in Chapter 5 below.

The synthetic statements of the empirical sciences, meanwhile, were held to be cognitively meaningful if and only if they were empirically testable in some sense (and their justification as knowledge claims derived from such successful tests). Roughly, if synthetic statements failed to be testable in principle they were considered to be cognitively meaningless, giving rise in philosophy only to pseudoproblems. (Their non-cognitive meaning provided ample material for analysis in biology, psychology, sociology and history.) Here the Vienna Circle appealed to a meaning criterion the correct formulation of which proved controversial and elusive first of all among its own ranks. Thus it must be noted that still before Wittgenstein’s famous dictum The sense of a proposition is its verification with its demand for conclusive verification (in principle) was adopted for a while as the criterion of significance in 1930, others less strict had been around that either, as with Carnap in 1928, demanded only non-conclusive testability (in principle) or that expressed a vaguely pragmatic conception of significance, as in Mach’s dictum Where neither confirmation nor refutation is possible, science is not concerned.³¹ By 1932 sharp differences had opened up between those who insisted on strict verificationism and those who favored relaxing the criterion again (soon along with relaxing the demand that all meaningful terms be reducible to observational ones). Differences also arose over whether an entire theory of meaning was to be built out of verificationism or whether it provided merely a criterion of empirical significance.

See Waismann (1967 [1979, 47–8]), Carnap (1928b [1967, 325–8]) and Mach (1883 [1960, 587]).

While some versions of Vienna Circle verificationism are plainly demolished, it is not clear then whether all attempts to provide criteria for empirical significance are derailed by the standard criticism showing that once indirect verification must be allowed for due to the irreducibility of the theoretical terms of science, it is no longer possible to rule out of court metaphysical terms by a formal criterion.³² If the proposal is limited to formal languages, a much later proposal of Carnap’s can be successfully defended.³³ The more pragmatic form of Vienna Circle verificationism represented by Neurath and Frank all along side-stepped the need for a precise formal criterion of significance by its exemplar-oriented understanding of the criterion of making a discernible difference. But be that as it may, different conceptions of the criterion of empirical significance were on the table at the time of the protocol-sentence debate.

For what has become the standard criticism, see Hempel (1950) and (1951), canonically combined in Hempel (1965, Ch. 4).

This has been argued by Creath (1976).

Part of the reason that verificationism became so important was that the Vienna Circle took what is known as the linguistic turn. First, contents of mind were considered amenable to philosophical analysis only once they were linguistically articulated. (This may be called the anti-intuitional aspect of the linguistic turn.) Second, it consisted, in the words of Bergmann who coined the phrase, in the turn from simply talking about the world to talk about the world by means of talking about a suitable language for talking about the world (1964, 177). (This may be called the metalinguistic aspect of the linguistic turn.) The linguistic turn, as understood here, thus consisted in the presuppositions that all knowledge required linguistic representation and that all philosophy concerned ways of representing, rather than the nature of the represented.

1.2.3 The Unity of Science

The claim that science accepts some kind of verifiability criterion amounts to the adoption of the unity of science thesis in one form or another (the thesis comes in different versions). According to the methodological version of that thesis, all of science abides by the same criteria: no basic methodological differences separate the natural from the social or cultural sciences (Geisteswissenschaften) as claimed by those who distinguish between explanation and understanding. Defenders of the separate Geisteswissenschaften had it that, whereas explanation in the natural sciences consists in the subsumption of a given phenomenon under a general law, understanding in the Geisteswissenschaften is accomplished through a process often called empathy. Given the impossibility of verifying empathetic knowledge as the re-experiencing of another’s mental states, statements claiming to represent such knowledge violated the Circle’s criteria of meaningfulness. (Any claim based on empathetic knowledge that, under a different construal of what’s involved in empathy, could also be supported empirically would of course constitute regular scientific knowledge without needing empathetic evidence in that specific anti-empiricist sense.) All members of the Circle endorsed this methodological version of the unity of science thesis.

Another, yet stronger, version was a linguistic thesis of the unity of science. This thesis held that all objects of scientific knowledge could in principle be comprehended by the same basic universal language. This form of the unity of science thesis is often associated with reductionism. However, the positivistic demand for the reducibility of all meaningful scientific terms to observational ones was already under pressure by 1932 and officially abandoned by 1935, giving way to a more liberal conception of physicalism.³⁴ Physicalism asserts that the universal language is the language which speaks of physical objects. While everybody in the Circle endorsed physicalism in this broad sense all along, the understanding of its precise meaning and importance varied. There was disagreement on whether the thesis of physicalism did (Carnap) or did not (Neurath) involve reference to the language of mathematical physics, and on whether the presumed truth of this thesis was merely contingent and philosophically irrelevant (Schlick) or, while not logically necessary, nevertheless philosophically decisive (Neurath).

See Hahn (1933 [1987, 41])—lectures given in 1932—and Carnap (1936b).

Finally, note that the linguistic version of the unity of science thesis must be clearly distinguished from the nomological version: whether all scientific laws could be reduced to those of physics was yet another matter (and doubted early on by Neurath).³⁵ Once the irreducibility of theoretical to observational terms had sunk in, the nomological version of the unity of science thesis was abandoned by all Vienna Circle theorists (notwithstanding its being championed still in 1958 by a young Hilary Putnam).³⁶

See Feigl (1939); on Neurath’s anti-reductionist attitude towards social science ever since the early 1930s, see Uebel (2007).

Note the difference between Carnap (1936b, 69) and (1938, 61) and compare Oppenheim and Putnam (1958).

1.2.4 The Point of Scientific Philosophy

The distinguishing trait of Viennese scientific philosophy is its desideratum that philosophical discourse has to meet the same standards of intelligibility as inquiries in the established empirical and formal sciences. Science, having long forsaken explanatory appeals to authorities whose workings were unfathomable, and having stressed instead the testability of its theories by any qualified individual, here served as the ideal exemplar of rationality. Its rationality consisted in the requirement that scientists follow two principles: first, that they make only testable claims and, second, that they adjust their beliefs to accord with intersubjectively available evidence. The claim that philosophers should follow science in this respect set the philosophy of the Vienna Circle and its sympathizers apart from most academic philosophies of their time. All participants of the Circle bemoaned the kind of philosophy they saw done and the inconclusive state of conflicting philosophical knowledge claims. The results of two thousand years of philosophical research compared badly with the progress of natural science in only a few hundred years. There did not even exist agreement as to what would constitute a solution to a philosophical problem. How different matters were in science: only those pronouncements were accepted into its body of theory which withstood experimental tests. Philosophy too, it was concluded, stood in need of the scientific method.

But philosophy not only failed to reach the standards of scientific inquiry, it also failed to understand science itself. If philosophy had to emulate the rationality of science to retain its claim to represent knowledge, as the Vienna Circle held, it was also required that the results of science be philosophically attended to. According to much academic philosophy, the knowledge claim of natural science was itself in jeopardy.³⁷ A new specifically antiscientific scepticism had arisen: the inability of traditional philosophy to account for the new physics of relativity theory was widely taken to challenge the authority of scientific rationality itself. Science, so it turned out, rested on the employment of concepts and principles which were not univocally supported by empirical evidence. Beyond the long-standing problem of the justification of basic principles of scientific explanation like induction, there now also existed the problem of how basic physical concepts like space and time were to be understood. Proponents of especially those intellectual and cultural positions which had been undermined by the advances of natural science in the 19th century, fastened on this crisis in scientific metatheory and sought to resurrect their dogmas on the alleged ruins of their adversary. In order to allow philosophy, by contrast, to firmly take its stand on the ground of empirical science, as their manifesto put it (Carnap, Hahn, Neurath 1929 [1973, 317]), the Vienna Circle had to establish the claim of science to provide knowledge.

On this point, see Frank (1941, 4), (1949a, 1–4) and Menger (1933b).

If philosophy is to be rational, it must resemble science in the right respects; hence it must come to understand what these respects are. Interpreters of the Vienna Circle, in turn, must understand aright the nature of the philosophical understanding the Circle sought. No member left the traditional picture of knowledge undisturbed: they rather saw the knowledge claim of the new physics as presenting a challenge to philosophy. In attempting to derive conclusions from philosophical experience, namely, from the history of philosophy, the philosophers of the Vienna Circle proceeded like scientists. (Neurath—responsible for the name Vienna Circle and partly responsible for the term logical empiricism—sometimes referred to scientific philosophy in this sense as empirical rationalism.)³⁸ Having discarded First Philosophy on account of its perceived futility, they inquired how it was possible for highly abstract and theoretical scientific claims to be meaningful so as to be truth-evaluatable not in order to call science into question, but to work towards an alternative to the traditional idea that knowledge was grounded on intuitive certainty. The differences which obtained between Schlick, Carnap and Neurath concerned their understanding of this non-traditional strategy, not whether this was the strategy to be preferred.

See Frank (1949a) for the recollection of Neurath’s intention to recall happy things