Paul Teller

Science is widely taken to aim, and often to succeed, in producing truths, a “mirror of nature”. Not so. Instead, science fashions models, understood broadly as representations that are never both completely precise and completely accurate. . This chapter discusses how the misconception arose and how it is now being corrected. The account begins with a tension between the founding metaphors of the Scientific Revolution, reading God’s book of nature and the clock metaphor. The former pre-frames laws and physics fundamentalism; the latter the discovery of mechanisms, how things work. Laws aimed at truth, but the world is too complicated for us to get generalizations exactly right. Likewise, the complexity of mechanisms always requires simplifying idealization. Further, different problems require different idealizations and so a pluralism of accounts. The pluralism is unproblematic as different problems about the same subject matter can be consistently dealt with using very different schemes of simplification.

If the semantic value of predicates are, as Williamson assumes, properties, then epistemicism is immediate. Epistemicism fails, so also this properties view of predicates. I use examination of Williamsons position as a foil, showing that his two positive arguments for bivalence fail, and that his efforts to rescue epistemicism from obvious problems fail to the point of incoherence. In Part II I argue that, despite the properties view’s problems, it has an important role to play in combinatorial semantics. We may separate the problem of how smallest parts of language get attached to the world from the problem of how those parts combine to form complex semantic values. For the latter problem we idealize and treat the smallest semantic values as properties (and referents). So doing functions to put to one side how the smallest parts get worldly attachment, a problem that would just get in the way of understanding the combinatorics. Attachment to the world has to be studied separately, and I review some of the options. As a bonus we see why, mostly, higher order vagueness is an artifact of taking properties as semantic values literally instead of as a simplifying idealization.

In the contemporary discussion of hidden variable interpretations of quantum mechanics, much attention has been paid to the “no hidden variable” proof contained in an important paper of Kochen and Specker. It is a little noticed fact that Bell published a proof of the same result the preceding year, in his well-known 1966 article, where it is modestly described as a corollary to Gleason's theorem. We want to bring out the great simplicity of Bell's formulation of this result and to show how it can be extended in certain respects

One can give a strong sense to the idea that a relation does not 'reduce' to non-relational properties by saying that a relation does not supervene upon the non-relational properties of its relata. That there are such inherent relations I call the doctrine of relational holism, a doctrine which seems to conflict with traditional ideas about physicalism. At least parts of classical physics seem to be free of relational holism, but quantum mechanics, on at least some interpretations, incorporates the doctrine in an all pervasive way

This essay endorses the conclusion of Sklar’s “Dappled Theories in a Uniform World” that he announces in his abstract, that notwithstanding recent attacks foundational theories are universal in their scope. But Sklar’s rejection of a “pluralist ontology” is questioned. It is concluded that so called “foundational” and “phenomenological” theories are on a much more equal footing as sources of knowledge than Sklar would allow, that “giving an ontology” generally involves dealing in idealizations, and that a transfigured “ficitonalism” provides an (in many respects) better model of scientific knowledge than the model of “foundational truths.”.

The practice of describing multiparticle quantum systems in terms of labeled particles indicates that we think of quantum entities as individuatable. The labels, together with particle indistinguishability, create the need for symmetrization or antisymmetrization (or, in principle, higher-order symmetries), which in turn results in “surplus formal structure” in the formalism, formal structure which corresponds to nothing in the real world. We argue that these facts show quanta to be unindividuatable entities, things in principle incapable of supporting labels, and so things which support no factual difference_if two of them are thought of as being switched. When thinking of the metaphysics of quanta, we should eschew the misleading labels of the tensor product Hilbert space formalism and prefer the ontologically more faithful description of the Fock space formalism. This conception eliminates puzzles about the quantum statistics of bosons

This paper challenges “traditional measurement-accuracy realism”, according to which there are in nature quantities of which concrete systems have definite values. An accurate measurement outcome is one that is close to the value for the quantity measured. For a measurement of the temperature of some water to be accurate in this sense requires that there be this temperature. But there isn’t. Not because there are no quantities “out there in nature” but because the term ‘the temperature of this water’ fails to refer owing to idealization and failure of specificity in picking out concrete cases. The problems can be seen as an artifact of vagueness, and so doing facilitates applying Eran Tal’s robustness account of measurement accuracy to suggest an attractive way of understanding vagueness in terms of the function of idealization, a way that sidesteps the problems of higher order vagueness and that shows how idealization provides a natural generalization of what it is to be vague.

Conventional scientific realism is just the doctrine that our theoretical terms refer. Conventional antirealism denies, for various reasons, theoretical reference and takes theory to give us only information about the word of the perceptual where reference, it would appear, is secure. But reference fails for the perceptual every bit as much for the perceptual as for the theoretical, and for the same reason: the world is too complicated for us to succeed in attaching specific referents to our terms. That would appear to leave us with a kind of latter day, representational idealism: All we have are representations. I argue that our representations tell us about an independent world without securing reference by showing that the world is very like the way it is represented in a range of different, often complementary modeling schemes. Though never exact, these representations are of something extra-representational because they present the world modally as going beyond what is represented explicitly.