1 / 50

Derivation Schemes for Topological Logics

Derivation Schemes for Topological Logics. Derived Logics. What Are They? Why Do We Need Them? How Can We Use Them? Colleague: Michael Westmoreland. History. 1936 Von Neumann and Birkhoff a lattice of propositions based on the closed subspaces of Hilbert space

sofia
Télécharger la présentation

Derivation Schemes for Topological Logics

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Derivation Schemes for Topological Logics

  2. Derived Logics • What Are They? • Why Do We Need Them? • How Can We Use Them? • Colleague: Michael Westmoreland

  3. History • 1936 Von Neumann and Birkhoff • a lattice of propositions based on the closed subspaces of Hilbert space • now known as “quantum logic” • based on measurement • non-Boolean (fails to meet distributive properties) • No satisfying way to do implication

  4. A Topological Logic • A proposition is an equivalence class of sets • S  S’ iff int(S) = int(S’) • [S] = [ (int S)c ] • [S]  [S’] = [ (int S)  (int S’) ] • [S]  [S’] = [ (int S)  (int S’) ] • Most Boolean properties hold • Law of noncontradiction: [S]  [S] = [int S]  [ (int S)c ]= [S  (int Sc) ] choosing canonical representation= [] • But not all:  [S] =? [S] No! •  [S] =  [ (int S)c ] So [S]  [int( (int Sc )c ] =  [ int Sc ] =  [S]

  5. Logic Properties • No tertium non datur: [S]  [S]  [U] where U is the universal set. • What about truth assignment? A measurement (open set) m verifies a proposition P iff mPiPiP. • Example: the real line with the standard topology. P = [ (-3, 5) ]. m = (0, 4) verifies P since (0,4)  (-3, 5), [-3, 5), [-3,5], (-3,5] • We speak of “verification” rather than truth. • Rationale: Let S be a classical system and P a proposition about S with P0 as the canonical representative of [P]. Then P0 = int Pj • Pj [P]. A measurement m that contains points of P0 but does not lie entirely in P0 would not verify P.

  6. More Properties • P = (-3, 5). m = (0,6) does not verify P. • Should we conclude P is false? The state of S could lie in P0 and still be consistent with the result of the measurement m. In fact, there is a more precise measurement, say m’ that lies entirely in P0 and the result of m. Hence, we cannot conclude that P is false. • New concept for assigning truth values: associated with a given measurement (set) , three possibilities: verifiability set, falsifiability set, indeterminate. • Twin Open Set Phase Space Logic (TOSPS) • A measurement m verifies P if m  P0 where P0 is the canonical rep of P. • A measurement falsifies a proposition if m  Cl(P0)c.

  7. Twin Open Set Phase Space Logic • Definition: P is a proposition in TOSPS logic if P = ( [V0], [F0] ), where V0 and F0 are disjoint open sets. • Definition: Let P = ( [V0], [F0] ) be a proposition in TOSPS logic and m be a measurement. P will be assigned the truth value true if mV0; false if mF0; indeterminate otherwise. • Logical Operators P = ( [PV], [PF] ) Q = ( [QV], [QF] ) PQ = ([int PV int QV], [int PF int QF] ) PQ = ( [int PV int QV], [int PF int QF] ) P = ( [int PF], [int PV] )

  8. Properties • P = ([PF], [QF]) = ( [PV], [PF] ) = P • P P = ( [PV], [PF] )  ( [PF], [PV] ) = ( [int PV int PF], [int PF int PV] )= ([], [U]) • P P = ([U], []) • DeMorgan’s laws • Ditributivity • All Boolean properties, but tertium non datur.

  9. Note:  fails to be truth functional • P = [(-1,2), (5,9)] Q = [(1,3), (8,11)] • P  Q = [(-1,3), (8,9)] The measurement m = (0, 2.5) assigns I to P, I to Q, and T to P  Q, since m  PV  QV • m = (0,4) assigns I to P  Q and I to P and I to Q, since m ⊈PV  QV

  10. Twin Open Set Logic Based on Exact Measurement (Discrete)

  11. Truth Tables for TOPSL

  12. Example • Example to illustrate lack of truth functionality for disjunction P = [(-2, 2), (5,9) ] Q = [(1, 3), (8, 11) ] P  Q = [(-2, 3), (8,9) ] Suppose m = (0, 2.5) m assigns “I” to each of P and Q, “T” to P  Q since m  PV QV Now suppose m = ( 0, 4) m assigns “I” to P  Q as well as to P and to Q since m  (PV QV)

  13. Applications to Billiard Ball Model of Computation

  14. OR-Gate

  15. AND-Gate

  16. NOT-Gate

  17. Derivation Gate Input the value of P and the value of P → Q

  18. Derivation • For a Boolean lattice, define P ≤ Q when P  Q is valid where ≤ is the lattice ordering • Modus Ponens

  19. Three Questions to Consider • What is a proper ordering for the propositions in twin open set logic? • What is a proper implication operator in twin open set logic? • What derivation method can be implemented given the answers to 1 and 2?

  20. Characterization Theorem Let (A, , , ) be a DeMorgan algebra. If we define an ordering ≼ on the algebra by P ≼ Q def P  Q = Q, then P  (P  Q) ≼ Q iff (A, , , ) is a boolean algebra . Reminder: TOPSL is a DeMorgan algebra.

  21. Proof Need: (A, , , ) satisfies the law of non contradiction. In any DeMorgan algebra satisfying our hypothesis, 0 ≼ P  P. Substituting Q = 0 in the modus ponens scheme, P  (P  0) ≼0

  22. Using distributivity, P  (P  0) ≼ 0 (P  P)  (P  0) ≼ 0 Since P  0 = 0 and Q  0 = Q for any Q, P  P ≼ 0 By antisymmetry of ≼, P  P = 0 and so (A, , , ) is boolean.

  23. Implications of the theorem: Any DeMorgan algebra in which • Entailment is given by  (), • The implication operator is given by P  Q, and • Modus ponens is satisfied must be a Boolean algebra.

  24. Non Standard Derivation • TOSPL is a DeMorgan algebra, but not a boolean algebra. • At least one of the three properties above must fail.

  25. Modus Ponens Fails • Ordering for TOSL (suggested by  or ) P  Q  PV  QV and QF  PF motivated by either P  Q = P  PV  QV and QF  PF P  Q = Q  PV  QV and QF  PF Q is more readily verified and less easily falsified than P.

  26. Implication • P→Q def P  Q • So P→Q = P  Q = [(PF  QV),(PV  QF)] • Previous Theorem tells us that modus ponens fails. Why does it?

  27. Theorem: With the ordering given by , it is not the case that P  (P→Q )  Q Proof: P  (P→Q ) = P  (P  Q ) = (P  P)  (P  Q) = [(PV PF), (PVPF)]  [(PV  QV), (PF  QF)] = [, (PVPF)]  [(PV  QV), (PF  QF)] = [(PV  QV),, ((PVPF)  (PF  QF))] = [(PV  QV), ((PV  QF)  PF)]

  28. For P  (P→Q ) ≼ Q, QF  (PV  QF)  PF But whenever PV  PF  X (the whole space), the containment fails. In any nondiscrete topology we have disjoint open sets PV and PF such that PV  PF  X and the claim is established.

  29. Need: a proposition that contains [(PV  QV), ((PV  QF)  PF)] One possibility [QV, ] Given P and P→Q [QV, ]

  30. Good Point: It works. Not so good: So does any proposition of the form [QV, Y] where Y is any open subset of int(QVC) Cannot falsify

  31. Modus Tollens P→Q = P  Q = (Q) P  P = Q → P Consider Q  (P→Q ) = [(PF  QF), ((PV  QF)  QV)] Analog to modus ponens: [PF, ]

  32. Another Possibility • For Modus Ponens: Given P and P→Q, Conclude [QV, PV  QF] =def QP • For Modus Tollens: Given Q  (P→Q ), Conclude [PF, PV  QF] =def PQ Now P  (P→Q ) ≼ QP

  33. Moreover, PV  QV  QV and PV  QF  (PV  QF)  PF thereby respecting entailment

  34. Non Standard Entailment P  Q def Pv Qv • Not antisymmetric, but is reflexive and transitive (a quasi ordering relation) Theorem:  satisfies: P  (P→Q )  Q

  35. What about falsifiability? P  Q def QF PF Does not give a valid modus ponens!

  36. Both Verifiability and Falsifiability Quasi ordering: Reminder: PS = PV PF P ≤ Q def

  37. theorem The quasi ordering defined gives P  (P→Q ) ≤ Q

  38. Non Standard Implication Instead of P→Q = P  Q P ↪ Q =def [PV  QV, QF\ ]

  39. Motivation sup(X | P  X ≤ Q) well defined for any orthonormal lattice. Propositions in TOSL make a lattice, but not orthonormal sup(X | P  X ≲ Q) where X = [XV, XF] and XV = sup(Y | PV Y  QV) and XF = inf(Y | QF PF Y)

  40. To get existence need: PF QV XV This blocks inf(Y | QF PF Y) Leading to XF = (QF \ )

  41. Theorem: P  (P ↪ Q) ≴ Q

  42. Why ? • We get the usual implication operator when considering the discrete twin logic. • Natural interpretation of implication when measurement P verifies P and P ↪ Q, whatever form ↪ may take.

  43. Discoveries • In any derivation scheme that is given by the lattice theoretic entailment, an implication P  Q that is equivalent to P  Q must be Boolean. • Define P  Q = P  Q = [(PF QV), (PV QF)] • m will assign a value of true to P  Q iff m assigns a value of true to either P or Q. i.e., m  (PF QV). • Alternately, m will assign a value of false to P  Q iff m  (PV QF). • m assigns indeterminate to P  Q iff m  (PV QF) and m  (PF QV)

  44. Derivation in Collision Models Replace modus ponens by P and P → Q yield [QV, PV QF] Replace modus tollens by Q and P → Q yield [QV, PV QF]

More Related