Limits through the Lens of Linear Operators

I've been tutoring students in Calculus for about three years, both at the college level and at the AP high school level. One thing I've noticed is that many of the trickier problems take advantage of the fact that limits, derivatives, and integrals are all linear operators. However, the notion of a linear operator doesn't come up until near the end of a first course in Linear Algebra, so this stuff isn't made clear until well after students have finished calculus.

Recently, I tried doing things the other way around. I was working with a very bright student who was struggling in AP Calculus. After a few sessions, I could tell that the student's core algebra skills were pretty solid, but that there still seemed to be some trouble with the mechanics of solving limit and derivative problems.

First I tried showing how constants can be factored out of limits, and that the operation of finding a limit can be distributed over addition. The student seemed to follow, but not necessarily buy in to what I was saying. Then I rushed on to derivatives and I said "See! The derivative is defined as a limit, so we should expect the same thing here." Well, of course that was a bad explanation, not just because it was hand-wavy, but particularly because it wasn't illuminating.

So I decided to back up and try teaching the mechanics of how linear operators behave in the abstract, with no reference to calculus.

The formal definition of a linear operator is a bit much for an AP Calc student, but the salient features can be presented in terms of Pre-Calculus concepts as follows:

Given two functions \(f(x)\) and \(g(x)\), a linear operator \(L\) will take \(L(a*f(x) + b*g(x))\) and give you \(a*L(f(x)) + b*L(g(x))\).

In this case, the student was a bit jarred about what to do with expressions like \(L(f(x))\), but I explained that \(L\) is an unknown function just like \(f(x)\) and that the important bit is to learn how it affects the algebra and order of operations. To cement the idea, I had the student do some practice problems like the following:

• Show that \(L(5x + 4x^2)\) can be simplified to \(5*L(x) + 4*L(x^2)\).
• Show that \(L(x + x^2) + L(2x) + L(2x^2)\) can be simplified to \(3L(x) + 3L(x^2)\).
• Show that \(L(a*f(x) + b*g(x)) - L(b*g(x) - c*h(x))\) can be simplified to \(L(a*f(x) + c*h(x))\).

Simple exercises like these can help a student grasp the concept of a linear operator in a mechanical way rather than in a conceptual way. Given that the concepts of linear operators don't become crucial until Linear Algebra, I think this sin is a forgivable one, especially in light of the opportunity for facilitating calculus exercises (and thus calculus concepts).

Once the student had a decent grasp of how to manipulate expressions with linear operators, I went back and reintroduced the limit concept from scratch with an emphasis on why we should expect limits to behave like linear operators. I wasn't super rigorous about this, but because of the way limits are usually taught (i.e. using intuition about long-term behavior rather than \(\epsilon-\delta\) neighborhood arguments) I really didn't have to.

After that, I set some more problems for the student, with the goal of making them appear very complex. The idea here was to promote confidence by demonstrating that nasty-looking limit problems could be broken down into separate components because of the fact that the limit is a linear operator. For example, I had the student prove statements similar to the following:

\[\begin{aligned} \lim_{x \to 0}\frac{1}{x} + \lim_{x \to 0}\frac{-1}{x} = & 0\\ \lim_{x \to 0}\frac{\sin(x)}{x}+\cos(x) = & 2\\ \lim_{x \to 0}\frac{\sin(x) + 1}{x} - \lim_{x \to 0}\left(\frac{1}{x} - \cos(x) \right)= & 2\end{aligned}\].

This was particularly successful, because being able to break these problems down into smaller components that can be simplified and cancelled helped the student realize that \(\lim_{x \to 0}\frac{sin(x)}{x}\) was the only part that really presented a challenge. Once I showed how to prove that the limit is 1, everything else seemed to become second nature.

In later sessions, I introduced the student to derivatives (and, when the time came, integrals) as examples of linear operators. I will discuss those experiences in future posts, but for now I will just note that the student in question quickly became a top performer and remained so for both semesters of the AP Calculus course. Mostly this was because of the student's stellar work ethic, but I am inclined to believe that this linear operator approach facilitated the student's ability to grasp new material more quickly and reinforced confidence by unmasking the hidden tricks in many of the harder textbook problems.