Week 2, Lecture 5: Monotone Convergence and Squeeze Theorem

Welcome to Lecture 5: Taming the Wild Sequences

Take a deep breath—you have successfully climbed over the steepest conceptual wall in this entire course. By mastering the formal $\epsilon\text{-}N$ definition and the Algebraic Limit Laws in Lecture 4, you built the unbreakable logical foundation of real analysis. You should be incredibly proud of that milestone.

As we close out Week 2, our focus on sequences shifts from building the foundation to actually using it. The algebraic tools you learned are powerful, but they have a catch: they only work if you already know the exact target limit, or if the sequence is perfectly cooperative. What happens when you face a sequence whose final destination is completely invisible, or an expression so chaotic that standard algebra shatters?

Don't worry—we aren't going back to scratchpad proofs. Today, we step outside the algebraic sandbox and hand you the advanced toolkit. You will learn to use the Monotone Convergence Theorem to mathematically guarantee a sequence converges even when you are flying blind, apply it to capture the elusive Euler's number ($e$), and master the Squeeze Theorem to force wildly oscillating sequences into submission. Let's dive in!

Imagine a sequence as a machine operating under two strict physical laws:

• The Engine (Monotonicity): The sequence is strictly increasing. Every new term steps higher than the last. It never reverses direction.
• The Glass Ceiling (Boundedness): There is a hard, impenetrable ceiling (an upper bound $M$) that the sequence is mathematically forbidden from crossing.

If an engine pushes upward forever, but is physically trapped beneath a ceiling, it cannot escape to infinity, and it cannot retreat. It has only one option: it must eventually crush flat against a final barrier.

📜 Historical Aside: Ghosts, Swiss Cheese, and the Monotone Convergence Theorem

The Monotone Convergence Theorem doesn't belong to just one person; it was forged in the mid-to-late 19th century by a group of mathematicians—primarily Karl Weierstrass, Richard Dedekind, and an earlier pioneer named Bernard Bolzano—who were trying to fix a massive logical crisis in calculus.

The Ghostly Foundations of Calculus

When Isaac Newton and Gottfried Leibniz invented calculus in the 1600s, they built it on "infinitesimals"—tiny, magical quantities that were smaller than any measurable number, but not quite zero. For 150 years, the calculations worked beautifully, but logically, it was a house of cards. Critics famously mocked infinitesimals as the "ghosts of departed quantities." You couldn't rigorously prove why limits worked; you just trusted the machinery.

The Swiss Cheese Number Line

By the 1800s, mathematicians wanted to tear down the ghostly magic and rebuild calculus on unbreakable logical fences. But they hit a terrifying roadblock: they realized their number line was full of microscopic holes. If your mathematical universe relies heavily on fractions, irrational numbers like $\sqrt{2}$, $\pi$, and $e$ effectively don't exist. The number line is basically Swiss cheese.

Imagine our exact sequence, $a_n = (1 + 1/n)^n$. It is strictly increasing, and we proved it is trapped below a ceiling of $3$. But if the number line has a "hole" where $e$ is supposed to be, the sequence climbs up, hits the ceiling, and just falls into the void. It has no physical number to land on.

The Fix: The Completeness Axiom

Weierstrass and Dedekind realized they had to mathematically outlaw these holes by formalizing the Completeness Axiom (stating there are absolutely no missing pixels on the real number line).

The Monotone Convergence Theorem is the direct, practical tool that enforces this law. It guarantees that if a sequence is marching steadily in one direction (monotonic) and is trapped by a wall (bounded), the exact spot where it gets crushed against the wall is a guaranteed, physical, existing real number. It will not fall through a hole.

Note: If you look this up and see the "Monotone Convergence Theorem" attributed to Beppo Levi in 1906, that is a different, much heavier theorem used for advanced Lebesgue integration. The sequence version here is the 19th-century Weierstrass/Dedekind foundation of the real numbers.

While the concept of continuous compound interest led Jacob Bernoulli to first observe this mathematical constant in 1683, it was the brilliant 18th-century Swiss mathematician Leonhard Euler who truly unlocked its profound power. Euler was the first to designate this constant with the letter $e$ (around 1727) and systematically map its properties across different branches of mathematics.

Euler didn't just find a number; he proved that $e$ represents the fundamental rate of growth for all natural continuous processes. By proving that it can be defined as the infinite limit of $\left(1 + \frac{1}{n}\right)^n$, Euler hardwired $e$ into the very foundations of calculus, physics, and probability. Today, we are going to reconstruct his exact proof to show that this sequence safely converges.

Unpacking the Algebra: Where do the factorials come from?

Before we formally define Euler's number as a limit, we have to deal with a formidable piece of algebra: expanding $a_n = \left(1 + \frac{1}{n}\right)^n$ using the Binomial Theorem. If textbooks throw this at you as a wall of symbols, it looks like magic. Let's break down what is actually happening behind the scenes.

1. What is a factorial really? (The Over-Counting Fix)

An exclamation mark like $2!$ or $3!$ is just a counting tool. When you expand an $n$-th power, you are choosing which parentheses to pull the $\frac{1}{n}$ from. Because choosing slot A then slot B is the exact same pair as slot B then slot A, you end up double-counting. The factorial in the denominator (like $2! = 2 \times 1$ or $3! = 3 \times 2 \times 1$) is literally just a correction factor to remove that duplication.

2. Decoding the "n choose k" Notation & Mechanically Building the Terms

The Binomial Theorem relies on a shorthand notation written with giant parentheses: $\binom{n}{k}$, which is read aloud as "n choose k". It simply answers the question: "If I have a total pool of $n$ items (parentheses), how many ways can I pick a group of $k$ of them?"

Mathematically, it is calculated using factorials like this:

Where does the Combinations Formula actually come from?

Let's completely step away from the $x$'s and $n$'s for a second. Let's build the formula $\frac{n!}{k!(n-k)!}$ from scratch using a concrete example so you can see exactly why every piece exists.

Imagine you have 5 different letters (A, B, C, D, E), and you need to grab a group of 3 letters.

Step 1: Making the picks (The Fundamental Counting Principle)

To pick 3 letters out of 5, you make three choices in a row:

• You have 5 options for your first pick.
• You have 4 options left for your second pick.
• You have 3 options left for your third pick.

Step 2: The First Factorial Trick (Chopping off the tail)

Mathematicians want to write $5 \times 4 \times 3$ using factorials. But a full $5!$ is too long: it is $5 \times 4 \times 3 \times \mathbf{2 \times 1}$. We need a way to mathematically "chop off" that $2 \times 1$ tail.

We do this by dividing by $2!$:

Translating to algebra: If you have $n$ total items (5) and you are picking $k$ of them (3), the unpicked "tail" you are chopping off is $(n-k)$ items long. So, to get your starting multiplication chain, the formula is: $\frac{n!}{(n-k)!}$.

Step 3: The Second Factorial Trick (Fixing the Over-Counting)

There is a major flaw in our 60-way count from Step 1. Our tree diagram treated order as sacred: it counted picking "A then B then C" as a completely different path than picking "C then B then A".

But if you are just grabbing a handful of letters to throw into a bag, the order doesn't matter. Once they are sitting inside the bag, {A, B, C} is the exact same group as {C, B, A}. Our list accidentally committed massive double-counting.

How many times did we over-count? ($3!$)

Let's isolate just one single group of three letters: A, B, and C. How many different ways can you shuffle or arrange that exact same group into a line?

• 3 options for the first position.
• 2 options for the second position.
• 1 option for the third position.

Multiplying those together gives $3 \times 2 \times 1 = \mathbf{3!}$ (or 6) different ways to write those same letters (ABC, ACB, BAC, BCA, CAB, CBA). Every single unique handful in our list has been bloated by a factor of exactly 6.

The Algebraic Rule: Because every handful of size $k$ can be internally rearranged in $k!$ different ways, our initial ordered count bloated every group by a factor of $k!$. To find the true number of unique groups, we must divide our previous formula by $k!$.

Step 4: Putting it all together

We started with $\frac{n!}{(n-k)!}$ to make our initial picks, and then we divided that entire thing by $k!$ to erase the duplicates. Stick them together in the denominator, and you get the final formula:

What is the actual goal of expanding?

Before we connect our combinations formula to algebra, we have to look at the big picture. When you "expand" a binomial like $(1+x)^5$, your goal is to turn a compacted equation into a long, stretched-out polynomial. If you multiplied all five brackets together by hand and grouped the matching pieces, your final answer would form a staircase of powers from $x^0$ up to $x^5$:

The grand goal of the Binomial Theorem is to figure out exactly what numbers (coefficients) go into those blank $[?]$ boxes without having to do hours of tedious multiplication. To do this, mathematicians don't look at the whole equation at once. They shine a spotlight on one single box at a time.

The "Vending Machine" Rule of Multiplying Brackets

Let's figure out what goes in the $x^3$ box (which is the step where $k=3$). An exponent of $5$ means multiplying $5$ identical brackets together. Let's label them A through E so we don't lose track:

Step 4: Putting it all together

We started with $\frac{n!}{(n-k)!}$ to make our initial picks, and then we divided that entire thing by $k!$ to erase the duplicates. Stick them together in the denominator, and you get the final combinations formula:

Step 4a The Big Picture: What is the actual goal of expanding?

When you "expand" a binomial like $(1+x)^5$, your goal is to turn a compacted equation into a long, stretched-out polynomial. If you multiplied all five brackets together by hand and grouped the matching pieces, your final answer would form a staircase of powers from $x^0$ up to $x^5$:

The grand goal of the Binomial Theorem is to figure out exactly what numbers (coefficients) go into those blank $[?]$ boxes without having to do hours of tedious multiplication. To do this, mathematicians don't look at the whole equation at once. They shine a spotlight on one single box at a time.

Step 4b The Distribution Rule: Selecting $1$ or $x$ from Each Bracket

When we multiply $n$ identical binomial factors together—like $(1+x)(1+x)\dots(1+x)$—algebraic distribution requires us to generate every possible product by choosing either $1$ or $x$ from every single bracket and multiplying those choices together.

To see how this works in practice, let's look at a concrete example: expanding $(x+1)^5$ (five brackets multiplied together), and get the coefficient for $x^3$:

Generalizing this to $n$ brackets: to get a power of $x^k$, we must choose $x$ from exactly $k$ brackets out of the $n$ total factors, and $1$ from the remaining $(n-k)$ brackets. Every such path evaluates to $x^k$, and the total number of times it appears is how many ways we can make that selection.

So, how many total $x^3$'s pop out to fill in our box? We just have to ask: "How many different ways can we pick 3 machines out of 5?"

This is the exact letter game we just solved! We use the combinations formula we built in Steps 1 through 3 to find the answer: $\binom{5}{3} = \frac{5!}{3!2!} = 10$ ways. Because there are $10$ unique groups of 3 machines, $x^3$ gets added to itself $10$ times. The number in our $[?]$ box is $10$.

Step 4c Putting It Into the Master Formula

When we generalize this logic from $5$ total machines to $n$ total machines, and from a specific step like $3$ to any step $k$ on the staircase, we get the complete blueprint for every single term in the expansion:

The Binomial Theorem is just this exact logic applied to every step of the staircase (from $k=0$ up to $k=n$) all added together:

Now, let's plug in $x = \frac{1}{n}$ to match our sequence $a_n = \left(1 + \frac{1}{n}\right)^n$ and watch how the algebra naturally cleans itself up for the first few terms:

3. The Final Collapsed Form

When you assemble all these pieces, the sprawling binomial expansion collapses into this clean, structured formula:

Look at what this does for us: every single term containing $n$ has $n$ in the denominator. As $n$ grows toward infinity, those correction terms vanish, leaving us with a pure sum of factorials that defines $e$!

$$a_n = 1 + 1 + \frac{1}{2!}\left(1 - \frac{1}{n}\right) + \frac{1}{3!}\left(1 - \frac{1}{n}\right)\left(1 - \frac{2}{n}\right) + \dots + \frac{1}{n!}\left(1 - \frac{1}{n}\right)\dots\left(1 - \frac{n-1}{n}\right)$$

$$a_n < 1 + 1 + \frac{1}{2!}(1) + \frac{1}{3!}(1)(1) + \dots + \frac{1}{n!}(1)\dots(1)$$

$$a_n < 1 + 1 + \frac{1}{2!} + \frac{1}{3!} + \frac{1}{4!} + \dots + \frac{1}{n!}$$

$$ \frac{1}{k!} \le \frac{1}{2^{k-1}} \quad \text{for all } k \ge 1 $$

Deep Dive: Why is this inequality true?

It is much easier to see why this is true if we flip the fractions upside down. In mathematics, taking the reciprocal of both sides of a positive inequality flips the sign. So, we just need to prove that $k! \ge 2^{k-1}$.

Let's write out exactly what both sides mean and stack them on top of each other term-by-term:

Look at the pairings. For the first two slots, the numbers are identical. But for every single slot after that, the building block in the top row (the factorial) is strictly larger than the $2$ in the bottom row. Because every individual piece of $k!$ is greater than or equal to the corresponding piece of $2^{k-1}$, the final multiplied total of $k!$ must be larger. Since $k!$ is a mathematically "heavier" denominator, dividing $1$ by it naturally results in a smaller fraction.

$$a_n < 1 + 1 + \frac{1}{2} + \frac{1}{4} + \dots + \frac{1}{2^{n-1}} < 1 + \frac{1}{1 - 1/2} = 3$$

Aside: The Algebraic Magic Trick Behind $S = \frac{a}{1-r}$

To calculate the exact finite total of an infinitely long geometric series, we use the standard formula $S = \frac{a}{1-r}$ (where $a$ is the starting term and $r$ is the common ratio). Instead of just memorizing the formula, look at the classic algebraic trick used to prove it.

Let $S$ equal an infinite sum:

If you multiply that entire equation by the ratio $r$, you shift every single term up by one power. If you then subtract that new equation from the original, almost every single term perfectly matches up and cancels out to zero:

You are left with a beautifully simple equation: $S - rS = a$. Factoring out the $S$ gives $S(1 - r) = a$, and dividing both sides by $(1 - r)$ leaves you with the standard formula: $S = \frac{a}{1 - r}$.

Applying it to our sequence

For our specific series ($1 + \frac{1}{2} + \frac{1}{4} + \dots$), the starting term is $a = 1$ and the multiplier is $r = \frac{1}{2}$. Plugging those directly into the engine gives:

The Physical Intuition: Imagine you are standing $2$ meters away from a wall. Your first step covers $1$ meter. Your next step covers $\frac{1}{2}$ a meter (half the remaining distance). Your next step covers $\frac{1}{4}$ of a meter. If you take an infinite number of these halving steps, you will eventually touch the wall at a total distance of exactly $2$ meters.

$$e = \lim_{n\to\infty} \left(1 + \frac{1}{n}\right)^n \approx 2.71828\dots$$

$$ \lim_{n \to \infty} \left(1 - \frac{1}{n}\right)^n = \frac{1}{e} \approx 36.8\% $$

When Algebra Fails: Handling Chaos

So far, we have relied on Algebraic Limit Laws. But those laws have a fatal weakness: they demand that every individual piece of your sequence behaves perfectly. If a sequence contains chaotic, oscillating terms—like $\sin(n)$—the standard limit laws immediately shatter because those terms never settle down on a single number.

If you cannot control a wild sequence algebraically, you don't fight it—you trap it.

Imagine you are walking a highly energetic dog down a long hallway. The dog is bouncing wildly, leaping from the floor to the ceiling. Tracking its exact path algebraically is impossible.

Now, imagine this hallway is actually a giant, slow-moving trash compactor. As you walk forward (as $n \to \infty$), the ceiling slowly grinds downward and the floor steadily pushes upward. You don't need to know the dog's exact frantic path to know where it will end up.

If both the floor and the ceiling eventually crush together to the exact same height, the chaotic bouncing is forced to stop. The dog is mathematically funneled into that exact same final position. This is the Squeeze Theorem.

📜 Historical Aside: Archimedes' Polygons and the Taming of Chaos

Unlike the Monotone Convergence Theorem, which was born in the 19th century to fix logical holes, the intuition for the Squeeze Theorem is over two thousand years old. It was originally invented by Archimedes of Syracuse around 250 BC.

The Ancient Geometric Trap

Archimedes didn't have algebra or limits; he had geometry. He wanted to calculate the exact value of $\pi$ (the circumference of a circle), but curved lines were impossible to measure directly. So, he built a trap. He drew a polygon perfectly inside the circle (the floor) and a polygon perfectly outside the circle (the ceiling).

By endlessly doubling the number of sides on both polygons, he forced the straight lines to hug the curve tighter and tighter. He physically "squeezed" the circle until the inner and outer boundaries crushed together, famously proving that $\pi$ was trapped strictly between $\frac{223}{71}$ and $\frac{22}{7}$. He called this the Method of Exhaustion.

The 19th Century Algebraic Upgrade

Fast forward to the 1820s. Mathematics was undergoing the same rigorous "arithmetization" led by Augustin-Louis Cauchy, Karl Gauss, and later Karl Weierstrass. Calculus was suddenly being used to study incredibly complex, chaotic phenomena like heat waves, acoustics, and oscillating signals (Fourier series).

Standard algebraic limit laws were completely useless against functions that vibrated infinitely, like $\frac{\sin(x)}{x}$. Cauchy and his contemporaries realized they could take Archimedes' ancient geometric trap and write it as an unbreakable algebraic law. They needed the Squeeze Theorem because it was the only mathematical tool that didn't care how chaotic or violent a sequence's journey was—it only cared about the boundaries. By mathematically translating Archimedes' polygons into sequences $A$, $B$, and $C$, they gained the power to tame oscillating chaos.

$$\lim_{n\to\infty} \frac{\sin(n)}{n}$$

$$-1 \le \sin(n) \le 1$$

$$\underbrace{-\frac{1}{n}}_{\text{Floor } (a_n)} \le \underbrace{\frac{\sin(n)}{n}}_{\text{Target } (b_n)} \le \underbrace{\frac{1}{n}}_{\text{Ceiling } (c_n)}$$

🏁 Lecture Wrap-Up: The Power of Boundaries

In this lecture, we evolved beyond the straightforward Algebraic Limit Laws. You learned that when a sequence’s formula is too complex, or its target is entirely unknown, we don't fight the algebra—we build physical traps.

• The Monotone Convergence Theorem provided an escape hatch, proving that an ever-climbing sequence bounded by a ceiling has no choice but to converge.
• The Anatomy of $e$: We used the MCT to securely construct Euler's number, pulling apart the Binomial Theorem and combinatorial logic ($n$ choose $k$) to see exactly how factorial denominators prevent over-counting.
• The Probability Limit: We flipped compound interest upside down with the lottery thought experiment, proving that infinite chances at infinite odds settle on a strict mathematical floor ($1/e$).
• The Squeeze Theorem gave us the ultimate trash compactor, allowing us to crush chaotic, oscillating sequences (like $\frac{\sin(n)}{n}$) into a single predictable limit by controlling only their outer boundaries.

Contextualizing your toolkit: You now possess the exact mathematical machinery that 19th-century giants like Weierstrass and Cauchy used to rigorously ground calculus. You are no longer just computing limits; you are leveraging the fundamental completeness of the real number line to prove that they must exist.

Ready to test your understanding?

Work through the progressive interactive problems to lock in these foundational concepts before moving on to Lecture 6.

Launch Interactive Problem Stepper ➔