0

I know that this is a trivial question, but I have never done any serious proofs before and therefore I am a complete novice when it comes to that part of math.

Anyhow, I used the "prove by contradiction" technique and my proof is the following:

Proof:

We use proof by contradiction. Suppose that there exists an integer $n > 0$ for which $a^n$ is even, with $a$ being an odd integer larger than $0$.

We notice that if it is indeed even $a^n$ can be written as $(2k)^n$ with $k$ being any positive integer.

Now let's look at possible values of $n$. If $n$ is $1$ then we already have a contradiction, because if $a^1$ is even, which is stated in the problem, $a$ has to be even as well. We conclude that $n$ has to be larger than $1$. Again, we write $a^n$ as $(2k)^n$ and we realize that, for us to get the value of $a$ and check if it's even we have to take the $n$th root of $a^n$. We do exactly that and we reach the following: $a = 2k$ which represents a contradiction.

Therefore, we have proved that if $a^n$ is even, so is $a$. $QED$

Any feedback on this proof and its validity would be highly appreciated.

  • 1
    There are several mistakes. Firstly, the statement itself is wrong: $2^3=8$ is even, though $n=3$ is odd. I will assume that the statement was mistyped and you meant for any $n>0$, if $a^n$ is even then $a$ is even. Now suppose we prove this statement by contradiction. Then our first assumption must be "suppose there EXISTS an $n$ for which $a^n$ even for some $a$ with $a$ odd". Then one can proceed. – Jihoon Kang Jul 08 '19 at 16:14
  • @JihoonKang yes you are completely correct. That was my mistake. I was supposed to write "$a$ has to be even as well". My apologies. – Luka Duranovic Jul 08 '19 at 16:23
  • So now to go about proving by contradiction. Your first assumption for your attempt at proving by contradiction is wrong. The statement you want to prove is a fact about EVERY $n$, so we only need to assume that there EXISTS an $n$ for which the statement does not hold. I.E suppose there is an $n$ for which $a^n$ is even for some $a$ with $a$ odd. – Jihoon Kang Jul 08 '19 at 16:27
  • @JihoonKang Thank you very much. I see what you mean. I will make the necessary edit right away. – Luka Duranovic Jul 08 '19 at 16:29
  • Hope it helps, you don't need to edit your post anymore as I was only pointing out a mistake with your proof attempt, not necessarily a mistake with your post. – Jihoon Kang Jul 08 '19 at 16:30
  • "We notice that if it is indeed even an can be written as (2k)n with k being any positive integer.". Why? Unless you justify this, it looks like you're assuming the conclusion. – Deepak Jul 08 '19 at 16:30
  • @Deepak Is justifying that really necessary. I only stated the fact that if a number is even it can be written as a product of 2 and an arbitrary integer k. By the way, thanks for your comment. – Luka Duranovic Jul 08 '19 at 16:33
  • Are you required to use a proof by contradiction? Using contraposition gives a natural and almost immediate proof, all you need to show is that every positive integer power of an odd number is odd. – Deepak Jul 08 '19 at 16:35
  • @Deepak No, I am not required. However, the authors hinted that using contradiction in this case may lead to an easy proof. – Luka Duranovic Jul 08 '19 at 16:37
  • @Luka No, it's not that simple. Writing $a^n = 2k$ is fine. But writing $a^n = (2k)^n$ is assuming the conclusion. How do you know that $2^n$ divides $a^n$ without assuming that $2$ divides $a$? – Deepak Jul 08 '19 at 16:37
  • @Deepak I see what you mean here. Thank you for pointing that out. Do you have any suggestions as to how to correct that? – Luka Duranovic Jul 08 '19 at 16:41
  • @Luka I can't immediately think of a way to show that without prime decomposition, which is elementary but unwieldy. Sorry, it's getting late where I am, I need to sleep. I answered with a simple proof by contraposition. Hope it helps. – Deepak Jul 08 '19 at 16:44
  • @Deepak Your help is much appreciated. I saw your proof and I might just use that for this example. – Luka Duranovic Jul 08 '19 at 16:46
  • @Luka Sorry my brain is fuzzy late at night. It just hit me after I turned in, so I got up to edit. I've edited my post to show you how to transform the proof by contraposition to one by contradiction. In proving P implies Q, the contrapositive is (NOT Q) implies (NOT P). The equivalent contradiction is showing (NOT Q) AND P leads to a contradiction. – Deepak Jul 08 '19 at 17:07

4 Answers4

1

You can prove that by induction in n.

To $n = 1$, we have $a^{1}$ is even, then, trivially, $a$ is even.

Suppose that if $a^{n}$ is even, then $a$ is even. Let's prove that $a^{n+1}$ is even then, $a$ is even.

Consider that $a^{n+1} = 2k$, $k \in \mathbb{N}.$ Observe that,

$$2k = a^{n+1} = a^{n}a$$

Then, $a^{n}$ is even or $a$ is even. If $a$ is even, we conclude our proof. If $a^{n}$ is even, we can use the induction hypothesis.

J.A.G
  • 739
0

In my opinion, a proof by contraposition is very natural here.

Required to prove: $a^n$ is even implies $a$ even.

Contrapositive is: $a$ is odd implies $a^n$ odd.

Let $a = 2k+1$. Now using binomial theorem, it's easy to see that $a^n$ is the sum of even numbers plus $1$, making it odd. The statement is proved by contraposition.

EDIT:

A proof by contradiction was requested. Actually proof by contraposition is easy to transform into a proof by contradiction.

In this case one has to show $a^n$ even implies $a$ even. Assume, to the contrary, that an odd $a=2k+1$ exists such that $a^n$ is even. Again, proceeding by binomial theorem, we can show $a^n$ is odd. Meaning there is a number that is simultaneously odd and even. Contradiction. Hence such an odd $a$ cannot exist and $a$ must be even.

Deepak
  • 26,801
  • I see how that would work here. Even though I was looking to use contradiction, I do agree that this might be the most natural option – Luka Duranovic Jul 08 '19 at 16:46
  • An alternative to use binomial theorem is prove by induction that if $a$ is odd, then $a^{n}$ is odd – J.A.G Jul 08 '19 at 16:49
0

A short proof

We claim that $a^n=even$ cannot hold for odd $a$.

Let $d$ denote the unit's digit of odd $a$ then $d=1,3,5,7 or 9$. This implies $a^n$ has $1,3,5,7 or 9$ at its unit's digit for all $n\in \mathbb N$.

Nitin Uniyal
  • 7,946
0

Your first mistake as

We notice that if it.. [$a^n$].... is indeed even an can be written as $(2k)^n$ with k being any positive integer.

That is assuming that $a$ is even as well. SO $a = 2k$ and $a^n = (2k)^n$. But this is assuming what you want to prove.

If you are assuming $a^n$ is even then you are assuming $a^n =2k$.

But if you are assuming is that $a$ is ODD so $a$ can be written as $a = 2j + 1$ and therefore $a^n = (2j + 1)^n$.

A good bit of contradiction would be to prove that if $a^n = (2j+1)^n$ that $a^n$ must be odd. A contradiction. Can you do that?

======

Actually my way when I was young would be a quasi induction bootstrap.

Lemma: And odd number times an odd number is odd.

Proof: If $a = 2k + 1$ and $b= 2j + 1$ then $ab = (2k+1)(2j+1) = 4jk + 2j + 2k + 1 = 2(2jk+j+k) + 1$ is odd.

So if $a$ is odd then $a^2 = a*a$ is odd. And $a^3 = a^2*a$ is odd. And continuing for as long as we need we get $a, a^2, a^3, a^4, a^5,......$ are all odd.

(I don't know if you've formally been introduced to Proof By Induction; but basically if you can prove something for a basic case, and you can prove a "bootstrapping" method to get from one case to the next, it makes sense that you can start from the base and bootstrap as many as you need to get to all the cases.)

I suppose another way to do this is $(2k + 1)^n = 2^nk^2 + n*2^{n-1}k^{n-1} + ...... + 2k + 1$ and prove that all but the last terms are even. That might be overkill and relies messily on the binomial theorem.

A rather sophisticated proof by contradiction might be:

Suppose $a$ is odd but $a^n$ is even. $a^1$ is odd. And $a^2=a*a$ is odd. But $a^n$ is even. That means sense the powers of $a$ start of as odd but sometime turn even then there must be some very first power of $a$ where $a^k$ is even but every lower power is odd. So $a^{k-1}$ is odd. ANd $a$ is odd but $a^k = a^{k-1}a$ would also have to be odd. So there can't be a first power where $a^k$ is even. SO all $a^n$ are odd.

fleablood
  • 124,253