I was reading Emil Artin's Galois Theory (2nd edition). On pp.39-40, Artin defines a number of things as follows.

Let $k$ be a field and $E=k(x_1,\ldots,x_n)$ be the field of all rational functions of $n$ indeterminates $x_1,\ldots,x_n$. Define a polynomial $$ f(t)=(t-x_1)\cdots(t-x_n)=t^n+a_1t^{n-1}+\cdots+a_n,\tag{3} $$ where $a_1=-\sum_ix_i,\,a_2=\sum_{i<j}x_ix_j$,... are elementary symmetric functions in $x_1,\ldots,x_n$. Then he puts $F_n(t)=f(t)$ and also for $i=n-1,\,n-2,\ldots,1$, $$ F_i(t)=\frac{f(t)}{(t-x_{i+1})(t-x_{i+2})\cdots(t-x_n)}=\frac{F_{i+1}(t)}{t-x_{i+1}}.\tag{5} $$ (If I am not mistaken, this simply means that $F_i(t)=(t-x_1)(t-x_2)..(t-x_i)$ for each $i$, but what matters below is the division on the RHS of $(5)$.) He then writes that (emphasis mine):

Performing the division [in $(5)$] we see that $F_i(t)$ is a polynomial in $t$ of degree $i$ whose highest coefficient is $1$ and whose coefficients are polynomials in the variables $a_1,a_2,\ldots,a_i$ and $x_{i+1},x_{i+2},\ldots,x_n$. Only integer enter as coefficients in these expressions.

Can someone explain the bold part? I have thought about using Euclidean algorithm, but the $a_i$s are polynomials in $x_1,\ldots,x_n$ rather than constants. I am not sure how the division is done.


You just perform polynomial division as if the symbols were numbers. For example, you can check that $$ (t^3+a_1t^2+a_2t+a_3)\div(t-x_3)=t^2+(a_1+x_3)t+[(a_1+x_3)x_3+a_2] $$ and $$ \left(t^2+(a_1+x_3)t+[(a_1+x_3)x_3+a_2]\right)\div(t-x_2)=t+(a_1+x_2+x_3). $$

  • $\begingroup$ That makes a lot of sense. Thank you very much! $\endgroup$ – William McGonagall Jan 23 '19 at 5:35

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.