Talk:True arithmetic

foundations

The use of usual in the original definition shows that the set N and hence the structure N' are undefined (here). One can construct an N in Set Theory. However, some concepts about N are probably used in setting up Set Theory. There is a foundational problem involved that should at least be mentioned in the article. G. Blaine (talk) 17:51, 3 August 2017 (UTC)

Dependence on metatheory

It seems to me that we could add in the page that "true arithmetic" is not really unique: it depends on the metatheory one works in. Thus saying « true arithmetic is the set of all true first-order statements about the arithmetic of natural numbers » is misleading. For instance if our metatheory were inconsistent we could prove that 0=1 in "true arithmetic" but if it is consistent (presumably ZF for instance) then we cannot prove this arithmetic sentence; thus true arithmetic depends on the metatheory. Another example is taking a Gödel encoding (call it ${\text{Con}}_{\text{ZF}}$ of "ZF is consistent" in the language of arithmetic, this is possible as ZF is recursively axiomatized. Now ${\text{ZF}}+{\text{Con}}_{\text{ZF}}$ and ${\text{ZF}}+\neg {\text{Con}}_{\text{ZF}}$ differ on some arithmetic statements (assuming ZF is consistent). I think one can formulate $\neg {\text{Con}}_{\text{ZF}}$ as a diophantine equation $\exists xP(x)=0$, with an explicit polynomial $P$, for a more concrete realization of the above. This is an incontestably arithmetic statement which will be true or false depending on the metatheory one chooses; the above metatheories are somewhat contrived so as to produce such ambiguous arithmetic statements, but it seems that we could obtain similar ambiguous statements (what we could call metatheory-dependent arithmetic truths/statements) by comparing "true arithmetics" obtained from reasonable set theories actually used by set theorists, used as metatheories. I would have to think about it a little more to find more striking examples, worked out in careful details, but i think that from the perspective above there is no single "true arithmetic" but rather some conventional truths in arithmetic, depending on the mathematical community's practice, their preferred choice of axioms; and that the incompleteness of any set theory we currently use as metatheory leads to a certain type of "incompleteness" of "true arithmetic" -of course not in the usual logical sense of "incomplete", so we can call this new property "mathematical practice-incomplete" or "meta-incomplete". If i don't edit the page, despite my criticism, it is because i would prefer to understand better the phenomenon and ask the opinion of some logicians/set theorists, to see if they find this nuance congenial, useful. Plm203 (talk) 13:44, 27 August 2023 (UTC)