## See if this definition of a function is better than others

Try the definitions from the Khan Academy, Math is Fun, Wikipedia, Paul's Online Math Notes, and there are some others. Most of them are obsessed with the input-output terminology, which is not used by professionals. I think my explanation is more visual.

## "Sending letters" is associated with arrows

Situation: *every* citizen from the city of sends letters to *some* citizen(s) of the city of . In Figure 1, the arrows show that citizen sends letters to three citizens of : they are , while citizen sends letters to only . Let us write if sends letters to . Here, is called an **argument** (some people say an input) and is called a **value** (some people say an output).

**Definition**. is called a **function** if for each argument, the corresponding value is unique.

The situation depicted in Figure 1 does not describe a function, because sends letters to more than one person. The visualization of a function is given in Figure 2, where arrows going from one citizen of to more that one citizen of are excluded.

We finalize the function definition with some remarks. In theory, of course, instead of cities, we talk about sets. Also, instead of "sending letters" we say "the function maps to ". is called an **image** of .

The set is called the **domain** of the function . Go to the beginning of this section and see that we said "*every* citizen from the city of ..." This means that our function is defined everywhere in the domain. In practice, when you are given a function, you may have to describe the domain by removing all elements where your function is not defined.

## Here is a detective story

One citizen of mysteriously dies. Before dying, he manages to tell those around him that he had visited his correspondent in and was probably poisoned during the visit. The local detective's task will be simpler if the poor guy had been receiving letters from only one person. We write if receives letters from . If nobody in receives letters from more than one sender, then satisfies the definition of a function.

**Definition**. We say that the **inverse function exists** if satisfies the definition of a function. Equivalently,

(1) whenever , their images should be different: .

In Figure 2, the inverse does not exist. In Figure 3, it does. The function takes us along the red arrow from to its image . The inverse function takes us back along the yellow arrow from to , which is called a **counterimage** (or **preimage**) of .

## How to find the derivative of the inverse

From now on the arguments and values will be real.

The direct function maps to and the inverse function maps back to . This gives an obvious identity: . Similarly, . Since this is true for all , we can differentiate this equation using the chain rule to get

Hence, which is often written as with

## Conditions sufficient for existence of the inverse

**Condition 1**. Let the function be *strictly monotone*. If it is increasing, for example, then this condition means that implies . Obviously, this implies (1) and the inverse exists.

**Condition 2** (condition sufficient for Condition 1). Suppose the derivative of does not change sign. For example, if the derivative is everywhere positive (think about the exponential function), then the function is strictly increasing and we can apply Condition 1.

**Example**. The function is not one-to-one on the real line. If we take as its domain , it will be one-to-one (it is strictly monotone and its derivative is positive on the set of positive numbers). On this set, its inverse is . The equations and take the form and . The equations and are equivalent.

You must be logged in to post a comment.