Tuesday, 10 September 2019

abstract algebra - Proving that left(mathbbQ[sqrtp1,dots,sqrtpn]:mathbbQright)=2n for distinct primes pi.



I have read the following theorem:




If p1,p2,,pn are distinct prime numbers, then(Q[p1,,pn]:Q)=2n.





I have tried to prove a more general statement but I have a problem at one point. (I still don't know how to prove the theorem above, too, because I don't know how not to use linear independence, which I do in the more general statement below.) Could you please help me overcome the obstacle I've encountered? I will post the intended proof and make it clear where I'm having trouble.



I want to prove the following statement:




Let n1. The set Bn:={pϵ11pϵ22pϵnn|(ϵ1,ϵ2,,ϵn){0,1}n} has 2n elements and is a Qbasis of Q[p1,p2,,pn].





The proof will be by induction.



For n=1, we have Bn={1,p1}. It is clear that p11, so the set has 2=21 elements. It is the basis of Q[p1] because the minimal polynomial of p1 over Q has degree 2, and there is a theorem that K[a] has a0,,ad1 as a basis, where d is the degree of the minimal polynomial of a over K.



Suppose the statement is true for n1, where n2. We have



(Bn=Bn1pnBn1) and (Bn1pnBn1=),




which is easy to see. It is also easy to see that card(Bn1)=card(pnBn1), and therefore



cardBn=2n.



Let



xBnqxx=0



for some {qx}xBnQ. Let p(x):=pnx for all xBn1. We have



xBnqxx=xBn1qxx+xpnBn1qxx=xBn1qxx+xBn1qp(x)pnx.



Therefore




xBn1qxx=pnxBn1qp(x)x,



and we can make the following division iff qp(x)0 for all xBn1 (because Bn1 is linearly indepentent over Q):



pn=xBn1qxxxBn1qp(x)x,




The right-hand side belongs to Q[p1,p2,,pn1], so we have



pnQ[p1,p2,,pn1].



Therefore we can write pn uniquely in the basis Bn1.



pn=yBn1cyy



for some {cy}yBn1Q.



After squaring this equation we will obtain



pn=yBn1c2yy2+2y,zBn1cyczyz.




The last sum must be zero because it is not in Q and because after reducing it, we obtain a representation of pn in the basis Bn1, which is unique. Thus



pn=yBn1c2yy2.



Unfortunately, I can't prove that cycz is always zero. This was my first thought, but clearly there's trouble with the possibility of reductions in
y,zBn1cyczyz.



Different pairs y,z may yield the same element of Bn1 in the product yz. This happens for example when y=53, z=52, and y=112, z=113.




If it were true that cycz is always zero, I would be able to continue my proof as follows. We would have only one y0 such that cy00 and we'd get



pn=c2y0y20.



Let cy0=kl. We can write
l2pn=k2y20.



But y20 is the product of some primes different from pn. Therefore the greatest power of pn that divides the right-hand side is even. However, the greatest power of pn that divides the left-hand side is odd. A contradiction.




The contradiction proves that qp(x)=0 for all xBn1. Hence (1) gives us that



xBn1qxx=0



and linear independence of Bn1 gives us that qx=0 for all xBn1.



This gives us that Bn is linearly independent. It generates the whole Q[p1,p2,,pn] because




Q[p1,p2,,pn]=(Q[p1,p2,,pn1])[pn].



This would end the proof.


Answer



HINT   An inductive proof follows easily from this



LEMMA   [K(a,b):K]=4  if  a, b, ab  all are not in K and 20 in K.




Proof    Let  L=K(b). Then [L:K]=2 via bK, so it is sufficient to prove [L(a):L]=2. It fails only if aL=K(b)  and then  a = r+s b  for  r,sK. But that is impossible since squaring yields (1):  a = r2+b s2+2rs b, which contradicts hypotheses as follows:



rs0    b  K   by solving (1) for b, using 20



 s=0     a  K   via  a = rK



 r=0    abK   via  a = s b, times b QED



Using the above as the inductive step one easily proves the following result of Besicovic.




THEOREM   Let Q be a field with 20, and  L=Q(S)  be an extension of Q generated by n square roots  S={a,b,} of elts  a,b,Q.
If every nonempty subset of S has product not in Q then each successive
adjunction  Q(a), Q(a,b), doubles the degree over Q, so, in total, [L:Q] = 2n. Hence the 2n subproducts of the product of S comprise a basis of L over Q.


No comments:

Post a Comment

real analysis - How to find limhrightarrow0fracsin(ha)h

How to find lim without lhopital rule? I know when I use lhopital I easy get $$ \lim_{h\rightarrow 0}...