# Common root of two polynomials, method of succesive elimination

I need to find $$x$$ such that $$P(x)=x^6-2x^5+2x+1=0$$ and $$Q(x)=3x^5-5x^4+1=0$$

Of course we know that the common roots of two polynomials are the roots of their gcd, so we could simply compute the gcd.

But i found another method: we make "combinations" of the equations in order to reduce the powers of x (method of successive subtraction):

So, if $$P(x)=0$$ and $$Q(x)=0$$ then we have $$xQ(x)-3P(x)=0$$ so $$P'(x)=x^5-5x-3=0$$

Then we have $$3P'(x)-Q(x)=0$$ so $$Q'(x)=x^4-3x-2=0$$

Finally it would follow that $$P'(x)-xQ'(x)=0$$ so $$x^2-x-1=0$$

The computations are simple but I'm interested in the logic, more exactly in the "equivalence"/reversibility of these computations.

Logically, clearly if x verifies $$P(x)=Q(x)=0$$ then $$x$$ must verify the last equation $$x^2-x-1=0$$. So if there exists a solution , it is a solution to this last equation. But how do i know for sure that the solutions of this last equation are solution to the first ones??? I mean, could i use this method and be wrong, so that not all the solutions of the last equation are solutions of the first? How do i know this "algorithm" stops at the right equation? For example if i would have "stopped" at $$Q'(x)$$ and could have solved it completely (even if it is 4-th degree) not all the solutions would be solutions for both the initial polynomials. What am i missing here?

• Note that what you computed is the gcd of P and Q: $gcd(P, Q) = x^2-x-1$. What you did is similar to the euclidean algorithm, only with differences instead of remainders. – Martin R Mar 8 at 7:54
• But how do i show this is the gcd? In euclidean alg the "termination condition" is when we obtain a zero remainder. Here with differences what is the stopping condition? – amarius8312 Mar 8 at 13:18