Topology Proof Synthesis

From Individual Theorems to a Proof Strategy

Earlier in the course, compactness and connectedness supplied different kinds of information. Compactness can ensure that a nonempty closed subset of a compact set has a minimum, while connectedness rules out gaps in a set or in the image of a continuous function. Continuity links the domain to its image and makes level sets closed. A topology proof often succeeds by assigning each hypothesis its own job rather than trying to use all the hypotheses in the same way.

This tutorial develops two conclusions by putting those roles together. The first is a first-crossing principle: if a continuous function starts below a level and ends above it, then there is a first point where it reaches that level, and the function stays below the level before that point. The second is a uniform sign margin: a continuous function on a nonempty compact connected set that never vanishes has one sign everywhere and stays a positive distance from zero.

A First-Crossing Principle

The Intermediate Value Theorem guarantees that an intermediate level is reached. That alone does not identify where the first crossing occurs. To do so, collect all crossing points into a level set. Continuity makes this set closed, and compactness then ensures that it has a least point.

Proof. Since \(f(a)<c<f(b)\), the Intermediate Value Theorem gives some \(x\in[a,b]\) such that \(f(x)=c\). Thus \(L\ne\varnothing\). The singleton \(\{c\}\) is closed in \(\mathbb{R}\), and continuity implies that \(L=f^{-1}(\{c\})\) is closed relative to \([a,b]\). The interval \([a,b]\) is compact, so \(L\), being a closed subset of a compact set, is compact. The Extreme Value Theorem applied to the identity function on \(L\) shows that \(L\) has a minimum \(r\). By the definition of \(L\), \(f(r)=c\).

We now show that \(f(x)<c\) whenever \(a\leq x<r\). Fix such an \(x\). If \(f(x)=c\), then \(x\in L\), contradicting the minimality of \(r\), since \(x<r\). If \(f(x)>c\), then \(f(a)<c<f(x)\), so the Intermediate Value Theorem applied on \([a,x]\) gives some \(z\in[a,x]\) with \(f(z)=c\). Since \(z\leq x<r\), this again contradicts the minimality of \(r\). The only remaining possibility is \(f(x)<c\). This proves the claim. \(\square\)

The proof uses a sequence of distinct tools. The Intermediate Value Theorem creates at least one crossing. Continuity makes the complete set of crossings closed. Compactness gives that set a least point. Finally, the Intermediate Value Theorem rules out the possibility that the function had risen above the level earlier. The first crossing is not chosen by guesswork; it is obtained as the minimum of a well-structured set.

When There Is More Than One Crossing

The word “first” matters: the theorem does not claim that the level is reached only once. A continuous function can cross a level repeatedly and still have a well-defined first crossing. The compactness argument selects the earliest point in the whole level set, not a unique point in that set.

A Uniform Margin from Zero

A second proof-synthesis result turns a qualitative statement into a quantitative one. The Nonvanishing Continuous Function Has Constant Sign Theorem says that a continuous function on a connected set that never equals zero has one sign throughout the set. On a compact set, the Positive Continuous Functions on Compact Sets Have a Positive Minimum Theorem can strengthen this: the absolute value cannot merely be positive point by point; it has a positive lower bound that works at every point.

Proof. By the Nonvanishing Continuous Function Has Constant Sign Theorem, either \(f(x)>0\) for every \(x\in K\), or \(f(x)<0\) for every \(x\in K\). Define \(g:K\to\mathbb{R}\) by \(g(x)=|f(x)|\). The function \(g\) is continuous because \(f\) and the absolute-value function are continuous. The hypothesis \(f(x)\ne0\) gives \(g(x)>0\) for every \(x\in K\). Since \(K\) is nonempty and compact, the Positive Continuous Functions on Compact Sets Have a Positive Minimum Theorem gives a number \(\delta>0\) such that \(g(x)\geq\delta\) for all \(x\in K\). If \(f\) is positive everywhere, then \(|f(x)|=f(x)\), so \(f(x)\geq\delta\). If \(f\) is negative everywhere, then \(|f(x)|=-f(x)\), so \(-f(x)\geq\delta\), or \(f(x)\leq-\delta\). These are the two asserted alternatives. \(\square\)

The roles of the hypotheses are visible in the proof. Connectedness gives a single sign; compactness supplies one positive lower bound for the magnitude. A pointwise claim \(f(x)\ne0\) does not by itself provide such a uniform bound on a noncompact domain. For instance, \(f(x)=x\) is positive on \((0,1)\), but no \(\delta>0\) satisfies \(x\geq\delta\) for every \(x\in(0,1)\).

A Reusable Synthesis Method

These arguments illustrate a practical way to organize proofs that involve topology. First identify the object whose existence or extremal position is needed. Then make a set of candidates, prove it is nonempty, and establish the closure and compactness needed to extract an extremum. Use connectedness when the conclusion concerns gaps or a change of sign. Continuity is the link that makes preimages of closed sets closed and supports intermediate values.

A common mistake is to invoke compactness when the argument actually needs connectedness, or to invoke connectedness when an extremum is required. A compact set can have gaps, so compactness alone cannot force a function to take every intermediate value. A connected set need not be compact, so connectedness alone cannot guarantee that a positive function stays a fixed distance from zero. Naming the precise job of each hypothesis both clarifies the proof and helps reveal when a claim is too strong.

There is also an important boundary issue in first-crossing arguments. The assumption \(f(a)<c<f(b)\) places the endpoint values strictly on opposite sides of the level, ensuring a crossing by the Intermediate Value Theorem. If an endpoint already has value \(c\), that endpoint may itself be the first crossing; if the endpoint values do not lie on opposite sides, a crossing need not exist. In each situation, the level set must be examined under the actual hypotheses rather than inferred from the picture.