Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement

Socialist primes are defined on the page for (2.6); (⋅p)\bigl(\tfrac{\cdot}{p}\bigr) is the Legendre symbol.

Let pp be a socialist prime and H={2,3,4,…,p−3}∖{p−12}H=\{2,3,4,\ldots,p-3\}\setminus\{\tfrac{p-1}{2}\}. The paper shows (pp. 3--4):

  • (3.1) (p. 3). There is a function ff on HH with (f(k))!≡−k!(modp)(f(k))!\equiv-k!\pmod p for all k∈Hk\in H, and ff is an involution of HH.
  • Parity and quadruples (p. 4). f(k)≡k(mod2)f(k)\equiv k\pmod 2, and HH splits into (p−5)/4(p-5)/4 quadruples Uk={k, f(k), p−1−k, p−1−f(k)}U_k=\{k,\,f(k),\,p-1-k,\,p-1-f(k)\}, each with ∏x∈Ukx≡1\prod_{x\in U_k}x\equiv1 and ∑x∈Ukx≡0(modp)\sum_{x\in U_k}x\equiv0\pmod p and all members of the same parity.
  • Quadratic characters (p. 4). The Legendre symbol (x!p)\bigl(\tfrac{x!}{p}\bigr) takes the same value for all x∈Ukx\in U_k. The print words this as "all members of UkU_k have the same quadratic residue modulo pp" (p. 4); the display before it, and the use made of it, concern the factorials x!x!. Consequently (2!⋅3!⋯p−32!p)=1\Bigl(\tfrac{2!\cdot3!\cdots\frac{p-3}{2}!}{p}\Bigr)=1, and also $\Bigl(\tfrac{2!\cdot4!\cdots((p-5)/2)!}{p}\Bigr)=1= \Bigl(\tfrac{3!\cdot5!\cdots((p-3)/2)!}{p}\Bigr)$.
  • Conclusion (p. 4). (p−14!p)=1\Bigl(\tfrac{\frac{p-1}{4}!}{p}\Bigr)=1: the residue (p−14)!\bigl(\tfrac{p-1}{4}\bigr)! is a quadratic residue modulo pp.

Proof pointer

Pp. 3--4. Since the factorial residues are distinct and miss only −(p−12)!-\bigl(\tfrac{p-1}{2}\bigr)!, each −k!-k! with k∈Hk\in H is (f(k))!(f(k))! for a unique f(k)∈Hf(k)\in H. Multiplying (3.1) by (p−1−f(k))! (p−1−k)!(p-1-f(k))!\,(p-1-k)! and using the reflection (2.3) relates (p−1−k)!(p-1-k)! and (p−1−f(k))!(p-1-f(k))! with sign (−1)k+f(k)+1(-1)^{k+f(k)+1}; distinctness then forces equal parity, and the four indices k,f(k),p−1−k,p−1−f(k)k,f(k),p-1-k,p-1-f(k) close up into UkU_k. The characters agree because (−1p)=1\bigl(\tfrac{-1}{p}\bigr)=1 (as p≡1(mod4)p\equiv1\pmod4) and (x!p)=(1/x!p)\bigl(\tfrac{x!}{p}\bigr)=\bigl(\tfrac{1/x!}{p}\bigr). The last step rewrites the product of factorials as a product of odd numbers, expresses it through (p−12)!\bigl(\tfrac{p-1}{2}\bigr)!, 2(p−1)/42^{(p-1)/4} and (p−14)!\bigl(\tfrac{p-1}{4}\bigr)!, and evaluates (((p−1)/2)!p)=−1\Bigl(\tfrac{((p-1)/2)!}{p}\Bigr)=-1 and (2(p−1)/4p)=−1\Bigl(\tfrac{2^{(p-1)/4}}{p}\Bigr)=-1 using p≡5(mod8)p\equiv5\pmod 8.

Read depth

Claims checked: the statements of Section 3 were read clause by clause on the arXiv v1 print, pp. 3--4, and the argument was followed. Nothing here is independently reviewed.

Dependencies

(2.3)–(2.5), including p≡5(mod8)p\equiv5\pmod 8.

Source. V. Andrejić and M. Tatarevic, On distinct residues of factorials, arXiv:1603.04086v1 (2016); published in Publ. Inst. Math. (Beograd) (N.S.) 100(114) (2016), 101--106. Labels and pages here are those of the arXiv v1 print; the edition read is named on the source card.

Bears on

  • Problem 478: more necessary structure for the extreme case ∣Ap∣=p−2\lvert A_p\rvert=p-2 (socialist primes; see the page for (2.6)); nothing about ∣Ap∣\lvert A_p\rvert in general.