Bergen Philosophy of Science Workshop 2013

Bergen Philosophy of Science Workshop 2013

BPSW 2013, 20-21 June

Department of Philosophy, University of Bergen
12/13 Sydnesplassen, Room 210

The talks are 40 min long followed by Q&A.
There is no registration/attendance fee. Abstracts below.

Thursday 20 June

12:45 Coffee
Welcome, Philosophy Dept. Chair Prof. Reidar Lie

Chair Sorin Bangu

13:00 - 14:00

Margaret Morrison (Univ. of Toronto)
Inconsistent Models: Problems and Perspectives  

14:15-15:15

Wendy Parker (Durham Univ. UK)
Simulation, Measurement & the Construction of Global Climate Datasets

15:15 - 15:30 Coffee break

15:30 - 16:30

Michal Walicki (Univ. of Bergen, Institute of Informatics)
The holism of truth and paradox
(joint work with Sjur Dyrkolbotn)


Friday 21 June

9:45 Coffee

Chair Michal Walicki

10:00 - 11:00

Alexander Paseau (Oxford Univ.)
Knowledge of Mathematics Without Proof

11:15 - 12:15

Rani L. Anjum (Norwegian Univ. of Life Sciences UMB)
Causation, Powers and Probability
(joint work with Stephen Mumford)

12:15 - 13:00 Lunch break on site

13:00 - 14:00

Colin Howson (Univ. of Toronto)
The Importance of Being Bayesian

14:00 Farewell

----

Abstracts

Margaret Morrison (Univ. of Toronto)
Inconsistent Models: Problems and Perspectives

One of the main stumbling blocks to theory unification is the problem
of having many incompatible models for the same phenomena. Not only is
this a problem for unification but it raises the issue of how to
epistemically assess the information these models contain.
Perspectivism is often seen as a way around this problem but a closer
look reveals that it only offers a solution in cases where the
incompatibility isn't really a problem after all. I discuss some of
the issues surrounding the use of inconsistent models and show that
the problem can persist even in the presence of a unified theory.

Wendy Parker (Durham Univ. UK)
Simulation, Measurement & the Construction of Global Climate Datasets

It is well known that computer simulation models are used to make
projections of future climate change. It is less well known that some
of the most-used "observational" datasets in climate science are
composed entirely of simulation output. I explain how such datasets
are produced (via a practice known as data assimilation) and consider
whether they are really so different from conventional observational
datasets. I argue that the differences are not as great as one might
suspect: in principle, these datasets can be high-quality measurements
of atmospheric properties, despite their genesis in simulation. In
arguing for this conclusion, I will present three core features of
measurement and explore intuitions about the nature of measurement
more generally. I will also argue that data assimilation is a special
case -- most simulation studies do not deliver measurements of
real-world systems.

Michal Walicki (Univ. of Bergen, Institute of Informatics)
The holism of truth and paradox (joint work with Sjur Dyrkolbotn)

Our main claim is that discourses, understood as relative and bounded
totalities of statements, provide the grounding for truth and paradox.
Single statements may be the carriers of truth-values but their
truth-claims become justifed or invalid only relatively to the actual
discourse. Truth-claims become invalid in the situation when no
coherent assignment of truth-values to all involved statements - of
the actual discourse - is possible and this amounts to a semantic
paradox. Only the totality of the actual discourse can justify the
conclusion of paradoxicality. (Typical examples, like the liar,
involve merely discourses consisting of single statements.) The
absence of paradox amounts exactly to the applicability of the
truth-concept: the possibility to distribute some truth-values among
all statements of the discourse. According to this view, paradox is
not any specific semantic value of statements but a failure of the
totality of a discourse, the limit suspending its aletheic
possibilities. Accepting thus paradoxes, the view is not threatened by
any revenge. The presented holism has only limited scope and functions
well along with truth of some statements understood as the
correspondence to the non-discursive facts. The presentation is based
on a series of informal examples and a formalisation, only hinted at,
is left for the interested readers.

Alexander Paseau (Oxford Univ.)
Knowledge of Mathematics Without Proof

Mathematicians do not claim to know a proposition unless they think
they possess a proof (or proof sketch) of it. For all their confidence
in the truth of a proposition with considerable non-deductive evidence
behind it (e.g. the Riemann Hypothesis), they maintain that strictly
speaking the proposition will remain unknown until such time as
someone has proved it. This paper challenges this conception of
knowledge, which is quasi-universal within mathematics. We present
four arguments to the effect that non-deductive evidence can in fact
yield knowledge of a mathematical proposition, showing in passing that
some of what mathematicians take to be deductive knowledge is in fact
non-deductive.

Rani L. Anjum (Norwegian Univ. of Life Sciences UMB)
Causation, Powers and Probability (joint work with Stephen Mumford)

Correlation data are often used for finding causation. But how does
causation relate to such data? Hume thought there was nothing more to
causation than regularities. Others think that causation is something
that underlies these correlations, for instance that a cause produce
its effect by necessitating it. An alternative to both views is
probabilistic causation. Instead of looking for perfect regularities,
one might say that a cause raises the probability of its effect.
Causal dispositionalism is an alternative that allows for both
probabilistic and non-probabilistic causation, while also throwing
some new light on the relation between causation and probability.

Colin Howson (Univ. of Toronto)
The Importance of Being Bayesian

Desirable, not to say indispensable, characteristics of a reliable
test, of a hypothesis are that it should minimise the chances of
incorrectly rejecting a true hypothesis and incorrectly accepting a
false one. These criteria are also known as the Neyman-Pearson
criteria of minimising (as far as possible) the chances of making type
I and type II errors, and in medicine of minimising false-negative and
false-positive rates of a diagnostic test. They are also the criteria
appealed to in the so-called No-Miracles argument for scientific
realism. They are popular among objectivists because they seem to
constitute a sound inductive strategy which makes no appeal to prior
probabilities. Unfortunately, they sanction demonstrably unsound
inferences. To ensure soundness priors are indispensable.