Dummit and Foote exercise 13.2.13 on algebraic extensions

6 replies · 3K views · last reply
Math Amateur
Gold Member
MHB
Messages
3,920
Reaction score
48

Homework Statement



I am reading Dummit and Foote, Chapter 13 - Field Theory.

I am currently studying Section 13.2 : Algebraic Extensions

I need some help with Exercise 13 of Section 13.2 ... ... indeed, I have not been able to make a meaningful start on the problem ... ...

As indicated above I need help in order to make a meaningful or significant start on the solution to this exercise .. ...Exercise 13 of Section 13.2 reads as follows:
Exercise 13 from a textbook on field theory: suppose F = Q(α1, α2, ..., αn) where each αi² is in Q; prove that the cube root of 2 is not an element of F.

Homework Equations

Definitions that may be relevant to solving this exercise include the following:

Two textbook definitions. First: if K is an extension of field F and alpha, beta, ... are elements of K, the smallest subfield of K containing F and these elements is denoted F(alpha, beta, ...) and is the field generated by them over F. Second: if K = F(alpha), then K is a simple extension of F and alpha is a primitive element.
A result which seems relevant is Lemma 16 plus the remarks that follow its proof ... Lemma 16 and the following remarks read as follows:
Typeset page from a textbook showing Lemma 16 about field generated by alpha and beta, its proof, and a sequence of extension fields F to K with degree multiplicativity.

The Attempt at a Solution



[/B]
As indicated above I need help in order to make a meaningful or significant start on the solution to this exercise .. ...

As indicated above I need help in order to make a meaningful or significant start on the solution to this exercise .. ...

One thought, though ... there must be some way to use ##\alpha_i^2 \in \mathbb{Q}## ... perhaps in establishing the dimension of ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_{k + 1} )## over ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_{k } )## ... and hence getting some knowledge of the dimension of ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_n )## over ##\mathbb{Q}## ... but even if we do gain such knowledge, how do we use it to show ##\sqrt [3]{2} \notin F## ... ...

... AND FURTHER ... anyway ... what is implied by ##\alpha_i^2 \in \mathbb{Q}## ... ... ?
Help will be much appreciated ...Peter
 
Attachments
  • Exercise 13 from Section 13.2 stating: Suppose F = Q(α1, α2, ..., αn) where αi^2 ∈ Q for i = 1, 2, ..., n. Prove that the cube root of 2 is not an element of F.
    D&F - Exercise 13 - Section 13.2 ... ....png
    3.8 KB · Views: 510
  • Two textbook definitions of field theory: the field generated by elements alpha, beta, ... over F as the smallest subfield of K containing F and those elements, and a simple extension K = F(alpha) with alpha a primitive element.
    D&F - Definition of Field Generated by Elements of K ..png
    18.7 KB · Views: 507
  • Screenshot of a textbook page with Lemma 16 on field extensions, F(α, β) = (F(α))(β), its proof, and the tower of fields F = F0 ⊆ F1 ⊆ ... ⊆ Fk = K with degree bound [K:F] ≤ n1 n2 ... nk.
    D&F - Lemma 16 and following remarks - Section 13.2 ....png
    48.4 KB · Views: 492
Last edited:
Math Amateur said:
One thought, though ... there must be some way to use ##2∈Q\alpha_i^2 \in \mathbb{Q}## ... perhaps in establishing the dimension of ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_{k + 1} )## over ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_{k } )## ... and hence getting some knowledge of the dimension of ##F( \alpha_1 , \alpha_2, \ ... \ ... \ \ \alpha_n )## over ##\mathbb{Q}## ... but even if we do gain such knowledge, how do we use it to show ##\sqrt [3]{2} \notin F## ... ...
Good idea. What can you say about ##\dim_\mathbb{Q} F\;##? And what about ##\dim_\mathbb{Q}\mathbb{Q}[\sqrt[3]{2}]\;##? Then what do you know about the fields in between, from ##\mathbb{Q}## to ##F\;:\;\mathbb{Q} \subseteq \mathbb{Q}[\alpha_1]\subseteq \mathbb{Q}[\alpha_1,\alpha_2]\subseteq \ldots \subseteq F\;##?
... AND FURTHER ... anyway ... what is implied by ##\alpha_i^2 \in \mathbb{Q}## ... ... ?
It means that ##\alpha_i## is a root of ##x^2-r## for some rational ##r \in \mathbb{Q}##, ergo the minimal polynomial of ##\alpha_i## because ##\alpha_i\notin \mathbb{Q}## which rules out that the degree could be less than two.
 
  • Like
Likes   Reactions: Math Amateur
fresh_42 said:
Good idea. What can you say about ##\dim_\mathbb{Q} F\;##? And what about ##\dim_\mathbb{Q}\mathbb{Q}[\sqrt[3]{2}]\;##? Then what do you know about the fields in between, from ##\mathbb{Q}## to ##F\;:\;\mathbb{Q} \subseteq \mathbb{Q}[\alpha_1]\subseteq \mathbb{Q}[\alpha_1,\alpha_2]\subseteq \ldots \subseteq F\;##?

It means that ##\alpha_i## is a root of ##x^2-r## for some rational ##r \in \mathbb{Q}##, ergo the minimal polynomial of ##\alpha_i## because ##\alpha_i\notin \mathbb{Q}## which rules out that the degree could be less than two.
Thanks for the help, Andrew ... ...

Just a simple clarification though ... ...

You write, in the context of ##\alpha_i^2 \in \mathbb{Q}##, the following ... ...

" ... ... It means that ##\alpha_i## is a root of ##x^2-r## for some rational ##r \in \mathbb{Q}##, ergo the minimal polynomial of ##\alpha_i## because ##\alpha_i\notin \mathbb{Q}## which rules out that the degree could be less than two ... ... "... BUT ..

... how do we know that ##\alpha_i \notin \mathbb{Q}## ... indeed maybe ##\alpha_i## is in ##\mathbb{Q}## ... then ##\alpha_i## will satisfy a polynomial of degree 1 in ##\mathbb{Q}## ... namely ##x - \alpha_i## ... ...

... so doesn't ##\alpha_i^2 \in \mathbb{Q}## mean that

## [ F( \alpha_1, \alpha_2, \ ... \ ... \ \ , \alpha_i ) \ : \ F( \alpha_1, \alpha_2, \ ... \ ... \ \ , \alpha_{i - 1} ) ] ##

is either 1 or 2 ... ...

can you clarify/comment ?{Note : I have to confess that although I think it may well be true that ## [ F( \alpha_1, \alpha_2, \ ... \ ... \ \ , \alpha_i ) \ : \ F( \alpha_1, \alpha_2, \ ... \ ... \ \ , \alpha_{i - 1} ) ] ## is either 1 or 2 (i.e a positive integer less than or equal to 2) I do not know the precise argument for why it cannot be larger than 2 ... say 3 r 4 or 37 or something ... can you explain ...}Peter
 
Last edited:
Hi Michael!

All cases in which ##\alpha_i \in \mathbb{Q}## can be omitted. Say ##\alpha_1 , \ldots \alpha_{k-1} \in \mathbb{Q}## and ##\alpha_k, \ldots , \alpha_n \notin \mathbb{Q}## then ##F=\mathbb{Q}[\alpha_1,\ldots , \alpha_n]=\mathbb{Q}[\alpha_k,\ldots , \alpha_n]## and we're back in the case where all ##\alpha_i \notin \mathbb{Q}##. (The case in which there isn't a proper extension at all is also trivial, since ##\sqrt[3]{2} \notin \mathbb{Q}## and nothing has to be shown.)

So key to this exercise lies in the formula for finite field extensions:
Let ##\mathbb{F} \subseteq \mathbb{K} \subseteq \mathbb{L}## be finite field extensions, then ##[\mathbb{L}:\mathbb{F}] = [\mathbb{L} : \mathbb{K}]\,\cdot \,[\mathbb{K}:\mathbb{F}]\;##.

In case you cannot use this equation, you might investigate how the roots of the minimal polynomial ##m_{\mathbb{Q}}(\sqrt[3]{2})(x)## can be or cannot be expressed by rationals and roots of the form ##x = \alpha_i = \sqrt{r_i}## with ##r_i \in \mathbb{Q}\;##.
 
  • Like
Likes   Reactions: Math Amateur
fresh_42 said:
Hi Michael!

All cases in which ##\alpha_i \in \mathbb{Q}## can be omitted. Say ##\alpha_1 , \ldots \alpha_{k-1} \in \mathbb{Q}## and ##\alpha_k, \ldots , \alpha_n \notin \mathbb{Q}## then ##F=\mathbb{Q}[\alpha_1,\ldots , \alpha_n]=\mathbb{Q}[\alpha_k,\ldots , \alpha_n]## and we're back in the case where all ##\alpha_i \notin \mathbb{Q}##. (The case in which there isn't a proper extension at all is also trivial, since ##\sqrt[3]{2} \notin \mathbb{Q}## and nothing has to be shown.)

So key to this exercise lies in the formula for finite field extensions:
Let ##\mathbb{F} \subseteq \mathbb{K} \subseteq \mathbb{L}## be finite field extensions, then ##[\mathbb{L}:\mathbb{F}] = [\mathbb{L} : \mathbb{K}]\,\cdot \,[\mathbb{K}:\mathbb{F}]\;##.

In case you cannot use this equation, you might investigate how the roots of the minimal polynomial ##m_{\mathbb{Q}}(\sqrt[3]{2})(x)## can be or cannot be expressed by rationals and roots of the form ##x = \alpha_i = \sqrt{r_i}## with ##r_i \in \mathbb{Q}\;##.
Thanks fresh_42 ... appreciate your help ...

Just reflecting on what you have said ... but obviously most helpful ...

Thanks again,

Peter
 
fresh_42 said:
Hi Michael!

All cases in which ##\alpha_i \in \mathbb{Q}## can be omitted. Say ##\alpha_1 , \ldots \alpha_{k-1} \in \mathbb{Q}## and ##\alpha_k, \ldots , \alpha_n \notin \mathbb{Q}## then ##F=\mathbb{Q}[\alpha_1,\ldots , \alpha_n]=\mathbb{Q}[\alpha_k,\ldots , \alpha_n]## and we're back in the case where all ##\alpha_i \notin \mathbb{Q}##. (The case in which there isn't a proper extension at all is also trivial, since ##\sqrt[3]{2} \notin \mathbb{Q}## and nothing has to be shown.)

So key to this exercise lies in the formula for finite field extensions:
Let ##\mathbb{F} \subseteq \mathbb{K} \subseteq \mathbb{L}## be finite field extensions, then ##[\mathbb{L}:\mathbb{F}] = [\mathbb{L} : \mathbb{K}]\,\cdot \,[\mathbb{K}:\mathbb{F}]\;##.

In case you cannot use this equation, you might investigate how the roots of the minimal polynomial ##m_{\mathbb{Q}}(\sqrt[3]{2})(x)## can be or cannot be expressed by rationals and roots of the form ##x = \alpha_i = \sqrt{r_i}## with ##r_i \in \mathbb{Q}\;##.
Thanks to Andrew and fresh_42 for all the help ...

Now to finish the exercise based on the help you gave ...Essentially then, it follows that ...

##[ F \ : \ \mathbb{Q} ] = 2^k## where ##0 \le k \le n##Now ... ... suppose ##\sqrt [3]{2} \in F ## ... (and try for a contradiction)Then since ##\mathbb{Q} \subseteq \mathbb{Q} ( \sqrt [3]{2} ) \subseteq F##

... we have that ...

##[ F \ : \ \mathbb{Q} ] = [ F \ : \ \mathbb{Q} ( \sqrt [3]{2} ) ] [ \mathbb{Q} ( \sqrt [3]{2} ) \ : \ \mathbb{Q} ]##

So that##2^k = [ F \ : \ \mathbb{Q} ( \sqrt [3]{2} ) ] . 3##... BUT ...##3## does not divide ##2^k## ... ... Contradiction!So ##\sqrt [3]{2} \notin F##
Can someone please confirm that the above is correct ... ...

Peter
 
Yes, this is correct. At some place, an extension of degree three (or at least ##3m##) would be needed.
 
  • Like
Likes   Reactions: Math Amateur