distances between formal...

44
General idea Axiomatic distance Conceptual distance Distances between Formal Theories Joint work with Mohamed Khaled (Bah¸ce¸ sehir University) and Michelle Friend (George Washington University) arXiv:1807.01501 Koen Lefever 1 Gergely Sz´ ekely 2 1 Vrije Universiteit Brussel 2 Alfr´ ed R´ enyi Institute for Mathematics Makkai80 Conference, Budapest 2019-06-20 Koen Lefever, Gergely Sz´ ekely Distances between Formal Theories

Upload: others

Post on 28-Jun-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Distances between Formal TheoriesJoint work with

Mohamed Khaled (Bahcesehir University) andMichelle Friend (George Washington University)

arXiv:1807.01501

Koen Lefever1 Gergely Szekely2

1Vrije Universiteit Brussel

2Alfred Renyi Institute for Mathematics

Makkai80 Conference, Budapest 2019-06-20

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 2: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Traditionally, empirical adequacy is the criterion to comparecompeting theories.

However, this is not always straightforward:

Early Copernican heliocentrism provided less accuratepredictions than the old Prolemaic geocentric model; see[Kuhn 1957].One theory may be better suited to explain one part of theempirical data, another theory may be better suited to explainother parts of reality.

We will here present another way to approach the comparisonof theories, independent of empirical adequacy.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 3: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

A motivating example

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 4: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Kin:={AxEField,AxEv,AxSelf,AxSymD,AxLine,AxTriv,AxNoAcc}

ClassicalKin:=Kin ∪ {AxEther,AbsTime,AxThExp+}

SpecRel:= Kin ∪ {AxPhc,AxThExp}

Theorem:

ClassicalKin ` Worldview transformations are Galileantransformations.

Theorem: (Andreka–Madarasz–Nemeti, 1998)

SpecRel ` Worldview transformations are Poincare transformations.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 5: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Definition

A translation is a function between formulas of languagespreserving the logical connectives, i.e. Tr(φ ∧ ψ) = Tr(φ) ∧ Tr(ψ),etc.

Definition

An interpretation of theory T1 in theory T2 is a translation Trwhich translates all tautologies and all axioms of T1 into theoremsof T2.

Definition

A definitional equivalence exists between two theories if thosetheories can be interpreted in each other and if all formulas fromboth theories translated into the other theory and back are logicalequivalent to the original formulas.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 6: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

There is an interpretation Tr of SpecRel in ClassicalKin.

SpecRel

Ether

FTL-IOb

ClassicalKinTr

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 7: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

There is no definitional equivalence between SpecRel andClassicalKin.

SpecRel

EtherFTL-IOb

ClassicalKinTr

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 8: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

How can we make SpecRel and ClassicalKin equivalent?

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 9: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Adding and removing concepts to make both theoriesequivalent:

Removing FTL observers from ClassicalKin

Adding a “primitive ether” to SpecRel

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 10: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

ClassicalKinSTL:=Kin ∪ {AxEther,AbsTime,AxThExpSTL,AxNoFTL}

AxNoFTL :

All inertial observers move slower than light with respect to theether frames.

¬∃m(IOb(m) ∧ ∃e[Ether(e) ∧ SpeedCK

e (m) ≥ ce]).

AxThExpSTL :

Inertial observers can move with any speed which is in the etherframe slower than that of light.

∃h(IOb(h)

)∧ ∀ex y

(Ether(e) ∧ space(x , y) < ce · time(x , y)

→ ∃k IOb(k) ∧W (e, k , x) ∧W (e, k , y)).

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 11: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

SpecRele :=SpecRel ∪ {AxPrimitiveEther}

AxPrimitiveEther :

There is a non-empty class of ether observers, stationary withrespect to each other, which is closed under trivial transformations.

∃e(E (e) ∧ ∀k

[[IOb(k) ∧ (∃T ∈ Triv)wSR

ek = T ]↔ E (k)])

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 12: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

There is an interpretation Tr ′+ of ClassicalKinSTL in SpecRele .SpecRele ` Tr ′+(ClassicalKinSTL)

SpecRel

SpecRele

E Ether

ClassicalKin

FTL-IOb

Tr

Tr ′+

4≡

Id

Tr+|SpecRel = Tr

Tr+(E (x)

)= Ether(x)

Tr+

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 13: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

Tr+ and Tr ′+ are a definitional equivalence between SpecRele andClassicalKinSTL.

Tr′+ ◦ Tr+(SpecRele )⇔ SpecRele and Tr+ ◦ Tr′+(ClassicalKinSTL)⇔ ClassicalKinSTL

SpecRel

SpecRele

E Ether

ClassicalKin

FTL-IOb

Tr

Tr ′+

4≡

Id

Tr+|SpecRel = Tr

Tr+(E (x)

)= Ether(x)

Tr+

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 14: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

There is an interpretation Tr∗ of ClassicalKinSTL in ClassicalKin.ClassicalKin ` Tr∗(ClassicalKin

STL)

Theorem:

There is an interpretation Tr ′∗ of ClassicalKin in ClassicalKinSTL.ClassicalKinSTL ` Tr ′∗(ClassicalKin)

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 15: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

Tr∗ and Tr ′∗ are a definitional equivalence between ClassicalKinSTL

and ClassicalKin.

Tr ′∗ ◦ Tr∗(ClassicalKinSTL)⇔ ClassicalKinSTL and

Tr∗ ◦ Tr ′∗(ClassicalKin)⇔ ClassicalKin

Ether

ClassicalKin

FTL-IOb

Ether

ClassicalKin

FTL-IOb

Tr ′∗4≡

Tr∗ClassicalKinSTL

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 16: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

Tr∗ ◦ Tr+ and Tr ′+ ◦ Tr ′∗ are a definitional equivalence betweenSpecRele and ClassicalKin.

SpecRel

SpecRele

E Ether

ClassicalKin

FTL-IOb

Tr

Tr ′+

4≡

Id

Tr+|SpecRel = Tr

Tr+(E (x)

)= Ether(x)

Tr+

Ether

ClassicalKin

FTL-IOb

Tr ′∗

4≡

Tr∗

ClassicalKinSTL

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 17: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Equivalence relations:

There are several ways in which theories can be equivalent to eachother:

Logical Equivalence (≡)⇓6⇑

Definitional Equivalence (4≡)

⇓6⇑Conceptual Equivalence (Morita Equivalence,definitional equivalence with different sorts)

⇓6⇑Categorical Equivalence

See [Barrett & Halvorson 2016].

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 18: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Equivalence relations gives us a trivial (discrete) distance:

Let X -be any set of theories and E any equivalence relation on X .The discrete distance on (X ,E ) is the following:

d(x , y) =

{0 if E (x , y),

1 if ¬E (x , y).

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 19: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

We are going to refine these trivial distances given byequivalence relations between theories.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 20: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

We define distances between two theories based on the minimumnumber of things (e.g., axioms or concepts) which need to beadded to or substracted from one theory to make it (logically,definitionally, etc.) equivalent to the other theory.

Two theories which are equivalent have a distance of zero.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 21: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Definition

A cluster network is a triple (X ,E ,S), where X is a class with anequivalence relation E and a symmetric relation S .

E (x , x ′): x and x ′ are basically the same objects.

S(x , x ′): x or x ′ can be reached from the other in one “step.”

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 22: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

A path between x and x ′ in cluster network (X ,E ,S) is a finitesequence of joining E -edges and S-edges connecting x and x ′.

The length of a path is the number of S-edges in the path.

Objects x , x ′ ∈ X are connected in (X ,E , S) iff there is a pathfrom one of them to the other.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 23: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Let X = (X ,E ,S) be a cluster network.

Definition

The step distance on X is the function dX : X × X → N ∪ {∞}defined as:

dX (x , x ′)def≡ min{k ∈ N : ∃ a path from x to x ′ of length k}

if x and x ′ are connected in X , and

dX (x , x ′)def≡ ∞ otherwise

for each x , x ′ ∈ X .

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 24: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance Motivating example General framework

Theorem:

Let dX : X × X → N ∪ {∞} be the step distance on clusternetwork X = (X ,E ,S). Then for each x , y , z ∈ X ,

(a) dX (x , y) ≥ 0, and dX (x , y) = 0 ⇐⇒ x E y .

(b) dX (x , y) = dX (y , x).

(c) dX (x , y) ≤ dX (x , z) + dX (z , y).

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 25: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Axiomatic distance

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 26: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Adding one axiom:

T ↽ T ′ iff there is ϕ ∈ Fm such that T ∪ {ϕ} ≡ T ′.

One axiom difference:

T T ′ iff either T ↽ T ′ or T ′ ↽ T .

Definition

Let T be a class of theories. We call the step distance on thiscluster network (T ,≡,) the axiomatic distance AdT on T .

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 27: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Example

Let T be a class of theories. Let T ,T⊥ ∈ T be two theoriesformulated in the same language. Suppose that T is consistentwhile T⊥ is inconsistent, then AdT (T ,T⊥) = 1.

Consequently, if T ,T ′ ∈ T are formulated on the same languageand an inconsistent theory T⊥ of that language is in T , thenAdT (T ,T ′) ≤ 2 since we have T ⇀ T⊥ ↽ T ′.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 28: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Let CON be the class of all consistent theories.

Proposition:

Let T ,T ′ ∈ CON having the same language L. Thenthe axiomatic distance AdCON(T ,T ′) ≤ 3.

Moreover, AdCON(T ,T ′) = 2 if T 6≡ T ′ and they are complete.

Corollary:

Let T ,T ′ ∈ CON. Then,

AdCON(T ,T ′) =∞

m

T and T ′ are formulated in different langauges.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 29: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Let CON be the class of all consistent theories.

Proposition:

Let T ,T ′ ∈ CON having the same language L. Thenthe axiomatic distance AdCON(T ,T ′) ≤ 3.

Moreover, AdCON(T ,T ′) = 2 if T 6≡ T ′ and they are complete.

Corollary:

Let T ,T ′ ∈ CON. Then,

AdCON(T ,T ′) =∞

m

T and T ′ are formulated in different langauges.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 30: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Let CON be the class of all consistent theories.

Proposition:

Let T ,T ′ ∈ CON having the same language L. Thenthe axiomatic distance AdCON(T ,T ′) ≤ 3.

Moreover, AdCON(T ,T ′) = 2 if T 6≡ T ′ and they are complete.

Corollary:

Let T ,T ′ ∈ CON. Then,

AdCON(T ,T ′) =∞

m

T and T ′ are formulated in different langauges.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 31: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

Fix a language L. Let T ∈ CON be any theory on language L andlet ∅L be the empty theory on L. Then:

(1) AdCON(∅L,T ) = 0 iff T is trivial (T ≡ ∅L).

(2) AdCON(∅L,T ) = 1 iff T is finitely axiomatizable andnon-trivial.

(3) AdCON(∅L,T ) = 2 iff T is not finitely axiomatizable, but hasa finitely axiomatizable consistent extension.

(4) AdCON(∅L,T ) = 3 iff T has no finitely axiomatizableconsistent extension.

Example

AdCON(∅L,PA) = 3, because PA (Peano Arithmetic) has no finitelyaxiomatizable consistent extension, see [Ryll-Nardzewski 1952].

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 32: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

Fix a language L. Let T ∈ CON be any theory on language L andlet ∅L be the empty theory on L. Then:

(1) AdCON(∅L,T ) = 0 iff T is trivial (T ≡ ∅L).

(2) AdCON(∅L,T ) = 1 iff T is finitely axiomatizable andnon-trivial.

(3) AdCON(∅L,T ) = 2 iff T is not finitely axiomatizable, but hasa finitely axiomatizable consistent extension.

(4) AdCON(∅L,T ) = 3 iff T has no finitely axiomatizableconsistent extension.

Example

AdCON(∅L,PA) = 3, because PA (Peano Arithmetic) has no finitelyaxiomatizable consistent extension, see [Ryll-Nardzewski 1952].

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 33: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

By giving weights to the axiom adding steps, e.g., considering theaddition of certain kinds of axioms as two or more steps, one couldincrease the resolution of axiomatic distance.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 34: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Conceptual distance

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 35: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Definition

Theory T ′ of language L′ is a one-concept-extension of theoryT of language L:

T T ′def⇐⇒ L′ = L ∪ {R} and (∀ϕ ∈ L)(T ′ |= ϕ iff T |= ϕ)

for some relation symbol R.

T ! T ′def⇐⇒ T T ′ or T ′ T

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 36: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Definition

Let T be a class of theories. The conceptual distance CdT is the

step distance on cluster network (T , 4≡,!), where4≡ is the

definitional equivalence.

If T is the class of all theories of FOL, we omit the subscript Tand simply write Cd.

Theorem:

For every n ∈ N ∪ {∞}, there are theories T and T ′ such thatCd(T ,T ′) = n.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 37: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Definition

Let T be a class of theories. The conceptual distance CdT is the

step distance on cluster network (T , 4≡,!), where4≡ is the

definitional equivalence.

If T is the class of all theories of FOL, we omit the subscript Tand simply write Cd.

Theorem:

For every n ∈ N ∪ {∞}, there are theories T and T ′ such thatCd(T ,T ′) = n.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 38: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Definition

The spectrum of theory T , in symbols I(T , κ), is the number ofits different models of T (up to isomorphism) of cardinality κ.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 39: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

If T1 and T2 are formulated in countable languages, then:

Cd(T1,T2) <∞

(∀ cardinal κ)[I(T1, κ) 6= 0 ⇐⇒ I(T2, κ) 6= 0

].

Corollary:

There are infinitely many theories that are, in terms of conceptualdistance, infinitely far from each other.

Proof.: If A and B are two finite models of different cardinality,then Cd

(Th(A),Th(B)

)=∞. Q.E.D.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 40: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

If T1 and T2 are formulated in finite languages, then:

Cd(T1,T2) <∞

m

(∀ cardinal κ)[I(T1, κ) 6= 0 ⇐⇒ I(T2, κ) 6= 0

].

Corollary:

There are infinitely many theories that are, in terms of conceptualdistance, infinitely far from each other.

Proof.: If A and B are two finite models of different cardinality,then Cd

(Th(A),Th(B)

)=∞. Q.E.D.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 41: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

If T1 and T2 are formulated in finite languages, then:

Cd(T1,T2) <∞

m

(∀ cardinal κ)[I(T1, κ) 6= 0 ⇐⇒ I(T2, κ) 6= 0

].

Corollary:

There are infinitely many theories that are, in terms of conceptualdistance, infinitely far from each other.

Proof.: If A and B are two finite models of different cardinality,then Cd

(Th(A),Th(B)

)=∞. Q.E.D.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 42: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

If T1 and T2 are formulated in finite languages, then[Cd(T1,T2) <∞ and I(T1, 1) = I(T2, 1) = 0

]⇓

Cd(T1,T2) ≤ 2.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 43: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

Theorem:

Cd(ClassicalKin, SpecRel) = 1

Proof.:

SpecRel SpecRele4≡ ClassicalKinSTL

4≡ ClassicalKin

Q.E.D.

Koen Lefever, Gergely Szekely Distances between Formal Theories

Page 44: Distances between Formal Theorieshomepages.vub.ac.be/~kolefeve/content/2019-Makkai80-Conceptual… · Koen Lefever, Gergely Sz ekely Distances between Formal Theories. General ideaAxiomatic

General idea Axiomatic distance Conceptual distance

References:H. Andreka, J. X. Madarasz and I. Nemeti, with contributions from A. Andai,G. Sagi, I. Sain, and Cs. Toke: “On the logical structure of relativity theories”,Alfred Renyi Institute of Mathematics, Hungar. Acad. Sci., (2002).

H. Andreka, J. X. Madarasz and I. Nemeti: “Defining new universes inmany-sorted logic”, Research Report, Alfred Renyi Institute of Mathematics,Hungar. Acad. Sci. (2008).

T. W. Barrett and H. Halvorson: “Morita Equivalence”, The Review ofSymbolic Logic 9 (3): 556-582 (2016).

M. Khaled, G. Szekely, K. Lefever and M. Friend: “Distances between FormalTheories”, arXiv:1807.01501 (2018).

T. Kuhn. “The Copernican Revolution: Planetary Astronomy in theDevelopment of Western Thought” Harvard University Press (1957).

K. Lefever: “Using Logical Interpretation and Definitional Equivalence toCompare Classical Kinematics and Special Relativity Theory”PhD Dissertation, Vrije Universiteit Brussel (2017).

K. Lefever and G. Szekely: “Comparing Classical and Relativistic Kinematics inFirst-Order Logic”, Logique et Analyse, ISSN: 2295-5836, p. 57-117, Vol 61, Nr241 (2018).

C. Ryll-Nardzewski, “The role of the axiom of induction in elementaryarithmetic”, Fundamenta Mathematicae, 39:239263 (1952).

Koen Lefever, Gergely Szekely Distances between Formal Theories