We can multiply $a$ and $n$ by adding $a$ a total of $n$ times.

$$ n \times a = a + a + a + \cdots +a$$

Can we define division similarly using only addition or subtraction?

  • 4
    $\begingroup$ Do you admit logarithms? If so, we can very easily define division using subtraction: $$a/b = \exp\left(\log \frac{a}{b}\right) = \exp(\log a - \log b).$$ $\endgroup$ – Emily Aug 24 '12 at 18:14
  • $\begingroup$ doesn't exponents and logarithms come we define multiplication and division $\endgroup$ – Monkey D. Luffy Aug 24 '12 at 18:19
  • $\begingroup$ Nope. We can define exponents and logarithms without requiring multiplication or division -- in a manner of speaking. We define $b^x$ as the supremum of a very specific subset of real numbers. This definition does not require that we define $b^x = b\cdot b \cdot b \cdots b$ some $x$ times. In fact, this definition works for any possible value of $x$. Logarithms can be defined in a similar manner. To justify this definition, we require that multiplication is an assumed property of the field of real numbers. We don't need to define exponentiation as repeated multiplication. $\endgroup$ – Emily Aug 24 '12 at 18:43
  • 1
    $\begingroup$ Reminded me this good old question. $\endgroup$ – user2468 Aug 25 '12 at 2:41
  • $\begingroup$ @Arkamis, what is this set whose supremum is $b^x$? Sounds very interesting. $\endgroup$ – goblin Dec 25 '14 at 7:06

To divide $60$ by $12$ using subtraction:

$$\begin{align*} &60-12=48\qquad\text{count }1\\ &48-12=36\qquad\text{count }2\\ &36-12=24\qquad\text{count }3\\ &24-12=12\qquad\text{count }4\\ &12-12=0\qquad\;\text{ count }5\;. \end{align*}$$

Thus, $60\div 12=5$.

You can even handle remainders:

$$\begin{align*} &64-12=52\qquad\text{count }1\\ &52-12=40\qquad\text{count }2\\ &40-12=28\qquad\text{count }3\\ &28-12=16\qquad\text{count }4\\ &16-12=4\qquad\;\text{ count }5\;. \end{align*}$$

$4<12$, so $64\div 12$ is $5$ with a remainder of $4$.

  • $\begingroup$ I remember re-implementing the built-in integer division and modulo functions of (insert language here) being a common programming exercise... :D $\endgroup$ – J. M. is a poor mathematician Aug 24 '12 at 23:45
  • 2
    $\begingroup$ @J.M. I had it on an exam; language was assembly for Zilog Z80 processors! $\endgroup$ – user2468 Aug 25 '12 at 3:37
  • $\begingroup$ Note this is horribly inefficient, computationally speaking. $\endgroup$ – Thomas Aug 25 '12 at 4:44
  • $\begingroup$ Seems a bit arbitrary that we always want to stop at zero. Nothing more than definition right? $\endgroup$ – Ovi Jun 3 '16 at 22:58

If $n$ is divisible by $b$ ($\frac{n}{b}$ is a whole number), then keep doing $n - b - b - b - b - b - \cdots - b$ until the value of that is $0$. The number of times you subtract $b$ is the answer. For example, $\frac{20}{4} \rightarrow 20 - 4 - 4 - 4 - 4 - 4$. We subtracted '$4$' five times, so the answer is $5$.


You can also use additions. One should use results from intermediate calculations to speed up.

Let us divide 63 by 12. $$ \begin{split} 12+12=24,&\qquad\textrm{count }1+1=2\\ 24+24=48,&\qquad\textrm{count }2+2=4\\ 48+24=72,&\qquad\textrm{count }4+2=6\textrm{ (exceeded 63)}\\ 48+12=60,&\qquad\textrm{count }4+1=5\textrm{ (so we try adding less)}\\ 63-60=3,&\qquad\textrm{(calculation of the remainder)}\\ \end{split} $$


You can define division as repeated subtraction:$${72\over 9}=72-9-9-9-9-9-9-9-9$$Subtracting by $9$ eight times is the same as subtracting by $72$ since $9\cdot8=72$. So, the answer is $8$. Also, this is why ${n\over a}=n-a-a-a-a\cdots$ for whatever whole number $a$ is other than zero.

If you have a remainder, then you just do this:$${13\over 2}=13-2-2-2-2-2-2-1$$as you just saw, subtracting by $2$ six times is the same as subtracting by $12$ since $2\cdot6=12$, but there's a remainder of $1$ being sutracted, so it's the same as subtracting by $13$ since $2\cdot6+1=13$, so the answer is $6$ R$1$ or $6.5$.

  • $\begingroup$ $\frac{n}a\neq n-a-a-a\ldots$. Certainly it is not true that $$\frac{72}{9}=72-9-9-9-9-9-9-9-9$$ $\endgroup$ – Milo Brandt Jan 24 '15 at 22:27
  • $\begingroup$ Okay, I edited it so it would make more sense. $\endgroup$ – ReliableMathBoy Jan 24 '15 at 22:33

protected by Zev Chonoles Sep 1 '16 at 3:50

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.