1
$\begingroup$

I find it hard to keep overview over the notations for sets of polynomials, evaluations of polynomials, and extensions of rings and so on.

Let $R$ be a ring, and $E/R$ an extension of $R$.

Occasionally one finds definitions like this

$$\boxed{R(\alpha) = \{ a_0 + a_1\alpha\ |\ a_i \in R\}}$$

for a specific number $\alpha \in E/R$ , i.e. $R(\alpha) \subseteq E/R$ is another set of numbers, for example in Lemmermeyer's Quadratische Zahlkörper:

enter image description here

On the other side, the ring of polynomials over $R$ is defined by

$$\boxed{R[X] = \{a_0 + a_1X + \dots + a_nX^n\ |\ n \in \mathbb{N}, a_i\in R\}}$$

with a unspecific variable $X$, i.e. $R[X]$ is a set of "open" algebraic expressions.

With

$$\boxed{R^k[X] = \{a_0 + a_1X + \dots + a_kX^k\ |\ a_i\in R\}}$$

one can write

$$\boxed{R[X] = \bigcup_{k=0}^{k< \infty}R^k[X]}$$

Next to $R[X]$ we find the definition of

$$\boxed{R(X) = \Big\{ \frac{P}{Q}\ |\ P, Q \in R[X], Q \neq 0\Big\}}$$

If $R[X]$ is a set of algebraic expressions (with a variable $X$), $R[\alpha] = R[X \rightarrow \alpha]$ is a set of numbers by setting the variable $X$ to the numerical value $\alpha$:

$$\boxed{R[\alpha] = \{a_0 + a_1\alpha^1 + \dots + a_n\alpha^n\ |\ n \in \mathbb{N}, a_i\in R\}}$$

or

$$\boxed{R[\alpha] = \{P(X\rightarrow \alpha)\ |\ P \in R[X]\}}$$

Accordingly:

$$\boxed{R[\alpha] = \bigcup_{k=0}^{k < \infty}R^k[\alpha]}$$

Occasionally:

$$\boxed{R[\alpha] = R^1[\alpha] = R(\alpha)}$$

e.g. for $R = \mathbb{Q}$, $\alpha=\sqrt{2}$ (more generally: $\alpha = \sqrt{d}$, $d$ square-free). [Thanks to user Servaes.]

Things would be easier if a consistent notation was used, especially $R^1[\alpha]$ instead of $R(\alpha)$ for the set $\{ a_0 + a_1\alpha\ |\ a_i \in R\}$.

Finally, there is one notation I'm desperately missing: for the minimal extension of a ring $R$ that contains all roots of a polynomial $P(X) \in R[X]$:

$$\boxed{R\langle P\rangle = R[\rho_1]\dots[\rho_k]}$$

for the roots $\rho_i$ of $P(X)$, i.e. $P(\rho_i) = 0$. It's essential (or isn't it?) that

$$\boxed{R[\rho_1]\dots[\rho_k] = R[\pi(\rho_1)]\dots[\pi(\rho_k)]}$$

for any permutation $\pi$ of the roots.

Note that $\mathbb{Z}[\frac{1}{2}]$ is an infinite-dimensional vector space over $\mathbb{Z}$ (with base $\{1,\frac{1}{2},\frac{1}{4},\frac{1}{8},\dots\}$, while $\mathbb{Z}^1[\sqrt{2}] = \mathbb{Z}[\sqrt{2}] $ is a two-dimensional one (with base $\{1,\sqrt{2}\}$) - like $\mathbb{Z}[\sqrt{-1}]$ (with base $\{1,i\}$)!

My question is:

Are there attempts to unify (= optimize) notation in the context of polynomials, evaluation of polynomials, and extensions of rings and fields that minimize confusion (that occasionally arises with established notation)?

$\endgroup$
  • $\begingroup$ Note that in the case of a squarefree integer $m$ we have $$\Bbb{Q}(\sqrt{m})=\Bbb{Q}[\sqrt{m}]=\Bbb{Q}^1[\sqrt{m}],$$ hence the notation is entirely consistent. As for the minimal extension containing all roots of a polynomial; I've often seen the notation $\Omega_P^R$ used in the context of field theory, but I guess it works just as well for rings. $\endgroup$ – Servaes Nov 3 '18 at 16:17
  • $\begingroup$ "You've seen": just occasionally? Isn't there an established notation? $\endgroup$ – Hans-Peter Stricker Nov 3 '18 at 16:31
  • $\begingroup$ ...do you have a question? $\endgroup$ – Eric Wofsey Nov 3 '18 at 17:01
  • $\begingroup$ @EricWofsey: Very good question! I forgot about it. $\endgroup$ – Hans-Peter Stricker Nov 3 '18 at 17:03
  • $\begingroup$ Each math author is welcome to define a notation (providing a definition) that "optimizes" the discussion of a topic. At the beginning of your post you have "$R(\alpha) = \{a_0 + a_1 \alpha \mid a_0,a_1 \in R \}$" which already involves some abbreviation that might cause confusion. This construction of a ring $R(\alpha)$ depends on some overring $S$ of $R$ where the arithmetic $a_0 + a_1 \alpha$ is already defined, and crucially, on $\alpha$ satisfying a monic quadratic polynomial over $R$. So it's helpful if those points are clear to a Reader, not so helpful otherwise. $\endgroup$ – hardmath Nov 4 '18 at 16:21
0
$\begingroup$

The definitions of $R(\alpha)$, $R[\alpha]$ an $R^k[\alpha]$ you give are all unambiguous and in widespread use (though $R_k[\alpha]$ is also used for the latter), so there is no ambiguity or possibility of confusion to be prevented. Note that in your particular case of a squarefree integer $m$ we have $$\Bbb{Q}(\sqrt{m})=\Bbb{Q}[\sqrt{m}]=\Bbb{Q}^1[\sqrt{m}],$$ hence the notation is entirely consistent.

As for the minimal extension containing all roots of a polynomial; I've often seen the notation $\Omega_P^R$ used in the context of field theory, but I guess it works just as well for rings.

$\endgroup$
  • $\begingroup$ I totally disagree that $R^k[\alpha]$ (or $R_k[\alpha]$) are "unambiguous and in widespread use". I have pretty much never seen either of those notations. And lacking context that indicates otherwise, I would assume they instead mean $S[\alpha]$ where $S=R^k$ or $S=R_k$ is a different ring. $\endgroup$ – Eric Wofsey Nov 3 '18 at 18:48

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.