Hyperintensional Formal Epistemology

We are holding a two-day conference on hyperintensional approaches in formal epistemology. The event takes place on September 16-17, 2022 in Berlin, as a satellite event of the GAP.11 conference that runs from September 12-15, 2022.

Confirmed speakers:

  • Sena Bozdag (LMU Munich, MCMP)
  • Johannes Korbmacher (Utrecht University)
  • Karolina Krzyżanowska (University of Amsterdam, ILLC)
  • Hannes Leitgeb (LMU Munich, MCMP)
  • Aybüke Özgün (University of Amsterdam, ILLC)
  • Timothy Williamson (University of Oxford)
  • Jonathan Mai (Goethe University Frankfurt)
  • Johannes Stern (University of Bristol)


Hyperintensional theories of content have recently received a great deal of attention. Influential accounts include Fine’s truthmaker semantics, Leitgeb’s HYPE framework, and two-component approaches inspired by Yablo’s work, modelling meanings as pairs of truth-conditions and subject matters. This workshop examines applications of hyperintensional frameworks in formal epistemology. On the one hand, the case for hyperintensionality seems especially forceful here, since intensional approaches notoriously face problems of logical omniscience: taking belief, knowledge, etc. to be closed under logical consequence, they seem adequate at most to highly idealized thinkers. On the other hand, without idealizing assumptions about the agents under consideration, there may not be enough structure to knowledge and belief to enable fruitful systematic theorizing.

Everyone is welcome to attend, but it is necessary to register for the event by sending an email to hamburgrelevance@gmail.com, stating your name and whether you wish to attend online or in person.

Provisional Programme

Day 1 (16/9):    

09.00-10.10    Johannes Korbmacher (Utrecht) – Parts and Probabilities

10.25-11.35    Aybüke Özgün (ILLC Amsterdam) – Indicative Conditionals: Probabilities and Relevance

11.50-13.00    Johannes Stern (Bristol) – Representation Sensitive State Semantics for Belief

Lunch Break

15.00-16.10    Sena Bozdag (MCMP Munich) – A Semantics For Hyperintensional Belief Revision Based On Information Bases

16.30-17.40     Timothy Williamson (Oxford) – Hyperintensional Epistemology and Degrees of Freedom (via Zoom)

Dinner and Drinks

Day 2 (17/9):   

09.00-10.10    Karolina Krzyżanowska (ILLC Amsterdam) – Conditionals, Relevance, and the Contribution of “Then”

10.25-11.35    Jonathan Mai (Frankfurt) – Impossibilities in Action

11.50-13.00    Hannes Leitgeb (Munich) – Formal Epistemology in Vector Space Models (via Zoom)


Aybüke Özgün, Indicative Conditionals: Probabilities and Relevance

In this talk, I will present a new account of indicative conditionals, giving acceptability and logical closure conditions for them. We start from Adams’ Thesis: the claim that the acceptability of a simple indicative equals the corresponding conditional probability. The Thesis is widely endorsed, but arguably false and refuted by empirical research. To fix it, we submit, we need a relevance constraint: we accept a simple conditional to the extent that (i) the conditional probability of the consequent given the antecedent is high, provided that (ii) the antecedent is relevant for the consequent. How (i) should work is well-understood. It is (ii) that holds the key to improve our understanding of conditionals. Our account has (i) a probabilistic component, using Popper functions; (ii) a relevance component, given via an algebraic structure of topics or subject matters. We present a probabilistic logic for simple indicatives, and argue that its (in)validities are both theoretically desirable and in line with empirical results on how people reason with conditionals. (This is joint work with Francesco Berto.)

Sena Bozdag, A Semantics for Hyperintensional Belief Revision based on Information Bases

The proposal of this presentation is a hyperintensional semantics for belief revision and a corresponding system of dynamic doxastic logic. The modality-free fragment of the semantics is based on Leitgeb’s HYPE framework. In the context of this work, HYPE-states are interpreted as possibly inconsistent and incomplete information states, while a fusion function represents information growth. We adjust the HYPE models to include a uniform preorder, to be understood as an epistemic preference ordering. Next, we introduce a static belief modality that is inspired by the base-generated revision theories, and a modal-shifting operator for belief revision. The resulting dynamic operation preserves the structure of a given state space along with the valuations, and it alters the preference ordering and the designated information state of the agent. In general, the focus of representation is shifted from the belief sets to collections of information. In this way, similar to the notion of evidence in epistemic frameworks, structured sets of information states keep track of an agent’s beliefs. Supported by the non-classical features of the proposed models, we obtain a non-classical logic of belief and belief revision. In particular, belief revision occurs in a non-explosive environment which also allows for a non-monotonic and hyperintensional belief dynamics. While the belief sets that are formed within this framework are not necessarily deductively closed, consistency is always preserved at the level of beliefs. The proposed system also satisfies CUT and cautious monotony, as desired properties of non-monotonic logics.

Jonathan Mai, Impossibilities in Action

It is common in epistemic modal logic to model the epistemic or doxastic states of agents via box operators in the normal logics S5 or KD45. However, this approach treats agents as logically omniscient by requiring their knowledge or belief to be closed under classical logical consequence. A promising way of avoiding logical omniscience consists in extending epistemic models with impossible states, that is states, where complex formulas are not evaluated recursively, but rather where they are treated semantically as atoms. However, this approach faces the dual problem of logical ignorance by modeling agents as not even minimally logically competent. In my talk I will present two epistemic logics that combine impossible states with epistemic action modalities. I will show that these logics avoid both the problem of logical omniscience and the problem of logical ignorance. The first logic makes use of dynamic realization modalities akin to the dynamic announcement operators from public announcement logic. The second uses the language of regular propositional dynamic logic in order to model deductions performed by agents via program modalities.

Hannes Leitgeb, Formal Epistemology in Vector Space Models

This talk will introduce and study logical systems in which formulas represent “effects” (e.g. of argumentation), such that these “effects” correspond formally to vectors. In a slogan: content is a vector. The logics involve deductive systems, hyperintensional semantics with appropriate notions of logical consequence, and extensions to similarity, inductive logic, and belief revision. The resulting systems may be interpreted in Bayesian terms, and they can be applied to logically reconstruct and address some well-known topics and problems from philosophy, cognitive psychology, computational linguistics, and machine learning.

Karolina Krzyżanowska, Conditionals, relevance, and the contribution of “then”

The oddity of missing-link conditionals, that is, conditionals whose antecedent is not relevant for its consequent, such as “if sharks are carnivorous, then Berlin is a European city,” has been traditionally explained away as a pragmatic phenomenon. Moreover, the requirement that the antecedent is a reason for, or an explanation of, the consequent has been analysed as the semantic contribution of the particle “then” rather than of the conditional itself. In my talk, I will present empirical evidence undermining this view. (This is joint work with Peter Collins and Ulrike Hahn.)

More abstracts will follow in due course.