The Cut Point Trick: A Fast Way to Tell Spaces Apart

The Question

One of the most useful beginner tools in topology is the cut point trick.

The goal is simple:

How can we prove that two spaces are not homeomorphic without using heavy machinery?

A cut point often gives the answer immediately.

This post explains:

• what a cut point is,
• why homeomorphisms preserve cut points,
• how to use the trick in examples,
• where the trick stops,
• how to prove $\mathbb{R}^n \not\cong \mathbb{R}^2$ with stronger tools,
• and why having the same number of cut points does not automatically mean two spaces are homeomorphic.

---

What Is a Cut Point?

Let $X$ be a connected topological space and let $p \in X$.

We say that $p$ is a cut point if removing it disconnects the space:

$$X \setminus \{p\} \text{ is disconnected.}$$

In plain language:

A cut point is a point that breaks the space when you remove it.

---

First Picture

Example 1: The Real Line

<-----------o----------->
            p

If we remove $p$ from $\mathbb{R}$, we get

$$\mathbb{R} \setminus \{p\} = (-\infty, p) \cup (p, \infty).$$

These are two separated pieces, so $p$ is a cut point.

In fact, every point of $\mathbb{R}$ is a cut point.

Example 2: The Circle

      .-'''-.
    .'       '.
   /           \
   \           /
    '.       .'
      '-._.-'

If we remove one point from $S^1$, the remaining space is still one connected arc.

In fact,

$$S^1 \setminus \{p\} \cong (0,1).$$

So:

The circle has no cut points.

That is already enough to prove that $\mathbb{R}$ and $S^1$ are not homeomorphic.

---

The Trick in One Sentence

So if one space has cut points and the other does not, then they cannot be homeomorphic.

More generally, if the "cut point behavior" is different, then the spaces are not topologically the same.

---

Why Homeomorphisms Preserve Cut Points

Suppose

$$h : X \to Y$$

is a homeomorphism, and let $x \in X$.

Then the restricted map

$$h\!\restriction_{X \setminus \{x\}} : X \setminus \{x\} \to Y \setminus \{h(x)\}$$

is also a homeomorphism.

Why? Because:

• restricting a continuous map to a subspace stays continuous,
• the inverse is still continuous after restriction,
• and the restriction is still bijective.

Therefore:

$$X \setminus \{x\} \text{ is disconnected } \Longleftrightarrow Y \setminus \{h(x)\} \text{ is disconnected.}$$

So:

$x$ is a cut point in $X$ if and only if $h(x)$ is a cut point in $Y$.

This is exactly why cut points are a topological invariant.

---

How To Use the Trick

The method is usually:

1. Assume or test whether $X$ and $Y$ might be homeomorphic.
2. Remove a point from each space.
3. Compare what happens.
4. If the behavior is different, conclude $X \not\cong Y$.

Very often this is much faster than computing a fundamental group or homology.

---

Example 1: The Line vs the Circle

Compare $\mathbb{R}$ and $S^1$.

In $\mathbb{R}$

Every point is a cut point.

<------o------>    remove o    <---     --->

In $S^1$

No point is a cut point.

circle minus one point = one connected open arc

Since the cut point structure is different,

$$\mathbb{R} \not\cong S^1.$$

---

Example 2: An Open Interval vs the Circle

The same argument shows

$$(0,1) \not\cong S^1.$$

Every point of $(0,1)$ is a cut point, but no point of the circle is.

This is one of the standard first applications of the trick.

---

Example 3: The Line vs the Plane

Compare $\mathbb{R}$ and $\mathbb{R}^2$.

In $\mathbb{R}$, removing a point breaks the line into two pieces.

In $\mathbb{R}^2$, removing one point does not disconnect the plane. You can still move around the missing point.

Plane with one missing point:

  you can go around the hole

      path  )   o   (

So

$$\mathbb{R} \not\cong \mathbb{R}^2.$$

More generally, for $n \ge 2$, no point of $\mathbb{R}^n$ is a cut point.

---

Example 4: The Interval vs the Letter T

This example is especially helpful because both spaces have many cut points, but not the same type.

Interval

o-----------o

If you remove an interior point, the interval splits into two components.

Letter T

-----o-----
     |
     |

If you remove the center point, the space splits into three components.

So the interval and the letter T are not homeomorphic.

This shows that it is not only the number of cut points that matters. Sometimes we must also look at how many connected components appear after deletion.

---

Example 5: The Letters X and Y

X-shape:            Y-shape:

\   /               \   /
 \ /                 \ /
 / \                  |
/   \                 |

In the X shape, the center is a point whose removal creates four components.

In the Y shape, the branching point creates three components.

Therefore these spaces are not homeomorphic.

This is a nice visual example of how cut points detect branching.

---

Important Warning

This is one of the most important ideas to remember.

If $X \cong Y$, then they must have matching cut point behavior.

But the converse fails:

Two spaces can have the same number of cut points and still fail to be homeomorphic.

A Simple Counterexample

Take:

• an open interval $(0,1)$,
• and the letter T.

Both have infinitely many cut points.

But they are not homeomorphic, because:

• in $(0,1)$ every cut point splits the space into exactly two components,
• while in T the center point splits the space into three components.

So matching the raw count of cut points does not settle the problem.

---

Where the Cut Point Trick Stops

The cut point trick is powerful, but it has limits.

It immediately proves

$$\mathbb{R} \not\cong \mathbb{R}^2,$$

because $\mathbb{R}$ has cut points and $\mathbb{R}^2$ does not.

But it does not distinguish $\mathbb{R}^2$ from $\mathbb{R}^3$ or, more generally, $\mathbb{R}^2$ from $\mathbb{R}^n$ for $n \ge 3$, because all these spaces have the same cut-point behavior:

removing one point still leaves them connected.

So to separate $\mathbb{R}^2$ from higher-dimensional Euclidean spaces, we need a stronger invariant.

---

A Stronger Proof: Why $\mathbb{R}^n$ Is Not Homeomorphic to $\mathbb{R}^2$

For $n = 1$, the cut point argument already works.

Now assume $n \ge 3$ and suppose, for contradiction, that there is a homeomorphism

$$h : \mathbb{R}^n \to \mathbb{R}^2.$$

Remove one point. Then restriction gives a homeomorphism

$$h\!\restriction_{\mathbb{R}^n \setminus \{0\}} : \mathbb{R}^n \setminus \{0\} \to \mathbb{R}^2 \setminus \{h(0)\}.$$

Hence the punctured spaces must have isomorphic fundamental groups:

$$\pi_1(\mathbb{R}^n \setminus \{0\}) \cong \pi_1(\mathbb{R}^2 \setminus \{h(0)\}).$$

Step 1: The punctured plane

The space $\mathbb{R}^2 \setminus \{0\}$ deformation retracts onto the unit circle $S^1$ by pushing each point radially to the circle:

$$r(x) = \frac{x}{\|x\|}, \qquad H_t(x) = \left((1-t) + \frac{t}{\|x\|}\right)x.$$

for $x \in \mathbb{R}^2 \setminus \{0\}$.

Therefore

$$\mathbb{R}^2 \setminus \{0\} \simeq S^1, \qquad \pi_1(\mathbb{R}^2 \setminus \{0\}) \cong \pi_1(S^1) \cong \mathbb{Z}.$$

Intuitively, loops around the missing point can wind around it one time, two times, three times, and so on.

Step 2: The punctured higher-dimensional space

For $n \ge 3$, the space $\mathbb{R}^n \setminus \{0\}$ deformation retracts onto the sphere $S^{n-1}$ by the same radial formula.

There is also a more precise statement, and it matches the idea many students guess at first:

$$\mathbb{R}^n \setminus \{x\} \cong S^{n-1} \times (0,\infty) \cong S^{n-1} \times \mathbb{R}.$$

So the correct notation is

$$\mathbb{R}^n \setminus \{x\},$$

not a quotient such as $\mathbb{R}^n / \{x\}$.

One explicit homeomorphism is

$$\Phi : \mathbb{R}^n \setminus \{x\} \to S^{n-1} \times (0,\infty), \qquad \Phi(y) = \left( \frac{y-x}{\|y-x\|}, \|y-x\| \right).$$

Its inverse is

$$\Phi^{-1}(u,r) = x + ru.$$

Since $(0,\infty) \cong \mathbb{R}$ via $r \mapsto \log r$, this gives

$$\mathbb{R}^n \setminus \{x\} \cong S^{n-1} \times \mathbb{R}.$$

So

$$\mathbb{R}^n \setminus \{0\} \simeq S^{n-1}, \qquad \pi_1(\mathbb{R}^n \setminus \{0\}) \cong \pi_1(S^{n-1}).$$

Now use the standard product formula for fundamental groups:

$$\pi_1(X \times Y, (x_0,y_0)) \cong \pi_1(X,x_0) \times \pi_1(Y,y_0)$$

for path-connected spaces $X$ and $Y$.

Applying this to $S^{n-1} \times \mathbb{R}$ gives

$$\pi_1(\mathbb{R}^n \setminus \{x\}) \cong \pi_1(S^{n-1} \times \mathbb{R}) \cong \pi_1(S^{n-1}) \times \pi_1(\mathbb{R}) \cong \pi_1(S^{n-1}) \times \{e\} \cong \pi_1(S^{n-1}),$$

since $\mathbb{R}$ is contractible and therefore has trivial fundamental group.

But when $n \ge 3$, we have $n-1 \ge 2$, and spheres of dimension at least $2$ are simply connected. Hence

$$\pi_1(\mathbb{R}^n \setminus \{0\}) \cong \pi_1(S^{n-1}) \cong \{e\}.$$

Contradiction

So we would get

$$\{e\} \cong \pi_1(\mathbb{R}^n \setminus \{0\}) \cong \pi_1(\mathbb{R}^2 \setminus \{h(0)\}) \cong \mathbb{Z},$$

which is impossible.

Therefore:

$$\mathbb{R}^n \not\cong \mathbb{R}^2 \qquad \text{for every } n \ne 2.$$

The Main Idea

The cut point trick studies what happens after deleting one point and asking whether the space stays connected.

This stronger proof studies what happens after deleting one point and asking:

what kinds of loops survive around the missing point?

That is the first place where algebraic topology goes beyond cut points.

---

A Quick Comparison of the Punctured Spaces

This table explains in one line why $\mathbb{R}^2$ cannot be homeomorphic to $\mathbb{R}^n$ for $n \ge 3$.

---

A Clean Theorem Statement

Here is the formal result behind the trick.

This is the entire logical engine of the method.

---

Why This Result Is So Useful

The cut point trick gives a quick obstruction to homeomorphism:

• it is visual,
• it is easy to explain,
• and it works before you learn more advanced invariants.

It is especially effective for:

• intervals,
• circles,
• graphs,
• tree-like spaces,
• and simple subsets of $\mathbb{R}^n$.

In many first topology courses, cut points are one of the earliest examples of a topological invariant that students can actually use right away.

---

Famous Results Connected to This Idea

Cut points are only the beginning. Once you start studying complements of points, curves, and spheres, you quickly reach some of the most famous theorems in topology.

1. Brouwer's Invariance of Domain

This theorem says:

If $U \subseteq \mathbb{R}^n$ is open and $f : U \to \mathbb{R}^n$ is continuous and injective, then $f(U)$ is open.

One major corollary is that dimension is a topological invariant for Euclidean spaces:

$$\mathbb{R}^n \cong \mathbb{R}^m \Longrightarrow n = m.$$

So while the cut point trick proves some non-homeomorphism results quickly, invariance of domain is the famous general theorem that settles the whole Euclidean story at once.

2. The Jordan Curve Theorem

If $C \subset \mathbb{R}^2$ is a simple closed curve, then

$$\mathbb{R}^2 \setminus C$$

has exactly two connected components, an inside and an outside, and each has $C$ as its boundary.

This is much deeper than the cut point trick, but philosophically it is related:

• remove a simple object from a space,
• study the complement,
• and use that information to understand the topology.

In that sense, cut points are a very first example of a much larger theme in topology.

---

A Good Way To Think About It

A homeomorphism is like a perfect rubber-sheet deformation:

• you may stretch,
• bend,
• and twist,

but you may not tear the space or glue new parts together.

Because of that, a point that acts like a "bridge" in one space must still act like a bridge in the other.

That bridge-point is exactly a cut point.

---

Summary Table

This table is often enough to solve many beginner problems.

---

Final Takeaway

The cut point trick is one of the fastest ways to prove that two spaces are not homeomorphic.

The core principle is:

Homeomorphisms preserve what happens when you delete a point.

So if deleting a point breaks one space but not the other, the spaces cannot be homeomorphic.

And even if both have cut points, the finer deletion behavior can still distinguish them.

That is why cut points are such a powerful first invariant in topology.

---

Further Reading

• James Munkres, Topology
• Gamelin and Greene, Introduction to Topology
• Allen Hatcher, Algebraic Topology, especially the early chapters on homotopy and the fundamental group
• Christian Schnell, MAT 530 lecture notes, for the Jordan curve theorem and Brouwer's invariance of domain