Exponentiation

Multiplication

Can it go any further downwards? That is, can we extend it to describe multiplication by any whole number (including zero) or even any integer (including the negative integers)? The pattern that I've written so far really doesn't continue; at each stage, we remove one more addition of a, one more ‘+ a’ disappears, but the last step, where 1a means a, doesn't really have any addition, and there is no more plus sign to remove. After all, a added to itself one time isn't really a added to anything at all!

However, I can tweak this table slightly to allow the pattern to continue one more step. We know that 0a is supposed to be 0, and you might think that this is what you get when you remove everything from the last line of the table. This really only works, however, if you think that a calculation begins with 0. In the case of repeated addition, that really is true, so let's make it explicit: to multiply a by (say) 3, start with 0, then add a to that 3 times: 0 + a + a + a instead of simply a + a + a. Then the table looks like this:

Can we continue this to multiplication by negative integers? Not exactly, but we can start a similar, reverse pattern by changing addition to the inverse operation of subtraction. The final table looks like this:

I've put a new column to the right, giving the simplified expression that you would usually use. (Why write in all those zeroes if you don't need them?) But the middle column is the one that really shows the pattern: start with 0, then add or subtract as necessary.

Definition

As the first example, consider (2/7)³, which means (2/7) · (2/7) · (2/7), which works out to 8/343. Again, there's really nothing special about 2/7; in general, the rule is that a³ means a · a · a. Here, a is called the base and 3 is the exponent; the result a³ is the power, or in full ‘a raised to the power of 3’. You can also read this as ‘a raised to the 3rd power’, ‘the 3rd power of a’, ‘a to the 3rd power’, ‘a to the 3rd’, ‘a to the 3’, or even ‘a cubed’. (This last one is a nickname that comes from geometric applications; there is also a nickname for a², which is ‘a squared’.)

Here is the table (analogous to the first table above) for raising numbers to a natural exponent:

What is a⁰ then? In other words, how do we extend this table one more line to make the analogue of the second table above? You might think that a⁰ is 0, just like 0a is 0, on the grounds that you start calculating at 0. But in fact that does not work for repeated multiplication; when you multiply, you really start calculating at 1!

To convince yourself of this, just imagine that you replace 1 with 0 in the table below. Then all of the results would simply be 0, which is not correct; only if you use 1 instead can you get the same answers as before.

What is the analogue for multiplication of subtraction? It is division, although there is an extra complication: you cannot divide by 0. Therefore, you cannot raise 0 to the power of a negative exponent.

Here is the final table, analogous to the final table in the previous section:

To form the usual (short) expressions in the right column, I can use rules for dividing fractions. For example, 1/a/a, or 1/a divided by a, means 1/a multiplied by the reciprocal of a, or (1/a) · (1/a), which works out to (1 · 1)/(a · a), or 1/(a · a).

Tricky bits

Sometimes people say that 0⁰ is also undefined, but this isn't really correct. The definition works perfectly well to produce 0⁰ = 1. Of course, you can always make an exception in a definition if you want something to be undefined, and in fact mathematicians argued about this for almost 200 years. But over the course of the last century, people have begun to actually use 0⁰, and they find that 0⁰ has to be defined as 1 for a bunch of other things (mostly in the mathematical field of combinatorics) to work.

(This is how such arguments usually get settled: people go back and forth for a while, getting nowhere as long as it's purely theoretical; but once somebody finds a way to use the ideas, then it's obvious what the right answer has to be. Sometimes it takes the textbooks a while to catch up, however; the textbook for this course still says that 0⁰ is undefined, but in doing so it is behind the times!)

Notice that the sign (positive or negative) of the exponent has nothing to do with the sign of the power. For example, 2² = 2 · 2 = 4, while 2⁻² = 1/(2 · 2) = 1/4, and these are both positive. This is because a negative exponent simply tells you to divide instead of multiplying, which has nothing to do with whether your result is negative. In particular, if the base is positive, then the power will also be positive, no matter what the exponent is.

On the other hand, if the base is negative, then the sign of the power depends on the parity (even or odd) of the exponent. For example, (−2)² = (−2) · (−2) = 4, while (−2)³ = (−2) · (−2) · (−2) = (−2) · 4 = −8. This is because multiplying (or dividing) by a negative number always switches the sign, so as you raise a negative number to a power, you keep going back and forth between negative and positive until the sign of the final answer depends on whether the number of times you flip is even or odd.

When you raise a fraction to a power, you can do the numerator and denominator separately. For example, (2/3)² = (2/3) · (2/3) = (2 · 2)/(3 · 3) = 4/9, and you get the same result if you do (2²)/(3²). Similarly, my original example (2/7)³ can be worked out simply as (2³)/(7³) = (2 · 2 · 2)/(7 · 7 · 7) = 8/343. This is because multiplying fractions amounts to simply multiplying the numerators and denominators separately. Furthermore, when repeatedly multiplying the same numbers (as we do in exponentiation), you'll never have to reduce the fraction as long as it was reduced originally.

When raising a fraction to a negative exponent, the easy thing to do is to take the reciprocal first. For example, (2/3)⁻² = (3/2)², since 3/2 is the reciprocal of 2/3, so (2/3)⁻² is simply 9/4. This is because raising to a negative exponent means division, which is simply multiplication by the reciprocal. As another example, (1/2)⁻³ = 2³, because 2 is the reciprocal of 1/2, so (1/2)⁻³ works out to 8. Conversely, 2⁻³ is 1/8.

Also, you should be careful with the order of operations; unless grouping symbols (like parentheses) get in the way, exponentiation always comes first. For example, −4² means −(4 · 4), which is −16, while (−4)² means (−4) · (−4), which is 16. Also, 2/3² means 2/(3 · 3), which is 2/9, while (2/3)² means (2/3) · (2/3), which is 4/9.

The permanent URI of this web page is http://tobybartels.name/MATH-0950/2017SU/exponents/.