1

I'm almost in the very beginning of my math journey. Please, have patience because I'm not a math guy (which I'm really willing to change now).

I'm trying to learn calculus for my math class at university. And I'm watching the 3Blue1Brown videos. Were are trying to find the derivative of $s(t) = t^3$. So my question is: how did he simplify this part:

$$\frac{ds}{dt} = \frac{(2+dt)^3 - (2)^3}{dt}$$

to get

$$ \frac{2^3 + 3(2)^2dt + 3(2)dt^2 + dt^3 - 2^3}{dt}$$

At first, I thought it would be binomial to the power of 3 rule, then monomial to the power of 3 simplification, and something related with the subtraction. But alas my knowledge is not enough to comprehend it.

UPD: I thought that the numerator had to be:

  1. $$(2 + dt)(2 + dt)(2 + dt) - (2)(2)(2)$$
  2. $$2(2 + dt) + dt(2 + dt)(2 + dt) - (2)(2)(2)$$

UPD: And this is where I stumbled upon I guess. I had to apply the binomial formula when binominal is being brought to the power of 3, right? Then $$(2+dt)^3$$ would look like $$a^3 - 3ab(a - b) - b^3$$

I'm really having hard time trying to understand it. What rules apply to this simplification? What topics do I need to know before trying to solve this?

  • It is by using the binomial formula ! Maybe you can show your work in your post so we can see where you're making a mistake in the calculations. Edit your post to do this. – Digitallis Jan 15 '23 at 18:47
  • I thought that the numerator had to be (2 + dt)(2 + dt)(2 + dt) - (2)(2)(2). And then I got stuck – ArtemNovikov Jan 15 '23 at 18:50
  • 1
    Edit your post to include this information. Most users won't read comments to answer your question. As for the math : good ! But what is the next step. Again edit your post to include this. – Digitallis Jan 15 '23 at 18:52
  • 1
    Please use $Mathjax$ for equations –  Jan 15 '23 at 19:03
  • 1
    This may helphttps://math.stackexchange.com/questions/4616029/why-will-there-always-be-this-term-in-xhn/4616334#4616334 – Ethan Bolker Jan 15 '23 at 19:06
  • I've just updated the question. Is it correct that I should've used the cube of a binomial formula on the step 2? – ArtemNovikov Jan 15 '23 at 19:12
  • 3
    Yes, the cube of the binomial is what you want. The last way you wrote it is correct but uncommon. You can see that it's the same as the binomial theorem if you apply the distributive law. Learning calculus when you are stuck at this algebra level will be hard. Consider some algebra refresher first. – Ethan Bolker Jan 15 '23 at 19:44
  • I can echo what Ethan has said. Any calculus teacher will tell you that the calculus is the easy part - the mistakes and the hard parts are always the algebra. – Ninad Munshi Jan 15 '23 at 20:05
  • 2
    I strongly recommend you resist the urge to apply various formulas until you're firmly grounded on being able to carry out the computation without formulas. I base this on the fact that you're casting about for a formula simply to expand a product of $3$ binomials. Sure, it's not a $10$ second calculation, but it's also not all that involved --- $(2+dt)(2+dt) = (2)(2) + (2)(dt) + (dt)(2) + (dt)(dt) = 4 + 4dt + dt^2.$ Note: $(dt)^2$ is being written as $dt^2,$ a notation I think is bad since it conflicts with what $dt^2$ usually means, which is the differential of $t^2.$ (continued) – Dave L. Renfro Jan 15 '23 at 20:19
  • 2
    Therefore, $(2+dt)^3 = (2+dt)(4 + 4dt + dt^2) = (2)(4) + (2)(4dt) + (2)(dt^2) + (dt)(4) + (dt)(4dt) + (dt)(dt^2) = 8 + 8dt + 2dt^2 + 4dt + 4dt^2 + dt^3 = 8 + 12dt + 6dt^2 + dt^3.$ – Dave L. Renfro Jan 15 '23 at 20:19
  • @DaveL.Renfro so it took me 8 months of pre-algebra and algebra 1 refreshers until I finally understood what you've been trying to explain to me. I've just re read all of the answers over here and your explanation finally seemed obvious to me. Thanks a lot! – ArtemNovikov Oct 14 '23 at 00:05
  • 1
    @ArtemNovikov: Wow, that's some serious dedication! Of possible interest is this Mathematics Educators answer. And maybe also this MSE answer, although it's a bit more algebra intensive. – Dave L. Renfro Oct 14 '23 at 00:20

1 Answers1

3

I think the issue here is actually 3B1B's presentation, which is surprising, as he's usually really good with explanations. I'm not sure why he decided to computer at a specific value of $t$. That always seems to muddle things up. (Also I agree with Dave Renfro in the comments, that using $dt^2$ is kind of an abuse of notation).

Instead of $t=2$, let's consider an indeterminate $t$. We still basically start the same way:

$$\frac{ds}{dt} = \frac{s(t + dt) - s(t)}{dt} = \frac{(t+dt)^3-t^3}{dt}$$

Then we expand the polynomial:

$$\frac{ds}{dt} = \frac{ \color{blue}{t^3 + 3t^2 dt + 3t(dt)^2 + (dt)^3} - t^3}{dt} = \frac{3t^2dt+3t(dt)^2 + (dt)^3}{dt}$$

Notice that the portion in $\color{blue}{\text{blue}}$ is the same as the $(2+dt)(2+dt)(2+dt)$, just using $t$ instead of the specific value $2$. Now we can factor out $dt$ on the top of the RHS and divide, giving:

$$\frac{ds}{dt} = 3t^2 + 3tdt + (dt)^2$$

If we now decrease $dt$ so that $dt \to 0$, we finally simply get:

$$\frac{ds}{dt} = 3t^2$$

For $t=2$, we can then calculate $ds/dt=3 \cdot 2^2 = 12$.

The huge thing to understand is that $dt$ is a "quantity" we can manipulate algebraically, but doesn't properly "exist" in the way that, say, $t^3$ does. After any algebra is done to it, it must be moved down to $0$. This is called the method of infinitesimals, and I think it's really useful to look at calculus from this perspective, but do understand that it's a non-standard way of looking at things, and usually isn't taught, which I feel is a huge shame.

Eric Snyder
  • 3,007
  • Teaching non-standard analysis would require teaching the axiomatic basis of $\Bbb R$ also. Most of the adults I have met in Ontario have only vague notions about $\Bbb R.$ – DanielWainfleet Jan 16 '23 at 04:36
  • @DanielWainfleet I think the average person learns at least a non-axiomatic basis for the reals, insofar as the standard teaching of math through high school includes integers, fractions, and irrationals. You don't need an axiomatic basis to understand it; you only need one to formally prove it. And just as we don't start by teaching kindergartners about successor functions before learning to count, we don't need to teach a formal basis for infinitesimals in order to understand them. – Eric Snyder Jan 18 '23 at 11:16