• Re: Provability in PA --- Cycle detected in directed graph of evaluation sequence

    From Mikko@mikko.levanto@iki.fi to sci.logic,comp.theory,sci.math,comp.ai.philosophy on Mon Oct 19 11:05:47 2026
    From Newsgroup: sci.math

    On 09/10/2026 18:15, olcott wrote:
    On 10/9/2026 3:50 AM, Mikko wrote:
    On 04/10/2026 16:22, olcott wrote:
    On 10/4/2026 6:30 AM, Mikko wrote:
    On 02/10/2026 14:55, olcott wrote:
    On 10/2/2026 4:37 AM, Mikko wrote:
    On 01/10/2026 17:47, olcott wrote:
    On 10/1/2026 5:03 AM, Mikko wrote:
    On 29/09/2026 19:31, olcott wrote:
    On 9/29/2026 11:01 AM, Python wrote:
    No, Pete. Calling an undecidable sentence "out-of-scope" does >>>>>>>>>> not make undecidability disappear.

    For a formal theory T, "G is undecidable in T" simply means >>>>>>>>>>
    -a-a-a T |-/- G-a-a-a and-a-a-a T |-/- ~G

    That is a statement entirely about proofs. No model theory is >>>>>>>>>> involved.

    And G is not "out-of-scope": it is a perfectly well-formed >>>>>>>>>> sentence of the language of T. What is out-of-scope,
    apparently, is merely anything for which your proposed
    semantics fails to supply an answer.

    Within PTS any expression that is neither provable nor refutable >>>>>>>>> never acquires any semantic meaning.

    There is one more thing one needs to consider: there is no complete >>>>>>>> method to find out whether a sentence or its negation can be
    proven.
    Consequently, it is not possible to construct a system that rejects >>>>>>>> all undecidable sentences.

    If back-chained inference from expression X to axioms
    has a loop it is rejected.

    Irrelevant. A direct proof is either a forward chain from premises >>>>>> to conclusion and an indirect proof is a similar chain from tne
    negation of the axioms and conclusion to a contradiction.

    It turns out that automated proof systems expose
    the correct details the most clearly. They used
    back-chained inference. In any case when a finite
    canonical proof is impossible then the expression
    never gets semantically grounded.

    Existence of a proof and existence of a meaning are irrelevant to

    The Goldbach conjecture is meaningful yet its truth value will be
    unknown unless and until it proof can be algorithmically compressed.
    It might be the case that its truth value is unknowable. I restricted
    my system to "the body of general knowledge that can be expressed in
    language" to handle these edge cases.

    each other. For example, the sentence reCx reCy (x * y = y * x) is

    The Commutative Law For Multiplication
    https://en.wikiversity.org/wiki/Basic_Laws_of_Algebra

    meaningful in many theories and provable in some of them and
    unporvable others.

    That is merely a limitation of the individual formal systems.
    My focus has always been on "the body of general knowledge
    that can be expressed in language". I never really cared that
    system X is too weak to fully express Y.

    Does that include cases where X is Peano arithmetic and Y is a
    proof that G is true about natural number?

    Now I have (with the help of Gemini) finally totally
    anchored my claim that in PTS the meaning of an
    expression is only derived by the existence of a
    canonical proof of it.

    Nevertheless, and arithmetic sentence has an arithmetic meaning,
    whether provable or not. For example, 4 < 3 is not provable in
    any natural number arithmetic but its meaning is clear.
    --
    Mikko

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Mikko@mikko.levanto@iki.fi to sci.logic,sci.math,comp.theory,comp.ai.philosophy on Mon Oct 19 11:12:25 2026
    From Newsgroup: sci.math

    On 09/10/2026 17:22, olcott wrote:
    On 10/9/2026 3:45 AM, Mikko wrote:
    On 02/10/2026 14:55, olcott wrote:
    On 10/2/2026 4:37 AM, Mikko wrote:
    On 01/10/2026 17:47, olcott wrote:
    On 10/1/2026 5:03 AM, Mikko wrote:
    On 29/09/2026 19:31, olcott wrote:
    On 9/29/2026 11:01 AM, Python wrote:
    No, Pete. Calling an undecidable sentence "out-of-scope" does >>>>>>>> not make undecidability disappear.

    For a formal theory T, "G is undecidable in T" simply means

    -a-a-a T |-/- G-a-a-a and-a-a-a T |-/- ~G

    That is a statement entirely about proofs. No model theory is >>>>>>>> involved.

    And G is not "out-of-scope": it is a perfectly well-formed
    sentence of the language of T. What is out-of-scope, apparently, >>>>>>>> is merely anything for which your proposed semantics fails to >>>>>>>> supply an answer.

    Within PTS any expression that is neither provable nor refutable >>>>>>> never acquires any semantic meaning.

    There is one more thing one needs to consider: there is no complete >>>>>> method to find out whether a sentence or its negation can be proven. >>>>>> Consequently, it is not possible to construct a system that rejects >>>>>> all undecidable sentences.

    If back-chained inference from expression X to axioms
    has a loop it is rejected.

    Irrelevant. A direct proof is either a forward chain from premises
    to conclusion and an indirect proof is a similar chain from tne
    negation of the axioms and conclusion to a contradiction.


    It turns out that automated proof systems expose
    the correct details the most clearly. They used
    back-chained inference. In any case when a finite
    canonical proof is impossible then the expression
    never gets semantically grounded.

    A sentence of natural number arithmetic has a well defined arithemtic
    meaning, whether grounded or not. Existence of a proof from some set
    of axioms is irrelevant to the arithmetic meaning.

    **On the Relation Between HeytingrCOs**
    **and GentzenrCOs Approaches to Meaning**

    "GentzenrCOs idea could be summarized by saying that
    -athe meaning of a sentence is determined by what
    -acounts as a canonical proof of it..."

    https://doi.org/10.1007/978-3-319-22686-6_2

    In reality there is much meaning that does not depend on proofs or
    provability. Perhaps you don't understand "Charlemagne received a
    horn of a unicorn from Harun ar-Rashid" without a proof but many
    others do.
    --
    Mikko
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Mikko@mikko.levanto@iki.fi to sci.logic,sci.math,comp.theory,comp.ai.philosophy on Tue Oct 20 11:20:31 2026
    From Newsgroup: sci.math

    On 09/10/2026 18:02, olcott wrote:
    On 10/9/2026 2:59 AM, Mikko wrote:
    On 07/10/2026 16:10, polcott wrote:
    On 10/7/2026 3:49 AM, Mikko wrote:
    On 06/10/2026 15:32, olcott wrote:
    On 10/6/2026 2:59 AM, Mikko wrote:
    On 04/10/2026 16:22, olcott wrote:
    On 10/4/2026 6:30 AM, Mikko wrote:
    On 02/10/2026 14:55, olcott wrote:
    On 10/2/2026 4:37 AM, Mikko wrote:
    On 01/10/2026 17:47, olcott wrote:
    On 10/1/2026 5:03 AM, Mikko wrote:
    On 29/09/2026 19:31, olcott wrote:
    On 9/29/2026 11:01 AM, Python wrote:
    No, Pete. Calling an undecidable sentence "out-of-scope" >>>>>>>>>>>>>> does not make undecidability disappear.

    For a formal theory T, "G is undecidable in T" simply means >>>>>>>>>>>>>>
    -a-a-a T |-/- G-a-a-a and-a-a-a T |-/- ~G

    That is a statement entirely about proofs. No model theory >>>>>>>>>>>>>> is involved.

    And G is not "out-of-scope": it is a perfectly well-formed >>>>>>>>>>>>>> sentence of the language of T. What is out-of-scope, >>>>>>>>>>>>>> apparently, is merely anything for which your proposed >>>>>>>>>>>>>> semantics fails to supply an answer.

    Within PTS any expression that is neither provable nor >>>>>>>>>>>>> refutable
    never acquires any semantic meaning.

    There is one more thing one needs to consider: there is no >>>>>>>>>>>> complete
    method to find out whether a sentence or its negation can be >>>>>>>>>>>> proven.
    Consequently, it is not possible to construct a system that >>>>>>>>>>>> rejects
    all undecidable sentences.

    If back-chained inference from expression X to axioms
    has a loop it is rejected.

    Irrelevant. A direct proof is either a forward chain from >>>>>>>>>> premises
    to conclusion and an indirect proof is a similar chain from tne >>>>>>>>>> negation of the axioms and conclusion to a contradiction.

    It turns out that automated proof systems expose
    the correct details the most clearly. They used
    back-chained inference. In any case when a finite
    canonical proof is impossible then the expression
    never gets semantically grounded.

    Existence of a proof and existence of a meaning are irrelevant to >>>>>>>
    The Goldbach conjecture is meaningful yet its truth value will be >>>>>>> unknown unless and until it proof can be algorithmically compressed. >>>>>>
    It truth value will be known as soon as its proof has been
    chekcked and
    found good. The proof can be algorithmically compressed as soon as >>>>>> it is
    found, even before the verification, though that is irrelevant to the >>>>>> knowledge of its truth. Thought it is possible that it will be
    found to
    be unprovable. But none of that affects its meaning and
    meaningfulness.
    What PTS says about its meaning is irrelevant.

    A finite proof might not exist.
    It may forever remain unknowable
    thus outside of the domain of knowledge.

    Maybe, maybe not. If it ever will be found it is in the domain
    of knowledge and has always been.

    Unless at least one mind contains a truth such as
    the ratio of the diameter to the circumference of
    a circle back in caveman days I construe this as
    outside of the domain of knowledge at the time.

    An obvious attempt to decive (this time a change of subject) does not
    work. In order to decive one needs to behave as a honest person would.

    It seems incorrect to call anything "knowledge" that
    no one has ever known. We can call it unknown truth.

    As Aristotle pointed out, words are sometimes used in an actual sense
    and sometimes in an potential sense. In actual sense "knowledge" means something that is known, in potential sense something that can be
    known. Both are important.
    --
    Mikko
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Mikko@mikko.levanto@iki.fi to sci.logic,comp.theory,sci.math,comp.ai.philosophy on Fri Oct 23 10:17:19 2026
    From Newsgroup: sci.math

    On 05/10/2026 16:28, olcott wrote:
    On 10/5/2026 3:19 AM, Mikko wrote:
    On 01/10/2026 17:47, olcott wrote:
    On 10/1/2026 5:03 AM, Mikko wrote:
    On 29/09/2026 19:31, olcott wrote:
    On 9/29/2026 11:01 AM, Python wrote:
    No, Pete. Calling an undecidable sentence "out-of-scope" does not >>>>>> make undecidability disappear.

    For a formal theory T, "G is undecidable in T" simply means

    -a-a-a T |-/- G-a-a-a and-a-a-a T |-/- ~G

    That is a statement entirely about proofs. No model theory is
    involved.

    And G is not "out-of-scope": it is a perfectly well-formed
    sentence of the language of T. What is out-of-scope, apparently,
    is merely anything for which your proposed semantics fails to
    supply an answer.

    Within PTS any expression that is neither provable nor refutable
    never acquires any semantic meaning.

    There is one more thing one needs to consider: there is no complete
    method to find out whether a sentence or its negation can be proven.
    Consequently, it is not possible to construct a system that rejects
    all undecidable sentences.

    If back-chained inference from expression X to axioms
    has a loop it is rejected.

    Irrelevant. Proofs are finite strings and strings don't contain
    loops.

    Unprovable and irrefutable in PA entails that
    G never acquires semantic meaning in PA.

    Every formula in the language of PA has an arithmetic meaning.
    --
    Mikko
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Tristan Wibberley@tristan.wibberley+netnews2@alumni.manchester.ac.uk to sci.logic,comp.theory,sci.math,comp.ai.philosophy on Sat Nov 7 12:20:50 2026
    From Newsgroup: sci.math

    On 05/10/2026 14:28, olcott wrote:
    On 10/5/2026 3:19 AM, Mikko wrote:
    On 01/10/2026 17:47, olcott wrote:
    On 10/1/2026 5:03 AM, Mikko wrote:
    On 29/09/2026 19:31, olcott wrote:
    On 9/29/2026 11:01 AM, Python wrote:
    No, Pete. Calling an undecidable sentence "out-of-scope" does not
    make undecidability disappear.

    For a formal theory T, "G is undecidable in T" simply means

    -a-a-a T |-/- G-a-a-a and-a-a-a T |-/- ~G

    That is a statement entirely about proofs. No model theory is
    involved.

    And G is not "out-of-scope": it is a perfectly well-formed
    sentence of the language of T. What is out-of-scope, apparently,
    is merely anything for which your proposed semantics fails to
    supply an answer.

    Within PTS any expression that is neither provable nor refutable
    never acquires any semantic meaning.

    There is one more thing one needs to consider: there is no complete
    method to find out whether a sentence or its negation can be proven.
    Consequently, it is not possible to construct a system that rejects
    all undecidable sentences.

    If back-chained inference from expression X to axioms
    has a loop it is rejected.

    Irrelevant. Proofs are finite strings and strings don't contain
    loops.


    Unprovable and irrefutable in PA entails that
    G never acquires semantic meaning in PA.

    Semantic meaning is only acquired by two types
    of relations between finite strings.
    (a) Stipulated relations provide the axiomatic basis.

    (b) Semantic relations specified syntactically are
    -a-a-a the only inference step that is allowed.

    The above is the way that
    "true on the basis of meaning expressed in language"
    has always worked in actual reality.

    This applies to

    ...We are therefore confronted with a proposition which
    asserts its own unprovability. rCa (G||del 1931:40/43)

    No, it does not. The quote is about a sentence and its interpretation,
    not about any inference.

    -a-a-a It would then be possible to reconstruct the
    -a-a-a antinomy of the liar in the metalanguage, by
    -a-a-a forming in the language itself a sentence x
    -a-a-a such that the sentence of the metalanguage
    -a-a-a which is correlated with x asserts that x is
    -a-a-a not a true sentence.
    -a-a-a https://liarparadox.org/Tarski_247_248.pdf

    Maybe, maybe mot. But whther it would be possible does not matter
    as it was not done.

    -a-a ... the diagonal lemma, with B(x) instantiated
    -a-a to -4True(x), yields a counterexample to this
    -a-a equivalence, by giving a "liar" formula S such
    -a-a that S rf| -4True(g(S)) holds in N.
    -a-a https://en.wikipedia.org/wiki/Tarski%27s_undefinability_theorem

    He implements this as the basis of his whole proof.

    Formalized as:
    x ree True if and only if p
    where the symbol 'p' represents the whole sentence x https://liarparadox.org/Tarski_275_276.pdf

    adapted to become this
    x ree Pr if and only if p-a // line 1 of the proof

    Here is the Tarski Undefinability Theorem proof
    (1) x ree Provable if and only if p-a-a-a-a-a-a // assumption (see above)
    (2) x ree True if and only if p-a-a-a-a-a-a-a-a-a-a-a-a-a // Tarski's convention T
    (3) x ree Provable if and only if x ree True. // (1) and (2) combined
    (4) either x ree True or x|a ree True;-a-a-a-a-a // axiom: ~True(x) re? ~True(~x)
    (5) if x ree Provable, then x ree True;-a // axiom: Provable(x) raA True(x) (6) if x|a ree Provable, then x|a ree True;-a // axiom: Provable(~x) raA True(~x)
    (7) x ree True
    (8) x ree Provable
    (9) x|a ree Provable

    This is the final resolution to the Liar Paradox that
    conclusively proves that it never had a truth value.

    No, it has a truth value: true.

    His system is (perhaps covertly) inconsistent, explodes, and
    consequently every sentence is true and every unary predicate is
    trivially a truth predicate.

    The error is in believing that the contensive ontology that "-4" is a
    negating connective rather than a holding connective. That is a
    NONtology. While truth might be monotonic wrt system extension, clearly ontology is not; parts of an ontology may be inverted rather than merely weakened, strengthened or sustained by system extension.

    The clue is in the fact that Tarski's theorem includes the condition
    that the system has negation. By Modus Tollens we find that if it has a
    truth predicate then it does not have negation. Since we have no
    negation in the face of explosion we may have a truth predicate, and for
    any useful system we must and we already know we have no negation upon explosion.

    Clearly he has a covert rule of the sort "If 'forall P. P' Then 'P'"
    (common in the pre-war European era it seems but not assumed by Russell)
    which gives him explosion.
    --
    Tristan Wibberley

    The message body is Copyright (C) 2026 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.
    --- Synchronet 3.22a-Linux NewsLink 1.2