# When can Galois theory actually help you find the roots of a polynomial?

I was told in another post that Galois theory could help you solve solvable polynomials, and that solvable polynomials had roots that could be expressed as functions of rational numbers, but that it couldn't help me solve a polynomial with rational roots. It cannot help me solve polynomials with roots that couldn't be expressed with radicals. So doesn't that leave me with only polynomials with roots that ARE expressed in radical that Galois theory can help me solve? In other words, does a polynomial's roots HAVE to contain at least one radical for Galois theory to help you solve it?

-
Did you read the exposition I had linked you to? – user21436 Mar 7 '12 at 18:43
I tried to read it, but I can't follow it too well because I don't even know abstract algebra very well. I'm asking this question because it seems like I'm getting conflicting information from different people. I was told that Galois theory can't help you solve polynomials with rational roots (not coefficients), but that it can help you solve solvable polynomials, and finally that polynomials with rational roots are solvable. – Kenny Mar 7 '12 at 19:02
I would advise you to wait and figure it out for yourself. The point is you might not understand it well enough to know if you're getting really conflicting answers. Patience is the key here and I have nothing to say. : ) – user21436 Mar 7 '12 at 19:07
This will be an unsatisfactory and circular answer, but I think it's worth saying anyhow: Galois theory has more to do with answering your question ("when can we solve for the roots of a polynomial?") than it is about actually finding the roots of polynomials. So to really understand when Galois theory is useful for finding roots, you will have to understand Galois theory itself. – Brett Frankel Mar 7 '12 at 19:25
Here's the link to the other post: math.stackexchange.com/q/117592/22405 – Brett Frankel Mar 7 '12 at 19:28

Let's look at an example. Does Galois Theory help you solve $x^2-3x+1=0$? In a sense, it does; you could use Galois Theory to analyze that equation and eventually to write down its solutions. But you'd be insane to do that: it's ever so much simpler to just write down the answers given by the quadratic formula. Moreover, there is no way that someone who doesn't already know the quadratic formula is going to be able to understand the Galois Theory approach to this equation.