categorial semantics for fill proof theory of fill
play

Categorial Semantics for FILL Proof Theory of FILL: problem and - PDF document

Categorial Semantics for FILL Proof Theory of FILL: problem and solutions ( , 1 , ) is a symmetric monoidal closed structure Remember: we need comma on the right to accommodate ` A B C iff A ( B C ) iff B ( A C )


  1. Categorial Semantics for FILL Proof Theory of FILL: problem and solutions ( ⊗ , 1 , ⊸ ) is a symmetric monoidal closed structure Remember: we need comma on the right to accommodate ` A ⊗ B ⊸ C iff A ⊸ ( B ⊸ C ) iff B ⊸ ( A ⊸ C ) Problem and existing solutions: ( A ⊗ 1 ) ⊸ A and A ⊸ ( A ⊗ 1 ) multiple conclusions single conclusion existing solutions ( ` , 0 ) is a symmetric monoidal structure Γ , A ⊢ B , ∆ Γ , A ⊢ B Γ , A ⊢ B , ∆ ( † ) ( A ` B ) ⊸ ( B ` A ) Γ ⊢ A ⊸ B , ∆ Γ ⊢ A ⊸ B , ∆ Γ ⊢ A ⊸ B , ∆ ( A ` 0 ) ⊸ A and A ⊸ ( A ` 0 ) unsound no cut-elimination cut-elimination interaction via either of † : side-conditions which ensure that A is “independent” of ∆ weak distributivity ( A ⊗ ( B ` C )) ⊸ (( A ⊗ B ) ` C ) Hyland, de Paiva 1993: type assignments to ensure that the Grishin(b) (( A ⊸ B ) ` C ) ⊸ ( A ⊸ ( B ` C )) variable typed by A not appear free in the terms typed by ∆ Bierman 1996: ( a ` b ) ` c ⊢ a , (( b ` c ) ⊸ d ) ` ( e ⊸ ( d ` e )) Collapse to (classical) MLL: if we add converse of Grishin(b) has no cut-free derivation in the Hyland and de Paiva calculus Grishin(a) ( A ⊸ ( B ` C )) ⊸ (( A ⊸ B ) ` C ) Display calculus for (an extension of) FILL Logical rules: introduced formula is always displayed X ⊢ A A ⊢ Y Structural Constant and Binary Connectives: Φ (id) p ⊢ p (cut) , < > X ⊢ Y Antecedent Structure: X a Y a ::= A | Φ | X a , Y a | X a < Y s Φ ⊢ X ( 1 ⊢ ) ( ⊢ 1 ) Φ ⊢ 1 Succcedent Structure: X s Y s ::= A | Φ | X s , Y s | X a > Y s 1 ⊢ X X ⊢ Φ Sequent: X a ⊢ Y s (drop subscripts to avoid clutter) ( ⊢ 0 ) ( 0 ⊢ ) 0 ⊢ Φ X ⊢ 0 Display Postulates: reversible structural rules A , B ⊢ X X ⊢ A Y ⊢ B ( ⊢ ⊗ ) ( ⊗ ⊢ ) A ⊗ B ⊢ X X , Y ⊢ A ⊗ B X a ⊢ Y a > Z s Z a < Y s ⊢ X s X ⊢ A , B A ⊢ X B ⊢ Y ( ` ⊢ ) ( ⊢ ` ) X a , Y a ⊢ Z s Z a ⊢ X s , Y s A ` B ⊢ X , Y X ⊢ A ` B Y a ⊢ X a > Z s Z a < X s ⊢ Y s X ⊢ A > B X ⊢ A B ⊢ Y ( ⊸ ⊢ ) ( ⊢ ⊸ ) A ⊸ B ⊢ X > Y X ⊢ A ⊸ B Display Property: For every antecedent (succedent) part Z of the A < B ⊢ X X ⊢ A B ⊢ Y ( − < ⊢ ) ( ⊢ − < ) sequent X ⊢ Y , there is a sequent Z ⊢ Y ′ (resp. X ′ ⊢ Z ) A − < B ⊢ X X < Y ⊢ A − < B obtainable from X ⊢ Y using only the display postulates, read upwards, one rule is a “rewrite” while other “constrains” thereby displaying the Z as the whole of one side Structural rules: no occurrences of formula meta-variables Categorial semantics for bi-intuitionistic linear logic BiILL ( ⊗ , 1 , ⊸ ) is a symmetric monoidal closed structure all sub-structural properties captured in a modular way A ⊗ B ⊸ C iff A ⊸ ( B ⊸ C ) iff B ⊸ ( A ⊸ C ) ( A ⊗ 1 ) ⊸ A and A ⊸ ( A ⊗ 1 ) X , Φ ⊢ Y X ⊢ Φ , Y (Φ ⊢ ) ( ⊢ Φ) X ⊢ Y X ⊢ Y ( − <, ` , 0 ) is a symmetric monoidal co-closed structure W , ( X , Y ) ⊢ Z W ⊢ ( X , Y ) , Z A ⊸ ( B ` C ) iff ( A − < B ) ⊸ C iff ( A − < C ) ⊸ B (Ass ⊢ ) ( ⊢ Ass) ( W , X ) , Y ⊢ Z W ⊢ X , ( Y , Z ) ( A ` 0 ) ⊸ A and A ⊸ ( A ` 0 ) X , Y ⊢ Z Z ⊢ Y , X (Com ⊢ ) ( ⊢ Com) Y , X ⊢ Z Z ⊢ X , Y interaction via either of W , ( X < Y ) ⊢ Z W ⊢ ( X > Y ) , Z Grishin(b) (( A ⊸ B ) ` C ) ⊸ ( A ⊸ ( B ` C )) (Grnb ⊢ ) ( W , X ) < Y ⊢ Z ( ⊢ Grnb) W ⊢ X > ( Y , Z ) dualGrishin(b) (( A ⊗ B ) − < C ) ⊸ ( A ⊗ ( B − < C )) (( A ⊸ B ) ` C ) ⊸ ( A ⊸ ( B ` C )) Collapse to (classical) MLL: if we add converse of either Soundness, completeness and cut-elimination From BiILL back to FILL Problem: Nice Display Calculus for BiILL ... is it sound for FILL ? Display calculus: must create antecedent < structures in its Thm: The sequent X ⊢ Y is derivable iff the formula-translation derivation of FILL -formulae in order to display and undisplay; τ a ( X ) ⊸ τ s ( Y ) is BiILL -valid and < is structural equivalent to − < , not in FILL Proof: the display calculus proof rules and the arrows of the free Question: is BiILL a conservative extension of FILL (that is, are BiILL-category are inter-definable. BiILL -derivable FILL -formulae FILL -derivable? we were not able to find a categorial proof Compare: to tense logic Kt say where there is a simple semantic Thm: If X ⊢ Y is derivable then it is cut-free derivable. proof that Kt is a conservative extension of K (same frames) Proof: The rules obey conditions C1-C8 given by Belnap (1982), hence the calculus enjoys cut-admissibility FILL BiILL BiILL dc BiILL dc 1 2 3 = ⇒ ⇐ ⇒ ⇐ ⇒ category category with cut no cut So we have a Display Calculus for BiILL ... is it sound for FILL ? � � � 7 � 4 � � 6 5 FILL dn ← − BiILL dn ⇐ ⇒ BiILL sn

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