Some mathematical facts feel too obvious to need proving. If a continuous function starts out negative and ends up positive, surely its graph has to cross the x-axis somewhere in between — you'd have to lift your pen off the paper to avoid it. Michael Spivak devotes an entire chapter of his Calculus to proving exactly this fact, and calls it, without irony, "Three Hard Theorems."
Why would something so visually obvious require real work to prove? Because "obvious from a picture" and "true for every continuous function, including the ones too strange to draw" are very different standards — and calculus insists on the second one.
Theorem 1: The Zero Must Be Somewhere
The first of the three theorems says: if f is continuous on [a, b] and f(a) is negative while f(b) is positive, then f(x) = 0 for some x in [a, b]. This is usually called the Intermediate Value Theorem once it's generalized slightly (a continuous function that takes on two values takes on everything in between).
Here's the twist that shows why continuity on the whole interval matters, not just most of it. Take the function that equals −1 for x below √2 and 1 from √2 onward, on the interval [0, 2]. It's negative at 0 and positive at 2 — but it never once equals 0. The function is continuous at every single point of [0, 2] except one: √2. That single missing point is enough to destroy the conclusion entirely. The theorem isn't fussy for no reason; drop continuity at even one point and it can fail completely.
Theorem 2: A Continuous Function on a Closed Interval Can't Run Away to Infinity
The second theorem: if f is continuous on the closed interval [a, b], then f is bounded above there — its graph never shoots up without limit. Again, the word "closed" is load-bearing. The function f(x) = 1/x is perfectly continuous on the open interval (0, 1), but it is not bounded above: as x creeps toward 0, f(x) grows past any number you name. The theorem requires the closed interval [0, 1], where the function would have to be defined at 0 too — and it simply isn't, which is exactly why the theorem doesn't apply and boundedness can fail.
Theorem 3: The Function Actually Reaches Its Highest Point
The third theorem goes further than the second: a continuous function on [a, b] doesn't just stay bounded — it actually attains a maximum value at some specific point in the interval. That's a meaningfully stronger claim. A function can be bounded above without ever touching its own ceiling (imagine a curve that climbs toward a value of 1 forever without quite reaching it). Theorem 3 rules that out for continuous functions on closed intervals: somewhere in [a, b], the function hits its highest point exactly.
Why These Are "Hard"
Every theorem in the chapter before this one was about continuity at a single point — a local property, and comparatively easy to reason about. These three theorems describe behavior across an entire interval — a global property — and Spivak is upfront that this is a different order of difficulty. In fact, he shows that Theorems 1 through 3 can't be proved using only the basic algebraic properties of numbers covered earlier in the book. Try to construct the proof of Theorem 1 directly — hunting for the very first point where the function crosses zero — and the argument breaks down at a suspiciously specific spot: you can show such a point is approached, but not that it's actually reached by any number you've been handed.
The missing ingredient turns out to be the single most important property of the real numbers: the fact that any set of numbers with an upper bound has a least upper bound. That property — completeness — is what the next chapter of Calculus is devoted to, and it's what finally makes Theorems 1 through 3 provable. Rational numbers alone aren't enough; the rationals have gaps (there's no rational number equal to √2, for instance), and those gaps are exactly where these theorems would fail.
The Payoff: Why √2 Actually Exists
Once Theorem 1 is available, it earns its keep almost immediately. Consider the function f(x) = x². It's continuous, f(0) = 0, and f(b) is bigger than any number α you like once b is large enough. So f starts below α and ends above it — and by Theorem 1 (in its slightly generalized form), f must actually equal α somewhere in between. In other words: every positive number has a square root. Not "approximately," not "as a limit" — an actual number x with x² = α, guaranteed to exist by the same theorem that guarantees a continuous curve crosses zero.
The same idea, pushed a little further, proves that every polynomial equation of odd degree has a real root, and pins down exactly which even-degree polynomials do. Three theorems that look like they should be free turn out to unlock a genuinely large piece of algebra.
This is the pattern that runs through all of Spivak's Calculus: the facts that look too obvious to prove are usually the ones hiding the most interesting mathematics. Chapters 7 and 8 — Three Hard Theorems and Least Upper Bounds — are where that pattern shows up most vividly.
If you want to see the complete proofs (including the clever argument for exactly where completeness comes in), Spivak's Calculus, 4th Edition walks through all of it, from the fallacious "proof" of Theorem 1 to the correct one, unusually honestly for a textbook.
Which "obvious" fact took you the longest to accept actually needed a proof? Tell us in the comments.
