classic like analytic tableaux for non classical logics
play

Classic-like Analytic Tableaux for Non-Classical Logics Joo Marcos - PowerPoint PPT Presentation

Classic-like Analytic Tableaux for Non-Classical Logics Joo Marcos UFRN, BR & RUB, DE PhDs in Logic IX Bochum 24 May 2017 Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction


  1. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux Please make it stop! . . . to Tableaux #2 T : ‚ α F : � α T : � α T : ‚� α F : ‚� α T : ‚‚ α T : α F : α T : α T : α F : α T : ‚ α T : ‚ α F : � α F : α ¸ F : ‚ α ¸ A more interesting exercise: Check that #2 does not allow for loops. Solution: Consider the non-canonical complexity measure # ℓ p ψ q ` 1 , if ϕ “ ‚ ψ ℓ p ϕ q “ ℓ p ψ q ` 2 , if ϕ “ � ψ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  2. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux Please make it stop! . . . to Tableaux #2 T : ‚ α F : � α T : � α T : ‚� α F : ‚� α T : ‚‚ α T : α F : α T : α T : α F : α T : ‚ α T : ‚ α F : � α F : α ¸ F : ‚ α ¸ A more interesting exercise: Check that #2 does not allow for loops. Solution: Consider the non-canonical complexity measure # ℓ p ψ q ` 1 , if ϕ “ ‚ ψ ℓ p ϕ q “ ℓ p ψ q ` 2 , if ϕ “ � ψ Note: This hints to a generalization of the Subformula Property ! João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  3. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux Should proof systems include a proof strategy? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  4. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux Should proof systems include a proof strategy? Does the rule application order matter? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  5. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux Should proof systems include a proof strategy? Does the rule application order matter? More about this at a later example! João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  6. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux As linear as can be João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  7. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux As linear as can be Signed Tableaux for Classical Logic: F : � α T : � α F : α Ñ β T : α Ñ β T : α F : α F : α T : β T : α F : α T : α ¸ F : β João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  8. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux As linear as can be Cut-based Signed Tableaux for Classical Logic: F : � α T : � α F : α Ñ β T : α Ñ β T : α Ñ β T : α T : α F : β F : α F : α T : α T : α F : α T : α ¸ F : β T : β F : α João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  9. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux As linear as can be Cut-based Signed Tableaux for Classical Logic: F : � α T : � α F : α Ñ β T : α Ñ β T : α Ñ β T : α T : α F : β F : α F : α T : α T : α F : α T : α ¸ F : β T : β F : α Exercises: Check again the provability of p p Ñ q q Ñ r $ p Ñ p q Ñ r q and of its converse. Check again the provability of $ p� p Ñ p q Ñ p . João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  10. Tableaux Classical Logic More on finite-valued logics A Non-classical example Through Suszko Reduction Analyticity generalized Comparing different logics Rule application strategies . . . and so much more! Cut-based tableaux As linear as can be Cut-based Signed Tableaux for Classical Logic: F : � α T : � α F : α Ñ β T : α Ñ β T : α Ñ β T : α T : α F : β F : α F : α T : α T : α F : α T : α ¸ F : β T : β F : α Exercises: Check again the provability of p p Ñ q q Ñ r $ p Ñ p q Ñ r q and of its converse. Check again the provability of $ p� p Ñ p q Ñ p . Are tableaux advantageous over truth-tables? The ‘average case’ is better, the ‘worst case’ much worse! Cut-based tableaux can polynomially simulate truth-tables. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  11. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! TWO is not enough! “ The philosophical logic simply had no appreciation for the finer conceptual distinctions because it did not operate with sharply delineated concepts and unambiguously determined symbols; rather it sank into the swamp of the fluid and vague speech used in everyday. ” João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  12. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! A 3-valued logic: Ł3 João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  13. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! A 3-valued logic: Ł3 The logical matrices of Ł3: (where D “ t 1 u ) 1 � Ñ 0 1 2 0 1 0 1 1 1 1 1 1 1 1 1 2 2 2 2 1 1 0 1 0 1 2 João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  14. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! A 3-valued logic: Ł3 The logical matrices of Ł3: (where D “ t 1 u ) 1 � Ñ 0 1 2 0 1 0 1 1 1 1 1 1 1 1 1 2 2 2 2 1 1 0 1 0 1 2 Tableaux through brute force : Examples of logical rules : Examples of closure rules : 1 1 2 : � α 2 : α Ñ β 0 : α 1 2 : α 1 2 : α 1 : α 1 1 : α 1 2 : α 2 : α 1 2 : β ¸ ¸ 0 : β João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  15. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! What about classic-like tableaux for Ł3? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  16. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! What about classic-like tableaux for Ł3? Finding an adequate bivalent semantics where V “ D Y U and D X U “ ∅ , h S V ¶ p x q “ T if x P D , b h “ ¶ ˝ h ¶ p x q “ F if x P U . ¶ Then: t T , F u Γ | ù b h α iff Γ | ù h α João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  17. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! What about classic-like tableaux for Ł3? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  18. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! What about classic-like tableaux for Ł3? The logical matrices of Ł3: (where D “ t 1 u ) 1 � Ñ 0 1 2 0 1 0 1 1 1 1 1 1 1 1 1 2 2 2 2 1 1 0 1 0 1 2 How the logical rules may now look like: New closure rules : F : α Ñ β T : �p α Ñ β q T : α T : α T : � α F : α F : � α T : α T : α F : β T : � β T : � β ¸ ¸ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  19. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  20. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  21. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 First step : (algebraic ˆ logical values) Pairwise distinguishing the truth-values in terms of ‘ binary prints ’. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  22. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 x ¶ p x q First step : (algebraic ˆ logical values) 0 F Pairwise distinguishing the truth-values 1 F 3 2 in terms of ‘ binary prints ’. F 3 1 T João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  23. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Example: Consider the separating formulas : θ 0 p ϕ q “ ϕ θ 1 p ϕ q “ � ϕ θ 2 p ϕ q “ ��p ϕ Ñ � ϕ q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  24. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Example: Consider the ¶ p Ă ¶ p Ă x ¶ p x q θ 1 p x q θ 1 p x qq θ 2 p x q θ 2 p x qq separating formulas : 0 1 1 F T T θ 0 p ϕ q “ ϕ 1 2 F F 1 T 3 3 2 1 1 θ 1 p ϕ q “ � ϕ F F F 3 3 3 1 0 0 T F F θ 2 p ϕ q “ ��p ϕ Ñ � ϕ q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  25. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Second step : Provide a bivalent description of the truth-tables using the binary prints of the truth-values. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  26. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Examples of logical rules : F : � α T : � α T : θ 1 p� α q F : α T : α T : θ 1 p α q F : θ 1 p α q F : α F : α T : α T : θ 2 p α q F : θ 2 p α q F : θ 1 p α q F : θ 1 p α q F : θ 1 p α q T : θ 2 p α q F : θ 2 p α q F : θ 2 p α q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  27. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Examples of T : θ 2 p α Ñ β q logical rules : F : α T : α T : α F : θ 1 p α q F : θ 1 p α q F : θ 1 p α q F : θ 2 p α q F : θ 2 p α q F : θ 2 p α q F : β F : β F : β T : θ 1 p β q T : θ 1 p β q F : θ 1 p β q T : θ 2 p β q T : θ 2 p β q T : θ 2 p β q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  28. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Third step : Take into account the unobtainable semantic scenarios. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  29. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Some binary prints do not correspond to any algebraic truth-value : ¶ p Ă ¶ p Ă ¶ p Ă x θ 0 p x qq θ 1 p x qq θ 2 p x qq ¸ T T T ¸ T T F ¸ T F T 1 T F F 0 F T T ¸ F T F 1 F F T 3 2 F F F 3 João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  30. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 Minimize the corresponding information: ¶ p Ă ¶ p Ă ¶ p Ă x θ 0 p x qq θ 1 p x qq θ 2 p x qq ¶ p Ă ¶ p Ă ¶ p Ă x θ 0 p x qq θ 1 p x qq θ 2 p x qq ¸ T T T ¸ T T ´ becomes ¸ T T F ¸ T ´ T ¸ T F T ¸ F T F ¸ F T F João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  31. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 The latter, in turn, originate the following closure rules : F : α T : α T : α T : α T : θ 1 p α q T : θ 2 p α q T : θ 1 p α q F : α F : θ 2 p α q ¸ ¸ ¸ ¸ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  32. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The logical matrices of Ł4: (where D “ t 1 u ) 1 2 � Ñ 0 1 3 3 0 1 0 1 1 1 1 1 2 1 2 1 1 1 3 3 3 3 2 1 2 1 2 1 1 3 3 3 3 3 1 2 1 0 1 0 1 3 3 The latter, in turn, originate the following closure rules : F : α T : α T : α T : α T : θ 1 p α q T : θ 2 p α q T : θ 1 p α q F : α F : θ 2 p α q ¸ ¸ ¸ ¸ 1 2 For reflection: What if we had used 0 , 3 , 3 and 1 as labels? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  33. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values Fourth step : Consider a strategy for rule application , in order to guarantee analyticity. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  34. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values Fourth step : Consider a strategy for rule application , in order to guarantee analyticity. Example: Consider a signed formula of the form T : ϕ , where ϕ is ��pp α Ñ β q Ñ �p α Ñ β qq . João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  35. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values Example: Consider a signed formula of the form T : ϕ , where ϕ is ��pp α Ñ β q Ñ �p α Ñ β qq . João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  36. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The corresponding labelled node can match the head of some Ł 4 -rule in 3 different ways, namely: Rule r T : θ 0 s is applied: Rule r T : θ 1 �s is applied: T : ��pp α Ñ β q Ñ �p α Ñ β qq T : θ 1 p�pp α Ñ β q Ñ �p α Ñ β qqq F : �pp α Ñ β q Ñ �p α Ñ β qq T : pp α Ñ β q Ñ �p α Ñ β qq T : θ 1 p�pp α Ñ β q Ñ �p α Ñ β qqq F : θ 1 ppp α Ñ β q Ñ �p α Ñ β qqq T : θ 2 p�pp α Ñ β q Ñ �p α Ñ β qqq F : θ 2 ppp α Ñ β q Ñ �p α Ñ β qqq Rule r T : θ 2 Ñs is applied: T : θ 2 p α Ñ β q F : α T : α T : α F : θ 1 p α q F : θ 1 p α q F : θ 1 p α q F : θ 2 p α q F : θ 2 p α q F : θ 2 p α q F : β F : β F : β T : θ 1 p β q T : θ 1 p β q F : θ 1 p β q T : θ 2 p β q T : θ 2 p β q T : θ 2 p β q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  37. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Adding more values The following rule application strategy , according to which: (1) θ k -rules, for k ą 0, are to be preferred over θ 0 -rules (2) if there are θ i and θ j such that θ i p ϕ i q “ ϕ “ θ j p ϕ j q , for some appropriate ϕ i and ϕ j , one should give preference to the θ -rule whose head is more ‘concrete’ (m.g.u.) guarantees the decrease of the following non-canonical complexity measure : ( ℓ 0) ℓ p θ k p ϕ qq “ ℓ p ϕ q , where k ą 0 and k is ‘minimal’ (see (2)) ( ℓ 1) ℓ p p q “ 0, where p is an atom ( ℓ 2) ℓ p� ϕ 1 q “ ℓ p ϕ 1 q ` 1 ( ℓ 3) ℓ p ϕ 2 Ñ ϕ 3 q “ ℓ p ϕ 2 q ` ℓ p ϕ 3 q ` 1 João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  38. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Deriving rules João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  39. Tableaux More on finite-valued logics Beyond two truth-values Through Suszko Reduction A bivalent approach Comparing different logics Further detail on the associated procedure . . . and so much more! Deriving rules On comparing logics How could one use the latter proof systems to compare the logics Ł3 and Ł4? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  40. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Is many-valuedness but a farse? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  41. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Is many-valuedness but a farse? ‘‘`a˜fˇt´eˇrffl 50 ”y´e´a˚r¯s ”wfle ¯sfi˚tˇi˜l¨l ˜f´a`c´e `a‹nffl ˚i˜l¨l´oˆgˇi`c´a˜l ¯p`a˚r`a`d˚i¯sfi`e `o˝f ”m`a‹n‹y ˚tˇr˚u˚t‚h¯s `a‹n`dffl ˜f´a˜l˙ sfi`e‚h`oˆoˆd¯s’’ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  42. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Is many-valuedness but a farse? ‘‘˛h`o“w ”wˆa¯s ˚i˚t ¯p`o¸ sfi¯sfi˚i˜b˝l´e ˚t‚h`a˚t ˚t‚h`e ˛h˚u‹m˜b˘u`g `o˝f ”m`a‹n‹y ˜l´oˆgˇi`c´a˜l ”vˆa˜lˇu`e˙ s ¯p`eˇr¯sfi˚i¯sfi˚t´e´dffl `o“vfleˇrffl ˚t‚h`e ˜l´a¯sfi˚t ˜fˇi˜fˇt›y ”y´e´a˚r¯s?’’ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  43. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Is many-valuedness but a farse? ‘‘Łu˛k`a¯sfi˚i`e›w˘i`cˇz ˚i¯s ˚t‚h`e `c‚h˚i`e¨f ¯p`eˇr¯p`eˇtˇr`a˚t´o˘rffl `o˝f `affl ”m`a`g›n˚i˜fˇi`c´e›n˚t `c´o“n`c´e˙ p˚tˇu`a˜l `d`e´c´eˇi¯p˚t ˜l´a¯sfi˚tˇi‹n`g `o˘u˚t ˚i‹nffl ”m`a˚t‚h`e›m`a˚tˇi`c´a˜l ˜l´oˆgˇi`c ˚t´o ˚t‚h`e ¯p˚r`e˙ sfi`e›n˚t `d`a‹y.’’ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  44. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Is many-valuedness but a farse? ‘‘`o˝b“v˘i`o˘u¯sfi˜l›y, `a‹n‹y ”m˚u˜lˇtˇi¯p˜lˇi`c´a˚tˇi`o“nffl `o˝f ˜l´oˆgˇi`c´a˜l ”vˆa˜lˇu`e˙ s ˚i¯s `affl ”m`a`dffl ˚i`d`e´affl’’ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  45. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to find separators João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  46. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to find separators The previously illustrated procedure may be fully automated. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  47. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to find separators Algorithm #1 . Starting from a logical matrix over a given signature, find out whether it is sufficiently expressive to separate the truth-values. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  48. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to find separators Algorithm #1 . Starting from a logical matrix over a given signature, find out whether it is sufficiently expressive to separate the truth-values. In case it is not, produce a minimal conservative extension that does the job. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  49. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce a bivalent characterization João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  50. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce a bivalent characterization Algorithm #2 . Starting from a sufficiently expressive logical matrix, extract an axiomatization (using a classical metalanguage) that describes a bivalent characterization of it (as previously explained in Steps 1 and 2). João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  51. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce a bivalent characterization Algorithm #2 . Starting from a sufficiently expressive logical matrix, extract an axiomatization (using a classical metalanguage) that describes a bivalent characterization of it (as previously explained in Steps 1 and 2). Make sure this characterization takes into account an appropriate well-founded notion of complexity that allows for a generalization of the compositionality principle. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  52. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce analytic tableaux João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  53. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce analytic tableaux Algorithm #3 . Given an appropriate bivalent characterization of a given logic, calculate the minimal closuring sequences and set up a rule application strategy (as previously explained in Steps 3 and 4). João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  54. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! Algorithm to produce analytic tableaux Algorithm #3 . Given an appropriate bivalent characterization of a given logic, calculate the minimal closuring sequences and set up a rule application strategy (as previously explained in Steps 3 and 4). Describe analytic tableaux based on the data obtained so far. João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  55. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! More values, or just different? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  56. Tableaux More on finite-valued logics The role of separators Through Suszko Reduction Bivalence is the key Comparing different logics Analytic tableaux . . . and so much more! More values, or just different? Again on the comparison of logics We have already discussed cases of logics L 1 and L 2 s.t. # p V 1 q ‰ # p V 2 q . What if # p V 1 q “ # p V 2 q , yet # p D 1 q ‰ # p D 2 q ? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  57. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Is FOUR more than enough? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  58. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  59. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J t f K truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  60. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J Syntax & Semantics : V “ t f , K , J , t u L 1 “ p V , ^ 1 , _ 1 , � 1 q t f L 2 “ p V , ^ 2 , _ 2 , � 2 q B “ p L 1 ♥L 2 , t , J , K , f q K FOUR BiLat “ CL BA truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  61. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J t f K truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  62. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J o -entailment: i ∆ iff Ű Ů ù o Γ | i r Γ u ď i i r ∆ u p -entailment: V j “ U j Y D j , and ù p Γ | j ∆ iff r Γ u� U j or r ∆ u� D j t f K truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  63. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J o -entailment: i ∆ iff Ű Ů ù o Γ | i r Γ u ď i i r ∆ u p -entailment: V j “ U j Y D j , and ù p Γ | j ∆ iff r Γ u� U j or r ∆ u� D j t f Contrast now: e ℓ U e ℓ “t f , Ku / D e ℓ “tJ , t u K truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  64. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J o -entailment: i ∆ iff Ű Ů ù o Γ | i r Γ u ď i i r ∆ u p -entailment: V j “ U j Y D j , and ù p Γ | j ∆ iff r Γ u� U j or r ∆ u� D j t f Contrast now: e ℓ U e ℓ “t f , Ku / D e ℓ “tJ , t u U n “t f , K , Ju / D n “t t u n K truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  65. b b b b b Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Variations around FDE info ( ď 2 ) J o -entailment: i ∆ iff Ű Ů ù o Γ | i r Γ u ď i i r ∆ u p -entailment: V j “ U j Y D j , and ù p Γ | j ∆ iff r Γ u� U j or r ∆ u� D j t f Contrast now: e ℓ U e ℓ “t f , Ku / D e ℓ “tJ , t u U n “t f , K , Ju / D n “t t u n b K U b “t f u / D b “tK , J , t u truth ( ď 1 ) João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  66. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  67. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u “ T ñ r α u “ T and r β u “ T [and 2 ] r α & β u “ T ð r α u “ T and r β u “ T João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  68. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u “ T ñ r α u “ T and r β u “ T [and 2 ] r α & β u “ T ð r α u “ T and r β u “ T [or 1 ] r α || β u “ T ñ r α u “ T or r β u “ T [or 2 ] r α || β u “ T ð r α u “ T or r β u “ T João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  69. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u “ T ñ r α u “ T and r β u “ T [and 2 ] r α & β u “ T ð r α u “ T and r β u “ T [or 1 ] r α || β u “ T ñ r α u “ T or r β u “ T [or 2 ] r α || β u “ T ð r α u “ T or r β u “ T João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  70. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u “ T ñ r α u “ T and r β u “ T [and 2 ] r α & β u “ F ñ r α u “ F or r β u “ F [or 1 ] r α || β u “ T ñ r α u “ T or r β u “ T [or 2 ] r α || β u “ F ñ r α u “ F and r β u “ F João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  71. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u “ T ñ r α u “ T and r β u “ T [and 2 ] r α & β u “ F ñ r α u “ F or r β u “ F [or 1 ] r α || β u “ T ñ r α u “ T or r β u “ T [or 2 ] r α || β u “ F ñ r α u “ F and r β u “ F João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  72. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u P D ñ r α u P D and r β u P D [and 2 ] r α & β u P U ñ r α u P U or r β u P U [or 1 ] r α || β u P D ñ r α u P D or r β u P D [or 2 ] r α || β u P U ñ r α u P U and r β u P U João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  73. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u P D ñ r α u P D and r β u P D [and 2 ] r α & β u P U ñ r α u P U or r β u P U [or 1 ] r α || β u P D ñ r α u P D r β u P D or [or 2 ] r α || β u P U ñ r α u P U and r β u P U CRs Operators Properties João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  74. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u P D ñ r α u P D and r β u P D [and 2 ] r α & β u P U ñ r α u P U or r β u P U [or 1 ] r α || β u P D ñ r α u P D r β u P D or [or 2 ] r α || β u P U ñ r α u P U and r β u P U CRs Operators Properties ù p ù o ù o [and 1 ], [and 2 ] | 1 , | 2 , | & P t^ 1 , ^ 2 u e ℓ ù p ù o ù o [or 1 ], [or 2 ] | 1 , | 2 , | || P t_ 1 , _ 2 u e ℓ João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  75. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u P D ñ r α u P D and r β u P D [and 2 ] r α & β u P U ñ r α u P U or r β u P U [or 1 ] r α || β u P D ñ r α u P D r β u P D or [or 2 ] r α || β u P U ñ r α u P U and r β u P U CRs Operators Properties ù p ù o ù o [and 1 ], [and 2 ] | 1 , | 2 , | & P t^ 1 , ^ 2 u e ℓ ù p ù o ù o [or 1 ], [or 2 ] | 1 , | 2 , | || P t_ 1 , _ 2 u e ℓ ù p ù p [and 1 ], [and 2 ] | b , | & “ ^ 1 n ù p ù p [or 1 ], [or 2 ] | b , | || “ _ 1 n João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  76. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! A closer look at ‘ands’ and ‘ors’ [and 1 ] r α & β u P D ñ r α u P D and r β u P D [and 2 ] r α & β u P U ñ r α u P U or r β u P U [or 1 ] r α || β u P D ñ r α u P D r β u P D or [or 2 ] r α || β u P U ñ r α u P U and r β u P U CRs Operators Properties ù p ù o ù o [and 1 ], [and 2 ] | 1 , | 2 , | & P t^ 1 , ^ 2 u e ℓ ù p ù o ù o [or 1 ], [or 2 ] | 1 , | 2 , | || P t_ 1 , _ 2 u e ℓ ù p ù p [and 1 ], [and 2 ] | b , | & “ ^ 1 n ù p ù p [or 1 ], [or 2 ] | b , | || “ _ 1 n ù p ù p [and 2 ] | b , | & “ ^ 2 n ù p ù p [or 1 ] | b , | || “ _ 2 n João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  77. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! How to find dissimilarities between these logics? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  78. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! How to find dissimilarities between these logics? Examples ù p � 2 p α _ 2 β q | n � 2 α ^ 2 � 2 β ù p � 2 p α _ 2 β q ­| b � 2 α ^ 2 � 2 β ù p α ^ 2 p β _ 2 γ q ­| n p α ^ 1 β q _ 1 p α ^ 2 γ q ù p α ^ 2 p β _ 2 γ q | b p α ^ 1 β q _ 1 p α ^ 2 γ q João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  79. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! How to find dissimilarities between these logics? Examples ù p � 2 p α _ 2 β q | n � 2 α ^ 2 � 2 β ù p � 2 p α _ 2 β q ­| b � 2 α ^ 2 � 2 β ù p α ^ 2 p β _ 2 γ q ­| n p α ^ 1 β q _ 1 p α ^ 2 γ q ù p α ^ 2 p β _ 2 γ q | b p α ^ 1 β q _ 1 p α ^ 2 γ q Note . Ingenuity is often required! João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  80. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! How to find dissimilarities between these logics? Examples ù p � 2 p α _ 2 β q | n � 2 α ^ 2 � 2 β ù p � 2 p α _ 2 β q ­| b � 2 α ^ 2 � 2 β ù p α ^ 2 p β _ 2 γ q ­| n p α ^ 1 β q _ 1 p α ^ 2 γ q ù p α ^ 2 p β _ 2 γ q | b p α ^ 1 β q _ 1 p α ^ 2 γ q Note . Ingenuity is often required! Can this task also be automated? João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  81. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Matrices in, tableaux out João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

  82. Tableaux More on finite-valued logics Another line of experimentation Through Suszko Reduction Same number of values, but other meanings for them Comparing different logics . . . and so much more! Matrices in, tableaux out Recall that: ¶ e ℓ p x q ¶ b p x q ¶ n p x q x f F F F K F T F J T T F t T T T João Marcos Classic-like Analytic Tableaux for Non-Classical Logics

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend