ILLC-LIRa February 2025

illc-lira@list.uva.nl

1 participants
8 discussions

LIRa session: Bartosz Więckowski (Thursday, 6 March 16:30)
by LIRa Seminar 28 Feb '25

28 Feb '25

Dear all, We will have our next LIRa session on Thursday, 6 March 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Bartosz Więckowski (Goethe Universität Frankfurt am Main) Date and Time: Thursday, March 6th 2025, 16:30-18:00 Venue: ILLC Seminar Room F1.15, Science Park 107 and online. Title: Proof-Theoretic Considerations on the Structure of Reasoning with Counterfactuals and Knowledge. Abstract. Combined reasoning with counterfactuals and knowledge/belief may involve constructions which: (i) embed knowledge (relativized to an agent a) in the antecedent (resp. consequent) of a counterfactual (e.g., 'If a knew that A, a would believe that B'), or (ii) prefix knowledge to a counterfactual (e.g., 'a knows that if A were the case, B would be the case'), or (iii) do both; where the embedding and prefixing can be iterated. Counterfactual inference is typically studied from a model-theoretic perspective. In this talk, we take a proof-theoretic perspective. We combine components from [1] and [2], so as to obtain intuitionistic subatomic natural deduction systems for combined reasoning with ('would'- and 'might'-) counterfactuals and knowledge (resp. belief) which are proof-theoretically well-behaved (normalization, subexpression/subformula property, internal completeness) and which admit the formulation of a semantically autarkic proof-theoretic semantics for elementary combined constructions of the aforementioned kinds. [1] Więckowski, B. (2021). Intuitionistic multi-agent subatomic natural deduction for belief and knowledge, Journal of Logic and Computation 31(3): 704-770. Special issue on External and Internal Calculi for Non-Classical Logics edited by A. Ciabattoni, D. Galmiche, N. Olivetti, and R. Ramanayake. https://doi.org/10.1093/logcom/exab013 [2] ---. (2024). Counterfactual assumptions and counterfactual implications. In Piecha, T. and Wehmeier, K. F., eds., Peter Schroeder-Heister on Proof-Theoretic Semantics, Outstanding Contributions to Logic, Vol. 29, pp. 399-423. Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-031-50981-0_15 Hope to see you there! The LIRa team

1 0

LIRa session: Michael Cohen (tomorrow, Thursday, 27 February 16:30)
by LIRa Seminar 26 Feb '25

26 Feb '25

Dear all, We will have our next LIRa session tomorrow, on Thursday, 27 February 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Michael Cohen (Tilburg University) Date and Time: Thursday, February 27th 2025, 16:30-18:00 Venue: ILLC Seminar Room F1.15, Science Park 107 and online. Title: Imperfect Recall and Monty Hall Abstract. Philosophical arguments for the rationality of Bayesian conditioning, such as diachronic Dutch book arguments, often refer to assumptions and concepts from epistemic logic without fully explicating them. In this talk, I use dynamic epistemic logic (DEL) to analyze the underlying logical assumptions of these philosophical arguments. I show that Perfect Recall is an important epistemic principle at the heart of the project of justifying Bayesian conditioning. Following the existing Bayesian literature, I use Monty Hall style cases to demonstrate the connection between Perfect Recall and the commitments of Bayesian conditioning. Along the way, we get an interesting DEL modelling of the Monty Hall puzzle. Hope to see you there! The LIRa team

1 0

LIRa session: Michael Cohen (Thursday, 27 February 16:30)
by LIRa Seminar 21 Feb '25

21 Feb '25

Dear all, We will have our next LIRa session on Thursday, 27 February 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Michael Cohen (Tilburg University) Date and Time: Thursday, February 27th 2025, 16:30-18:00 Venue: ILLC Seminar Room F1.15, Science Park 107 and online. Title: Imperfect Recall and Monty Hall Abstract. Philosophical arguments for the rationality of Bayesian conditioning, such as diachronic Dutch book arguments, often refer to assumptions and concepts from epistemic logic without fully explicating them. In this talk, I use dynamic epistemic logic (DEL) to analyze the underlying logical assumptions of these philosophical arguments. I show that Perfect Recall is an important epistemic principle at the heart of the project of justifying Bayesian conditioning. Following the existing Bayesian literature, I use Monty Hall style cases to demonstrate the connection between Perfect Recall and the commitments of Bayesian conditioning. Along the way, we get an interesting DEL modelling of the Monty Hall puzzle. Hope to see you there! The LIRa team

1 0

LIRa session: Gerhard Schurz (tomorrow, Thursday, 20 February 16:30, online-only)
by LIRa Seminar 19 Feb '25

19 Feb '25

Dear all, We will have our next LIRa session tomorrow, on Thursday, 20 February 16:30. To attend, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Gerhard Schurz (Heinrich-Heine-Universität Düsseldorf) Date and time: Thursday, February 20th 2025, 16:30-18:00 Venue: online. Title: Meta-Induction and Its Applications in Social Epistemology Abstract. Meta-induction has been developed as a new approach to Hume's problem of justifying induction. The approach concedes the force of Hume’s sceptical arguments against the possibility of a non-circular justification of the reliability of induction. What it demonstrates is that one can nevertheless give a non-circular justification of the optimality of induction based on meta-induction. The method of meta-induction tracks the success records of all accessible methods of prediction or action (including advice from experts, peers, or algorithms) and combines them in a success-dependent way, in order to obtain a superior method that is provably optimal among all prediction methods accessible to the epistemic agent. Since meta-induction can be viewed as a form of social learning, it has important applications in social epistemology. In the second part of the talk a form of social learning called local meta-induction is presented. In local meta-induction it is assumed that individuals can access only the success records of the individuals in their immediate epistemic neighborhood. It is shown that local meta-inductive learning can spread reliable information over the entire population, and has clear advantages compared to success-independent social learning methods such as peer-imitation and authority-imitation. Hope to see you there! The LIRa team

1 0

LIRa session: Gerhard Schurz (Thursday, 20 February 16:30)
by LIRa Seminar 14 Feb '25

14 Feb '25

Dear all, We will have our next LIRa session on Thursday, 20 February 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Gerhard Schurz (Heinrich-Heine-Universität Düsseldorf) Date and time: Thursday, February 20th 2025, 16:30-18:00 Venue: online. Title: Meta-Induction and Its Applications in Social Epistemology Abstract. Meta-induction has been developed as a new approach to Hume's problem of justifying induction. The approach concedes the force of Hume’s sceptical arguments against the possibility of a non-circular justification of the reliability of induction. What it demonstrates is that one can nevertheless give a non-circular justification of the optimality of induction based on meta-induction. The method of meta-induction tracks the success records of all accessible methods of prediction or action (including advice from experts, peers, or algorithms) and combines them in a success-dependent way, in order to obtain a superior method that is provably optimal among all prediction methods accessible to the epistemic agent. Since meta-induction can be viewed as a form of social learning, it has important applications in social epistemology. In the second part of the talk a form of social learning called local meta-induction is presented. In local meta-induction it is assumed that individuals can access only the success records of the individuals in their immediate epistemic neighborhood. It is shown that local meta-inductive learning can spread reliable information over the entire population, and has clear advantages compared to success-independent social learning methods such as peer-imitation and authority-imitation. Hope to see you there! The LIRa team

1 0

LIRa session: Maria Aloni (tomorrow, Thursday, 13 February 16:30)
by LIRa Seminar 12 Feb '25

12 Feb '25

Dear all, We will have our next LIRa session tomorrow, on Thursday, 13 February 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Maria Aloni (ILLC, University of Amsterdam) Date and Time: Thursday, February 13th 2025, 16:30-18:00 Venue: ILLC Seminar Room F1.15, Science Park 107 and online. Title: Nothing is logical. Abstract. People reason in ways which often deviate from classical logic. An influential idea introduced by Grice is that these deviations are not logical mistakes but rather consequences of pragmatic enrichments: the fact that language users often interpret linguistic expressions as conveying more than what is literally said (Grice 1989). In the Gricean tradition, pragmatic enrichments (also known as implicatures) are derived as the product of rational interactions between cooperative language users. In this project we explore a different explanation for some of the non-classical inferences people draw in conversation: they are not the result of Gricean reasonings but rather a consequence of a tendency in human cognition to avoid empty representations (neglect-zero bias). In the talk I will first introduce a team-based modal logic which models the neglect-zero tendency and makes predictions on its impact on reasoning and interpretation; and then present the results of some experiments we conducted to test these predictions. Hope to see you there! The LIRa team

1 0

LIRa session: Maria Aloni (Thursday, 13 February 16:30)
by LIRa Seminar 07 Feb '25

07 Feb '25

Dear all, We will have our next LIRa session on Thursday, 13 February 16:30. To attend online, please use our recurring zoom link: https://uva-live.zoom.us/j/89230639823?pwd=YWJuSnJmTDhXcWhmd1ZkeG5zb0o5UT09 (Meeting ID: 892 3063 9823, Passcode: 421723) You can find the details of the talk below. Speaker: Maria Aloni (ILLC, University of Amsterdam) Date and Time: Thursday, February 13th 2025, 16:30-18:00 Venue: ILLC Seminar Room F1.15, Science Park 107 and online. Title: Nothing is logical. Abstract. People reason in ways which often deviate from classical logic. An influential idea introduced by Grice is that these deviations are not logical mistakes but rather consequences of pragmatic enrichments: the fact that language users often interpret linguistic expressions as conveying more than what is literally said (Grice 1989). In the Gricean tradition, pragmatic enrichments (also known as implicatures) are derived as the product of rational interactions between cooperative language users. In this project we explore a different explanation for some of the non-classical inferences people draw in conversation: they are not the result of Gricean reasonings but rather a consequence of a tendency in human cognition to avoid empty representations (neglect-zero bias). In the talk I will first introduce a team-based modal logic which models the neglect-zero tendency and makes predictions on its impact on reasoning and interpretation; and then present the results of some experiments we conducted to test these predictions. Hope to see you there! The LIRa team

1 0

Joint CLS-DIP-Nihil-LIRa session: Raffaella Bernardi (tomorrow, Thursday, 6 February 16:30)
by LIRa Seminar 05 Feb '25

05 Feb '25

Dear all, We will start the new year with a Joint CLS-DIP-Nihil-LIRa session on Thursday, 6 February 16:30. To attend online, please use this (different than usual!) zoom link: https://uva-live.zoom.us/j/86741669758?pwd=vbWKzB6DbCZJEwqHg014qF5PlEGSWw.1 You can find the details of the talk below. Speaker: Raffaella Bernardi (University of Trento) Date and Time: Thursday, February 6th 2025, 16:30-18:00 Venue: room L0.06 in Lab 42, Science Park Amsterdam and online (note: different zoom link than usual, see above) Title: The interplay between language and reasoning. Abstract. Large Language Models, and ChatGPT in particular, have recently grabbed the attention of the community and the media. Having reached high language proficiency, attention has been shifting toward its reasoning capabilities. It has been shown that ChatGPT can carry out some simple deductive reasoning steps when provided with a series of facts out of which it is tasked to draw some inferences. In this talk, I will argue for the need for models whose language generation is driven by an implicit reasoning process and a communication goal. To support my claim, I will present two papers recently produced within my group: one evaluates LLMs' formal reasoning skills and the other focuses on LLMs’ information-seeking strategies; to this end, we take syllogisms and the 20-Questions game as test beds. These tasks have been used extensively in cognitive sciences to study human reasoning skills, hence they provide us with a variety of experiments to inspect the language and reasoning interplay in LLMs. Leonardo Bertolazzi, Albert Gatt, Raffaella Bernardi: A Systematic Analysis of Large Language Models as Soft Reasoners: The Case of Syllogistic Inferences, EMNLP 2024 https://arxiv.org/abs/2406.11341 Davide Mazzaccara, Alberto Testoni, Raffaella Bernardi: Learning to Ask Informative Questions: Enhancing LLMs with Preference Optimization and Expected Information Gain, EMNLP 2024 https://arxiv.org/abs/2406.17453 Hope to see you there! The LIRa team

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

ILLC-LIRa February 2025