# Cardinality of infinity (3)

1. Apr 7, 2008

### kingwinner

1. The problem statement, all variables and given/known data
1) Find the cardnality of the set of constructible angles whose cosine is
a) irrational
b) rational

2. The attempt at a solution
For the other questions I have posted my attempt, but for this one I really have no clue...
Could someone please give me some general hints?

Last edited: Apr 7, 2008
2. Apr 7, 2008

### morphism

Isn't this problem simply equivalent to the problem of determining the cardinality of the sets of constructible rational and irrational numbers, respectively? (Since an angle is constructible iff its cosine is.)

3. Apr 7, 2008

### kingwinner

Hi,

I don't think they are equivalent.
cos(x) has to be between -1 and 1 inclusive, so it can't be e.g. 2, 3 , -4, etc. while the set of constructible numbers have these as elements.

Last edited: Apr 7, 2008
4. Apr 7, 2008

### HallsofIvy

Staff Emeritus
That is not relevant. What he said was that an angle is constructible if and only if its cosine is. Yes, 2, 3, etc. are constructible and so are cos(2), cos(3), etc. And a number is constructible if and only if it is algebraic of order a power of 2 which includes all rational numbers. Now- there are at most a countable number of possible polynomials, with integer coefficients, of degree n, each of which has at most n zeros so the set of all numbers algebraic of order n is countable. What does that tell you about the cardinality of the set of all algebraic numbers?

5. Apr 7, 2008

### kingwinner

"And a number is constructible if and only if it is algebraic of order a power of 2 which includes all rational numbers."
I haven't learnt this in my class, so I can't take this for granted...there is so little tool I can use :(

Last edited: Apr 7, 2008
6. Apr 8, 2008

### kingwinner

So I think the problem is equivalent to the problem of finding the cardnality of the sets of consturctible rational and irrational numbers between -1 and 1 inclusive. To justify this, we need to first prove the equivalence.
e.g. Is there a 1-1 and onto map from {consturctible irrational numbers between -1 and 1} and {consturctible angles whose cosine is irrational}? I can't think of one...for example cos360o=cos720o, how can it be one-to-one, then?

I know that {all consturctible numbers} is countably infinite, but how about {consturctible irrational numbers between -1 and 1} and {consturctible rational numbers between -1 and 1 inclusive}? How to find their cardinalities?

Thanks!

7. Apr 8, 2008

### kingwinner

An angle is constructible iff its cosine is, but the trouble is that the correspondence is not one-to-one. There are infintely many angles whose cosine is the same value.

8. Apr 8, 2008

### morphism

That doesn't matter. x is a constructible angle iff cos(x) is constructible. So if x is a constructible angle whose cosine is rational, that means cos(x) is a constructible rational number, etc. etc.

9. Apr 9, 2008

### kingwinner

But if cos(theta) and theta does not have one-to-one correspondence, then we CANNOT say that:
|{constructible angles whose cosine is irrational}| = |{constructble irrational numbers between -1 and 1}|
which seems to create a lot of trouble...

10. Apr 9, 2008

### HallsofIvy

Staff Emeritus
It seems to me that a large part of this problem is know what numbers are constructible. Since you say that you do not know that "a number is constructible if and only if it is algebraic of order a power of 2" what do you know about constructible numbers? Since you are asking about the "cardinality of the set of constructible angles whose cosine is rational", are you aware that all rational numbers are constructible?

11. Apr 9, 2008

### kingwinner

Yes, I am aware that all rationals are constructible.

12. Apr 9, 2008

### HallsofIvy

Staff Emeritus
Good! So you know the answer to part b! Now since you know that cos(x) is irrational if and only if x is, and, between 0 and $2\pi$ there at most two angles that have the same cardinality, you know that the answer to part a is just the cardinality of the constructible numbers.

13. Apr 9, 2008

### kingwinner

"cos(x) is irrational if and only if x is"

"between 0 and $2\pi$ there at most two angles that have the same cardinality"
Why only between 0 and 2pi? Constructible angles can be e.g.720 degrees, 1440 degrees, etc.

14. Apr 9, 2008

### kingwinner

Still struggling to understand...

15. Apr 10, 2008

### kingwinner

If I restrict theta to be between 0 and 180 degrees, then cos(theta) and theta would have one-to-one corrspondence. Thus |{constructible angles between 0 and 180 degrees whose cosine is irrational}| = |{constructble irrational numbers between -1 and 1}|

{constructble numbers} is countably infinite, so {constructble irrational numbers between -1 and 1} as a subset of it must be countable. But how can we justify that {constructble irrational numbers between -1 and 1} is an infinite set?

Now, assuming {constructble irrational numbers between -1 and 1} is countably infinte, then {constructible angles between 0 and 180 degrees whose cosine is irrational} is countably infinite since they have the same cardinality

{constructible angles whose cosine is irrational} is the union of {constructible angles between 180k and 180k+180 degrees whose cosine is irrational} where k is from 0 to infinity
This is a countable union, and a countable union of countable sets is countable, so {constructible angles whose cosine is irrational} is countable.

Is this a valid proof?

16. Apr 10, 2008

### HallsofIvy

Staff Emeritus
Yes, once you have cleared up the question of whether the set of all constructible numbers between -1 and 1 is infinite.

17. Apr 10, 2008

### kingwinner

How can I prove this? Any hints?

18. Apr 10, 2008

### kingwinner

The irrationals are dense in the reals, so they must be infintely many irrationals between -1 and 1. But here they must be constructbile, too! How can I prove that this set is also infinte?