Periodic Reporting for period 4 - LoC (The Logic of Conceivability: Modelling Rational Imagination with Non-Normal Modal Logics)
Période du rapport: 2021-09-01 au 2022-06-30
The Logic of Conceivability (LoC) project focuses on the logic of mental simulation: how we reason when we consider hypothetical scenarios and wonder what would follow if they happened. Our mind represents non-actual scenarios to extract information from them: since we cannot experience beforehand which situations are or will be actual, we explore them in our imagination, leaving our perceptions offline: ‘What will happen if there's a disorderly Brexit?’; 'What will I do if I cannot repay my mortgage?'. The cognitive importance of this activity is hard to overestimate.
But what is its logic? The orthodox logical treatment of representational mental states like conceiving, imagining, and believing, comes from modal logic’s so-called 'possible worlds semantics': the modal analysis of knowledge, belief, information, provided by modal logic, was taken up by philosophy, linguistics, and Artificial Intelligence. However, the approach faces problems. By systematically addressing them, the LoC project aims at yielding a paradigm shift in our understanding of the logic of human suppositional thought.
One purely logical problem is that mainstream epistemic logics model cognitive agents as 'logically omniscient', thus as disconnected from the reality of human, fallible minds. The cognitive agents represented in this way conceive (imagine, believe, etc.) all the logical consequences of what they conceive (etc.). They know all logical truths, and are perfectly consistent in their beliefs. Real people are not like that.
One major philosophical problem concerns the entailment from conceivability to possibility, e.g. in ‘thought experiments’ of theoretical philosophy: how does conceiving a scenario give evidence of its possibility? Imagination seems to be totally unconstrained; but if we can imagine whatever we like after supposing an initial scenario, e.g. via free association of ideas, then such exercises seem to give us no hints on what is or is not likely to happen, on what may and may not take place.
LoC addresses such issues via the techniques of non-classical modal logics. One aim is to make logically precise the distinction, taken from cognitive psychology, between Fast Thinking (associative, context-sensitive) and Slow Thinking (rule-based, analytic). Another is to show how logical omniscience is avoided, and evidence of possibility is achieved, in different manners in the Fast and Slow Way. Finally, the LoC project aims at connecting the abstract theoretical modelling of cognitive agents performed within mathematical logic with the reality of empirical results from the psychology of reasoning.
Based at the Department of Philosophy of the University of St Andrews, and at the Institute for Logic, Language and Computation of the University of Amsterdam, and advised by a Board of researchers from Europe, the US, and Australia, LoC is delivering high-impact outputs for top peer-reviewed journals and publishers, and knowledge dissemination results for non-specialists.
But what is its logic? The orthodox logical treatment of representational mental states like conceiving, imagining, and believing, comes from modal logic’s so-called 'possible worlds semantics': the modal analysis of knowledge, belief, information, provided by modal logic, was taken up by philosophy, linguistics, and Artificial Intelligence. However, the approach faces problems. By systematically addressing them, the LoC project aims at yielding a paradigm shift in our understanding of the logic of human suppositional thought.
One purely logical problem is that mainstream epistemic logics model cognitive agents as 'logically omniscient', thus as disconnected from the reality of human, fallible minds. The cognitive agents represented in this way conceive (imagine, believe, etc.) all the logical consequences of what they conceive (etc.). They know all logical truths, and are perfectly consistent in their beliefs. Real people are not like that.
One major philosophical problem concerns the entailment from conceivability to possibility, e.g. in ‘thought experiments’ of theoretical philosophy: how does conceiving a scenario give evidence of its possibility? Imagination seems to be totally unconstrained; but if we can imagine whatever we like after supposing an initial scenario, e.g. via free association of ideas, then such exercises seem to give us no hints on what is or is not likely to happen, on what may and may not take place.
LoC addresses such issues via the techniques of non-classical modal logics. One aim is to make logically precise the distinction, taken from cognitive psychology, between Fast Thinking (associative, context-sensitive) and Slow Thinking (rule-based, analytic). Another is to show how logical omniscience is avoided, and evidence of possibility is achieved, in different manners in the Fast and Slow Way. Finally, the LoC project aims at connecting the abstract theoretical modelling of cognitive agents performed within mathematical logic with the reality of empirical results from the psychology of reasoning.
Based at the Department of Philosophy of the University of St Andrews, and at the Institute for Logic, Language and Computation of the University of Amsterdam, and advised by a Board of researchers from Europe, the US, and Australia, LoC is delivering high-impact outputs for top peer-reviewed journals and publishers, and knowledge dissemination results for non-specialists.
The LoC project has been carried out by a team composed by one PhD candidate, Tom Schoonen, and four postdoctoral researchers: an epistemologist (Dr. Peter Hawke), two mathematical logicians (Drs. Aybüke Özgün and Thomas Ferguson), and a psychologist of reasoning (Dr. Karolina Krzyzanowska) to address the LoC research questions via a combined effort.
Two kinds of logics for realistic cognitive agents have been developed: (a) systems using so-called non-normal worlds model how such agents can have inconsistent beliefs without thereby trivializing their belief system, and can be moderately rational in spite of being unable to think through all the logical consequences of what they suppose, imagine, or conceive. (b) Systems that stick to traditional possible worlds semantics, but enrich it with algebraic structures of 'topics' or 'subject matters' (intuitively: what the agent's thoughts are about) that model how agents can have distinct attitudes towards logically equivalent claims. Systems of the kind (a) were predicted to be used in the initial LoC project description; systems of the kind (b) were developed after the start of the LoC project and delivered a number of surprising results.
The team has produced a very large number of publications presenting such systems and applying them to the modelling of mental simulation, information processing, belief revision. These have appeared the most selective peer-reviewed journals of logic, philosophy, and cognition, such as the Journal of Philosophical Logic, Cognitive Psychology, Mind, Synthese, Philosophical Studies, The Australasian Journal of Philosophy, the Review of Symbolic Logic.
Berto has published two monographs with Oxford University Press, the world's most prestigious publisher for philosophy: (1) 'Impossible Worlds' (2019), with Prof. Mark Jago, focuses on logics of the kind (a) above; (2) 'Topics of Thought' (2022), with chapters co-authored with Hawke and Özgün, focuses on logics of the kind (b) above. Both are Open Access:
https://global.oup.com/academic/product/impossible-worlds-9780198812791?cc=gh&lang=en&
https://global.oup.com/academic/product/topics-of-thought-9780192857491?cc=gh&lang=en&
The team has presented the results of its research in non-peer-reviewed popular publications, blogs, and via popular talks. A list of the project outputs so far can be found here:
https://projects.illc.uva.nl/conceivability/The-Outputs/
The team has given dozens of invited and contributed talks at conferences throughout Europe and the US.
Two kinds of logics for realistic cognitive agents have been developed: (a) systems using so-called non-normal worlds model how such agents can have inconsistent beliefs without thereby trivializing their belief system, and can be moderately rational in spite of being unable to think through all the logical consequences of what they suppose, imagine, or conceive. (b) Systems that stick to traditional possible worlds semantics, but enrich it with algebraic structures of 'topics' or 'subject matters' (intuitively: what the agent's thoughts are about) that model how agents can have distinct attitudes towards logically equivalent claims. Systems of the kind (a) were predicted to be used in the initial LoC project description; systems of the kind (b) were developed after the start of the LoC project and delivered a number of surprising results.
The team has produced a very large number of publications presenting such systems and applying them to the modelling of mental simulation, information processing, belief revision. These have appeared the most selective peer-reviewed journals of logic, philosophy, and cognition, such as the Journal of Philosophical Logic, Cognitive Psychology, Mind, Synthese, Philosophical Studies, The Australasian Journal of Philosophy, the Review of Symbolic Logic.
Berto has published two monographs with Oxford University Press, the world's most prestigious publisher for philosophy: (1) 'Impossible Worlds' (2019), with Prof. Mark Jago, focuses on logics of the kind (a) above; (2) 'Topics of Thought' (2022), with chapters co-authored with Hawke and Özgün, focuses on logics of the kind (b) above. Both are Open Access:
https://global.oup.com/academic/product/impossible-worlds-9780198812791?cc=gh&lang=en&
https://global.oup.com/academic/product/topics-of-thought-9780192857491?cc=gh&lang=en&
The team has presented the results of its research in non-peer-reviewed popular publications, blogs, and via popular talks. A list of the project outputs so far can be found here:
https://projects.illc.uva.nl/conceivability/The-Outputs/
The team has given dozens of invited and contributed talks at conferences throughout Europe and the US.
The two approaches (a) and (b) above have provided realistic models of the logic of mental simulation, including suppositional thinking and belief revision. Such models represent, and make predictions on, how human agents will reason in such contexts of imaginative thought.
Approach (a) has been presented in the aforementioned OUP book by Berto and Jago which is becoming the reference book for non-normal worlds semantics and its applications. It has also been combined with the techniques of Dynamic Epistemic Logic, which model the dynamics of reasoning, belief update, and revision.
Approach (b) has also been intensively developed by the team, and technical logical results (completneess proofs, axiomatizations of formal systems in the (b) family) concerning it have been achieved, mostly thanks to the work of the two mathematical logicians, Özgün and Ferguson.
Both approaches have been picked up by logicians, philosophers of logic, mind, and cognition. Thanks to their being Open Access, a number of papers have attracted thousands of downloads and have already started to gather good amounts of citations -- three examples with download/citation counts here:
https://link.springer.com/article/10.1007/s10670-017-9875-5
https://link.springer.com/article/10.1007/s11098-017-0937-y
https://link.springer.com/article/10.1007/s10670-019-00128-z
The Impossible Worlds approach is also expanding -- as shown by the curve of citations on Google Scholar:
https://scholar.google.co.uk/citations?view_op=view_citation&hl=en&user=8unCUCUAAAAJ&citation_for_view=8unCUCUAAAAJ:A7-hzOuI2KQC
The 'Impossible Worlds' book has been Oxford University Press' most downloaded Open Access book of 2021:
https://twitter.com/OUPPhilosophy/status/1453347529127432200
Approach (a) has been presented in the aforementioned OUP book by Berto and Jago which is becoming the reference book for non-normal worlds semantics and its applications. It has also been combined with the techniques of Dynamic Epistemic Logic, which model the dynamics of reasoning, belief update, and revision.
Approach (b) has also been intensively developed by the team, and technical logical results (completneess proofs, axiomatizations of formal systems in the (b) family) concerning it have been achieved, mostly thanks to the work of the two mathematical logicians, Özgün and Ferguson.
Both approaches have been picked up by logicians, philosophers of logic, mind, and cognition. Thanks to their being Open Access, a number of papers have attracted thousands of downloads and have already started to gather good amounts of citations -- three examples with download/citation counts here:
https://link.springer.com/article/10.1007/s10670-017-9875-5
https://link.springer.com/article/10.1007/s11098-017-0937-y
https://link.springer.com/article/10.1007/s10670-019-00128-z
The Impossible Worlds approach is also expanding -- as shown by the curve of citations on Google Scholar:
https://scholar.google.co.uk/citations?view_op=view_citation&hl=en&user=8unCUCUAAAAJ&citation_for_view=8unCUCUAAAAJ:A7-hzOuI2KQC
The 'Impossible Worlds' book has been Oxford University Press' most downloaded Open Access book of 2021:
https://twitter.com/OUPPhilosophy/status/1453347529127432200