8
$\begingroup$

I was reading my textbook right now on elementary number theory, and this question came up. It got me wondering: How can subtraction be performed on two elements $x,y\in\mathbb{N}$ if subtraction is defined by $x-y=x+(-y)$, and $\forall a\in\mathbb{N}, 0<a<\infty$?

I realize that $1-10=-9$, which is not in $\mathbb{N}$. But the question just bothered me, because I wonder if you can technically subtract two elements of $\mathbb{N}$?

$\endgroup$
2
  • 5
    $\begingroup$ "if subtraction is defined by $x−y=x+(−y)$" Then you have to define what you mean by $-y$. If you mean the element such that $y + (-y) = 0$, your first problem is that in the natural numbers, $-y$ does not exist for any $y\ne0$. Then you cannot even do $4-2$, because there is no $-2$. $\endgroup$
    – user856
    Commented Mar 12, 2013 at 16:10
  • $\begingroup$ Sure, you can subtract two elements of $\bf N$. You just may end up with something (a nonpositive integer) which is not in $\bf N$. Such is the very definition of a set not being closed under an operation. Of course, for this to be an operation that make sense, we need to view $\bf N$ as a subset of a bigger thing, namely $\bf Z$. Not an issue. $\endgroup$
    – anon
    Commented Mar 19, 2013 at 2:37

6 Answers 6

13
$\begingroup$

Regular subtraction is not well-defined on the natural numbers. In natural number contexts one often deals instead with truncated subtraction, which is defined: $$a\dot-b = \begin{cases}0,&\text{if $a\le b$}\\ a-b&\text{if $a\ge b$}\end{cases}$$

For example, one can define a truncated subtraction in Peano arithmetic as follows: $$\begin{array}{rcrl} 0 & \dot- & n & = 0 \\ Sn & \dot- & 0 & = Sn \\ Sn & \dot- & Sm & = n\dot- m \end{array} $$

One can similarly define it in the context of Church numerals, or in the context of total recursive functions.

This is often sufficient for whatever purposes one needs subtraction.

$\endgroup$
0
6
$\begingroup$

When one subtracts a natural number from another, say $1-2=-1$, one may not get a natural number. However, this does not forbid us from defining substraction. Substraction would be closed if we enlarge the set to the set of all integers.

$\endgroup$
4
$\begingroup$

Subtraction is partially defined in $\Bbb N$, i.e. it is defined only for some pairs of natural numbers. You cannot define it as $x-y=x+(-y)$, though, since $-y\notin \Bbb N$.

You can define $x-y=z$ for $x,y\in \Bbb N$ just as the $z\in \Bbb N$, if it exists, such that $y+z=x$.

$\endgroup$
1
  • $\begingroup$ OK, I'll remove my comments now so that they don't confuse anyone. $\endgroup$
    – Tara B
    Commented Mar 12, 2013 at 16:35
4
$\begingroup$

No, subtraction is not closed on the set of natural numbers.

One can define the difference between $a$ and $b$, $a, b \in \mathbb N\,$ in terms of the magnitude of the difference: taking the absolute value: $|a - b|$ for $a, b \in \mathbb N$, but the problem with "normal subtraction" is that $\,a - b = a + (-b)$. And here, $-b$ is the additive inverse of $b$: and since here we have $b \in \mathbb N$, unless $b = 0$ (if $\mathbb N$ includes $0$), $-b \notin \mathbb N$.

  • The additive inverse of an integer $n$ is the number such that for any $n \in \mathbb Z$, $\,n + -n = -n + n = 0,\;$ where $\,0\,$ is the additive identity.

  • Hence, we have the integers, which are closed under subtraction, (or rather closed under inverses), and hence defining subtraction on the integers presents no problems.

  • However, for all $n \in \mathbb N$, n\neq 0,\; -n \notin \mathbb N$. That, essentially, is what is meant when we say that the set of natural numbers is not closed under subtraction (...because it is not closed under inverses).

$\endgroup$
3
$\begingroup$

When we write for the natural numbers $a-b=c$ we mean that the number $c$ is the natural number (if it exists) that verify $c+b=a$ so the subtraction is defined by the sum operation. Now if we write for example $1-10=c$, the question is are there a natural number $c$ s.t $c+10=1$? The answer is obviously NO! so the set of natural number is not closed under subtraction.

$\endgroup$
0
2
$\begingroup$

You can prove that for any natural numbers $a,b$ there is at most one $c$ such that $b+c=a$. That means that we can define $z=a-b$ to mean $z+b=a$. This means that if $z=a-b$ and $w=a-b$ then $z=w$ - that is, subtraction is well-defined when it is defined.

You could also just define the integers based on the naturals, and define subtraction there for all pairs of integers.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .