Thus far we have still not located precisely what it is about ethics that appears to make it scientifically problematic. In this chapter, I shall argue, using the analyses of the last chapter, that it is a certain thesis held by moral objectivists about moral norms and the reasons they generate that fails to pass scientific muster. According to this thesis, there are norms that generate reasons whose authority over us is objective, and moral objectivists accept this thesis with respect to moral norms and reasons. However, moral objectivism is merely a version of normative objectivism, which is a general position that assumes this thesis with respect to one or more types of norms (which can be epistemic, political, and so on). One can be a normative objectivist but not a moral objectivist, and one can be a normative objectivist about morality but refuse to accept the thesis for any other type of norm. I shall argue that the consistent naturalist must oppose all forms of normative objectivism, where moral objectivism is just a particularly popular version of this more general view.

The first task is to develop a minimalist and metaphysically neutral theory of the normativity of reasons, since all parties accept that there are such things as reasons that oblige us.

Given that only the Gibbardian account is congenial to naturalism, shouldn't that be the account to prefer, even if, as I've discussed, it is an account that fails to accord with our intuitions and implies an error theory of our judgments of practical reason? In this section, I want to argue that this is not so, on the grounds that the naturalists' conception of science, as well as their argument against objective moral theory, actually assumes a Kantian conception of the authority of the imperatives of reason constitutive of its methods. If this conclusion is right, it will establish that any science-based argument against the idea of objective normative authority is self-refuting. For if we conclude on scientific grounds that such authority doesn't exist, we do so on the basis of the rational methods of science that turn out to assume this same authority. This refutation only works if it illicitly assumes what it claims to refute.

To begin, consider why naturalists are so convinced that we ought to believe what scientists tell us, rather than what, say, astrologers or magic-users or mystics tell us. What makes only the scientists authorities about the world, and these other people (at best) merely colorful and amusing cultural phenomena?

Is the only possible answer to this question one that makes reference merely to the way in which our culture has made us (and taught us) that scientists are “authorities”? Such an answer is Gibbardian in the sense that it explains our sense that “we ought to believe what scientists say” as deriving from a norm whose content is a cultural creation, and whose authority over us is (merely) a psycho-social phenomenon.

Jean Hampton died on April 2, 1996, three days after going into a coma induced by a massive brain hemorrhage. The main intellectual project she was actively engaged on just before her death was the completion and revision of a draft of a book with the working title A Theory of Reasons. The draft she was working from contained nine chapters and a preface, introduction, and preliminary material. She had significantly revised the first three chapters in the weeks prior to her unexpected and premature death, which prevented her from completing her work on the rest of the manuscript.

Jean thought of this as her most important work, and had asked me to ensure its publication in the event she was unable to do so herself. With the encouragement of Terry Moore of Cambridge University Press, I have undertaken to honor her request. The book is closely based on computer files of the most recent versions of the preface, introduction, and preliminary material, and nine chapters of A Theory of Reasons as they were left at Jean's death.

The material has been lightly edited to remove obvious typographical and other minor errors. More substantial editorial interventions are noted explicitly, mainly in footnotes and appendices – the words of the editor in the text are enclosed in brackets.

there is no question that Rudolf Carnap's first major book, Der logische Aufbau der Welt (The logical structure of the world; hereafter Aufbau), published in 1928, is a central document of analytic philosophy. Its place has been secured by Carnap's general importance in setting the agenda for analytic philosophy from the late 1920s until his death in 1970. The Aufbau itself has been seen as perhaps the crucial document in the formation of the project of logical positivism. It is also Carnap's most sustained attempt to provide a general epistemology of empirical knowledge. Because of its historical role in shaping analytic philosophy generally and logical positivism in particular, many of the standard-bearers of analytic philosophy have had occasion to engage in interpretations of this book– often as a way of motivating their own philosophical enquiries. Among them we find Hillary Putnam, David Lewis, and, most importantly, Nelson Goodman and W. V. Quine. There is no better place to turn for a preliminary account of the epistemology put forward in Carnap's book than to the text itself. To fix certain points of reference, therefore, let us, without further ado, rehearse some of the principal themes of Carnap's epistemological project.

FUNDAMENTALS OF THE EPISTEMOLOGY OF THE AUFBAU

Der logische Aufbau der Welt begins with these words (§1): “The aim of the present investigations is the establishment of an epistemo-logical [erkenntnismassig-logischen] system of objects or concepts, the ‘constitutional system’” Two important points are then made. First, “object” is taken “in its widest possible sense, namely, for anything about which a statement can be made” (§1).

the neo-Kantian themes uncovered in Chapter Five provide an entry point to Carnap's thinking about epistemological and methodological issues throughout the 1920s. Carnap's dissertation, Der Raum (Space; hereafter Raum), written and defended in 1921 and published in 1922, shows him to be an unabashed, if unorthodox, neo-Kantian about space. His views show many similarities to the account of physical methodology we saw in Cassirer, especially as regards the distinction between the universal and relativized synthetic a priori. There are, of course, differences, which stem largely from Carnap's commitment to Russellian logicism and from his more detailed understanding of the technical physical and mathematical issues. In this chapter and the next, our goal is to use the framework for thinking about neo-Kantianism to investigate Carnap's evolving thought in his pre-Aufbau period. We shall see that in Carnap's work, also, there is an attempt to tie the methodological synthetic a priorito a logic of objective knowledge. Conventionalism in mathematical physics, combined with a desire to have a scientific understanding of philosophy itself, leads ultimately to the problems Carnap finds in explaining the relation among objectivity, intersubjectivity, and logical form in the Aufbau.

This chapter is given over entirely to an outline of the theories about space and geometry that Carnap presents in his dissertation. Chapter Seven provides a brief examination of the main themes of the essays on the methodology of physics that he published during the years (1922–6) when he was writing the Aufbau. At no point in any of these essays does Carnap endorse anything that looks like strict empiricism.

in this chapter we shall examine more closely the epistemological problem of the constitutional system of the Der logische Aufbau der Welt, as well as some of the technical means that Carnap proposes to use to solve this problem. Our first order of business will be to examine the import of the epistemological vocabulary used to motivate the projects of the work. This vocabulary emphasizes the subjective-objective distinction. It is incumbent upon us, then, to try to acquire a better sense of Carnap's use of these terms. Most particularly, we must try to understand what is at stake in the claim that structure yields objectivity.

The main purpose of this chapter is to present Carnap's major motivating ideas as an introduction to his thought on objectivity. In Chapter Three, we shall examine his particular solutions to the problem of objectivity in the Aufbau. The point is to emphasize the crucial interpretative issues raised by the work on their own terms, while freeing ourselves from the need to cram their significance into ready-made philosophical pigeonholes. The puzzles that we will be left with at the end of these chapters and the question of how they arise within the philosophical context of Carnap's book will be our major concern in the remainder of the book.

KNOWLEDGE VERSUS EXPERIENCE: THE PROBLEM OF SUBJECTIVE ORIGINS

Let us recall the form of the epistemological problem that Carnap presents at the opening of his book: Knowledge, for any given agent, begins in the stream of experience of that individual.

In our own times the use of hypothetical examples in connection with the debate over personal identity and related issues has become controversial. In particular, some have argued that because these examples are of (possibly) impossible situations, the consideration of them cannot shed light on what matters in survival. Since in the present study I intend to make use of hypothetical examples of (possibly) impossible situations, I want now to defend the uses to which I shall put them. To do that, I first specify a fission example that is a modified version of one originally presented by Shoemaker (1984, p. 119). Then I use this example to explain why fission examples, in particular, and hypothetical examples of (possibly) impossible situations, in general, are not only a legitimate but perhaps an indispensable tool in revealing what matters to people in survival.

Imagine, then, that you have a health problem that will result soon in your sudden and painless death unless you receive one or the other of two available treatments. The first is to have your brain removed and placed into the empty cranium of a body that, except for being brainless, is qualitatively identical to your own. The second is to have your brain removed, divided into functionally identical halves (each capable of sustaining your full psychology), and then to have each of these halves put into the empty cranium of a body of its own, again one that is brainless but otherwise qualitatively identical to your own.

i ended Chapter Eight with remarks about the development of Carnap's thought, especially his rejection of epistemology. I shall end our consideration of Carnap's early philosophy of empirical knowledge by examining more closely some of the major themes of this development as it occurred in the 1930s. This will indicate, I hope, that the interpretative framework within which we have been considering Carnap's early work can illuminate further developments as well. At the close, I shall indicate that from Carnap's point of view there can be no empiricism without dogmas, if we insist on using this term to characterize a view that claims methodological ineliminability of the analytic-synthetic distinction. To disagree with Carnap on this matter is to disagree with him about what empiricism is or could be.

FROM EPISTEMOLOGY TO THE LOGIC OF SCIENCE

In the mid-1930s Carnap rejected epistemology in no uncertain terms. This is most explicit in an address presented at the 1935 “Congress for the Unity of Science,” in Paris, entitled “Von Erkenntnistheorie zur Wissenschaftslogik” (From epistemology to the logic of science; hereafter VEW). In this essay, Carnap invited his audience to view current developments as a move to a third stage of scientific philosophy. In the first stage, scientific philosophy had rejected metaphysics. This ushered in a “transition from speculative philosophy to epistemology” (VEW, p. 36). The second stage had involved the rejection of the synthetic a priori and the consequent adoption of empiricism in epistemology.

this study examines a particular portion of Rudolf Carnap's philosophical career, from a particular point of view. The period covered is roughly the fifteen years from Carnap's first publication– his 1922 dissertation, Der Raum – to the full flowering of his theory of the logical syntax of scientific language in his 1934 book, Die logische Syntax der Sprache and his seminal 1936–7 paper, “Testability and Meaning.” Although in the final chapter I speak to some central issues in the analyticity debate, I make no claim to deal with the details of Carnap's semantic period, for by the syntax period his general philosophical orientation had already been set, and those details are largely irrelevant for the story I want to tell. That is the story of Carnap's thinking about what it is to have an epistemology of empirical knowledge. The principal text for the story is Carnap's first book, Der logische Aufbau der Welt (The logical structure of the world; hereafter Aufbau). This was his most detailed and influential statement of a project in general epistemology. It contrasts both with the methodological focus of his earliest essays and with the rejection of epistemology that marks his syntax period.

The account given is somewhat novel. It rejects the easy assimilation of Carnap's epistemological views to those of Bertrand Russell's “external world program.” This rejection is guided by contextual and historiographic concerns. The Russellian perspective fails to engage with the text of the Aufbau in anything like its own terms. Rather, it imposes on it a philosophical perspective concerning epistemology that stands quite at odds with Carnap's own views of what epistemology is and what it is for.

From Plato until John Locke personal identity was explained in the West primarily by appeal to the notion of a spiritual substance or soul. From Locke until the late 1960s it was explained both in this Platonic way and by appeal to physical and/or psychological relations between a person at one time and one at another (and theorists assumed that how earlier and later persons are related to each other, through intervening persons, by itself determines whether the two are the same person). Since the 1960s there have been three major developments: First, so-called intrinsic relational views have been largely superseded by extrinsic relational (or closestcontinuer, or externalist) views, according to which what determines whether a person at one time and one at another are identical is not just how the two are related to each other but also how they are related to every other person. Second, the traditional metaphysical debate over personal identity has spawned a closely related but relatively novel debate over egoistic survival values. This debate has been over the question of whether - from what in actual, as opposed to hypothetical, circumstances would pass for a self-interested point of view - identity or other relations that do not suffice for identity do and/or should matter primarily in survival. And, third, some theorists have replaced the traditional threedimensional view of persons with a four-dimensional view, according to which the relata of the identity relation are not (whole) persons at short intervals of time but, rather, appropriately unified aggregates of personstages that collectively span a lifetime.

Ever since John Locke, Western personal identity theorists have been preoccupied with looking at life backwards, through the lens of memory. Locke, for instance, used his prince and cobbler example to make the point that because someone might remember the experiences had by someone who inhabited a different body, people could switch bodies - a consideration thought by many to refute bodily continuity theories. As recently as the mid-1960s, bodily continuity theorists have tried to turn the tables by arguing that only genuine memories can sustain personal identity and only memories brought about by their normal physical causes can be genuine (Martin and Deutscher, 1966).

Beginning in the late 1960s, however, and increasingly since then, forward-looking perspectives have come to the fore, ushered in, as we have seen, by the consideration of fission examples. What makes fission examples so theoretically interesting is that they have seemed - and still do seem - to many to support the idea that people could lose their identities yet obtain what either should or does matter primarily to them in their so-called self-interested concern to survive. So, unlike the traditional (pre-1970s) debate, which, insofar as it influenced contemporary theory, focused almost exclusively on specifying the conditions under which identity is preserved, with memory as the primary focus of concern, more recently there have been the additional problems of trying to discover what should or does matter primarily in survival (with some still convinced it is identity), and then specifying the conditions under which that (whatever it is) obtains.

the look back at neo-Kantian epistemology and Carnap's own work in the 1920s leading to Der logische Aufbau der Welt has served to highlight important themes that we shall examine in this chapter in relation to the 1928 book itself. First, the point of view of the epistemological project of the Aufbau is in some respects closer to, for example, Ernst Cassirer's general epistemology in Substance and Function than to Carnap's own conventionalist methodology of science. Three aspects of the Aufbau are particularly noteworthy in this regard: First, the stress on the epistemological centrality of objectivity is more pronounced than it was in Carnap's methodological work; second, the role of logic is much greater here than it had been previously in Carnap's work; third, simply as a general epistemology, the Aufbau has the broad scope of Cassirer's logic of objective knowledge, rather than the more narrow scope of Carnap's work on physics. These connections, of course, give rise ultimately to great divergences between the project of the Aufbau and Cassirer's logical idealism. The principal reason for this is that Carnap's formal logic allows him to present a technical project that Cassirer had never attempted. This technical project no longer allows the expression of Carnap's earlier rejection of strict empiricism. But this does not mean that he now endorses strict empiricism; it means, rather, that he can no longer distinguish the genuinely epistemological aspects of the project of empiricism from the epistemological aspects of neo-Kantianism.

Beyond this, however, we now have all the pieces we need to understand the circumstance that has driven our curiosity.

ON THE TASK OF PHYSICS

Carnap's first postdissertation publication, his 1923 Kant-Studien essay, “Über die Aufgabe der Physik und die Anwendung des Grundsatzes der Einfachstheit” (On the task of physics and the use of the axiom of simplicity; hereafter UAP), brings with it many divergences from the view put forward in the dissertation. Carnap makes no mention of intuition and presents a more general framework for physical conventionalism. His one reference to the Kantian synthetic a priori, a halfhearted nod in the direction of the relativized a priori, reveals his reluctance to engage Kantian terminology or to wear a neo-Kantian mantle. Nonetheless, the essay begins with a clear rejection of strict empiricism, worth quoting in full (UAP, p. 90):

Carnap backs this rejection of pure empiricism up with an argument against reductionism that makes clear that the conventional choices underlying mathematical physics are not to be viewed as logical definitions on the basis of experiential primitives in the manner of Russell.

we have had occasion to note Carnap's advocacy of Kantian themes in his early philosophical thought. But, of course, Carnap is not adopting Kantianism in all its details– too many aspects of Kant's own philosophical thought were rendered implausible, if not simply exposed as mistaken, by scientific and mathematical advances in the nineteenth and early twentieth centuries. Carnap's early views are much more substantially informed by the views of the scientific neo-Kantians of the early twentieth century. These neo-Kantians both took considerable insight from Kant's own work and significantly changed some central Kantian tenets. Thus, for an understanding of Carnap, it is the work of the neo-Kantians that is of primary importance. Nevertheless, the best place to start in examining the work of the neo-Kantians is with the metaphysical and epistemological positions of Kant himself and the problems that arose for these positions in the evolution of scientific, mathematical, and logical theorizing in the years between Kant and Carnap. With this in hand we will be better able to see the continuities and discontinuities between neo-Kantian and Kantian thought. Thus, in this chapter we shall reconstruct some themes in Kant's account of the synthetic a priori and its role in objective, theoretical knowledge. In the next, we shall see how the twentieth-century neo-Kantians take up these themes in the post-Einsteinian era.

THE KANTIAN PROBLEMATIC

In this section I will briefly review some of the major themes in Kant's account of scientific and metaphysical knowledge, as presented in the Critique of Pure Reason (Kant 1781/1787/1965) and related documents. I will then explain how post-Kantian scientific advances rendered some details of Kant's account implausible.