# Why are limits considered as a central idea of calculus beyond just mechanical computations?

I had my first encounter with Calculus a decade ago. Back then it was purely mechanical. Formulas and rules of derivation and integration were being written on the board without deriving it and were told to compute a bunch of derivatives and integrals without even having a notion of functions, limits and other essential concepts. Today i have a burning desire to relearn all the forgotton math, calculus in particular.

I'm having a bit of concern with the notion of Limits and Continuity. Is it really essential that i have solid understanding of limits and continuity? I could teach myself differential calculus after analytically computing limits as well as L'Hopital's rules and up to partial derivatives and basic integrals up to trig substitution. Frankly speaking I still don't have a solid understanding of Limits.

Please explain what a limit is in layman's terms, preferbly with an easy to understand analogy. I looked in wikipedia and searched for a lot of youtube videos, but couldn't make sense. Also explain Continuity in simple terms.

I would also like to know a little bit about Linear and quadratic approximations.

Many people agree that it's indeed possible to keep differentiating and integrating functions without even knowing what a limit is! Big question is why is it considered as a central idea of calculus?

• It is indeed possible (sadly enough) to be able to keep differentiating functions without even remembering the notions of limits and continuity, if you treat the differentiation formulae as some rules handed down from on high. $\epsilon-\delta$ is now of course the standard way to consider limits. I am personally fond of Bressoud's treatment of $\epsilon-\delta$ as a "game" in his book A Radical Approach to Real Analysis. Commented Aug 31, 2011 at 14:49
• On continuity: this easily follows when you're on solid ground with limits. If the limit of a function at some point is the same whether you approach from the right or from the left, then it it is continuous at that point. Commented Aug 31, 2011 at 14:51
• Arturo Magidin gave a good explanation of limits in this answer. Commented Aug 31, 2011 at 15:25
• It is true that you can differentiate and integrate many functions without knowing what a limit is. It is not true that differentiating and integrating many functions is the same as calculus. Commented Aug 31, 2011 at 15:51
• @J.M.: We all know what you mean, but for the sake of correctness: A function is continuous at $x = a$ if $\displaystyle\lim_{x \to a^-} f(x) = \displaystyle\lim_{x \to a^+} f(x)$ and their common value is $f(a)$. Commented Aug 31, 2011 at 16:30

Here is a (possibly over simplified) way to look at continuity and differentiability for functions from $\mathbb{R}$ to $\mathbb{R}$.

Continuity means that there are no "breaks" or "jumps" in the graph of the function. If I were to draw the graph of the function, I would be able to do it without lifting my pen. Continuity also means that around every point, I can choose an interval small enough so that the function varies as little as I want.

A limit will be when we look at the behaviour around a point, but not necessarily at the point itself. It is very important to note that for a continuous function, $\lim_{x\rightarrow a} f(x)=f(a)$ for every point. (This can actually be an alternate definition for continuity)

Differentiability means that at every point, if we zoom in really really close, the function looks like a straight line segment. It means that the function will be continuous, and that there will be no sharp points in the graph of the function.

Hope that helps,

Remark: This answer may be a bit murky, and I personally suggest learning the $\epsilon$-$\delta$ to fully understand things for yourself.

• On the matter of differentiability and "zooming": see this. Commented Aug 31, 2011 at 14:55
• @alok I am not sure if this would help everyone, but I found it easier to understand the mechanics of the $N_0$-$\epsilon$-$\delta$ business easier in the context of limits of sequences than for functions. (For example, how to prove that the sequence $(3n^2+5n+2)/n^2$ converges to $3$ as $n \to \infty$ from first principles?) Doing such examples would make you more comfortable with limits in general, and $\epsilon$-$\delta$ in particular. Note that the two concepts aren't the same, but they have a lot of similarities. Commented Aug 31, 2011 at 15:00
• @Srivatsan: That is a good point, perhaps post it as an answer? Commented Aug 31, 2011 at 15:02
• @alok: it depends; what end point did you have in mind for yourself? Commented Aug 31, 2011 at 15:23
• @alok Re "concept of limits somewhat akeward. I simply can't grasp it." I specifically did not prescribe to you the plug-and-chug (nice phrase!) problems, of which there are many (and I wouldn't call these the fundamentals of limits). I rather asked you to understand the concept of limits and the idea behind $\epsilon$-$\delta$, by doing simple examples the hard way. (If your question is whether such an understanding is important for you, I can't say that.) Commented Aug 31, 2011 at 15:46

Consider a curve/line/function $L$, not necessarily straight nor continuous. Take a point $a$ that lies in the line $L$.

Consider two other points $a^-$ and $a^+$, both also lies in $L$. The point $a^-$ lies to the left of $a$ (i.e. $a^- < a$) while the point $a^+$ lies to the right of $a$ (i.e. $a < a^+$).

Take another two points $b^-$ and $b^+$, both also lies on $L$. However, now $b^-$ lies between $a^-$ and $a$ (i.e. $a^- < b^- < a$), while $b^+$ lies between $a$ and $a^+$ (i.e. $a < b^+ < a^+$).

Take another two points $c^-$ and $c^+$, both also lies on $L$. However, now $c^-$ lies between $b^-$ and $a$ (i.e. $b^- < c^- < a$), while $c^+$ lies between $a$ and $b^+$ (i.e. $a < c^+ < b^+$).

If we can repeat this process indefinitely, and at each time $L(x^-)$ and $L(x^+)$ are getting closer and closer to the same point (i.e. as $x^- \to a$ then $L(x^-) \to L(a)$ and as $x^+ \to a$ then $L(x^+) \to L(a)$) as we take more and more steps, then we say that "at point $a$, the line $L$ has the limit $L(a)$", i.e.

$$\lim_{x \to a} \space L(x) = L(a)$$

Next, there are a similar notion of one-sided limits. If you only do this process of taking a point closer and closer from one side, then you have a one-sided limit. There is also the notion of limits of infinities, of what happens to the value of the function as they tend to infinity.

In summary, limit is the study of the function's behavior as they try to reach a certain point.

You may like to go through this Khan academy video to understand limits in very simple terms.In general Khan academy videos on calculus is a great resource for an autodidact.

However,if you want to learn the subject properly with good guidance, I like to suggest you to the $18.01$SC OCW scholar course on single variable calculus.

Remark: If you are not familiar with the notion of the limit of a sequence this answer does not repply to your question.

I assume you are familiar with the notion of the limit of a sequence $(x_{n})$. The limit of a function $f(x)$ when $x$ tends to $a$ is $b$ if and only if for any sequence $(x_{n})$ of values of $x$ approaching $a$ but different from $a$ the corresponding sequence $y_{n}=f(x_{n})$ of values of the function approaches $b$.

[Edited] A function $f(x)$ is continuous at $a$ if and only if (see figure)

1. $a$ is on its domain and
2. the difference $k=f(x)-f(a)$ tends to $0$ when the difference $h=x-a$ approaches $0$. (In other words, when for any sequence $h_1=x_1-a,h_2=a-x_2$, $h_3=a-x_3,\dots$ approaching $0$ but different from $0$, the corresponding sequence $k_1=f(x_1)-f(a),k_2=f(x_2)-f(a),k_3=f(x_3)-f(a),\dots$ tends to $0$).