Erdős Problem #674
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: proved (not automatically the live result).
Subject areas: number theory.
Read this status correctly
The status above is reproduced from GCL's immutable external-catalogue snapshot. It is not a new claim of current mathematical resolution, a source-semantic review, an independent proof check, or GCL certification. A published proof and a GCL replay are separate events.
The question in plain language
Can three integers greater than one satisfy the surprising equation
Erdős Problem #674 asks about integer solutions. The answer to the existence question is yes: explicit solutions, indeed an infinite family, were known from Ko's work. The deeper historical issue of classifying all nontrivial solutions is different from finding one.
A small example
The maintained catalogue gives the exact integers
All three bases involve only the prime factors \(2\) and \(3\). To check the enormous-power equation we therefore need only match prime valuations.
The exponent of \(2\) in \(x^x y^y\) is \(12x+8y=49\,268\,736\), while in \(z^z\) it is \(11z=49\,268\,736\). The exponent of \(3\) in the left-hand side is \(6x+8y=31\,352\,832\), while in the right-hand side it is \(7z=31\,352\,832\). Because no other prime divides either side, equality of these two valuations proves the entire identity exactly.
What is known
The catalogue attributes infinitely many examples to Ko. Its parameterization can be put in a readable form. Choose an integer \(r\geq2\) and let
Define
All quantities are positive integers. The condition \(r\geq2\) ensures \(v-r\geq0\). For \(r=2\), \(u=4\), \(v=3\) and \(T=2^8 3^6\), recovering the concrete example above.
A direct verification of the family. The desired equation is equivalent, by taking logarithms, to
Substituting the three definitions and dividing by the positive number \(T\) gives
Since \(u-v=1\), \(u=2^r\) and
the expression becomes zero term by term. Exponentiation proves \(x^x y^y=z^z\) for every \(r\geq2\). This is an independent elementary verification of the existing catalogue-listed formula, not a newly discovered infinite family.
An additional property follows immediately:
This relationship describes the Ko examples, but does not imply that every possible solution has it.
The source reports partial structural results by Schinzel, Mills, Dem'janenko and Uchiyama, including restrictions involving \(4xy\) compared with \(z^2\) and the prime divisors of \(x,y,z\). Their precise hypotheses matter. Results classifying certain regimes cannot automatically be combined into a complete classification without inspecting their original statements and proofs.
GCL's current state
The authenticated source return for original issue #1945 checks the explicit prime-exponent equality and separates existence from classification. It retains the maintained catalogue's external PROVED (LEAN) status while recording that GCL's pinned Formal Conjectures provider contains no admitted #674 statement file and that no external Lean proof was replayed here.
This is an unaccepted draft documenting an established result. A correct elementary proof of the stated Ko family does not certify the unresolved historical classification questions.
What remains
The first documentary task is direct inspection of Ko's original 1940 paper and the follow-up Schinzel, Mills, Dem'janenko and Uchiyama papers. Produce an exact table of what each theorem asserts, especially its assumptions on coprimality, common prime factors, and the cases \(4xy<z^2\), \(4xy=z^2\) and \(4xy>z^2\).
If formal certification is required, locate an exact external Lean project, its dependencies and axioms; compare the theorem target against the verified infinite family and do not confuse the existence result with a universal classification assertion.
The article requires independent source-critical review, a protected production merge, browser mathematics rendering and separate V2 admission before it becomes part of the accepted Compendium.
Sources and claim boundaries
-
Maintained Erdős Problem #674, published example, Ko family, historical classification references and external formal status.
-
Catalogue revision history, preserving the original formula and later updates.
-
GCL issue-bound source audit #674, independently verified finite prime valuations.
-
Ko [Ko40], Schinzel [Sc58], Mills [Mi59], Dem'janenko [De75b], Uchiyama [Uc84]: original paper pages not independently inspected.
Boundary: explicit known infinite-family formula verified algebraically, but no novel solution, complete classification theorem, GCL Lean certificate, protected publication, deployed MathJax readback or editorial admission.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1945 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
- Formal-statement coverage: No statement was confirmed in GCL's pinned formal-provider snapshot. Presence does not mean the statement or proof has been checked.
- External mathematical status: consult current problem page and its dated bibliography.