Exercise 15.2.9

This exercise is concerned with the proof of Corollary 15.2.7.

(a)
Suppose that P ( u ) , Q ( u ) [ u ] are relatively prime and Q ( 0 ) = 1 . Prove that uP ( u 4 ) and Q ( u 4 ) have no common roots in any extension of .
(b)
Fix x in and m > 0 in , and let P m ( u ) , Q m ( u ) [ u ] be as in Theorem 15.2.5. Thus φ ( mx ) Q m ( φ 4 ( x ) ) = φ ( x ) P m ( φ 4 ( x ) ) . Prove that Q m ( φ 4 ( x ) ) 0 when φ ( x ) 0 .
(c)
Show that φ ( 2 ϖ n ) 0 when n > 2 is in and conclude that Q m ( φ 4 ( 2 ϖ n ) ) 0 .

Answers

Proof. (a)

The ring is principal, thus is a UFD with field of fractions . By Gauss’s Lemma (Theorem A.3.2, or Theorem A.5.8), if P , Q [ u ] are relatively prime in [ u ] , then P , Q are relatively prime in [ u ] .

Since P ( u ) , Q ( u ) are relatively prime in [ u ] , there are some polynomials A , B [ u ] such that A ( u ) P ( u ) + B ( u ) Q ( u ) = 1 , thus the substitution u u 4 gives A ( u 4 ) P ( u 4 ) + B ( u 4 ) Q ( u 4 ) = 1 . Reasoning by contradiction, suppose that uP ( u 4 ) and Q ( u 4 ) have a common root α in some extension of . Since Q ( 0 ) = 1 , α 0 , thus P ( α 4 ) = 0 . Then P ( α 4 ) = Q ( α 4 ) = 0 implies 1 = A ( α 4 ) P ( α 4 ) + B ( α 4 ) Q ( α 4 ) = 0 : this is a contradiction.

So uP ( u 4 ) and Q ( u 4 ) have no common roots in any extension of . (b) If m is odd, then φ ( mx ) Q m ( φ 4 ) ( x ) ) = φ ( x ) P m ( φ 4 ( x ) ) . Reasoning by contradiction, suppose that, for some x , Q m ( φ 4 ( x ) ) = 0 . Then φ ( x ) P m ( φ 4 ( x ) = 0 , so that α = φ ( x ) is a common root of Q m ( u 4 ) and u P m ( u 4 ) . Since P m , Q m are relatively prime, and Q m ( 0 ) = 1 , this is impossible by part (a).

If m is even, then φ ( mx ) Q m ( φ 4 ( x ) ) = φ ( x ) P m ( φ 4 ( x ) ) φ ( x ) . Suppose that, for some x , Q m ( φ 4 ( x ) ) = 0 . Then φ ( x ) P m ( φ 4 ( x ) ) φ ( x ) = 0 . If φ ( x ) = 0 , then 1 φ 4 ( x ) = 0 , thus φ 4 ( x ) = 1 , and φ ( x ) = ± 1 . If φ ( x ) { 1 , 1 } , then φ ( x ) P m ( φ ( x ) ) = 0 , so that α = φ ( x ) is a common root of Q m ( u 4 ) and u P m ( u 4 ) , which is impossible by part (a).

We have proved that Q m ( φ 4 ( x ) ) 0 when φ ( x ) { 1 , 1 } .

(Misprint in the sentence of part (b) ? If φ ( x ) = 0 , then Q m ( φ 4 ( x ) ) = Q m ( 0 ) = 1 0 , so there is no need to suppose φ ( x ) 0 .) (c) For all x , φ ( x ) = 0 if and only if x = for some k .

We must verify that φ ( 2 ϖ n ) { 1 , 1 } . If n > 2 , then 0 < 2 ϖ n < ϖ . This proves that 0 < φ ( 2 ϖ n ) < 1 , thus φ ( 2 ϖ n ) { 1 , 0 , 1 } .

By our version of part (b), this implies that

Q m ( φ 4 ( 2 ϖ n ) ) 0 .

Note: If we read the proof of Theorem 15.2.5, it is obvious that the denominators Q n ( φ 4 ( x ) ) never vanish, for all x , because 1 + φ 2 ( nx ) φ 2 ( x ) 0 . □

User profile picture
2022-07-19 00:00
Comments