8

I've run into a small problem while working through Enderton's Elements of Set Theory. I'm doing the following problem:

Call a natural number even if it has the form $2\cdot m$ for some $m$. Call it odd if it has the form $(2\cdot p)+1$ for some $p$. Show that each natural number number is either even or odd, but never both.

I've shown most of this, and along the way I've derived many of the results found in Arturo Magidin's great post on addition, so any of the theorems there may be used. It is the 'never both' part with which I'm having trouble. This is some of what I have:

Let $$ B=\{n\in\omega\ |\neg(\exists m(n=2\cdot m)\wedge\exists p(n=2\cdot p+1))\}, $$ the set of all natural numbers that are not both even and odd. Since $m\cdot 0=0$, $0$ is even. Also $0$ is not odd, for if $0=2\cdot p+1$, then $0=(2\cdot p)^+=\sigma(2\cdot p)$, but then $0\in\text{ran}\ \sigma$, contrary to the first Peano postulate. Hence $0\in B$. Suppose $k\in B$. Suppose $k$ is odd but not even, so $k=2\cdot p+1$ for some $p$. Earlier work of mine shows that $k^+$ is even. However, $k^+$ is not odd, for if $k^+=2\cdot m+1$ for some $m$, then since the successor function $\sigma$ is injective, we have $$ k^+=2\cdot m+1=(2\cdot m)^+\implies k=2\cdot m $$ contrary to the fact that $k$ is not even.

Now suppose $k$ is even, but not odd. I have been able to show that $k^+$ is odd, but I can't figure a way to show that $k^+$ is not even. I suppose it must be simple, but I'm just not seeing it. Could someone explain this little part? Thank you.

Later
  • 750
yunone
  • 22,783

6 Answers6

4

HINT $\ $ Here's the inductive step: $\rm\ 2m \ne 2n+1\ \Rightarrow\ 2m+1 \ne 2(n+1)$

Bill Dubuque
  • 282,220
  • @yunone: $\rm\ a^{+} = b^{+}\ \Rightarrow\ a = b\ $ is a Peano axiom. – Bill Dubuque Jan 10 '11 at 04:30
  • Sorry, my last comment was at the post before the edit. I believe I see now. So suppose $k=2m\neq 2n+1$. Then since $\sigma$ is injective, $k^+=(2m)^+\neq(2n+1)^+\Rightarrow k^+=2m+1\neq 2n+1^+=2n+2=2(n+1)$. So $k^+$ is odd, but not even. Thanks! – yunone Jan 10 '11 at 04:33
2

Here is a complete proof. We assume that $n^+=m^+$ implies $n=m$ (*) and say that $n$ is even if it is $m+m$ for some natural number $m$ and odd if it is $(m+m)^+$ for some natural number $m$.

We know that $\phi=\phi+\phi$ so $\phi$ is even and $\phi\neq p^+$ for any $p$ and so is therefore not odd. Now assume $k$ is not both odd and even. Now consider $k^+.$
Suppose $k^+$ is odd. Then $k^+=(n+n)^+$ for some $n$. So by (*), $k=n+n$ so $k$ is even.
Suppose $k^+$ is even. Then $k^+=(m+m)$ for some $m$. We know $m\neq\phi$ as otherwise $k^+=\phi$, so $m=p^+$ for some $p$. Hence $k^+=(p^++p^+)=(p^++{p)}^+$. So by (*), $k=p+p^+=(p+p)^+$, which is odd. Hence if $k^+$ is even and odd then $k$ is odd and even, a contradiction to our induction hypothesis. Hence $k^+$ is not both even and odd, and we have the result by induction.

Porky
  • 1,054
  • 1
    May I ask where the $k \neq \emptyset$ hypothesis comes in (i.e. where is it used in the subsequent proof) in the second line of your second paragraph? Also, why are we allowed to do that? Doesn't proof by induction explicitly require that we let $k \in A$ where $A$ is the set we want to prove inductive? – EE18 Apr 17 '24 at 14:43
  • Yes, you are right. It's not required and I have updated. – Porky Apr 23 '24 at 01:06
0

Suppose there exists some $n \in \mathbb{N}$ which is both even and odd. Then $n= 2m = 2p+1$. So $2m = 2p+1$ or $2(m-p) = 1$. Contradiction.

0

Referring to OP's question on proving "if $k$ is not odd, then $k^+$ is not even", here is an attempt.

We prove the contrapositive of the statement, i.e. we prove "if $k^+$ is even, then $k$ is odd". Let $k^+=2n$ for some $n\in\mathbb{N}$. Note that $n\neq 0$; for otherwise $k^+=2\cdot 0=0$, contradicting the fact that $0$ is not the successor of any natural number. Since $n\neq 0$, we can write $n=m^+$ for some $m\in\mathbb{N}$, as every natural number other than $0$ is the successor of some other natural number. Then $$k^+=2m^+=2m+2=2m+1^+=(2m+1)^+.$$ Since the successor operator $\sigma:\mathbb{N}\to\mathbb{N}$ is injective, $k=2m+1$.

Ho Man-Ho
  • 753
0

enter image description here

N is subset of Z , so it can entirely said as no integer is both odd and even

Master
  • 21
0

A proof by contradiction showing a natural number cannot be both even and odd...

Assume there exists a natural number $n$ such that $n$ is both even and odd. By definition of even, there exists an integer $k$ such that $n=2k$, and by definition of odd, there exists an integer $p$ such that $n=2p+1$. Then,

$ \begin{array}{llll} n=n & \text{by the reflexive property of equality} \\ 2k=n & \text{by substitution with $n=2k$} \\ 2k=2p+1 & \text{by substitution with $n=2p+1$} \\ \frac{1}{2} 2k=\frac{1}{2} (2p+1) & \text{by multiplication property of equality} \\ \frac{1}{2} 2k=\frac{1}{2}2p+\frac{1}{2}1 & \text{by distributive law} \\ 1k=1p+\frac{1}{2}1 & \text{by multiplicative inverse law} \\ k=p+\frac{1}{2} & \text{by multiplicative identity law} \\ \end{array} $

Thus, $k=p+\frac{1}{2}$ is not an integer. But we have already stated $k$ is an integer. Hence, we have contradiction and our assumption that $n$ is both even and odd is in fact false. Therefore, if $n$ is a natural number, then $n$ is not both even and odd.