I am reading Section 7: Groups and Homomorphisms, Chapter 1: Foundation, textbook Analysis I by Herbert Amann and Joachim Escher.

First of all, I am so sorry for posting many screenshots. Since the information is too complicated for me to summarize, I have no way but to do so. There is a proof of Remark 8.20(c) that I could not understand. I have been stuck at this proof for two weeks in spite of re-reading the proof many times. Please help me get over it!

enter image description here enter image description here

I got stuck at the below arguments in Remark 8.20(c).

Clearly, $p = \sum_{\alpha} p_\alpha X^\alpha$ can be written in the form $$\sum_{j=0}^n q_j X^j_m$$ for suitable $n \in \Bbb N$ and $q_j \in K[X_1,\cdots,X_{m-1}]$. This suggests a proof by induction on the number of indeterminates: For $m = 1$, the claim is true by Remark 8.19(d).

Here is the Remark 8.19(d):

enter image description here

where the homomorphism (8.22) is

enter image description here

My questions:

  1. For $m=1$, $m-1=0$. What is $K[X_0]$? The index of $X$ starts from $1$ and the authors say $K[X_1,\cdots,X_{m-1}]$.

  2. What is $X^j_m$?

Previously, the authors define

enter image description here

In my understanding, $X^j_m = \begin{cases}1, &j=m \\ 0, &j\neq m \end{cases}$. As a result, $X^j_m \in K$.

  1. I can not see how Remark 8.19(d) helps to prove the case where $m=1$. Please elaborate on this points!

Thank you for your help!

  1. The induction argument using $K[X_1,\dots,X_{m-1}]$ is only used for $m>1$. If you had to interpret it when $m=1$ you would probably say that $K[X_1,\dots,X_{m-1}]=K$, because the indexing set is empty in that case, but this is not necessary here, you can just ignore it when $m=1$ if it confuses you.

  2. You are mixing different notations (granted, you are not helped by the notations in the book which are incredibly confusing in my opinion). When they define $X_\beta^\alpha$ as $0$ or $1$, what they mean is that $X^\alpha$ is a function from $\mathbb{N}^m$ to $R$, and that its value on $\beta\in \mathbb{N}^m$ is $1$ when $\alpha=\beta$. On the other hand, $X_m^j$ is just the $j$th power of the element $X_m$ in the ring $R[X_1,\dots,X_m]$. If you want to be really formal, $X_m$ is the element $X^\alpha$ for $\alpha=(0,\dots,0,1)$, with the $1$ in the $m$th slot. So if you really want to see $X_m^j$ as a function $\mathbb{N}^m\to R$ (which I don't think is not such a good idea), it is the function which sends $\beta\in \mathbb{N}^m$ to $1$ if $\beta = (0,\dots,0,j)$, and to $0$ otherwise.

  3. I'm not sure I understand this question, remark 8.19(d) is exactly the case $m=1$, so I would think the link is quite clear. The morphism described in (8.22) is precisely the morphism described in (8.30) when $m=1$ (since $R[X]$ is $R[X_1,\dots,X_m]$ when $m=1$), so saying that it is injective (which is remark 8.19(d)) is the case $m=1$ of remark 8.20(d).

| cite | improve this answer | |
  • $\begingroup$ Thank you so much for your detailed answer! As you explained, $X^j_m \in K[X_1,\dots,X_m]$. As the authors mention, $q_j \in K[X_1,\cdots,X_{m-1}]$. As a result, $q_j$ and $X^j_m$ belong to two different rings of polynomials. I could not understand how the product $q_j X^j_m$ makes sense. Please elaborate more on this point! $\endgroup$ – LAD Apr 20 '19 at 14:58

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.