TiLPS

The Tilburg Center for Logic, Ethics, and Philosophy of Science studies knowledge, reasoning, and value in all their forms.

tilburg university

Research Seminar in Epistemology and Philosophy of Science

This research seminar is a forum for all members of TiLPS with research interests in epistemology or philosophy of science to present and discuss their work. Sometimes we also have a guest speaker or discuss a recently published article. Master students and Research Master students may take this research seminar for credit. Please contact Matteo Colombo for more information. Papers for discussion and some background reading will be available from this website at least one week in advance.


Upcoming seminars

Thursday, 25 January 2018, 12:45-14:00

Room: D 119

Speaker: Seamus Bradley (TiLPS)

Title: tba

Abstract: tba

Thursday, 22 February 2018, 12:45-14:00

Room: D 119

Speaker: Colin Elliot (TiLPS)

Title: tba

Abstract: tba

Thursday, 5 April 2018, 12:45-14:00

Rom: D 119

Speaker: Naftali Weinberger (TiLPS)

Title: tba

Abstract: tba


Recent seminars

Thursday, 14 September 2017, 12:45-14:00

Room: D 119

Speaker: Jan Sprenger (TiLPS)

Title: Trivalent Semantics for Indicative Conditionals: A Resurrection Attempt

The semantics of indicative conditionals are a notoriously difficult philosophical problem. Many philosophers take the view that they don't have classical truth conditions: the truth value of 'If A, then B', cannot be determined as a function of the truth values of A and B. Focus has therefore shifted to the acceptability and/or probability of indicative conditionals.

Unfortunately, this view severs the ties between the semantics and the epistemology of indicative conditionals. I propose a principled solution that goes back to De Finetti: to adopt a trivalent semantics, based on reading indicative conditionals as conditional predictions of the consequence. The advantages of this view are spelled out, and it is defended against the standard objections (e.g., invalidity of classical inference schemes).

Thursday, 28 September 2017, 12:45-14:00

Room: D 119

Speaker: Tom Sterkenburg (MCMP Munich)

Title: The Meta-Inductive Justification of Induction

Abstract: Gerhard Schurz has recently proposed a Reichenbachian justification of induction based on results from machine learning. These results concern the design of 'meta-inductive' prediction methods that never perform much worse than a given pool of competing methods. In my talk, I will explain Schurz's proposal and discuss challenges and extensions.

Thursday, 12 October 2017, 12:45-14:00

Room: D 119

Speaker: Matteo Colombo (TiLPS)

Title and abstract: t.b.a.

Thursday, 26 October 2017, 12:45-14:00

Room: D 119

Speaker: Silvia Ivani (TiLPS)

Title and abstract: t.b.a.

Thursday, 9 November 2017, 12:45-14:00

Room: D 313

Speaker: Nathan Wildman (TiLPS)

Title: Virtual Digitalism versus Virtual Fictionalism

Abstract: Recently, Chalmers (forthcoming) has advanced several arguments in favour of virtual digitalism, according to which the various virtual objects that make up virtual reality environments are "genuinely real entities". However, much about virtual digitalism – including whether it provides an adequate ontology for virtual objects – remains unclear. With that in mind, this talk has three aims. The first is to make some progress towards clarifiying the content of Chalmers' virtual digitalism. I do so by distinguishing and contrasting two variants of the view: strong virtual digitalism, which identifies virtual objects with digital data structures, and weak virtual digitalism, which takes virtual objects to be dependent upon said structures. The second aim is to critique both variants of Chalmers' view by demonstrating that the arguments Chalmers offers in favour of virtual digitalism mistakenly entail that VR users are, if he adopts strong v-digitialism, identical to, or, if weak v-digitalism, dependent upon, digital data structures. The upshot is that, regardless of which variant Chalmers adopts, virtual digitalism is problematic. The final aim is to sketch an alternative position according to which virtual objects are merely fictional. To do so, I develop a broadly Waltonian view, virtual walt-fictionali​sm. After exploring how the view works, I conclude that, prima facie, it is far preferable to virtual digitalism.

Thursday, 16 November 2017, 12:45-14:00

Room: D 119

Speaker: Filip Buekens (TiLPS)

Title: The value of truth is extrinsic to it

Abstract: t.b.a.

Thursday, 23 November 2017, 12:45-14:00

Room: D 119

Speaker: Noah van Dongen (TiLPS)

Title: Addicted to p-values: Part II

Abstract: Published scientific results turn out to have a low level of replicability and there is an ongoing debate on the merits of current statistical methods and potential replacements. I would like to contribute to this debate through a literature investigating of a ‘what if’ question, 'What if science’s use of the currently ubiquitous measure of evidence (i.e., the p-value) is like an addiction, instead of an (mindless) ritual?'

This presentation is a follow-up of my presentation of March 2nd 2017. I briefly portray the use of the p-value as a measure of research evidence from a scientist-centered perspective and provide an explanation of the misuse and fallacies concerning this measure by comparing it to addictive behavior. If my interpretation is correct, in part or in full, the current initiatives in reforming scientific practice might fail or be less successful than desired. I conclude with an outline for a potentially more reliable hypothesis testing method that does confirm to scientists' desires. 

Thursday, 7 December 2017, 12:45-14:00

Room: D 119

Speaker: Naftali Weinberger (TiLPS)

Title: Race is not a cause, but a normative baseline

Abstract: Would President Obama have been treated differently had he been white? This question seemingly concerns whether his race was a cause of how he was treated. But demographic variables such as race pose a problem for causal inference. Specifically, it is unclear how to understand causal counterfactuals regarding what it would mean for someone to have been a different race. In some scenarios, this challenge can be addressed through a suitable disambiguation. For instance, in many cases we care not about what if someone would have been a different race, but rather with what if she were perceived as being a different race. But if we adopt this strategy across the board, we abandon the possibility of modeling race as a general explanation for a range of ways in which some minorities are systematically disadvantaged. In my talk, I propose a way of incorporating the role of race into models for measuring discrimination without treating race itself as a cause. The concept of race is useful insofar as (A) the various manifestations of race that are loosely referred to as its ‘effects’ are the result of a systematic social discrimination and (B) in evaluating whether certain individuals were harmed by an action or policy, we consider ways in which they would have been better off in the absence of this discrimination. As clause (B) indicates, the usefulness of the concept is not merely an empirical matter, but depends on normative questions regarding whether harms resulting from racism should be treated differently from those resulting from other factors promoting inequality.