Homepage Solution manuals Tom Leinster Galois Theory Exercise 1.1.3 (Conjugacy)

Exercise 1.1.3 (Conjugacy)

Let z,z . Then z and z are conjugate over if and only if z = z or z = z¯. Both proofs of ‘if’ contain little gaps: ‘It follows by induction’ in the first proof, and ‘it’s easy to see’ in the second. Fill them.

Answers

Solution: We show both both parts (i) and (ii) seperately

(i)
Follows from induction that for any polynomial p over , p(w)¯ = p(w¯):
Let p(w) = c0 + c1w1 + c2w2 + ... + cnwn where wn and cn .
p(w)¯ = c0 + c1w1 + c2w2 + ... + cnwn¯ = c0¯ + c1w1¯ + c2w2¯ + ... + cnwn¯ = c0 + c1w1¯ + c2w2¯ + ... + cnwn¯ = p(w¯)
(ii)
Checking that r is the zero polynomial:
Lemma. If r(x) = a0 + a1x1 + ... + anxn and r(x) = 0,x0, then a0 = 0. Since is a field, and must contain a zero.

By this lemma we have that x, r(x) = 0 and therefore
r(x) = 0 = x(a1 + a2x + ... + anxn1) = a1 + a2x + ... + anxn1 and x0 a1 = 0. Hence, we repeat the Lemma and show that all a1,...,an = 0. Therefore r(x) is the zero polynomial.

User profile picture
HN
2022-10-25 23:13
Comments
  • (i) is correct, but Leinster asked for an induction proof. In (ii), $z$ is a fixed complex number, so the solution is not related to this lemma. Moreover, the degree of $r$ is less than $2$ : that is the key (see the other solution).
    richardganaye2024-05-20

Proof.

(i)
We prove by induction that, for any polynomial p ( t ) [ t ] , and for any complex number w , p ( w ) ¯ = p ( w ¯ ) . Consider the proposition, where n = 0 , 𝒫 ( n ) : p [ t ] , deg ( p ) n w , p ( w ) ¯ = p ( w ¯ ) .

𝒫 ( 0 ) : If deg ( p ) 0 , where p [ t ] , then p = a is a constant complex. Thus p ( w ) ¯ = a ¯ = p ( w ¯ ) . This shows that 𝒫 ( 0 ) is true.

Assume now the induction hypothesis 𝒫 ( n ) : for any polynomial p such that deg ( p ) n , and for any complex w , p ( w ) ¯ = p ( w ¯ ) .

(We suppose that the induction proof of w n ¯ = w ¯ n ( w , n ) is already done :

w 0 ¯ = 1 ¯ = 1 = w ¯ 0 ; if w n ¯ = w ¯ n for some n , then w n + 1 ¯ = w w n ¯ = w ¯ w n ¯ = w ¯ w ¯ n = w ¯ n + 1 .)

Let now q be a polynomial of degree n + 1 . Then q ( t ) = a t n + 1 + p ( t ) , where a , p [ t ] satisfies deg ( p ) n ( t is a variable). Using the induction hypothesis on p , we obtain for any w ,

q ( w ) ¯ = a w n + 1 + p ( w ) ¯ = a w n + 1 ¯ + p ( w ) ¯ = a w ¯ n + 1 + p ( w ¯ ) (induction hypothesis) = q ( w ¯ ) .

This proves 𝒫 ( n + 1 ) , and the induction is done.

To conclude,

p [ t ] , w , p ( w ) ¯ = p ( w ¯ ) .

Therefore, if p [ t ] , and w , p ( w ) = 0 p ( w ¯ ) = 0 .

(ii)
Write z = x + iy with x , y , some fixed complex nomber. Let p [ t ] such that p ( z ) = 0 . We will prove that p ( z ¯ ) = 0 . If y = 0 , then p ( z ¯ ) = p ( x ) = p ( z ) = 0 , so we can suppose that y 0 .

Let m ( t ) = ( t x ) 2 + y 2 [ t ] . Then m ( z ) = 0 . The Euclidean division of p ( t ) by m ( t ) gives

p ( t ) = m ( t ) q ( t ) + r ( t ) , q ( t ) [ t ] , r ( t ) [ t ] , deg ( r ) < 2 .

Thus r ( t ) = at + b , a , b . Since p ( z ) = 0 and m ( z ) = 0 , we obtain az + b = 0 , that is

0 = a ( x + iy ) + b = ( ax + b ) + iay .

Here ax + b , y . This implies ax + b = 0 and ay = 0 . Since y 0 , we obtain a = 0 , and b = ay = 0 . Therefore r ( t ) = at + b = 0 is the null polynomial. This gives p ( t ) = m ( t ) q ( t ) . But m ( z ¯ ) = 0 , so p ( z ¯ ) = 0 .

User profile picture
2024-05-20 10:28
Comments
  • Thanks for your corrections on this and my other posts Richard. As there is no content moderation on this website, I will leave up my solutions, though there may be further errors. Hopefully your corrections serve as a benefit to any future readers. Cheers.
    HN2024-05-24
  • Your message is encouraging. THANKS. Errors become part of the history of the exercise and can be retained. I too could be wrong.
    richardganaye2024-05-28