Effectivity for existence of rational points is undecidable (Q6979233)

Hilbert's well-known 10th problem asked for designing an algorithm that recognizes whether a given Diophantine equation has integer solutions or not. Since natural numbers are definable inside integers by a Diophantine formula, the problem was reduced to natural solutions. A negative answer, that is, the nonexistence of such an algorithm, was proved by showing that every computably enumerable set is Diophantine. Since some computably enumerable sets are algorithmically undecidable, no algorithm can decide whether a given Diophantine equation has a solution in natural numbers or not. \N\NA big related problem that has resisted any solution so far is the 10th problem for rational numbers (instead of natural or integer numbers). We do not know if there exists a Diophantine definition, so-called positive existential interpretation, for natural or integer numbers in the field of rational numbers, or not. \N\NThe main result of this paper is that there is a Diophantine definition for natural numbers in the field of the rational numbers when the language is extended from the language \(\{0,1,+,\times,=\}\) of fields to contain the ``height conditions'', the predicates \(H_{m,n}\) with the interpretation that \(H_{m,n}(\vec{x},\vec{y})\) holds for \(\vec{x}\in\mathbb{Q}^m\) and \(\vec{y}\in\mathbb{Q}^n\) when \(h_m(\vec{x})\leq h_n(\vec{y})\) is true, where \(h_k(\vec{z})\) is defined for \(z\in\mathbb{Q}^k\) by \(h_k(\vec{z})=\log \max \{d,|dz_1|,\dots,|dz_k|\}\), where \(d\geq 1\) is the least common denominator of \(z_i\)'s (when \(\vec{z}=\langle z_1,\dots,z_k\rangle\)). \N\NIn fact, this result, and one of its consequences to the effect that the positive existential theory of \(\mathbb{Q}\) over the extended language is undecidable, holds when one restricts \(H_{m,n}\)'s to \(m,n\leq 3\). The proof first shows the Diophantine definability of the structure \((\mathbb{N};0,1,+,B,=)\) in \((\mathbb{Q};0,1,+,\times,=,\{H_{m,n}\}_{m,n\leq 3})\), where \(B\) is the binary relation that holds between \(x\) and \(y\) when \(x\) is a square and \(y\) is its consecutive square, i.e., when for some \(z\), we have \(x=z^2\) and \(y=(z+1)^2\). Then it is shown (in Lemma 4.4) that multiplication on \(\mathbb{N}\) is definable in \((0,1,+,B,=)\) by \(x\times y =z \) if and only if \(\text{ exists } u,v,w: \sigma(x,u)\wedge\sigma(y,v)\wedge\sigma(x+y,w)\wedge(w=u+2z+v)\), where \(\sigma(a,b)\), with the interpretation \(b=a^2\), is defined by \(\text{ exists } c: B(b,c)\wedge(c=b+2a+1)\). \N\NThe paper is technical, and reading it requires familiarity with its prerequisite references.