Lecture 6 Three kinds of context

Slides:



Advertisements
Similar presentations
Hyperintensionality and Impossible Worlds: An Introduction
Advertisements

Introducing Formal Methods, Module 1, Version 1.1, Oct., Formal Specification and Analytical Verification L 5.
Hoare’s Correctness Triplets Dijkstra’s Predicate Transformers
First Order Logic Resolution
L41 Lecture 2: Predicates and Quantifiers.. L42 Agenda Predicates and Quantifiers –Existential Quantifier  –Universal Quantifier 
1 Ontology as a logic of intensions Marie Duží, Martina Číhalová, Marek Menšík VSB–Technical University Ostrava, Department of Computer Science FEI, Silesian.
1 Conditional XPath, the first order complete XPath dialect Maarten Marx Presented by: Einav Bar-Ner.
March 10: Quantificational Notions Interpretations: Every interpretation interprets every individual constant, predicate, and sentence letter of PL. Our.
EE1J2 - Slide 1 EE1J2 – Discrete Maths Lecture 6 Limitations of propositional logic Introduction to predicate logic Symbols, terms and formulae, Parse.
Synthetic concepts a priori
Logical Equivalence & Predicate Logic
1st-order Predicate Logic (FOL)
CSE S. Tanimoto Lambda Calculus 1 Lambda Calculus What is the simplest functional language that is still Turing complete? Where do functional languages.
Uni-Log 1  -reduction by value A plea for  -reduction by value Marie Duží.
1 Predicate (Relational) Logic 1. Introduction The propositional logic is not powerful enough to express certain types of relationship between propositions.
Albert Gatt LIN3021 Formal Semantics Lecture 4. In this lecture Compositionality in Natural Langauge revisited: The role of types The typed lambda calculus.
Hazırlayan DISCRETE COMPUTATIONAL STRUCTURES Propositional Logic PROF. DR. YUSUF OYSAL.
CS 285- Discrete Mathematics Lecture 4. Section 1.3 Predicate logic Predicate logic is an extension of propositional logic that permits concisely reasoning.
Transparent intensional logic, -rule and Compositionality Marie Duží VSB-Technical University Ostrava
We will now study some special kinds of non-standard quantifiers. Definition 4. Let  (x),  (x) be two fixed formulae of a language L n such that x is.
Albert Gatt LIN3021 Formal Semantics Lecture 3. Aims This lecture is divided into two parts: 1. We make our first attempts at formalising the notion of.
Lecture 041 Predicate Calculus Learning outcomes Students are able to: 1. Evaluate predicate 2. Translate predicate into human language and vice versa.
PPL Syntax & Formal Semantics Lecture Notes: Chapter 2.
Propositional Logic. Assignment Write any five rules each from two games which you like by using propositional logic notations.
Chapter Eight Predicate Logic Semantics. 1. Interpretations in Predicate Logic An argument is valid in predicate logic iff there is no valuation on which.
Lecture 6 Modality: Possible worlds
Transparent Intensional Logic TIL
Natural language processing Lecture 7
Unit – 3 :LAMBDA CALCULUS AND FUNCTIONAL PROGRAMMING
Introduction to Logic for Artificial Intelligence Lecture 2
Lecture 4 Natural Language Processing
12. Principles of Parameter Estimation
CSE15 Discrete Mathematics 01/23/17
Natural Language processing
Chapter 3 The Logic of Quantified Statements
Rationale Behind the Precise Formulation of the Four Quantifier Rules
Lecture 3 Natural Language Processing
DIFFERENTIATION RULES
7/3/2018 EMR 17 Logical Reasoning Lecture 11.
Predicate Calculus Discussion #14 Chapter 2, Section 1 1/20.
Discussion #14 Predicate Calculus
Procedural Semantics of TIL
Trigonometric Identities
CS201: Data Structures and Discrete Mathematics I
CS201: Data Structures and Discrete Mathematics I
Predicate Calculus Discussion #14 Chapter 2, Section 1.
CSE 311: Foundations of Computing
1st-order Predicate Logic (FOL)
Week #2 – 4/6 September 2002 Prof. Marie desJardins
TIL: notional attitudes of seeking and finding
Discrete Mathematics CMP-200 Propositional Equivalences, Predicates & Quantifiers, Negating Quantified Statements Abdul Hameed
Foundations of TIL; method of analysis
Predicates and Quantifiers
This Lecture Substitution model
CS201: Data Structures and Discrete Mathematics I
Logic Logic is a discipline that studies the principles and methods used to construct valid arguments. An argument is a related sequence of statements.
CSE S. Tanimoto Lambda Calculus
Predicate Calculus. Predicate Calculus Topics Predicates Variables and instantiation Quantifiers Logical expressions Bound and free variables.
Natural Language Processing (Transparent Intensional Logic)
Lecture 4 Natural Language Processing
Logical and Rule-Based Reasoning Part I
12. Principles of Parameter Estimation
Lecture 3 Natural Language Processing
CS201: Data Structures and Discrete Mathematics I
Natural language processing Lecture 7
Natural language processing Logic of attitudes
Marie Duží Natural Language processing de dicto, de re, ambiguities in natural language Marie Duží
Lecture 9; Anaphora resolution; -reduction by value
Natural Language Processing (Transparent Intensional Logic) TIL
1st-order Predicate Logic (FOL)
Presentation transcript:

Lecture 6 Three kinds of context Marie Duží http://www.cs.vsb.cz/duzi/

Exercise 3 - examples The temperature in Amsterdam is rising. The temperature in Amsterdam = The temperature in Prague. –––––––––––––––––––––––––––––––––––––––––––––––––– The temperature in Prague is rising. Types. Terms ‘temperature in Amsterdam’ and ‘temperature in Prague’ denote magnitudes of type . Temperature_in/(), Prague, Amsterdam/, =/(), Rising/(): property of a function (here magnitude), that it is rising at a given argument, i.e., its derivative has a positive value, wt [0Temperature_inwt 0Amsterdam], wt [0Temperature_inwt 0Prague]   wt [0Risingwt wt [0Temperature_inwt 0Amsterdam]] wt [0= wt [0Temperature_inwt 0Amsterdam]wt wt [0Temperature_inwt 0Prague]wt] ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– wt [0Risingwt wt [0Temperature_inwt 0Prague]] The argument is invalid, because the second premise specifies a contingent identity of the values of the two (distinct) magnitude; however, Rising is a property of the whole magnitude. We could sbstitude only if the second premise specified the identity of magnitudes, which is not so. The Closures wt [0Temperature_inwt 0Amsterdam] and wt [0Temperature_inwt 0Prague] occur in the first premise and in the conclusion with de dicto supposition, while in the second premise with de re supposition.

Exercise 3 The Mayor of Ostrava visited Brno. –––––––––––––––––––––––––––––––––––––––– The Mayor of Ostrava exists. wt [0Visitwt wt [0Mayor_ofwt 0Ostrava]wt 0Brno] ––––––––––––––––––––––––––––––––––––––––––––––––––– wt [0Existwt wt [0Mayor_ofwt 0Ostrava]] Types. Visit/(); Exist/() Exist = wt u [0x [x = uwt]]; u v , x v , =/(): identity of individuals Proof. 1. [0Visitwt wt [0Mayor_ofwt 0Ostrava]wt 0Brno] assumption 2. [0Impwt 0[wt [0Mayor_ofwt 0Ostrava]wt]] Def. Composition 3. [0Empty x [x = wt [0Mayor_ofwt 0Ostrava]wt] 4. [0x [x=wt [0Mayor_ofwt 0Ostrava]wt]] 5. [0Existwt wt [0Mayor_ofwt 0Ostrava]] Def. Exist Imp/(): the property of a construction that it is v-improper in a given w, t Empty/(()): the class of empty classes of individuals

Two principles de re a) Existential presupposition The Mayor of Ostrava did/did not visit Brno. ––––––––––––––––––––––––––––––––––––––– The Mayor of Ostrava exists. b) Substitution of co-referential terms The Mayor of Ostrava is Mr. X The Mayor of Ostrava visited Brno. X visited Brno

Exercise 3 For all numbers x holds that dividing x by 0 is improper. Improper/(1): the class of constructions v-improper for every valuation v. [0Improper 0[0: x 00]] – constructs T regardless of valuation of x, because [0: x 00] is v-improper for every valuation. The variable x is o-bound, which is stronger than -binding, it occurs hyperintenally (not in execution mode) Does it make sense to quantify over x occurring hyperintensionally? [0x [0Improper 0[0: x 00]]] – constructs T, OK, but for instance, [x [0Improper 0[0: x 00]] 05] constructs [0Improper 0[0: x 00]] rather than [0Improper 0[0: 05 00]] Hence quantifying into hyperintensional context is of no effect

Examples There is a number y such that for any number x dividing x by y is v-improper. [0Improper 0[0: x 00]]  ??? [0y [0Improper 0[0: x y]]] But the Composition [0: x y] is not v-improper for any valuation v; for instance, it is v(5/x, 1/y)-proper! The existential quantifier has no effect here, because the variable y occurs hyperintensionally in 0[0: x y], it is not free for quantification If we wanted to prove the above argument, it would not go: [0Improper 0[0: x 00]] premise - T [y [0Improper 0[0: x y]] 00] constructs F !!!

hyperintensional context The whole construction C is an object of predication (argument); hence its product, i.e. the function f constructed by C (if any) is irrelevant The construction C does not occur in the execution mode; it is just displayed (by Trivialization) All the subconstructions of C (including variables) are also displayed, occur hyperintensionally as well How then to operate in a hyperintensional context? How to quantify into a hyperintensional context? Solution, way in: Substitution method !!!

hyperintensional context Substitution method [0Improper 0[0: x 00]]  !!! [0y [0Improper [0Sub [0Tr y] 0z 0[0: x z]]]] The variable y is now free in the Composition [0Improper … ], we can -bind it, and also the -quantifier functions here as it should Proof. [0Improper 0[0: x 00]] premise [y [0Improper [0Sub [0Tr y] 0z 0[0: x z]]] 00] -abstraction (constructs T, because [0Sub [0Tr y] 0z 0[0: x z]] v(0/y)-constructs the construction [0: x 00]) [0Empty y [0Improper [0Sub [0Tr y] 0z 0[0: x z]]]]

Exercise 3 a) [0Sub [0Tr x] 0y 0[0Sin y]] b) 2[0Sub [0Tr x] 0y 0[0Sin y]] Sub/(nnnn); Tr/(n); Sin/(); x, y v . Valuation v(/x), i.e. valuation that assigns the number  to the variable x ad a) [0Sub [0Tr x] 0y 0[0Sin y]] =* 0[0Sin 0] ad b) 2[0Sub [0Tr x] 0y 0[0Sin y]] = 20[0Sin 0] = [0Sin 0] = 00 =*/(nn): identity of constructions =/(): identity of numbers

De dicto vs. de re wt [0Want_to_bewt 0Tom 0Pope] de dicto wt [0= 0Popewt 0Francisco] de re De dicto is a special case of an intensional occurence (of empirical notions) De re is a special case of an extensional occurence (of empirical notions) in the sub-construction that is right after the first wt, i.e. world w and time t in which we are to evaluate

Three kinds of context Let C be a subconstruction of D. Then The occurrence of C in D is hyperintensional, if the whole construction C is an argument of a function constructed within D. We say that such an occurrence of C is displayed in D. Then all the subconstructions of this occurrence of C have also a hyperintensional occurrence. Exmaple. Tom calculates Sin(). wt [0Calculatewt 0Tom 0[0Sin 0]] hyperint Calculate/(1) Trivialization of a construction C (0C) raises the context up to the hyperintensional level Dually, a Double Execution brings the context down to an intensional or extensional level: 20C = C

Three kinds of context If an occurence of C in D is not hyperintensional, then the construction C occurs in the execution mode; we say that C is a constituent of D. A constituent C of D can occur intensionally or extensionally. Intensional occurrence of C v f: the whole function f v-constructed by C (f can be even a nullary function, i.e., an atomic object like a number or an individual) is an object of predication/argument of another function constructed within D. Extensional occurrence of C v f/(): -value of the function f v-constructed by C (where f is at least unary) is an object of predication/argument of another function constructed within D.

wt [0Calculatewt 0Tom 0[0+ 03 05]] Three kinds of context wt [0Calculatewt 0Tom 0[0+ 03 05]] Constituents of [0Počítáwt 0Tom 0[0+ 03 05]], i.e. those subconstructions that must be executed whenever the whole Composition is to be executed are these: [0Calculatewt 0Tom 0[0+ 03 05]] [[0Calculate w]t] [0Calculate w], 0Calculate, w, t, 0Tom, 0[0+ 03 05]. The subconstruction [0+ 03 05] is not a constituent, it is just an object of which we predicate that Tom wants to find out what this construction produces.

Three kinds of context Example. Sine is a periodic function. Types. Periodic/(()): the set of periodic unary functions, Sine/(). [0Periodic 0Sine] intensional Sin()=0: [0= [0Sine 0] 00] extensional

Intensional vs. extensional context -Closure makes up a generic intensional context; it raises the context up to the intensional level. Composition brings the context down to the extensional level. Higher context is dominant over a lower one. [0: x 00] constituent 0: occurs here extensionally; the Composition is v-improper for any valuation v x [0: x 00] constituent 0: occurs here intensionally; Closure is not v-improper for any valuation; it constructs a degenerate function

Intensional vs. extensional context [x y [0: x y] [0Cotg 0]] The occurrence of 0Cotg is here extensional. The Composition is improper, it fails to construct anything, because [0Cotg 0] is improper; the function Cotg is not defined at the argument . -reduction by name is not strictly equivalent !!! [y [0: [0Cotg 0] y]] The occurrence of 0Cotg is here intensional due to -generic context (y); the Composition produces a degenerate function; it is not improper -reduction by value is strictly equivalent !!! 2[0Sub [0Tr [0Cotg 0]] 0x 0[y [0: x y]] The occurrence of 0Cotg is here extensional. The Composition is improper, it fails to construct anything, because [0Cotg 0] is improper; partiality is „propagated up“.

Substitution Intensional context: Extensional context: The whole function (intension) f is an object of predication For an intensional occurence of C that v-constructs a function f we can substitute a construction D such that D v-constructs the same function f. Hence C=D, i.e. C and D are equivalent, i.e. v-congruent for any valuation v Extensional context: The value (if any) of the v-constructed function (intension) f at a given argument is an object of predication For an extensional occurrence of C we can substitute a construction D such that D v-constructs the same value (even of a different function) at a given argument a. Hence C =v D, i.e. C and D are v-congruent for a certain valuation v(a/x)

Substitution Hyperintensional context: The whole construction C is an object of predication For a hyperintensional occurrence of C we can substitute the same construction C. Only the same construction C? But then we don’t substitute, it turns out to be a too strong demand We can substitute a construction D that is procedurally isomorphic with D Hence 0C =i 0D, where =i/(nn) is the relation of procedural isomorphism How to define this relation?

Identity of procedures How hyper is hyper? Identity of procedures Functions in extension (sets, mappings, PWS-intensions) are extensionally individuated: x [f(x) = g(x)]  f = g Procedures are (hyper)intensionally individuated; Church’s ‘functions in intension’ Procedures are equivalent  produce the same object Procedures are identical  consist of the same constituents But: each procedure can be refined, ad infinitum, when are the constituents identical ??? Does it matter? Hyperintensional contexts – only synonymous terms can be substituted Synonymous terms  have the same meaning, are assigned the same meaning construction, procedure (?)

How hyper is hyper? Carnap’s intensional isomorphism Church’s synonymous isomorphism Any two terms or expressions whose respective meanings are procedurally isomorphic are deemed semantically indistinguishable, hence synonymous. Thus procedurally isomorphic constructions can be mutually substituted in any context, including hyperintensional ones. Church’s Alternatives; (A2) logical equivalence (A1) includes - and -conversion; (A1’) + -conversion (A0) includes -conversion and meaning postulates for atomic constants such as ‘bachelor’, ‘fortnight’, ‘prime’. Procedural isomorphism; (A½) Quid-relation (Materna); - and -conversion (Duží, Jespersen, Materna 2010) (A¾); -, -conversion and restricted -conversion (Duží, Jespersen 2013) (A1’’); modification of Church’s (A1), -conversion ‘by value’ (Duží, Jespersen 2014) (A0’) ???

Procedural isomorfismus Strict criterion: We exclude -conversion and unrestricted -conversion ‘by name’ In the logic of partial functions such as TIL these conversions are not equivalent transformations Different constituents Loss of analytic information

Problems with - (/)-reduction non-equivalence arises when drawing an extensional occurrence of a constituent into (hyper/) intensional context Example: [x [y [0+ x y]] [0Cotg 0]] is an improper construction; it does not construct anything, because there is no value of the cotangent function at  but its -reduced Composition [y [0+ [0Cotg 0] y]] constructs a degenerated function The improper construction [0Cotg 0] has been drawn into the intensional context of the Closure [y [0+ x y]]. De re attitudes: Tilman believes of the Pope that he is wise wt [he [0Believewt 0Tilman w*t* [0Wisew*t* he]] 0Popewt]  wt [0Believewt 0Tilman w*t* [0Wisew*t* 0Popewt]]

Procedural isomorphism -conversion The set of positive numbers: x [0> x 00], y [0> y 00], z [0> z 00],… which if them? Any of them restricted -conversion by name: The Mayor of Ostrava is rich: wt [0Richwt w’t’ [0Mayor_ofw’t’ 0Ostrava]wt] =r wt [0Richwt [0Mayor_ofwt 0Ostrava]] We don’t use -bound variables in a natural language -conversion by value

Procedural isomorphism Obviously it is wise to define several criteria of procedural isomorphism (hence of synonymy) depending on the discourse under scrutiny Natural language  we don’t use -bound variables -conversion, -conversion by value, restricted -conversion by name + meaning postulates for syntactically simple terms (Church’s A0) In the language of mathematics and in a programming language variables can play a significant role The most restrictive criterion: only meaning postulates for syntactically simple terms (Church’s A0)

Existential quantification Extensional context – only in this case partiality plays a crucial role: the application of a function f() to the argument a   can fail if f is not defined at a (x  ): [0Proper [f a]] | [Empty x [f x]] | [0 x [f x]] Intensional context – we can derive that there is a function f (an object of predication) but not that there is a value of f Tom is seeking a unicorn | unicorns exist Not to turn logic to magic 

Existential generalization Tom is seeking a unicorn ------------------------------------ Tom is seeking something (there is something such that Tom is seeking it) wt [0Seekwt 0Tom 0Unicorn] wt [0p [0Seekwt 0Tom p]] Seek/(()) : the relation-in-intension of an individual to a property the instances of which the individual wants to find, Unicorn/(), p v ()

Existential generalization Tom calculates tg(/2) ----------------------------------- Tom calculates something (Tom calculates something but not a non-existing number) wt [0Calculatewt 0Tom 0[0Tg [0/ 0 02]] wt [0c [0Calculatewt 0Tom c]] Calculate/(1) : the relation-in-intension of an individual to a construction, c v 1