Abstract
We investigate a modal logic of model change, introducing a new operator which allows for changing the valuation at a particular state in a model. After investigating some properties of the logic, and aspects of its expressive power, we show it to be undecidable by way of a reduction using memory logic.
This chapter is in its final form and it is not submitted to publication anywhere else.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Each ± can be independent, so we might have \(p_0 \wedge \lnot p_1 \wedge \dots \).
- 2.
For this section, we will use \(\models _\mathsf {LFC}\) to refer to the definition of truth for LFC, introduced in Definition 7.
- 3.
The notations \(V_\mathfrak {G}^x\) and \(V_\mathfrak {G}^y\) are intended to indicate possibly different modifications to \(V_\mathfrak {G}\).
- 4.
Since is defined, we need not show this case.
References
Areces, Carlos, Raul Fervari, Guillaume Hoffmann, and Mauricio Martel. 2017. Undecidability of Relation-Changing Modal Logics. In Dynamic Logic. New Trends and Applications, vol. 23, 1–16. Lecture Notes in Computer Science. Cham: Springer. ISBN: 978-3-319-73578-8 978-3- 319-73579-5.
Areces, Carlos. 2007. Hybrid Logics: The Old and The New. In Proceedings of LogKCA-07, 15–29.
Areces, Carlos, Diego Figueira, Santiago Figueira, and Sergio Mera. 2011. The Expressive Power of Memory Logics. The Review of Symbolic Logic 4 (02): 290–318.
Aucher, Guillaume, Philippe Balbiani, Luis Fariñas del Cerro, and Andreas Herzig. 2009. Global and Local Graph Modifiers. Electronic Notes in Theoretical Computer Science 231(25): 293–307; In Proceedings of the 5th Workshop on Methods for Modalities (M4M5 2007).
Aucher, Guillaume, Johan van Benthem, and Davide Grossi. 2018. Modal Logics of Sabotage Revisited. Journal of Logic and Computation 28 (2): 269–303.
Balbiani, Philippe, Andreas Herzig, and Nicolas Troquard. 2013. Dynamic Logic of Propositional Assignments: A Well-Behaved Variant of PDL. In Proceedings of the 2013 28th Annual ACM/IEEE Symposium on Logic in Computer Science, 143–152. LICS ’13. Washington: IEEE Computer Society. ISBN: 978-0-7695-5020-6.
Benthem, van, and Johan. 1997. Modal Foundations for Predicate Logic. Logic Journal of the IGPL 5: 259–286.
Benthem, van, Patrick Girard Johan, and Olivier Roy. 2009. Everything Else Being Equal: A Modal Logic for Ceteris Paribus Preferences. Journal of Philosophical Logic 38 (1): 83–125.
Blackburn, Patrick, and Jerry Seligman. 1995. Hybrid Languages. Journal of Logic, Language and Information 4 (3): 251–272.
Bockmayr, Alexander, and Heike Siebert. 2013. Bio-Logics: Logical Analysis of Bioregulatory Networks. In Programming Logics, vol. 7797, eds. Andrei Voronkov and Christoph Weidenbach, 19-34. Berlin: Springer. ISBN: 978-3-642-37650-4 978-3-642-37651-1.
de Bruin, B.P. 2004. Explaining Games: On the Logic of Game Theoretic Explanations. Amsterdam: Institute for Logic, Language and Computation. ISBN: 978-90-5776-129-4.
der Hoek, Van, and Wiebe, and Michael Wooldridge. 2005. On the Logic of Cooperation and Propositional Control. Artificial Intelligence 164(1–2): 81–119.
Gutierrez, Julian, Paul Harrenstein, and Michael Wooldridge. 2015. Iterated Boolean Games. Information and Computation 242: 53–79.
Harrenstein, Paul, Wiebe van der Hoek, John-Jules Meyer, and Cees Witteveen. 2001. Boolean Games. In Proceedings of the 8th Conference on Theoretical Aspects of Rationality and Knowledge, 287–298. Burlington: Morgan Kaufmann Publishers Inc.
Harrenstein, Paul, Wiebe Van der Hoek, John-Jules Meyer, and Cees Witteveen. 2003. A Modal Characterization of Nash Equilibrium. Fundamenta Informaticae 57(2-4): 281–321.
Mera, Sergio Fernando. 2009. Modal Memory Logics. PhD Thesis, Universidad de Buenos Aires.
Miller, Joseph S., and Lawrence S. Moss. 2005. The Undecidability of Iterated Modal Relativization. Studia Logica 79, no. 3 (1): 373–407.
Seligman, Jeremy, and Declan Thompson. 2015. Boolean Network Games and Iterated Boolean Games. In Logic, Rationality, and Interaction, eds. Wiebe van der Hoek, Wesley H. Holliday, and Wen-fang Wang, 353–365. Lecture Notes in Computer Science. Berlin: Springer. ISBN: 978-3-662-48561-3.
Tiomkin, M.L., and J.A. Makowsky. 1985. Propositional Dynamic Logic with Local Assignment. Theoretical Computer Science 36 (1): 71–87.
van Benthem, Johan. 2011. Logical Dynamics of Information and Interaction. Cambridge: Cambridge University Press. ISBN: 978-0-521-76579-4.
van Benthem, Johan. 2018. Constructive Agents. Indagationes Mathematicae 29(1): 23–35.
Acknowledgements
I would like to thank Carlos Areces for very helpful input, especially in connection with memory logics. Ben Sparkes encouraged investigating similarities with hybrid logics, which led to many of the results in this paper. I also received very helpful feedback from Johan van Benthem, the members of the Logic Workshop at Stanford, and the participants of the Fourth Asian Workshop in Philosophical Logic. Finally, the anonymous reviewers provided very useful feedback.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
1.1 Proof of Fact 4
Fact 8
Let \(A\subseteq \mathsf {Prop}\). If \(\psi \in \mathsf {Bool}(\mathsf {At}(\varphi ))\) and \(\mathfrak {F}, V, s \models \psi \) and \(\mathfrak {F}, V^s_A, s \models \psi \) then
Proof
Recall that \(\psi \) is of the form \(\pm p_0 \wedge \dots \pm p_n\), where \(p_0, \dots p_n \in \mathsf {At}(\varphi )\). Since \(\psi \) has no modal operators and its evaluation only depends on the values of propositional variables in \(\mathsf {At}(\varphi )\), we know \(V(s) \cap \mathsf {At}(\varphi ) = V^s_A(s) \cap \mathsf {At}(\varphi )\). From this, we conclude \(V(t) \cap \mathsf {At}(\varphi ) = V^s_A(t) \cap \mathsf {At}(\varphi )\) for all t. Observing that \(\varphi \) also only depends on the values of propositional variables in \(\mathsf {At}(\varphi )\) gives us the result. \(\square \)
Fact 9
iff .
Proof
For simplicity, we will omit \(\mathfrak {F}\) below. We have the following equivalences:
Equation (1) follows from the definition of , and (2) via basic logic. Taking \(\varphi _{V,s}\) to characterise V at s, equivalence (3) holds, and since \(V\models \psi _{V,s}\) by the definition of \(\psi _{V,s}\), we get (4). Equations (5) and (6) then follow from the truth definitions. In the remainder of the proof, we will establish that (6) holds iff .
(\(\Rightarrow \)) Suppose \(V^s_B, s\not \models p\) or \(V^s_B, s\not \models \psi _{V,s}\) or \(V^s_B, s\models \varphi \) for all B. Let \(A = V(s) \cup \{p\}\). We have that \(V^s_A, s \models p\) trivially, and \(V^s_A, s \models \psi _{V,s}\) since \(V, s \models \psi _{V,}\) and \(p \notin \mathsf {At}(\psi _{V,s})\), so we must have \(V^s_A, s \models \varphi \). Hence \(V^s_A, s \models p \wedge \psi _{V,s} \wedge \varphi \) and so . From this it quickly follows that from the definition of .
(\(\Leftarrow \)) Suppose . By similar reasoning to the move from (2) to (4), we know that . Thus, \(V^s_C, s \models p \wedge \psi _{V,s} \wedge \varphi \) for some C. Take arbitrary B and suppose \(V^s_{B}, s \models p\) and \(V^s_{B}, s \models \psi _{V,s}\). Since \(V^s_B\) and \(V^s_C\) agree on p and \(\psi _{V,s}\) at s, we know that \(V^s_{B}, s \models \varphi \) iff \(V^s_C, s \models \varphi \), by Fact 8. Since \(V^s_C, s\models \varphi \), we have \(V^s_B, s \models \varphi \). Since B was arbitrary, we can conclude that for all B, either \(V^s_B, s\not \models p\) or \(V^s_B, s\not \models \psi _{V,s}\) or \(V^s_B, s\models \varphi \). \(\square \)
Fact 4
iff \(\mathfrak {F}, V, s \models [+p] \varphi \), for every formula \(\varphi \) and all frames \(\mathfrak {F}\), valuations V and points s.
Proof
By induction on the complexity of \(\varphi \). For simplicity, we will omit reference to \(\mathfrak {F}\) in what follows. Let \(A = V(s) \cup \{p\}\).
For \(\varphi \) propositional there are two cases. If \(\varphi = p\), then iff , which is the case since . It’s trivial to see that \(V^s_A, s \models p\). The second case is when \(\varphi = q \ne p\). In this case, we conclude that iff and . Since we know and , we conclude that iff \(V,s\models q\). By the definition of truth, \(V,s\models q\) iff \(q\in V(s)\) iff \(q\in V(s)\cup \{p\}\) (since \(q\ne p\)). This holds iff \(V^s_A, s \models q\) iff \(V, s \models [+p]q\).
Next consider the case where \(\varphi = \lnot \psi \) for some \(\psi \). Using Fact 9, iff . The required result follows quickly using the inductive hypothesis. Similar arguments to those given in Fact 9 can be used to establish the case for \(\varphi = \psi \wedge \chi \) for some \(\psi \) and \(\chi \).
Now we take for some \(\chi \). Using a similar argument to Fact 9, we know that iff , where \(\psi _{V,s} \in \mathsf {Bool}(\mathsf {At}(\chi )\setminus \{p\})\) characterises V at s. By definition, this holds iff for some B. This immediately establishes that since we can move to \(V^s_B\). Hence, . For the other direction, notice that \(V^s_A, s \models p\) and \(V^s_A, s \models \psi _{V,s}\) since \(V, s \models \psi _{V,s}\) and \(p \notin \mathsf {At}(\psi _{V,s})\).
Finally,Footnote 4 suppose \(\varphi = \lozenge \chi \). Unpacking definitions, we know that iff \(V^s_B, s \models p \wedge \psi _{V,s} \wedge \lozenge \chi \) for some B, where \(\psi _{V,s} \in \mathsf {Bool}(\mathsf {At}(\chi )\setminus \{p\})\) characterises V at s. So \(V^s_B, s \models p \wedge \psi _{V,s}\) and by the definition of \(V^s_A\) and \(\psi _{V,s}\) we know \(V^s_A, s \models p \wedge \psi _{V,s}\). It follows that \(V^s_B\) and \(V^s_A\) agree on all propositional variables in \(\mathsf {At}(\lozenge \chi )\), and so by Fact 8 we can conclude that \(V^s_B, s \models \lozenge \chi \) iff \(V^s_A, s \models \lozenge \chi \).
Initially, we found that iff \(V^s_B, s \models p \wedge \psi _{V,s} \wedge \lozenge \chi \), and now we have shown that \(V^s_B, s \models \lozenge \chi \) iff \(V^s_A, s \models \lozenge \chi \). From these, we establish that iff \(\mathfrak {F}, V, s \models [+p]\lozenge \chi \). \(\square \)
1.2 Bisimulation
We can define a natural notion of bisimulation for LFC. Like the Ehrenfeucht–Fraïssé games, this is similar to the approach for ML in Areces et al. (2011).
Definition 20
Let \(\mathfrak {F}_1 = (W_1, R_1)\) and \(\mathfrak {F}_2 = (W_2, R_2)\).
A relation Z between \(W_1 \times \mathsf {Prop}^{W_1}\) and \(W_2 \times \mathsf {Prop}^{W_2}\) is a bisimulation if the following clauses hold: Â Â
- Non-empty:
-
\(Z \ne \emptyset \).
- Agree:
-
If \((s_1, V_1) Z (s_2, V_2)\) then \(V_1(s_1) = V_2(s_2)\).
- Zig:
-
If \((s_1, V_1) Z (s_2, V_2)\) and \(R_1s_1t_1\) then there is \(t_2\in A_2\) with \(R_2s_2t_2\) and \((t_1, V_1) Z (s_t, V_2)\).
- Zag:
-
If \((s_1, V_1) Z (s_2, V_2)\) and \(R_2s_2t_2\) then there is \(t_1\in A_1\) with \(R_1s_1t_1\) and \((t_1, V_1) Z (s_t, V_2)\).
- Change:
-
If \((s_1, V_1) Z (s_2, V_2)\) then \((s_1, {V_1}^{s_1}_A) Z (s_2, {V_2}^{s_2}_A)\) for every \(A \subseteq \mathsf {Prop}\).
This is the standard definition of bisimulation, but with valuations included. This is because we must keep track not only of which state we’re in, but what the current valuation is. The last clause simply ensures that a change of valuation at bisimilar points preserves bisimilarity (assuming the valuation is changed uniformly).
Fact 5
(Correctness of bisimulation). Let \(\mathfrak {F}_1 = (W_1, R_1)\) and \(\mathfrak {F}_2 = (W_2, R_2)\). For every formula \(\varphi \in \mathcal {L}_{\mathsf {LFC}}\), \(s_1 \in W_1\) and \(s_2 \in W_2\), if there is a bisimulation Z with \((s_1, V_1) Z (s_2, V_2)\) then \(\mathfrak {F}_1, V_1, s_1 \models \varphi \) iff \(\mathfrak {F}_2, V_2, s_2 \models \varphi \).
Proof
By induction on \(\varphi \). The Boolean and \(\lozenge \) cases are standard.
The final case is where . iff \(\mathfrak {F}_1, {V_1}^{s_1}_A, s_1 \models \psi \) for some \(A \subseteq \mathsf {Prop}\). Since \((s_1, V_1) Z (s_2, V_2)\), by Change we know \((s_1, {V_1}^{s_1}_A) Z (s_2, {V_2}^{s_2}_A)\). Applying the inductive hypothesis, \(\mathfrak {F}_1, {V_1}^{s_1}_A, s_1 \models \psi \) iff \(\mathfrak {F}_2, {V_2}^{s_2}_A, s_2 \models \psi \). But this holds exactly when , as required. \(\square \)
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Thompson, D. (2020). Local Fact Change Logic. In: Liu, F., Ono, H., Yu, J. (eds) Knowledge, Proof and Dynamics. Logic in Asia: Studia Logica Library. Springer, Singapore. https://doi.org/10.1007/978-981-15-2221-5_5
Download citation
DOI: https://doi.org/10.1007/978-981-15-2221-5_5
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-15-2220-8
Online ISBN: 978-981-15-2221-5
eBook Packages: Religion and PhilosophyPhilosophy and Religion (R0)