Let $R$ be an integral domain which is not a field.

Does $R$ necessarily have an irreducible element?

I suspect the answer is no, but I couldn't find an example showing that...

  • 3
    $\begingroup$ On the other hand, the answer is yes if $R$ is noetherian. $\endgroup$ – user18119 Feb 7 '13 at 21:51
  • 1
    $\begingroup$ Because if $R$ is Noetherian then the family of principal ideals has a maximum $(a)$, and thus $(a)$ is irreducible? $\endgroup$ – Lior B-S Feb 8 '13 at 7:30
  • 2
    $\begingroup$ Yes, this is correct. One can even show that any non-zero element is a finite product of irreducible elements in a similar way. $\endgroup$ – user18119 Feb 8 '13 at 14:47

It is proven here that the ring of all algebraic integers has no irreducible elements.


Hint $\ $ There are rings, not fields, closed under square roots (e.g. the ring of all algebraic integers). In such a ring there are no irreducibles since every element factors: $\rm\ a = \sqrt{a} \sqrt{a}.$

A concrete example: $ $ let $\rm\, D = \Bbb Q[x,x^{1/2},x^{1/4},x^{1/8},\ldots].\,$ Every element $\ne0$ is writable in the form $\rm\ d_i x^{i}\! + d_j x^j\! +\cdots\! + d_k x^k\! = x^i\, f\ \,$ for $\rm\,\ f\in D,\,\ d_i\in\Bbb Q,\,\ d_i\ne 0,\:$ and $\rm\:i < j < \ldots < k,\:$ e.g.

$$\rm a\, x^{1/8}\! + b\, x^{1/4}\! + c\, x^{1/2} =\, x^{1/8} (a + b\, x^{1/8}\! + c\, x^{3/8})\, =\, x^{1/8}\, f,\quad a\ne 0$$

All exponents of $\rm\,x\,$ in terms of $\rm\,f\,$ remain of form $\rm\,n/2^m\!,\,$ being differences of such. We can force all the cofactors $\rm\,f\,$ to be units by adjoining inverses of all such elements, as follows. Since $\rm\,f\,$ has "constant" term $\rm\:a\ne 0,\,\ f\,$ is not in the maximal ideal $\rm\, M = (x,x^{1/2},x^{1/4},x^{1/8},\ldots).$ Therefore, localizing at $\rm\,M,\,$ i.e. adjoining inverses of all elements $\rm\not\in M,\,$ turns all the $\rm\,f\,$ cofactors into units, yielding a domain $\rm\,\bar D,\,$ not a field $\rm(x^{-1}\!\not\in \bar D),\,$ where all elements have form $\rm\,x^k\, u,\,$ for some unit $\rm\,u.\,$ Since $\rm\ a = x^k = (x^{k/2})^2 = \sqrt{a}\sqrt{a},\ $ this domain has no irreducibles.

  • $\begingroup$ Of course YACP example, the ring of algebraic integers, is closed under square roots... $\endgroup$ – Lior B-S Feb 8 '13 at 7:31
  • 1
    $\begingroup$ This is also remarked in the first sentence of this answer. $\endgroup$ – Martin Brandenburg Feb 8 '13 at 10:19

Another class of examples is given by valuation rings whose value group is divisible.

  • $\begingroup$ Sorry, I don't quite understand your first paragraph. How do you get, for example, a holomorphic square root of $f(z)=z$ on an open disk around zero? $\endgroup$ – Nils Matthes Feb 7 '13 at 17:33
  • $\begingroup$ You are right, this doesn't work. The answer by Math Gems contains what I wanted to say. $\endgroup$ – Martin Brandenburg Feb 7 '13 at 20:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.