(I know the existence of Galois group of $x^4-2$, so please don't start tagging this question as duplicate)

The degree of the field extension is 8, so the only group that the Galois group can be isomorphic to is $D_4$, the symmetry group of the square.

Now, I know about the definition of $D_4$, which is $D_4= \langle a,b\mid a^4=1,b^2=1,ba=a^{-1}b\rangle$. Why then are the elements of $D_4$ $a, a^2, a^3, b, ab, a^2b, a^3b$ and id? Would it make any sense to make a Cayley table of the automorphisms and $D_4$? How do I know that all the automorphisms are in $D_4$?

Furthermore, how can I know the Galois group is isomorphic to $D_4$ in another way?

I know it's a lot of questions, but could someone please answer something, it might clear up a lot for me. :)

  • $\begingroup$ The question, "How can I know the Galois group is isomorphic to D4 in another way?" really does seem to be a duplicate of the indicated question, and so perhaps it should be excised. The question "Why then are the elements of $D_4$ $a, a^2, a^3, b, ab, a^2b, a^3b$ and id?" also is a duplicate (or near-duplicate) of a question that has been asked in the last few days, and so probably should be removed, too. $\endgroup$ Oct 31 '15 at 22:55
  • $\begingroup$ As for the remaining questions, one could certainly identify elements $a, ab,$ etc., of $D_4$ with explicit automorphisms in $\operatorname{Gal}(x^4 - 2)$, and indeed this is probably instructive if you haven't done that sort of exercise before. But this identification is not canonical---it's only defined up to automorphism of $D_4$. (Here, $\operatorname{Aut}(D_4) \cong D_4$, but NB the analogous statement is not true for most dihedral groups.) $\endgroup$ Oct 31 '15 at 23:00

One of the many questions in your post is why the elements of $D_4$ are those 8 expressions in $a$ and $b$? From your definition of $D_4$ (as presentation) it is clear that elements are expressions of the type: $$a^{k_1}ba^{k_2}b\ldots ba_{k_n}\qquad(*)$$ where $k_j$'s are integers between 1 to 3. Allowing $k_1$ and $k_n$ to be zero takes care of generating expressions starting/ending in $b$.

Now the condition $ba= a^{-1}b$ can be used straighten all the expressions: that is, if $b$ happens to be ahead of a we can make the $a$ to come ahead, but at a price. The should be replaced by its inverse, which is the same as $a^3$. Thus all the $a$'s can migrate to the left leaving all the $b$'s to the right. As $b^2$ is is identity, an expressions of the type $(*)$ with en even number of $b$'s will simplify to a mere power of $a$, and expressions with an odd number of $b$'s end as a power of $a$ followed by a single $b$.


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.