# How do we add numbers?

How do we compute sums in general? How can we tell the result of the operation $A+B$?

Even when we talk about very basic numbers like $\Bbb{N}$ I find it hard to understand the algorithm we use to compute a sum. We know things like $2+3=5$ but what about larger numbers?

How do we generalize this technique?

To clarify what I'm asking, I'm looking for an algorithm, or a way to compute a sum of two numbers that doesn't need the notion of multiplication. I've read similar threads where methods like representing $1234$ as $1 \cdot 10^3 + 2 \cdot 10^2 + 3 \cdot 10^1 + 4 \cdot 10^0$ and then operating applying the distributive, commutative and associative laws, but that only works if you've defined multiplication (and maybe even exponentiation before!).

• Are you asking about pencil-and-paper algorithms to add large numbers? – hardmath May 28 '15 at 1:20
• We note that $+$ is associative and commutative so you can add numbers in whichever order you like. Then we make use of decimal notation which tells us for instance that $1234 = 1000+200+30+4$. – user137731 May 28 '15 at 1:29
• Possible duplicate? – user137731 May 28 '15 at 1:33
• @Bye_World Thanks, but no, I clarified why that wasn't a duplicate just now. – YoTengoUnLCD May 29 '15 at 20:56
• I agree. If you bar any notion of multiplication, you can't even use $1234$ to really represent that number. It might denote it, but it's just shorthand for $\text{SSSSSS}\ldots\text{SSS}0$, where there are $1234$ S's. Speaking of which, you can shuttle S back and forth between two numbers in a sum; that is, $\text{SSSS}0+\text{S}0 = \text{SSS}0+\text{SS}0$. Continue until one of the numbers is $0$. Does that serve as an algorithm for performing addition? – Brian Tung May 29 '15 at 21:03

## 2 Answers

It very much depends on how the numbers $A$ and $B$ are encoded - if you're given a pile of $A$ rocks and a pile of $B$ rocks, you might not have any better options than to start counting. Of course a more common method is radix representation (Wikipedia link), e.g. binary, decimal; then you can use the standard "carrying" algorithm (Wikipedia link). More complex algorithms are often implemented in computers' adder circuits (Wikipedia link).

• (Well, if a pile of rocks is considered an acceptable encoding of a number, then we can just lump the piles together with a bulldozer to compute the sum!) – Chris Culter May 28 '15 at 1:56
• Haha that's a good point :) – Zev Chonoles May 28 '15 at 2:12

A more general approach:

In general the calculation of a anything can be traced back to its definition. The axioms concerning addition in the naturals can be found here: http://en.wikipedia.org/wiki/Peano_axioms. In particular one aims to derive faster algorithms from those definitions and prove that those algorithms agree with the most fundamental calculation of the sum.