TermsforClassical Sequents Proof Invariants & Strong - - PowerPoint PPT Presentation

termsforclassical sequents
SMART_READER_LITE
LIVE PREVIEW

TermsforClassical Sequents Proof Invariants & Strong - - PowerPoint PPT Presentation

TermsforClassical Sequents Proof Invariants & Strong Normalisation Greg Restall melbourne logic seminar 6 may 2016 My Aim To introduce a new invariant for classical propositional proofs and to show how they can be used. Greg Restall


slide-1
SLIDE 1

TermsforClassical Sequents

Proof Invariants & Strong Normalisation Greg Restall

melbourne logic seminar · 6 may 2016

slide-2
SLIDE 2

My Aim

To introduce a new invariant for classical propositional proofs and to show how they can be used.

Greg Restall Terms for Classical Sequents 2 of 66

slide-3
SLIDE 3

Today's Plan

Background Preterms Derivations Terms Eliminating Cuts Strong Normalisation Further Work

Greg Restall Terms for Classical Sequents 3 of 66

slide-4
SLIDE 4

background

slide-5
SLIDE 5

When is π1 the same proof as π2?

p p

∨R

p p ∨ q

∧L

p ∧ q p ∨ q p ∧ q

∧E

p

∨I

p ∨ q p p

∧L

p ∧ q p

∨R

p ∧ q p ∨ q

R L E I L R

Greg Restall Terms for Classical Sequents 5 of 66

slide-6
SLIDE 6

When is π1 the same proof as π2?

p p

∨R

p p ∨ q

∧L

p ∧ q p ∨ q p ∧ q

∧E

p

∨I

p ∨ q p p

∧L

p ∧ q p

∨R

p ∧ q p ∨ q q q

∨R

q p ∨ q

∧L

p ∧ q p ∨ q p ∧ q

∧E

q

∨I

p ∨ q q q

∧L

p ∧ q q

∨R

p ∧ q p ∨ q

Greg Restall Terms for Classical Sequents 5 of 66

slide-7
SLIDE 7

When is π1 the same proof as π2?

p ∨ q [p]1

∨I

q ∨ p [q]1

∨I

q ∨ p

∨E1

q ∨ p

∨I

(q ∨ p) ∨ r p ∨ q [p]1

∨I

q ∨ p

∨I

(q ∨ p) ∨ r [q]1

∨I

q ∨ p

∨I

(q ∨ p) ∨ r

∨E1

(q ∨ p) ∨ r Are these different proofs, or different ways of presenting the same proof ?

Greg Restall Terms for Classical Sequents 6 of 66

slide-8
SLIDE 8

Girard, Lafont and Taylor: ProofsandTypes, Chapter 2

Natural deduction is a slightly paradoxical system: it is limited to the intuitionistic case (in the classical case it has no particularly good properties) but it is only satisfactory for the (∧, ⇒, ∀) fragment of the language: we shall defer consideration of ∨ and ∃ until chapter 10. Yet disjunction and existence are the two most typically intuitionistic connectors! The basic idea of natural deduction is an asymmetry: a proof is a vaguely tree-like structure (this view is more a graphical illusion than a mathematical reality, but it is a pleasant illusion) with one or more hypotheses (possibly none) but a single conclusion. The deep symmetry of the calculus is shown by the introduction and elimination rules which match each other exactly. Observe, incidentally, that with a tree-like structure, one can always decide uniquely what was the last rule used, which is something we could not say if there were several conclusions.

Greg Restall Terms for Classical Sequents 7 of 66

slide-9
SLIDE 9

LambdaTerms and Proofs

[x : p ⊃ (q ⊃ r)] [z : p]

⊃E

xz : q ⊃ r [y : p ⊃ q] [z : p]

⊃E

yz : q

⊃E

(xz)(yz) : r

⊃I

λz (xz)(yz) : p ⊃ r

⊃I

λyλz (xz)(yz) : (p ⊃ q) ⊃ (p ⊃ r)

⊃I

λxλyλz (xz)(yz) : (p ⊃ (q ⊃ r)) ⊃ ((p ⊃ q) ⊃ (p ⊃ r))

Greg Restall Terms for Classical Sequents 8 of 66

slide-10
SLIDE 10

Contraction and weakening are managed by variables

[x : p]

⊃I

λy x : q ⊃ p

⊃I

λxλy x : p ⊃ (q ⊃ p)

E E I

Greg Restall Terms for Classical Sequents 9 of 66

slide-11
SLIDE 11

Contraction and weakening are managed by variables

[x : p]

⊃I

λy x : q ⊃ p

⊃I

λxλy x : p ⊃ (q ⊃ p) x : p ⊃ (p ⊃ q) [y : p]

⊃E

xy : p ⊃ q [y : p]

⊃E

(xy)y : q

⊃I

λy (xy)y : p ⊃ q

Greg Restall Terms for Classical Sequents 9 of 66

slide-12
SLIDE 12

Classical Sequent Derivations

p p

¬R

p, ¬p

∨R

p ∨ ¬p p p

¬L

p, ¬p

∧L

p ∧ ¬p

L R L R

Greg Restall Terms for Classical Sequents 10 of 66

slide-13
SLIDE 13

Classical Sequent Derivations

p p

¬R

p, ¬p

∨R

p ∨ ¬p p p

¬L

p, ¬p

∧L

p ∧ ¬p p p q q r r

∨L

q ∨ r q, r

∧R

p, q ∨ r p ∧ q, r

∧L

p ∧ (q ∨ r) p ∧ q, r

∨R

p ∧ (q ∨ r) (p ∧ q) ∨ r

Greg Restall Terms for Classical Sequents 10 of 66

slide-14
SLIDE 14

Sequents and Terms X Y X A, Y X, A Y

Where do you put the variables, and where do you put the terms?

Greg Restall Terms for Classical Sequents 11 of 66

slide-15
SLIDE 15

Our Choice x1 : A1, . . . , xn : An y1 : B1, . . . , ym : Bm

Each premise and conclusion is decorated with variables. The sequent gets the term, connecting inputs & outputs.

Greg Restall Terms for Classical Sequents 12 of 66

slide-16
SLIDE 16

Our Choice x1 : A1, . . . , xn : An y1 : B1, . . . , ym : Bm

Each premise and conclusion is decorated with variables. The sequent gets the term, connecting inputs & outputs.

Greg Restall Terms for Classical Sequents 12 of 66

slide-17
SLIDE 17

Our Choice

π(x1, . . . , xn)[y1, . . . , ym]

x1 : A1, . . . , xn : An y1 : B1, . . . , ym : Bm

Each premise and conclusion is decorated with variables. The sequent gets the term, connecting inputs & outputs.

Greg Restall Terms for Classical Sequents 12 of 66

slide-18
SLIDE 18

Example 1

x⌢x

x : p x : p

y⌢y

y : q y : q

z⌢z

z : r z : r

∨L

Lw⌢y Rw⌢z

w : q ∨ r y : q, z : r

∧R

x⌢Fv Lw⌢Sv Rw⌢z

x : p, w : q ∨ r v : p ∧ q, z : r

∧L

Fu⌢Fv LSu⌢Sv RSu⌢z

u : p ∧ (q ∨ r) v : p ∧ q, z : r

∨R

Fu⌢FLt LSu⌢SLt RSu⌢Rt

u : p ∧ (q ∨ r) t : (p ∧ q) ∨ r

Greg Restall Terms for Classical Sequents 13 of 66

slide-19
SLIDE 19

Example 2

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fy x⌢Sy

x : p y : p ∧ p

z⌢z

z : p z : p

∧L

Fw⌢z

w : p ∧ p z : p

Cut

x⌢F• x⌢S• F•⌢z

x : p z : p

Greg Restall Terms for Classical Sequents 14 of 66

slide-20
SLIDE 20

preterms

slide-21
SLIDE 21

Variables and Cut Points

▶ For each formula A, xA 1 , xA 2 , . . . are variables of type A.

For each formula , , , are cut points of type . – We use ; as schematic letters for variables and cut points, ommitting type superscripts where possible.

Greg Restall Terms for Classical Sequents 16 of 66

slide-22
SLIDE 22

Variables and Cut Points

▶ For each formula A, xA 1 , xA 2 , . . . are variables of type A. ▶ For each formula A, •A 1 , •A 2 , . . . are cut points of type A.

– We use ; as schematic letters for variables and cut points, ommitting type superscripts where possible.

Greg Restall Terms for Classical Sequents 16 of 66

slide-23
SLIDE 23

Variables and Cut Points

▶ For each formula A, xA 1 , xA 2 , . . . are variables of type A. ▶ For each formula A, •A 1 , •A 2 , . . . are cut points of type A.

– We use x, y, z, u, v, w, . . .; •, ⋆, ∗, ♯, ♭ as schematic letters for variables and cut points, ommitting type superscripts where possible.

Greg Restall Terms for Classical Sequents 16 of 66

slide-24
SLIDE 24

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes.

If is an node, then L is an node and R is a node. If is an node, then F is an node and S is a node. If is an node, then A is an node and C is a node. If is a node, then N is an node. For each complex node L , R , F , S , A , C and N , is its immediate subnode, and the subnodes of are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-25
SLIDE 25

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes. ▶ If n is an A ∧ B node, then Ln is an A node and Rn is a B node.

If is an node, then F is an node and S is a node. If is an node, then A is an node and C is a node. If is a node, then N is an node. For each complex node L , R , F , S , A , C and N , is its immediate subnode, and the subnodes of are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-26
SLIDE 26

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes. ▶ If n is an A ∧ B node, then Ln is an A node and Rn is a B node. ▶ If n is an A ∨ B node, then Fn is an A node and Sn is a B node.

If is an node, then A is an node and C is a node. If is a node, then N is an node. For each complex node L , R , F , S , A , C and N , is its immediate subnode, and the subnodes of are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-27
SLIDE 27

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes. ▶ If n is an A ∧ B node, then Ln is an A node and Rn is a B node. ▶ If n is an A ∨ B node, then Fn is an A node and Sn is a B node. ▶ If n is an A ⊃ B node, then An is an A node and Cn is a B node.

If is a node, then N is an node. For each complex node L , R , F , S , A , C and N , is its immediate subnode, and the subnodes of are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-28
SLIDE 28

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes. ▶ If n is an A ∧ B node, then Ln is an A node and Rn is a B node. ▶ If n is an A ∨ B node, then Fn is an A node and Sn is a B node. ▶ If n is an A ⊃ B node, then An is an A node and Cn is a B node. ▶ If n is a ¬A node, then Nn is an A node.

For each complex node L , R , F , S , A , C and N , is its immediate subnode, and the subnodes of are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-29
SLIDE 29

Nodes and Subnodes

▶ A variable x of type A and a cut point • of type A are both A nodes. ▶ If n is an A ∧ B node, then Ln is an A node and Rn is a B node. ▶ If n is an A ∨ B node, then Fn is an A node and Sn is a B node. ▶ If n is an A ⊃ B node, then An is an A node and Cn is a B node. ▶ If n is a ¬A node, then Nn is an A node. ▶ For each complex node Ln, Rn, Fn, Sn, An, Cn and Nn, n is its immediate

subnode, and the subnodes of n are also subnodes of the original node.

Greg Restall Terms for Classical Sequents 17 of 66

slide-30
SLIDE 30

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type.

In , is in input position, and is in output position. Positions generalise to subnodes as follows:

If L , R , F , S

  • r C

are in input position, is also in input position. If L , R , F , S

  • r C

are in output position, is also in output position. – L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-31
SLIDE 31

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position.

Positions generalise to subnodes as follows:

If L , R , F , S

  • r C

are in input position, is also in input position. If L , R , F , S

  • r C

are in output position, is also in output position. – L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-32
SLIDE 32

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

If L , R , F , S

  • r C

are in input position, is also in input position. If L , R , F , S

  • r C

are in output position, is also in output position. – L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-33
SLIDE 33

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position.

If L , R , F , S

  • r C

are in output position, is also in output position. – L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-34
SLIDE 34

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-35
SLIDE 35

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position. If A

  • r N

is in input position, is in output position. If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-36
SLIDE 36

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position.

▶ If An or Nn is in input position, n is in output position.

If A

  • r N

is in output position, is in input position. – A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-37
SLIDE 37

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position.

▶ If An or Nn is in input position, n is in output position. ▶ If An or Nn is in output position, n is in input position.

– A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-38
SLIDE 38

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position.

▶ If An or Nn is in input position, n is in output position. ▶ If An or Nn is in output position, n is in input position.

– A and N reverse position.

The inputs (outputs) of a linking are the variables in input (output) position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-39
SLIDE 39

Linkings, Inputs and Outputs

▶ A linking is a pair n⌢m of nodes of the same type. ▶ In n⌢m, n is in input position, and m is in output position. ▶ Positions generalise to subnodes as follows:

▶ If Ln, Rn, Fn, Sn or Cn are in input position, n is also in input position. ▶ If Ln, Rn, Fn, Sn or Cn are in output position, n is also in output position.

– L, R, F, S and C each preserve position.

▶ If An or Nn is in input position, n is in output position. ▶ If An or Nn is in output position, n is in input position.

– A and N reverse position.

▶ The inputs (outputs) of a linking are the variables in input (output)

position of that linking.

Greg Restall Terms for Classical Sequents 18 of 66

slide-40
SLIDE 40

Example Linkings x of type ((p ⊃ q) ⊃ p) ⊃ p AAA C CA C

Greg Restall Terms for Classical Sequents 19 of 66

slide-41
SLIDE 41

Example Linkings x of type ((p ⊃ q) ⊃ p) ⊃ p AAAx⌢Cx CA C

Greg Restall Terms for Classical Sequents 19 of 66

slide-42
SLIDE 42

Example Linkings x of type ((p ⊃ q) ⊃ p) ⊃ p AAAx⌢Cx CAx⌢Cx

Greg Restall Terms for Classical Sequents 19 of 66

slide-43
SLIDE 43

Preterms

▶ A preterm is a finite set of linkings.

The inputs of a preterm are the inputs of its linkings. Its outputs are the outputs of its linkings.

Greg Restall Terms for Classical Sequents 20 of 66

slide-44
SLIDE 44

Preterms

▶ A preterm is a finite set of linkings. ▶ The inputs of a preterm are the inputs of its linkings.

Its outputs are the outputs of its linkings.

Greg Restall Terms for Classical Sequents 20 of 66

slide-45
SLIDE 45

Preterms

▶ A preterm is a finite set of linkings. ▶ The inputs of a preterm are the inputs of its linkings. ▶ Its outputs are the outputs of its linkings.

Greg Restall Terms for Classical Sequents 20 of 66

slide-46
SLIDE 46

derivations

slide-47
SLIDE 47

Annotating Derivations: Identity

x⌢y

Σ, x : A y : A, ∆

Greg Restall Terms for Classical Sequents 22 of 66

slide-48
SLIDE 48

Annotating Derivations: Conjunction

π(x, y)

Σ, x : A, y : B ∆

∧L

π(Fz, Sz)

Σ, z : A ∧ B ∆

R

F S

Greg Restall Terms for Classical Sequents 23 of 66

slide-49
SLIDE 49

Annotating Derivations: Conjunction

π(x, y)

Σ, x : A, y : B ∆

∧L

π(Fz, Sz)

Σ, z : A ∧ B ∆

π[x]

Σ x : A, ∆

π′[y]

Σ′ y : B, ∆′

∧R

π[Fz] π′[Sz]

Σ, Σ′ z : A ∧ B, ∆, ∆′

Greg Restall Terms for Classical Sequents 23 of 66

slide-50
SLIDE 50

Excursus on Weakening and Variables

[x : p]

⊃I

λy x : q ⊃ p

⊃I

λxλy x : p ⊃ (q ⊃ p)

L

F S

can be

L

F

In a premise the indicated and display all

  • f the

and inputs to the proof term. There might be none.

Greg Restall Terms for Classical Sequents 24 of 66

slide-51
SLIDE 51

Excursus on Weakening and Variables

[x : p]

⊃I

λy x : q ⊃ p

⊃I

λxλy x : p ⊃ (q ⊃ p)

π(x, y)

Σ, x : A, y : B ∆

∧L

π(Fz, Sz)

Σ, z : A ∧ B ∆ can be

π(x)

Σ, x : A ∆

∧L

π(Fz)

Σ, z : A ∧ B ∆ In a premise the indicated and display all

  • f the

and inputs to the proof term. There might be none.

Greg Restall Terms for Classical Sequents 24 of 66

slide-52
SLIDE 52

Excursus on Weakening and Variables

[x : p]

⊃I

λy x : q ⊃ p

⊃I

λxλy x : p ⊃ (q ⊃ p)

π(x, y)

Σ, x : A, y : B ∆

∧L

π(Fz, Sz)

Σ, z : A ∧ B ∆ can be

π(x)

Σ, x : A ∆

∧L

π(Fz)

Σ, z : A ∧ B ∆ In a premise π(x, y) the indicated x and y display all

  • f the x and y inputs to the proof term.

There might be none.

Greg Restall Terms for Classical Sequents 24 of 66

slide-53
SLIDE 53

Annotating Derivations: Negation

π[x]

Σ x : A, ∆

¬L

π[Nz]

Σ, z : ¬A ∆

π(x)

Σ, x : A ∆

¬R

π(Nz)

Σ z : ¬A, ∆

Greg Restall Terms for Classical Sequents 25 of 66

slide-54
SLIDE 54

Annotating Derivations: Disjunction

π(x)

Σ, x : A ∆

π′(y)

Σ′, y : B ∆′

∨L

π(Lz) π′(Rz)

Σ, Σ′, z : A ∨ B ∆, ∆′

π[x, y]

Σ x : A, y : B, ∆

∨R

π[Lz, Rz]

Σ z : A ∨ B, ∆

Greg Restall Terms for Classical Sequents 26 of 66

slide-55
SLIDE 55

Annotating Derivations: Conditional

π[x]

Σ x : A, ∆

π′(y)

Σ′, y : B ∆′

⊃L

π[Az] π′(Lz)

Σ, Σ′, z : A ⊃ B ∆, ∆′

π(x)[y]

Σ, x : A y : B, ∆

⊃R

π(Az)[Cz]

Σ z : A ⊃ B, ∆

Greg Restall Terms for Classical Sequents 27 of 66

slide-56
SLIDE 56

Example Annotation

x⌢x

x : p x : p

y⌢y

y : q y : q

z⌢z

z : r z : r

∨L

Lw⌢y Rw⌢z

w : q ∨ r y : q, z : r

∧R

x⌢Fv Lw⌢Sv Rw⌢z

x : p, w : q ∨ r v : p ∧ q, z : r

∧L

Fu⌢Fv LSu⌢Sv RSu⌢z

u : p ∧ (q ∨ r) v : p ∧ q, z : r

∨R

Fu⌢FLt LSu⌢SLt RSu⌢Rt

u : p ∧ (q ∨ r) t : (p ∧ q) ∨ r

Greg Restall Terms for Classical Sequents 28 of 66

slide-57
SLIDE 57

Annotating Derivations: Cut

π[x]

Σ x : A, ∆

π′(y)

Σ′, y : A ∆′

Cut

π[•] π′(•)

Σ, Σ′ ∆, ∆′

Greg Restall Terms for Classical Sequents 29 of 66

slide-58
SLIDE 58

Example Annotation, with Cut

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

Cut

Ly⌢• Ry⌢•

  • ⌢Fz
  • ⌢Sz

y : p ∨ p z : p ∧ p

Greg Restall Terms for Classical Sequents 30 of 66

slide-59
SLIDE 59

When is π1 the same proof as π2 (revisited)?

z⌢z

z : p z : p

∨R

z⌢Ly

z : p y : p ∨ q

∧L

Fx⌢Ly

x : p ∧ q y : p ∨ q p ∧ q

∧E

p

∨I

p ∨ q

z⌢z

z : p z : p

∨R

Fx⌢z

x : p ∧ q z : p

∧L

Fx⌢Ly

x : p ∧ q y : p ∨ q

R

R

L

S R

E I R

S

L

S R

Greg Restall Terms for Classical Sequents 31 of 66

slide-60
SLIDE 60

When is π1 the same proof as π2 (revisited)?

z⌢z

z : p z : p

∨R

z⌢Ly

z : p y : p ∨ q

∧L

Fx⌢Ly

x : p ∧ q y : p ∨ q p ∧ q

∧E

p

∨I

p ∨ q

z⌢z

z : p z : p

∨R

Fx⌢z

x : p ∧ q z : p

∧L

Fx⌢Ly

x : p ∧ q y : p ∨ q

w⌢w

w : q w : q

∨R

w⌢Ry

w : q y : p ∨ q

∧L

Sx⌢Ry

x : p ∧ q y : p ∨ q p ∧ q

∧E

q

∨I

p ∨ q

w⌢w

w : q w : q

∨R

Sx⌢w

x : p ∧ q w : q

∧L

Sx⌢Ry

x : p ∧ q y : p ∨ q

Greg Restall Terms for Classical Sequents 31 of 66

slide-61
SLIDE 61

When is π1 the same proof as π2 (revisited)?

p ∨ q [p]1

∨I

q ∨ p [q]1

∨I

q ∨ p

∨E1

q ∨ p

∨I

(q ∨ p) ∨ r

x⌢x

x : p x : p

∨R

x⌢Rz

x : p z : q ∨ p

y⌢y

y : q y : q

∨R

y⌢Lz

y : q z : q ∨ p

∨L

Lw⌢Rz Rw⌢Lz

w : p ∨ q z : q ∨ p

∨R

Lw⌢RLu Rw⌢LLu

w : p ∨ q u : (q ∨ p) ∨ r

Greg Restall Terms for Classical Sequents 32 of 66

slide-62
SLIDE 62

When is π1 the same proof as π2 (revisited)?

p ∨ q [p]1

∨I

q ∨ p

∨I

(q ∨ p) ∨ r [q]1

∨I

q ∨ p

∨I

(q ∨ p) ∨ r

∨E1

(q ∨ p) ∨ r

x⌢x

x : p x : p

∨R

x⌢Rz

x : p z : q ∨ p

∨R

x⌢RLu

x : p u : (q ∨ p) ∨ r

y⌢y

y : q y : q

∨R

x⌢LLu

y : q z : q ∨ p

∨R

y⌢LLu

y : q u : (q ∨ p) ∨ r

∨L

Lw⌢RLu Rw⌢LLu

w : p ∨ q u : (q ∨ p) ∨ r

Greg Restall Terms for Classical Sequents 33 of 66

slide-63
SLIDE 63

Sequentialisable Preterms

A preterm is sequentialisable iff it is the conclusion of some derivation.

Greg Restall Terms for Classical Sequents 34 of 66

slide-64
SLIDE 64

terms

slide-65
SLIDE 65

Nonsequentialisable Preterms

Lx⌢Fy Rx⌢Sy

x : p ∨ q y : p ∧ q

This is connected, but it is not connected enough.

Greg Restall Terms for Classical Sequents 36 of 66

slide-66
SLIDE 66

Switching Example

Lx⌢Fy Rx⌢Sy

x : p ∨ q y : p ∧ q

L F R S L F R S L F R S L F R S

Greg Restall Terms for Classical Sequents 37 of 66

slide-67
SLIDE 67

Switching Example

Lx⌢Fy Rx⌢Sy

x : p ∨ q y : p ∧ q

Lx⌢Fy ✚

Rx⌢✚

Sy

x : p ∨ − y : p ∧ −

Lx⌢

  • Fy ✚

Rx⌢Sy

x : p ∨ − y : − ∧ q

Lx⌢Fy Rx⌢✚

Sy

x : − ∨ q y : p ∧ −

Lx⌢

  • Fy

Rx⌢Sy

x : − ∨ q y : − ∧ q

Greg Restall Terms for Classical Sequents 37 of 66

slide-68
SLIDE 68

Switchings

▶ The switchings of a preterm π are found by selecting for each pair of

subterms Ln and Rn in input position; Fn and Sn in output position, An in

  • utput position and Cn in input position; or the cut point • (in both input

and output position), one item of the pair to keep, and the other to delete. A linking in a switching of a preterm survives if and only if neither side of the link involves a deletion. A preterm is spanned if every switching has at least one surviving linking.

Greg Restall Terms for Classical Sequents 38 of 66

slide-69
SLIDE 69

Switchings

▶ The switchings of a preterm π are found by selecting for each pair of

subterms Ln and Rn in input position; Fn and Sn in output position, An in

  • utput position and Cn in input position; or the cut point • (in both input

and output position), one item of the pair to keep, and the other to delete.

▶ A linking in a switching of a preterm π survives if and only if neither

side of the link involves a deletion. A preterm is spanned if every switching has at least one surviving linking.

Greg Restall Terms for Classical Sequents 38 of 66

slide-70
SLIDE 70

Switchings

▶ The switchings of a preterm π are found by selecting for each pair of

subterms Ln and Rn in input position; Fn and Sn in output position, An in

  • utput position and Cn in input position; or the cut point • (in both input

and output position), one item of the pair to keep, and the other to delete.

▶ A linking in a switching of a preterm π survives if and only if neither

side of the link involves a deletion.

▶ A preterm is spanned if every switching has at least one surviving linking.

Greg Restall Terms for Classical Sequents 38 of 66

slide-71
SLIDE 71

Example

Fu⌢FLt LSu⌢SLt RSu⌢Rt This has two pairs for switching: LSu/RSu in input position. FLt/SLt in output position. F FL LS SL RS R F FL LS SL RS R F FL LS SL RS R F FL LS SL RS R

Greg Restall Terms for Classical Sequents 39 of 66

slide-72
SLIDE 72

Example

Fu⌢FLt LSu⌢SLt RSu⌢Rt This has two pairs for switching: LSu/RSu in input position. FLt/SLt in output position. Fu⌢✟

FLt ✟✟

LSu⌢SLt RSu⌢Rt Fu⌢FLt ✟✟

LSu⌢✟

SLt RSu⌢Rt Fu⌢✟

FLt LSu⌢SLt ✘✘

RSu⌢Rt Fu⌢FLt LSu⌢✟

SLt ✘✘

RSu⌢Rt

Greg Restall Terms for Classical Sequents 39 of 66

slide-73
SLIDE 73

Terms A preterm π is a term when it is spanned.

Greg Restall Terms for Classical Sequents 40 of 66

slide-74
SLIDE 74

Sequentialisable Preterms are Terms By induction on the derivation sequentialising π.

Greg Restall Terms for Classical Sequents 41 of 66

slide-75
SLIDE 75

Sequentialisable Preterms are Terms: Identity

x⌢y

Σ, x : A y : A, ∆

Greg Restall Terms for Classical Sequents 42 of 66

slide-76
SLIDE 76

Sequentialisable Preterms are Terms: Conjunction

π(x, y)

Σ, x : A, y : B ∆

∧L

π(Fz, Sz)

Σ, z : A ∧ B ∆

π[x]

Σ x : A, ∆

π′[y]

Σ′ y : B, ∆′

∧R

π[Fz] π′[Sz]

Σ, Σ′ z : A ∧ B, ∆, ∆′

Greg Restall Terms for Classical Sequents 43 of 66

slide-77
SLIDE 77

Sequentialisable Preterms are Terms: Negation

π[x]

Σ x : A, ∆

¬L

π[Nz]

Σ, z : ¬A ∆

π(x)

Σ, x : A ∆

¬R

π(Nz)

Σ z : ¬A, ∆

Greg Restall Terms for Classical Sequents 44 of 66

slide-78
SLIDE 78

Sequentialisable Preterms are Terms: Disjunction

π(x)

Σ, x : A ∆

π′(y)

Σ′, y : B ∆′

∨L

π(Lz) π′(Rz)

Σ, Σ′, z : A ∨ B ∆, ∆′

π[x, y]

Σ x : A, y : B, ∆

∨R

π[Lz, Rz]

Σ z : A ∨ B, ∆

Greg Restall Terms for Classical Sequents 45 of 66

slide-79
SLIDE 79

Sequentialisable Preterms are Terms: Conditional

π[x]

Σ x : A, ∆

π′(y)

Σ′, y : B ∆′

⊃L

π[Az] π′(Cz)

Σ, Σ′, z : A ⊃ B ∆, ∆′

π(x)[y]

Σ, x : A y : B, ∆

⊃R

π(Az)[Cz]

Σ z : A ⊃ B, ∆

Greg Restall Terms for Classical Sequents 46 of 66

slide-80
SLIDE 80

Sequentialisable Preterms are Terms: Cut

π[x]

Σ x : A, ∆

π′(y)

Σ′, y : A ∆′

Cut

π[•] π′(•)

Σ, Σ′ ∆, ∆′

Greg Restall Terms for Classical Sequents 47 of 66

slide-81
SLIDE 81

Terms are Sequentialisable By induction on the number of pairs for switching in π. Except … x⌢y u⌢v

Greg Restall Terms for Classical Sequents 48 of 66

slide-82
SLIDE 82

eliminating cuts

slide-83
SLIDE 83

Conjunction Cut Reduction

π[x]

Σ x : A, ∆

π′[y]

Σ′ y : B, ∆

∧R

π[Fz] π′[Sz]

Σ, Σ′ z : A ∧ B, ∆, ∆

π′′(u, v)

Σ′′, u : A, v : B ∆′′

∧L

π′′(Fw, Sw)

Σ′′, w : A ∧ B ∆′′

Cut

π[F•] π′[S•] π′′(F•, S•)

Σ, Σ′, Σ′′ ∆, ∆′, ∆′′ reduces to

Cut Cut

Greg Restall Terms for Classical Sequents 50 of 66

slide-84
SLIDE 84

Conjunction Cut Reduction

π[x]

Σ x : A, ∆

π′[y]

Σ′ y : B, ∆

∧R

π[Fz] π′[Sz]

Σ, Σ′ z : A ∧ B, ∆, ∆

π′′(u, v)

Σ′′, u : A, v : B ∆′′

∧L

π′′(Fw, Sw)

Σ′′, w : A ∧ B ∆′′

Cut

π[F•] π′[S•] π′′(F•, S•)

Σ, Σ′, Σ′′ ∆, ∆′, ∆′′ reduces to

π[x]

Σ x : A, ∆

π′[y]

Σ′ y : B, ∆

π′′(u, v)

Σ′′, u : A, v : B ∆′′

Cut

π′[⋆] π′′(u, ⋆)

Σ′, Σ′′, u : A ∆′, ∆′′

Cut

π[∗] π′[⋆] π′′(∗, ⋆)

Σ, Σ′, Σ′′ ∆, ∆′, ∆′′

Greg Restall Terms for Classical Sequents 50 of 66

slide-85
SLIDE 85

Identity Cut Reduction

π[x]

Σ x : A, ∆

y⌢z

Σ′, y : A z : A, ∆′

Cut

π[•] •⌢z

Σ, Σ′ z : A, ∆, ∆′ reduces to

Greg Restall Terms for Classical Sequents 51 of 66

slide-86
SLIDE 86

Identity Cut Reduction

π[x]

Σ x : A, ∆

y⌢z

Σ′, y : A z : A, ∆′

Cut

π[•] •⌢z

Σ, Σ′ z : A, ∆, ∆′ reduces to

π[z]

Σ, Σ′ z : A, ∆, ∆′

Greg Restall Terms for Classical Sequents 51 of 66

slide-87
SLIDE 87

Difficult Cases: Contraction

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

Cut

Ly⌢• Ry⌢• •⌢Fz •⌢Sz

y : p ∨ p z : p ∧ p

L

L R

L

L R

R

L F R F L S R S

Greg Restall Terms for Classical Sequents 52 of 66

slide-88
SLIDE 88

Difficult Cases: Contraction

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

Cut

Ly⌢• Ry⌢• •⌢Fz •⌢Sz

y : p ∨ p z : p ∧ p

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

∧R

Ly⌢Fz Ry⌢Fz Ly⌢Sz Ry⌢Sz

y : p ∨ p z : p ∧ p

Greg Restall Terms for Classical Sequents 52 of 66

slide-89
SLIDE 89

Difficult Cases: Contraction

x⌢x

x : p x : p

x⌢x

x : p x : p

∨L

Ly⌢x Ry⌢x

y : p ∨ p x : p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

Cut

Ly⌢• Ry⌢• •⌢Fz •⌢Sz

y : p ∨ p z : p ∧ p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

x⌢x

x : p x : p

x⌢x

x : p x : p

∧R

x⌢Fz x⌢Sz

x : p z : p ∧ p

∨L

Ly⌢Fz Ry⌢Fz Ly⌢Sz Ry⌢Sz

y : p ∨ p z : p ∧ p

Greg Restall Terms for Classical Sequents 53 of 66

slide-90
SLIDE 90

Difficult Cases: Weakening

π

Σ ∆

π

Σ x : A, ∆

π′

Σ ∆

π′

Σ, y : A ∆

Cut

π π′

Σ ∆

Mix

Greg Restall Terms for Classical Sequents 54 of 66

slide-91
SLIDE 91

Difficult Cases: Weakening

π

Σ ∆

π

Σ x : A, ∆

π′

Σ ∆

π′

Σ, y : A ∆

Cut

π π′

Σ ∆

π

Σ ∆

π′

Σ ∆

Mix

π π′

Σ ∆

Greg Restall Terms for Classical Sequents 54 of 66

slide-92
SLIDE 92

Back to Sequentialisation

x⌢y u⌢v

Mix

Greg Restall Terms for Classical Sequents 55 of 66

slide-93
SLIDE 93

Back to Sequentialisation

x⌢y u⌢v

x⌢y

x : A y : A

u⌢v

u : B v : B

Mix

x⌢y u⌢v

x : A, u : B y : B, v : A

Greg Restall Terms for Classical Sequents 55 of 66

slide-94
SLIDE 94

Sequentialisation: Terms with No Switchings

The term contains no Ln, Rn, Cn and • in input position

  • r Fn, Sn, An and • in output position.

It has a derivation using the linear rules L, L, R, R and R and mixes. F L NR L S R

Greg Restall Terms for Classical Sequents 56 of 66

slide-95
SLIDE 95

Sequentialisation: Terms with No Switchings

The term contains no Ln, Rn, Cn and • in input position

  • r Fn, Sn, An and • in output position.

It has a derivation using the linear rules ∧L, ¬L, ¬R, ∨R and ⊃R and mixes. F L NR L S R

Greg Restall Terms for Classical Sequents 56 of 66

slide-96
SLIDE 96

Sequentialisation: Terms with No Switchings

The term contains no Ln, Rn, Cn and • in input position

  • r Fn, Sn, An and • in output position.

It has a derivation using the linear rules ∧L, ¬L, ¬R, ∨R and ⊃R and mixes. Fy⌢Lz NRz⌢Lz Sy⌢Rz y : p ∧ ¬p z : p ∨ ¬p

Greg Restall Terms for Classical Sequents 56 of 66

slide-97
SLIDE 97

Terms with No Switchings: Example

x⌢x

x : p x : p

∨R

x⌢Lz

x : p z : p ∨ ¬p

∧L

Fy⌢Lz

y : p ∧ ¬p z : p ∨ ¬p

u⌢u

u : p u : p

¬R

Nv⌢u

u : p, v : ¬p

∨R

NRz⌢Lz

z : p ∨ ¬p

v⌢v

v : ¬p v : ¬p

∨R

v⌢Rz

v : ¬p z : p ∨ ¬p

∧L

Sy⌢Rz

y : p ∧ ¬p z : p ∨ ¬p

Mix

NRz⌢Lz Sy⌢Rz

y : p ∧ ¬p z : p ∨ ¬p

Mix

Fy⌢Lz NRz⌢Lz Sy⌢Rz

y : p ∧ ¬p z : p ∨ ¬p

Greg Restall Terms for Classical Sequents 57 of 66

slide-98
SLIDE 98

Terms with Switchings

By induction on the number of switched pairs. Take a switched pair at the adjacent to variables or cut points (peel away unswitched steps if there aren’t any).

π[x](−)

Σ x : A, ∆

π[−](y)

Σ′, y : B ∆′

⊃L

π[Az](Lz)

Σ, Σ′, z : A ⊃ B ∆, ∆′

Greg Restall Terms for Classical Sequents 58 of 66

slide-99
SLIDE 99

Back to Eliminating Cuts: Cuts can be Complicated

π[x, u]

x : A ∧ B, u : A

π′[x, v]

x : A ∧ B, v : B

∧R

π[x, Fx] π′[x, Sx]

x : A ∧ B

π′′(y, z, x)

y : A, z : B, x : A ∧ B

∧L

π′′(Fx, Sx, x)

x : A ∧ B

Cut

π[•, F•] π′[•, S•] π′′(F•, S•, •)

Σ ∆

Greg Restall Terms for Classical Sequents 59 of 66

slide-100
SLIDE 100

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m.

conjunction: for each F /S , add new cut points and . For any add for each link with as input. For any add for each link with as

  • utput.

S F F S F S S F N S F S F FN S SN F N S F F S

Greg Restall Terms for Classical Sequents 60 of 66

slide-101
SLIDE 101

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

S F F S F S S F N S F S F FN S SN F N S F F S

Greg Restall Terms for Classical Sequents 60 of 66

slide-102
SLIDE 102

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

Sz⌢F• Fz⌢S• F•⌢Sx S•⌢Fx Ny⌢•

  • ⌢v

S F S F FN S SN F N S F F S

Greg Restall Terms for Classical Sequents 60 of 66

slide-103
SLIDE 103

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

Sz⌢F• Fz⌢S• F•⌢Sx S•⌢Fx Ny⌢•

  • ⌢v

Sz⌢⋆ Fz⌢∗ ⋆⌢Sx ∗⌢Fx FN S SN F N S F F S

Greg Restall Terms for Classical Sequents 60 of 66

slide-104
SLIDE 104

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

Sz⌢F• Fz⌢S• F•⌢Sx S•⌢Fx Ny⌢•

  • ⌢v

Sz⌢⋆ Fz⌢∗ ⋆⌢Sx ∗⌢Fx FNy⌢Sx SNy⌢Fx Ny⌢v S F F S

Greg Restall Terms for Classical Sequents 60 of 66

slide-105
SLIDE 105

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

Sz⌢F• Fz⌢S• F•⌢Sx S•⌢Fx Ny⌢•

  • ⌢v

Sz⌢⋆ Fz⌢∗ ⋆⌢Sx ∗⌢Fx FNy⌢Sx SNy⌢Fx Ny⌢v Sz⌢Fv Fz⌢Sv

Greg Restall Terms for Classical Sequents 60 of 66

slide-106
SLIDE 106

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

▶ negation: for each N•, add a new cut point ⋆. For any •⌢n add l(n) for each link

l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as output. disjunction: for each L /R , add new cut points and . For any add for each link with as input. For any add for each link with as

  • utput.

conditional: for each A /C , add new cut points and . For any add for each link with as input. For any add for each link with as

  • utput.

Greg Restall Terms for Classical Sequents 61 of 66

slide-107
SLIDE 107

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

▶ negation: for each N•, add a new cut point ⋆. For any •⌢n add l(n) for each link

l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as output.

▶ disjunction: for each L•/R•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

conditional: for each A /C , add new cut points and . For any add for each link with as input. For any add for each link with as

  • utput.

Greg Restall Terms for Classical Sequents 61 of 66

slide-108
SLIDE 108

Cut Reductions

Given a term π(•)[•] and a cut-point •, the •-reduction of π is found by:

▶ atomic: replace each pair n⌢• and •⌢m by n⌢m. ▶ conjunction: for each F•/S•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

▶ negation: for each N•, add a new cut point ⋆. For any •⌢n add l(n) for each link

l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as output.

▶ disjunction: for each L•/R•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

▶ conditional: for each A•/C•, add new cut points ⋆ and ∗. For any •⌢n add l(n) for

each link l(•) with n as input. For any n⌢• add l[n] for each link l[•] with n as

  • utput.

Greg Restall Terms for Classical Sequents 61 of 66

slide-109
SLIDE 109

strong normalisation

slide-110
SLIDE 110

Any reduction for π terminates in a unique∗ term π∗ There is some terminating reduction process. Proof reduction is confluent.

– If and then there is a where and .

Greg Restall Terms for Classical Sequents 63 of 66

slide-111
SLIDE 111

Any reduction for π terminates in a unique∗ term π∗

▶ There is some terminating reduction process.

Proof reduction is confluent.

– If and then there is a where and .

Greg Restall Terms for Classical Sequents 63 of 66

slide-112
SLIDE 112

Any reduction for π terminates in a unique∗ term π∗

▶ There is some terminating reduction process. ▶ Proof reduction is confluent.

– If π ⇝• π′ and π ⇝⋆ π′′ then there is a π′′′ where π′ ⇝⋆ π′′′ and π′′ ⇝• π′′′.

Greg Restall Terms for Classical Sequents 63 of 66

slide-113
SLIDE 113

further work

slide-114
SLIDE 114

To Do List

▶ Are these genuine invariants? (Can we show that if two derivations have

the same term, some set of permutations permute one to the other?) Categories (The class of single input, single output terms with composition by defined by Cut + reduction is a category. What are its properties?) Apply terms to theories of warrants. Extend beyond propositional logic.

Greg Restall Terms for Classical Sequents 65 of 66

slide-115
SLIDE 115

To Do List

▶ Are these genuine invariants? (Can we show that if two derivations have

the same term, some set of permutations permute one to the other?)

▶ Categories (The class of single input, single output terms with composition by

defined by Cut + reduction is a category. What are its properties?) Apply terms to theories of warrants. Extend beyond propositional logic.

Greg Restall Terms for Classical Sequents 65 of 66

slide-116
SLIDE 116

To Do List

▶ Are these genuine invariants? (Can we show that if two derivations have

the same term, some set of permutations permute one to the other?)

▶ Categories (The class of single input, single output terms with composition by

defined by Cut + reduction is a category. What are its properties?)

▶ Apply terms to theories of warrants.

Extend beyond propositional logic.

Greg Restall Terms for Classical Sequents 65 of 66

slide-117
SLIDE 117

To Do List

▶ Are these genuine invariants? (Can we show that if two derivations have

the same term, some set of permutations permute one to the other?)

▶ Categories (The class of single input, single output terms with composition by

defined by Cut + reduction is a category. What are its properties?)

▶ Apply terms to theories of warrants. ▶ Extend beyond propositional logic.

Greg Restall Terms for Classical Sequents 65 of 66

slide-118
SLIDE 118

thank you!

https://consequently.org/presentation/2016/ terms-for-classical-sequents-logicmelb/ @consequently on Twitter