Archive for February, 2017


“Daedal Data: The Problem of Empirical Adequacy” (2/24/17)

February 25, 2017

Nora Boyd

Abstract: Whatever else our theories about the natural world are, they ought to be consistent with the evidence produced by our interactions with it–our theories ought to be at least empirically adequate. This is the minimal commitment of empiricism. Yet the central notions of evidence and empirical adequacy have not been satisfactorily elucidated.  Prominent accounts of evidence treat it as detachable from the manner in which it was produced.  However, considered as detached results, the corpus of empirical evidence appears to be contradictory and discontinuous.  Empirically derived parameter values evolve, sometimes radically, over time and the very concepts used to interpret evidence change between epistemic contexts.  It would be a fool’s errand to try to make our theories adequate with respect to evidence in this sense.  In this talk, I lay the groundwork for a new empiricist philosophy of science by furnishing a non-detached characterization of evidence and an epistemology of empirical adequacy appropriate to it.  I illustrate these accounts using case studies from astrophysics and cosmology, including observations of the Hulse-Taylor pulsar, historical observations of supernovae, and the history of measurements of the Hubble parameter.


“Cuverian Functionalism and the impossibility of evolution” (2/17/17)

February 22, 2017

Aaron Novick

Abstract: Ernst Mayr famously argued that all major disputes in the history of evolutionary theory could be understood as disputes between (good) population thinkers and (bad) typological thinkers. The basic difference between these two camps concerned the metaphysics of variation. In elaborating his conception of typological thinking, Mayr made the striking claim that a typologist must either deny evolution altogether or advocate a saltational theory of evolution (i.e. evolution by large leaps).
Enter contemporary developmental genetics. During the Cambrian radiation, the basic body plans of all extant animal phyla evolved. In the subsequent 500+ million years, those body plans have changed rather little. Eric Davidson and Douglas Erwin have developed an explanation of this remarkable evolutionary stasis in terms of recursively wired gene regulatory network subcircuits (called “kernels”) that serve special developmental functions. Interestingly, their explanation crucially invokes the saltation-or-non-evolution dichotomy that Mayr treated as a consequence of typological thinking. Is typology making a revival?
My goal in this paper is to elucidate the nature of this new explanation of deep conservation. I defend three claims. (1) I argue, on conceptual grounds, that what generates the saltation-or-non-evolution dichotomy is not typological metaphysics but a rejection of the possibility of gradual evolution, and I show historical “typologists” (e.g. Georges Cuvier and Otto Schindewolf) rejected gradualism for reasons entirely independent of a belief in typological metaphysics. (2) I show that the appearance of this dichotomy in contemporary developmental genetics rests on a form of functionalism similar to that of Cuvier, albeit modified. I show, further, that these very Cuvierian aspects of contemporary developmental genetics lead to a revival of Schindewolf’s notion of a graded hierarchy of variability. (3) But Cuvier was a famous opponent of all evolutionary (transformist) theories, and Schindewolf a famous opponent of gradual, Darwinian evolution. Does the revival of their ideas signal radical changes to evolutionary theory as we know it? To mitigate this fear (or dampen this excitement, as the case may be), I show how developmental genetics and standard evolutionary theorizing can be smoothly integrated.

“Success Without Representation” (2/10/17)

February 12, 2017

Shahin Kaveh

Abstract: I will argue that it is possible to explain the success of mature physical theories in such a way that assumes a robust connection between the theory’s parameters and the inner workings of the system, but without assuming that these parameters refer to or resemble any entities, properties, or structures within the system. In short, the idea is that the theory’s parameters track a contextually-defined state of the system, characterized by certain constraints and resolution limits, and that tracking is what makes the theory successful. َSome of the conclusions I will be arguing for are: 1) tracking is a relation to the system that is weaker than “representation” but stronger than “capturing the observables”, 2) the idea of tracking satisfies the realist and the anti-realist’s intuitions simultaneously and thus stands the test of both realist and anti-realist objections, 3) even if tracking parameters in some sense represent the system, we still cannot infer the inner composition of the system from them, and 4) even if tracking parameters represent, what makes them successful is that they track the system, not that they represent it.