• Re: Provability in PA

    From Mikko@mikko.levanto@iki.fi to sci.logic,comp.theory,comp.ai.philosophy,sci.math.symbolic,sci.math on Tue Oct 6 11:32:26 2026
    From Newsgroup: comp.ai.philosophy

    On 01/10/2026 17:17, olcott wrote:
    On 10/1/2026 4:54 AM, Mikko wrote:
    On 29/09/2026 18:48, olcott wrote:
    On 9/29/2026 4:25 AM, Mikko wrote:
    On 28/09/2026 15:10, olcott wrote:
    On 9/28/2026 4:31 AM, Mikko wrote:
    On 27/09/2026 05:59, olcott wrote:
    On 9/24/2026 6:24 AM, Mikko wrote:
    On 23/09/2026 16:39, olcott wrote:
    On 9/23/2026 4:45 AM, Mikko wrote:
    On 22/09/2026 18:01, olcott wrote:

    Provability in PA requires a sequence of syntactic
    inference steps that remain totally in PA and do not
    leap outside of PA into any model of arithmetic.

    That is not specific to PA. That is what formal provability means >>>>>>>>>> for every formal theory.

    If that was true then everyone would understand that
    Wittgenstein is correct and G||del is wrong.

    That is clearly false. The meaning of the word "provability" (or >>>>>>>> amy
    word) has no consequences beyond meanings of words.

    The rest of the quoted message and my response to it are irrelevant >>>>>>>> to my comment above.

    -a-a-a 'True in Russell's system' means, as was said:
    -a-a-a proved in Russell's system; and 'false in Russell's
    -a-a-a system' means: the opposite has been proved in
    -a-a-a Russell's system.
    https://www.liarparadox.org/Wittgenstein.pdf

    With that defimition the words "true" and "false" lose a part of >>>>>>>> their
    useful meanings. In order to say "proved in Russell's system" >>>>>>>> one can
    say "proved in Russell's system". A statement proved in Russell's >>>>>>>> system can also be called a "theorem of Russell's system".

    G||del's result is only achieved by going outside of Peano
    Arithmetic into the standard model of Arithmetic.

    No, it is only necessary to go to some system where the results >>>>>>>> make
    sense. G||del's most important results are not about numbers but >>>>>>>> about
    formal theories. A map of Italy does not help if you have problems >>>>>>>> with finding your way in Tokyo.
    Proof Theoretic Semantics does this differently:
    -a-a-a What is the appropriate notion of truth for sentences >>>>>>>>> -a-a-a whose meanings are understood in epistemic terms
    -a-a-a such as proof or ground for an assertion? It seems
    -a-a-a that the truth of such sentences has to be identified >>>>>>>>> -a-a-a with the existence of proofs or grounds.

    Truth as an Epistemic Notion --- Dag Prawitz (2011)
    https://link.springer.com/article/10.1007/s11245-011-9107-6

    Truth theoretic semantics is not the kind of semantics we
    usually want.
    Many useful things can be done with no semantics at all.

    "true on the basis of meaning expressed in language" is
    reliably computable for the entire body of general knowledge.
    There is no coherent sense of "incompleteness" or undecidability >>>>>>> within such a system.

    Forst of all, that is a change of topic, apparently in order to make >>>>>> people to forget what you said earlier. WHich is in fact a good idea, >>>>>> there is no need to remember your errors. But it is useful to
    remember
    that most of what you say is erroneous.

    Whether "true on the basis of meaning expressed in language" is
    computable depends on a language. For an uninterpreted formal
    language nothing in the language is true, so the computation is
    trivial.

    The body of general knowledge expressed in formalized English
    can be formalized in CycL. That "cats are animals" is one
    example of "true on the basis of meaning expressed in language"

    An ordinary dictionary like https://en.wiktionary.org/wiki/cat can
    tell that cats are animals. If you want to get anybody interested
    you need more impressive examples.

    Just that one example refutes Quine's
    Two Dogmas of Empiricism
    https://www.ditext.com/quine/quine.html

    Quine's "Two Dogams of Empiricism" is an example of what most
    people consider uninteresting.

    Math and logic people don't care about this because they
    take alternative foundations as inherently incorrect.

    Math and logic people don't take alternative foundations as incorrect.

    Most philosophers take Quine ideas to be an established
    verified fact, thus shutting off making
    "true on the basis of meaning expressed in language"
    reliably computable before it has a chance to begin.

    Philosophers are looking for wisdom. One should not care too much about
    what say those who havn't found yet.

    Anyway, "true on the basis of meaning expressed in language" does
    not cover the truths we usually need to know. An example is the
    whether tomorrow: it is not computable. A prediction can be computed >>>>>> but sometimes the truth is different.

    The scope of the body of general knowledge that is
    "true on the basis of meaning expressed in language"
    Tomorrow's weather is outside of this scope.

    And so are most of the things that people might want to find out.

    The best that any system of knowledge can possibly
    provide is elements of the body of knowledge. Requiring
    psychic ability is a bad requirement.

    Most people care about real world knowledge, which tends to be
    uncertain but is often useful anyway.

    I am not sure what you mean when we limit the
    scope to the body of general knowledge. Unknowns
    are definitely excluded.

    Uncertain is not quite the same as unknown. The difference can be
    quantified with Shannon's information theory.
    --
    Mikko
    --- Synchronet 3.22a-Linux NewsLink 1.2