3

Let $R = \mathbb{Z}/(p^{k}\mathbb{Z})$, where $p$ be any prime number, and $k > 1$ be any integer. Now let us consider an equation $x^r = p$ in $R$ and $\pi$ be the root of this equation, where $r \neq 1$. Now, it is clear that $\pi \notin R$, because if we assume $\pi \in R$, then $$\pi^r = p \\ \implies v_p(\pi^r) = v_p(p) \\ \implies v_p(\pi) = \frac{1}{r}$$

Now note that $R = \{x \in \Bbb{Z}_p : x = a_0 + a_{1}p + \cdots a_{k-1}p^{k-1} ~\&~ 0 \leq a_{i} < p \}$. So $v_{p}(R^\times) = \Bbb{N} \cup \{0\}$. Hence, our assumption is wrong and $\pi \notin R$.

I want to create an extension ring by adjoining $\pi$ to this ring $R$. I am a little confused with how the ring looks after adjoining an element. I read from Artin's algebra, and according to this book, we can denote the extension ring as $R[\pi]$ where $R[\pi] = \{f(\pi): f \in R[x]\}$.

Can we write something like $R[\pi]$ is the same as the ring $R[x]/\langle x^{r} - p\rangle$? My question is, are these two notations the same? If isomorphic, then how can I prove this isomorphism?

One can consider one homomorphism $\phi :R[x] \to R[\pi]$, by $\phi(f(x)) = f(\pi)$. Can we show that $\operatorname{Ker}(\phi)=\langle x^{r} - p\rangle$?

I want to understand how to adjoin an element to a ring that is not a PID or a field (for example $\mathbb{Z}/(p^{k}\mathbb{Z})$).

Edit: After all this discussion and from the answer, I can write that: If we assume $\pi = x + \langle x^r - p \rangle$, then my $\pi$ will not be ambiguous.

Afntu
  • 2,393

1 Answers1

7

The notation $R[\pi]$ is unclear because you haven't defined the meaning of this symbol $\pi$; this could be interpreted as referring to the polynomial ring. The unambiguous way to adjoin an element satisfying $x^r = p$ is to write $R[x]/(x^r - p)$, although you could probably also informally write $R[\sqrt[r]{p}]$ and be understood.

because we can't claim that, the kernel is an ideal so the principal ideal and generated by the smallest power polynomial.

This is irrelevant. This ring is just defined as the quotient by a principal ideal. It doesn't matter that there are some other ideals that are not principal.

Note that whether this constitutes an "extension ring" of $R$ is less clear if $R$ is not a field, since in general there is no guarantee that the natural map from $R$ is injective. In fact if you try to adjoin some arbitrary elements satisfying some arbitrary relations you may get the zero ring. So you have to check that it's nonzero if you want to get any use out of it.

Qiaochu Yuan
  • 468,795
  • Thanks for this answer, but I have some questions. To attach the element, we need to specify what is the element $\pi$, but to add an element satisfying $x^r = p$ is to write $R[x]/(x^r - p)$ is ok. I mean, how to prove this, or is this the definition? – Afntu Jun 22 '24 at 06:15
  • It would be better to say $\pi = p^{1/r}$, and then we can define clearly $R[\pi]$? – Afntu Jun 22 '24 at 06:17
  • Also, if possible, suggest some texts or material from where I can understand how to attach an element to a ring, which I asked in the question. – Afntu Jun 22 '24 at 06:21
  • 2
    Just to add to this good answer, even in the case of fields, these symbol strings where you supposedly adjoin a certain "root" of an element are just not well-defined. Read https://math.stackexchange.com/q/4231699/96384 for how even "$\mathbb Q_5(\sqrt[3]{2})$" is ambiguous. – Torsten Schoeneberg Jun 23 '24 at 05:29
  • 2
    To clarify @Torsten's remark it may be ambiguous only if one seeks to adjoin a specific (vs. generic) root of a polynomial. – Bill Dubuque Jun 23 '24 at 05:48
  • So, can we say that $\pi$ is always in the higher ring $R[x]/\langle x^r - p \rangle$. So, we can consider this ring to be the ring extension of $R$. But we can't define $R[\pi]$, as $\pi$ is not clear, so also we can't say that $R[\pi]$ is isomorphic to $R[x]/\langle x^r - p \rangle$. – Afntu Jun 23 '24 at 10:17
  • 1
    Once you have the quotient ring, you can just define $\pi$ to be the class of $x$ and then the homomorphism you proposed does work I think – Keplerto Jun 23 '24 at 11:12