JRC for Logic
Tsinghua University – University of Amsterdam Joint Research Centre for Logic
JRC for Logic
Tsinghua University – University of Amsterdam Joint Research Centre for Logic

Tsing Ch’a Sessions 清茶会

To promote interdisciplinary interaction between different faculty members and students on the campus, a weekly meeting has been organized by our postdoc Jialiang Yan since September 2023, called Tsing Ch’a Sessions (清茶会). Its slogan “know thyself and let others know you better.”

■Current Sessions

This talk advocates for a cross-domain approach to temporal semantics. I will focus on the semantics of the verbal-suffix le, a central puzzle in the study of the temporal system of Mandarin. It is argued that instead of a dedicated temporal morpheme, le should be construed as a cross-domain functional particle underspecified for the associated semantic domain. The key motivation for this characterization comes from the parallel uses of le in run-of-the-mill VPs (1) and comparatives (2) that are subject to similar licensing requirements. Le is responsible for the past interpretation in (1), which is absent in (2).
图片
Such cross-domain behavior arguably calls for a simple semantic characterization. Indeed, I will show that the sentential meanings of (1) and (2) can be derived in parallel with le imposing a simple ordering relation, respectively on the timeline and the degree scale (of height). The licensing condition of le – here reflected as the requirement for measure phrases – can also be captured uniformly as a semantic constraint against a prejacent that satisfies the property I call ‘Atomic Distributivity’. The latter is essentially a domain-neutral, strict version of the ‘subinterval property’ (Bennett & Partee 1972). I will further demonstrate that this characterization delivers a satisfactory account for other temporal behaviors of le, and can be extended to its association with the spatial domain, when interacting with so-called as ‘extent predicates’ (Gawron 2005, 2009).

About the speaker:清华大学外文系水木学者博士后（即将入职）。主要研究领域为语义学和句法-语义接口。2025年于纽约大学语言学系获博士学位，2025至2026年在巴黎让尼可研究所（Institut Jean-Nicod）从事博士后研究。

In multi-player games, where agents form coalitions to reach collective goals, a natural notion is coalitional power: what a coalition can achieve regardless of any other agents’ actions. Coalitional power can be captured both abstractly by effectivity functions and concretely by strategic game forms, thanks to the representation theorem of truly playable effectivity functions by strategic game forms (see [1]). This representation theorem has empowered a host of logics for reasoning about coalitional power, such as those based on Coalition Logic (CL). However, the theorem only applies to effectivity functions satisfying superadditivity, which precludes dependence between different coalitions’ powers or actions. Given the ubiquity of such dependence, it is natural to study effectivity functions without superadditivity.
By weakening the superadditivity condition of true playability, we define the class of dependence-friendly playable effectivity functions. We will show that these effectivity functions can be represented by strategic game forms allowing dependence, a generalization of strategic game forms. To capture reasoning about these classes of effectivity functions and strategic game forms, we define Coalition Logic Allowing Dependence (CLD), for which we provide a neighbourhood semantics based on dependence-friendly playable effectivity functions and a Kripke semantics based on strategic game forms allowing dependence. We will demonstrate the finite model property of CLD with respect to its neighbourhood models, and in virtue of our representation theorem, prove that CLD is sound and weakly complete with respect to its neighbourhood and Kripke frames.
This is joint work with Chenwei Shi.
References
[1] Valentin Goranko, Wojciech Jamroga, and Paolo Turrini. Strategic games and truly playable effectivity functions. Autonomous Agents and Multi-Agent Systems, 26:288–314, 2013.

About the speaker: 王威，清华大学2023级逻辑学博士研究生，研究兴趣为策略推理、时态逻辑、认知逻辑、大模型逻辑推理及形式语义学。个人网页：https://www.weiwang.cool

Recent advances in the understanding of the micro-physiological mechanisms underlying carcinogenesis have increased the importance of personalized reasoning and decision-making in clinical oncology. In this rapidly evolving context, the purpose of personalized cancer medicine is to counterbalance standardization in patient care by tailoring and adapting interventions to individual clinical histories and therapeutic preferences of each patient. It is perhaps surprising, then, that models for personalized decision-analysis under limited evidence are still mostly confined to econometrics, while similarly methodologically justified approaches to integrate and evaluate evidence in clinical decision-making remain comparatively underused.
Starting from a basic template for decision-making in cancer treatment, this presentation frames and discusses some of the main methodological challenges to manage evidence and uncertainty in personalized oncology. Particular attention will be devoted to clinical decisions in which guideline-based procedures must be combined with patient-specific features emerging from multiomic clinical records. This approach thus departs from evidencebased medicine, which tends to evaluate the effectiveness of medical interventions through fixed epistemic hierarchies privileging the use of randomized controlled trials and metaanalyses.
The proposed framework combines a substantive account of how clinical evidence is generated with the assessment of effectiveness hypotheses in rare cancers.

About the speaker: Giovanni Sanavio a third year Ph.D. candidate conducting his research within the “Reasoning with Data” (ReDa) group based at the Department of Philosophy of the University of Milan, and part of the joint doctoral program “HuME” shared with IUSS Pavia and the SNS of Pisa. His research has thus far focused on clinical reasoning, evidence evaluation and belief revision for bed-side decision making. More generally, his interests include philosophy of science and medicine, formal epistemology and inductive logic.

In his 1996 doctoral thesis, Maurice Pagnucco introduced the first AGM-like abductive expansion operation. Building on his operation, together with a taxonomy—inspired by Atocha Aliseda—that serves to highlight and formalize the main components of abductive reasoning, the main aim of this presentation is to introduce a new paraconsistent AGM-like abductive expansion operation—capable of assimilating contradictory explanatory hypotheses without leading to trivialization and the resulting absurd epistemic state—, with its postulates and its transitively relational partial meet construction. To a large extent, the formal development we will present was only made possible by the recent creation of the paraconsistent logic RCbr, by Marcelo E. Coniglio, Martin Figallo and Rafael R. Testa, an LFI (Logics of Formal Inconsistencies) that establishes properties especially relevant to belief revision contexts, in particular, the ability to be self-extensional—i.e., to satisfy the replacement property.

The transitively relational partial meet abductive expansion operation introduced in this talk is part of a system called AGM_{p_{abd}}. Despite not assigning, in the postulates and construction, any relevant epistemic role to the paraconsistent operators of negation \neg and consistency \circ, this operation is, to the best of my knowledge, the first of its kind in the AGM literature.

 ■Past Sessions

Click HERE to check the past sessions from 2020-2024.

Click HERE to check the past sessions from 2025.