I'm writing the 4th book in the series of Althea's MathMysteries, Althea and the Mysteries of Calculus. (I've written the first two, and they still need more beta readers before we publish them. If you're interested, email me at altheasmathmysteries on gmail. The 3rd isn't written yet.)
I taught calculus almost every semester for 28 years, and then I retired and started writing more. But writing this book is making me think about how I could have taught calculus better. I had already departed from the textbook in many ways, but I'm seeing today how chained to it I was, at least once in a while.
I taught the conventional proof for the derivatives of sine and cosine. I made a (very dense) four-page handout, walking students through the steps necessary. This involves a bit of review of trig, and some geometry, and some sophisticated use of inequalities. It's a lot. I'll guess now that maybe one student a year could really follow all that.
I knew there was another way, and I briefly showed my students the cool way to see this geometrically, if you allow for infinitesimals. But for some reason, I would stumble when explaining that way. So I never threw out the conventional proof.
This post is all about limit-based proofs versus infinitesimal-based proofs in a beginning course in calculus. It will help to know a little of the history.
- In the late 1600s, Newton and Leibniz put together what we now call calculus. Newton spoke of fluxions when he wanted to discuss infinitely small quantities. But what the heck is an infinitely small quantity?! Calculus was a huge boon to science, and there was much anxiety about whether its foundations were logically sound.
- It took other mathematicians desperately seeking a solution to this issue 150 years (early 1800s) to develop a logically sound foundation for calculus. That tells you how hard a problem it was. And their solution was so logically complicated that you have to be a lover of logic problems to actually get it. Their solution was limits. Back to that in a moment. First let's move another 150 years forward.
- In 1960, Abraham Robinson developed a number system called hyperreals (very carefully, one logial step at a time). And in that number system, infinitesimals are fine. But the textbooks had been written long ago, and no one was going to throw out those limits. (So sad.)
Here's a definition (written almost like a poem, because that helps):
limit (x -> c) f(x)=L (read this as "the limit, as x approaches c, of f(x) is L")
means
for any epsilon > 0
there exists a delta > 0
such that
if | x-c | < delta
then | f(x)-L | < epsilon
Yeah, right. I'm guessing that bit sounds perfectly alien to 80% of the people who read my blog. (Does anyone still read blogs?) And it would be over 99% if you all hadn't already self-selected as math lovers.
It might be hard to understand how that solved the problem of infinitely small quantities. But it does. You want an example? Ok.
limit (x -> 2) (x^2 - 4) / (x - 2) = 4. We can see that by multiplying out the top and factoring out x-2. If we don't do that, we have 0/0, which some of us like to call indeterminate. We know that 0/0 is undefined, but what we really want to know is what happens when x is infinitely close to 2. And then we don't have 0/0, we actually have x+2. And 2+2=4, so we have an answer. And that answer actually tells the slope of the y=x^2 parabola when x=2. That's one of the things calculus can do, and the limit stuff makes sure that it works.
But the limit stuff involves lots of hard algebra. And it's often hard to see where a step in the reasoning came from. So let's look at how infinitesimals solve this problem. That four-page handout for sine and cosine derivatives becomes a one-page diagram with a bit of explanation. I got the online handout, by Alex Alemi, years ago. [It just took me well over an hour to find it tonight online. Does he not see this as valuable? I found other stuff he's written much more quickly. Hmm.]
If you go look at that handout, it might take you a bit, but the connections between sine and cosine and their derivatives make visual sense once you get it. After writing some dialogue between Althea and her friends, as they try to figure out what's going on, I wrote their product rule adventures the next day, and then today they blasted through the quotient rule. (In the conventional proofs, both of these have non-intuitive algebra steps.) After their adventures with the sine and cosine derivatives, Aiden starts to talk about their time in Infinitesimal Land.
Excerpt from Althea and the Mysteries of Calculus
Perhaps I need to study a good calculus through infinitesimals textbook before finishing Althea's story. Any recommendations?