Calculus I · Patterns That Go Somewhere

Where this is heading

The word “approaches” is the hinge of everything that follows. In Chapter 2 we ask what a function approaches as its input closes in on a point — the same idea, lifted from lists to curves. In Chapter 3 we use it to do the thing this whole book is named for: take the slope of a straight line between two points on a curve and watch what it approaches as the points slide together. That limit is the slope of the curve itself — the derivative. Every limit you compute now is practice for that moment.

And ee, which appeared here as nothing more than the destination of a climbing list, will turn out to be one of the two or three most important numbers in all of mathematics. You met it first as a sequence. That’s fitting: in this book, the deep things arrive quietly, as patterns that go somewhere.


Problems

Six tiers, each with a different job. Fluency and Practice are doing* — drill until the moves are automatic, then multi-step work where you pick the tool. Concept & counterexample and ◆ Diamond are thinking — traps, constructions, and Layer-2 proofs that test whether you understand, not just compute. Teach it is teaching — explain, write up, or fix reasoning for a peer, which is the surest way to find the gaps in your own. Looking ahead hands you Chapter 2 early. Do all of Fluency; do not skip Concept or Teach it — that is where the understanding actually sets.*

Fluency — do these fast, until they’re automatic

  1. First four terms of: (a) an=3n−1a_n=3n-1; (b) an=(12)na_n=\left(\tfrac12\right)^n; (c) an=nn+1a_n=\dfrac{n}{n+1}; (d) an=(−1)nna_n=\dfrac{(-1)^n}{n}.
  2. Arithmetic, geometric, or neither — and give dd or rr where it applies: (a) 7,10,13,16,…7,10,13,16,\ldots; (b) 3,−6,12,−24,…3,-6,12,-24,\ldots; (c) 1,12,13,14,…1,\tfrac12,\tfrac13,\tfrac14,\ldots; (d) 2,5,10,17,…2,5,10,17,\ldots.
  3. Use a formula, not a list: (a) the 1515th term of the arithmetic sequence with a1=4,d=5a_1=4,\ d=5; (b) the 66th term of the geometric sequence with a1=2,r=3a_1=2,\ r=3.
  4. Guess the limit (or say “no limit”): (a) 1n\tfrac1n; (b) 2n−1n\tfrac{2n-1}{n}; (c) (−1)n(-1)^n; (d) (0.8)n(0.8)^n; (e) n2n^2; (f) (−1)nn\tfrac{(-1)^n}{n}.
  5. Divide by the highest power: (a) lim⁡3n+1n+2\displaystyle\lim\frac{3n+1}{n+2}; (b) lim⁡n2−12n2+n\displaystyle\lim\frac{n^2-1}{2n^2+n}; (c) lim⁡5nn2+1\displaystyle\lim\frac{5n}{n^2+1}; (d) lim⁡2n3n2+1\displaystyle\lim\frac{2n^3}{n^2+1}.
  6. For each rr, does rn→0r^n\to0, →1\to1, blow up, or oscillate? r=13r=\tfrac13; r=−12r=-\tfrac12; r=1r=1; r=−1r=-1; r=2r=2.

Practice — choose the right tool

  1. lim⁡n→∞n2−3n2n2+1\displaystyle\lim_{n\to\infty}\frac{n^2-3n}{2n^2+1}.
  2. An ∞−∞\infty-\infty that hides a finite answer: lim⁡n→∞(n2+n−n)\displaystyle\lim_{n\to\infty}\left(\sqrt{n^2+n}-n\right). (Rationalize, then divide by the top power.)
  3. An ∞⋅0\infty\cdot0: lim⁡n→∞n(1+1n−1)\displaystyle\lim_{n\to\infty} n\left(\sqrt{1+\tfrac1n}-1\right). (Mirror Fight one from the chapter exactly.)
  4. The ee-family: (a) lim⁡(1+3n)n\displaystyle\lim\left(1+\frac3n\right)^n; (b) lim⁡(1−2n)n\displaystyle\lim\left(1-\frac2n\right)^n; (c) lim⁡(1+1n)3n\displaystyle\lim\left(1+\frac1n\right)^{3n}.
  5. Use the gauge “(exponent)×(base−1)→k⇒(\text{exponent})\times(\text{base}-1)\to k \Rightarrow limit =ek=e^k”: evaluate or classify (1+1n)n2\left(1+\frac1n\right)^{n^2}, (1+1n2)n\left(1+\frac{1}{n^2}\right)^{n}, and (1+5n)n\left(1+\frac5n\right)^{n}. State kk first each time.
  6. A self-defined sequence: a1=1a_1=1, an+1=12+ana_{n+1}=\sqrt{12+a_n}. Compute a2,a3a_2,a_3, then find the limit by letting LL satisfy the rule. (You’ll prove it converges in the Diamond tier.)
  7. A geometric sequence has a1=1,r=0.9a_1=1,\ r=0.9. About how many terms until it drops below 0.0010.001? (Use the idea behind the rn→0r^n\to0 proof, or trial and error.)
  8. Decompose, then take the limit: lim⁡n→∞n+(−1)nn\displaystyle\lim_{n\to\infty}\frac{n+(-1)^n}{n}. (Split into two pieces, one obvious, one you’ve already handled.)

Concept & counterexample — think; almost no arithmetic. State true/false and justify, or build the example asked for.

    1. If an→La_n\to L, then every term is within 0.010.01 of LL. (b) If a sequence is bounded, it converges. (c) If an→0a_n\to0, the sequence is eventually decreasing. (d) A sequence can converge to two different limits.
  1. Construct a bounded sequence that does not converge.
  2. Construct an increasing sequence that does not converge — and explain in one line which hypothesis of the Monotone Convergence principle your example is missing.
  3. Construct a sequence that converges to 55 but whose terms are sometimes above and sometimes below 55.
  4. Find the error. A student writes: “(−1)n(-1)^n gets arbitrarily close to 11 infinitely often and to −1-1 infinitely often, so it has two limits, 11 and −1-1.” Identify precisely what’s wrong, and state what is true.
  5. Is it possible? A sequence with an→0a_n\to0 whose running totals a1+a2+⋯+ana_1+a_2+\cdots+a_n nonetheless grow without bound. (Name an example if yes — this is a quiet preview of infinite series.)
  6. The Monotone Convergence principle needs two hypotheses, “increasing” and “bounded above.” Give one sequence showing the conclusion fails if you drop “bounded above,” and one showing it fails if you drop “increasing.” (This proves both hypotheses are load-bearing.)

◆ Diamond — Layer 2: prove it

  1. ◆ Prove 1n→0\dfrac1n\to0 from the ε\varepsilon–NN definition.
  2. ◆ Prove 1n2→0\dfrac{1}{n^2}\to0 from the definition.
  3. ◆ Prove nn+1→1\dfrac{n}{n+1}\to1 from the definition. (First show |nn+1−1|=1n+1\left|\tfrac{n}{n+1}-1\right|=\tfrac{1}{n+1}.)
  4. ◆ Prove 2n+1n→2\dfrac{2n+1}{n}\to2 from the definition.
  5. ◆ Suppose an→5a_n\to5. Using the limit laws, find lim⁡(3an−1n)\displaystyle\lim\left(3a_n-\tfrac1n\right), naming the law at each step.
  6. ◆ Prove sin⁡nn→0\dfrac{\sin n}{n}\to0 via the Squeeze Theorem: name the two walls, cite the theorem, and explain in one sentence why sin⁡n\sin n never needed a limit of its own.
  7. ◆ Prove the uniqueness of limits: a sequence cannot converge to two different values L≠ML\ne M. (Hint: if it did, take ε=|L−M|3\varepsilon=\tfrac{|L-M|}{3} and derive a contradiction — a term can’t be within ε\varepsilon of both.) Connect this to your answer in 15(d).
  8. ◆ For the sequence in problem 12, prove it is increasing and bounded above by 44, completing the argument that it converges.
  9. ◆ Prove Bernoulli’s inequality (1+h)n≥1+nh(1+h)^n\ge1+nh for h>−1h>-1, n≥1n\ge1, by induction.
  10. ◆◆ Prove (1+1n)n<3\left(1+\tfrac1n\right)^n<3 for every nn. Steps: expand by the binomial theorem; bound the kk-th term by 1k!\tfrac{1}{k!}; bound 1k!≤12k−1\tfrac{1}{k!}\le\tfrac{1}{2^{k-1}}; sum the geometric bound to get below 1+(1+12+14+⋯)=31+\left(1+\tfrac12+\tfrac14+\cdots\right)=3.
  11. ◆◆ Adapt the nn→1\sqrt[n]{n}\to1 proof to show 2n+3nn→3\sqrt[n]{2^n+3^n}\to3. (Factor out the dominant 3n3^n, then squeeze.)

Teach it — the surest test of mastery is explaining it

  1. A friend has only ever heard “a limit is what a sequence gets closer and closer to.” In three or four sentences written to them, explain why that phrasing is unsafe, using a single concrete sequence that “gets closer” for a while but does not converge.
  2. Write a complete, textbook-quality worked solution to lim⁡n→∞3n2+2nn2−5\displaystyle\lim_{n\to\infty}\frac{3n^2+2n}{n^2-5}, narrating each step in words a classmate could follow without you in the room.
  3. Debug your classmate. They argue: “(1+1n)n→1\left(1+\tfrac1n\right)^n\to1, because 1+1n→11+\tfrac1n\to1 and 11 to any power is 11.” Pinpoint the exact flawed step and write the one-sentence correction. (Name the indeterminate form they walked into.)
  4. Design a problem. Invent a self-defined sequence an+1=c+ana_{n+1}=\sqrt{c+a_n} whose limit is exactly 55. State the cc you need, then write the four-step solution key (bounded, increasing, converges, solve for LL) as if for a teacher’s answer book.
  5. In two or three sentences, teach the difference between “bounded” and “convergent.” Give one sequence that is bounded but not convergent, and say in plain words why one property does not imply the other.

Looking ahead — Chapter 2 in your hands early

  1. In Chapter 2 the input switches from a whole number nn to a real number xx sliding toward a point. Predict the value of lim⁡x→2x2−4x−2\displaystyle\lim_{x\to2}\frac{x^2-4}{x-2} by canceling the common factor, exactly as you’d simplify a sequence ratio. (Next chapter makes this rigorous — and explains why the cancel is legal.)
  2. You met ee as lim⁡(1+1n)n\lim\left(1+\tfrac1n\right)^n. Locate it from the other direction: with a calculator, compute 2h−1h\tfrac{2^h-1}{h} and 3h−1h\tfrac{3^h-1}{h} for small hh (say h=0.001h=0.001). One is below 11, one above. Between which two whole-number bases must the “slope =1=1” base — which Chapter 5 will reveal is ee — lie?
  3. Your recursive sequences found their limit by demanding L=(the rule applied to L)L=(\text{the rule applied to }L) — a fixed point. In Chapter 8 you’ll iterate x←x−f(x)f′(x)x \leftarrow x-\tfrac{f(x)}{f'(x)} to hunt a root. Predict: if that iteration settles to a value LL, what equation must LL satisfy?

Questions to talk about

No answer key. These are for arguing over, with someone who read the same chapter.

  1. 1, 1/2, 1/3, 1/4 … creeps toward zero without ever arriving. Why is 'never arrives' not the same as 'does not get there'? Try saying precisely what the list is doing.
  2. The chapter says the part that separates people who know about limits from people who can use them is knowing what to do when the obvious move fails. Where else is that the real dividing line?
  3. A recursive sequence finds its limit by demanding L equals the rule applied to L — a fixed point. What is appealing about defining an answer as the thing that does not change when you act on it?

All 12 chapters of Calculus I · Foundations · All courses