The Central Idea
A mathematical knot is a closed loop in three-dimensional space.
Topologically, we model a knot as an embedding
\[ K:S^1\hookrightarrow\mathbb{R}^3 \]or, equivalently for many purposes, into the three-sphere \(S^3\).
The loop has no loose ends.
The central question is:
Can one embedded loop be continuously deformed into another without cutting the loop or allowing it to pass through itself?
Why Ordinary Rope Knots Are Not Quite the Same
A physical rope usually has two ends.
Those ends can often be pulled through a tangle until the rope becomes straight.
Mathematical knot theory removes this escape route by joining the ends together.
Once the loop is closed, some tangles can no longer be undone without cutting or passing strands through one another.
The Unknot
The simplest knot is the unknot.
It is any embedded loop that can be continuously deformed into an ordinary round circle.
Thus an apparently complicated loop may still be an unknot if all of its apparent tangles can be removed by allowed deformations.
Knot Equivalence
Two knots are considered equivalent when one can be continuously moved through three-dimensional space until it becomes the other without:
- cutting the loop;
- gluing new pieces together;
- allowing one strand to pass through another.
The formal notion underlying this equivalence is ambient isotopy.
Ambient Isotopy
An ambient isotopy continuously deforms the entire surrounding space while carrying one embedded knot to another.
Informally, imagine that space itself is made of flexible material.
The knot may stretch, bend, rotate, and slide as the surrounding space deforms, but strands cannot pass through one another.
Why Ambient Isotopy Matters
Simply moving the points of a loop continuously is not enough if the motion allows self-intersections.
Knot equivalence must preserve the fact that the loop remains embedded throughout the deformation.
Ambient isotopy captures that requirement.
Knot Diagrams
Knots live in three dimensions, but they are often studied through two-dimensional projections called knot diagrams.
At each crossing, the diagram records which strand passes over and which passes under.
Without this over-under information, a planar drawing would not contain enough information to reconstruct the knot type.
Crossings Are Features of a Diagram
A crossing in a knot diagram is usually not a self-intersection of the knot in three-dimensional space.
One strand lies above the other.
The apparent intersection occurs only in the two-dimensional projection.
Different Diagrams Can Represent the Same Knot
A single knot can have many different diagrams.
A diagram may acquire extra twists or crossings even though the underlying knot type remains unchanged.
Therefore the number of crossings in one particular drawing is not by itself a topological invariant.
The Reidemeister Moves
Three local diagram transformations, called the Reidemeister moves, generate the allowed changes between diagrams of equivalent knots.
They are commonly labeled:
- Reidemeister I;
- Reidemeister II;
- Reidemeister III.
Reidemeister Move I
Reidemeister I adds or removes a small twist containing one crossing.
The move changes the appearance of the diagram but not the knot type.
It shows immediately that raw crossing count is not invariant under all diagram moves.
Reidemeister Move II
Reidemeister II adds or removes two crossings formed by two nearby strands.
One strand can be pulled across the other in a way that creates or cancels the pair without changing the knot.
Reidemeister Move III
Reidemeister III slides one strand past a crossing of two other strands.
The total number of crossings remains the same, but their local arrangement changes.
Reidemeister's Theorem
Two knot diagrams represent equivalent knots if and only if one can be transformed into the other by a finite sequence of Reidemeister moves together with planar deformation.
This theorem converts a three-dimensional isotopy question into a two-dimensional diagrammatic one.
The Trefoil Knot
The simplest nontrivial knot is the trefoil.
Its minimal diagram has
\[ \boxed{3} \]crossings.
Unlike the unknot, the trefoil cannot be reduced to a simple circle by Reidemeister moves.
Crossing Number
The crossing number of a knot is the smallest number of crossings appearing in any diagram of that knot.
It is written
\[ c(K). \]For the unknot,
\[ c(K)=0. \]For the trefoil,
\[ \boxed{c(K)=3}. \]Why Minimal Crossing Number Is an Invariant
Individual diagrams may have many unnecessary crossings.
But the smallest achievable crossing number over all diagrams depends only on the knot type.
Therefore crossing number is a knot invariant.
The Figure-Eight Knot
Another fundamental example is the figure-eight knot.
Its minimal crossing number is
\[ \boxed{4}. \]It is not equivalent to the trefoil because their minimal crossing numbers differ.
Knot Invariants
A knot invariant is a quantity or algebraic object that has the same value for equivalent knots.
Examples include:
- crossing number;
- unknotting number;
- knot group;
- Alexander polynomial;
- Jones polynomial;
- knot genus.
Different invariants capture different aspects of knot structure.
Invariants as Obstructions
If two knots have different values of a valid knot invariant, then they cannot be equivalent.
\[ \boxed{ I(K_1)\ne I(K_2) \Longrightarrow K_1\not\sim K_2 }. \]Matching invariant values, however, may not be enough to prove equivalence.
Knot Polynomials
Several important invariants assign a polynomial to a knot.
If equivalent knots always receive the same polynomial, then different polynomials prove that the knots are distinct.
Polynomial invariants are one of the major bridges between topology and algebra.
The Alexander Polynomial
The Alexander polynomial is one of the classical knot invariants.
It associates to a knot a Laurent polynomial, defined up to multiplication by certain units.
The details belong to algebraic topology, but the conceptual role is simple:
convert geometric knot structure into algebra that can be compared.
The Jones Polynomial
The Jones polynomial is another influential knot invariant.
It can distinguish many knots that simpler invariants cannot.
Its discovery created important connections among knot theory, statistical mechanics, quantum topology, and representation theory.
Knot Genus
Knot theory also uses the word genus, but here it refers to a different construction from the genus of a closed surface.
A knot bounds an orientable surface in three-dimensional space called a Seifert surface.
The genus of the knot is the smallest genus among all such spanning surfaces.
Seifert Surfaces
A Seifert surface for a knot is an orientable surface whose boundary is exactly the knot.
Symbolically, if \(S\) is such a surface and \(K\) is the knot, then
\[ \boxed{\partial S=K}. \]The unknot bounds a disk, so its knot genus is \(0\).
The Trefoil Has Knot Genus One
The trefoil does not bound an embedded disk.
It does bound a once-punctured torus.
Therefore its knot genus is
\[ \boxed{1}. \]Unknotting Number
The unknotting number of a knot is the smallest number of crossing changes needed to turn the knot into the unknot.
It is often written
\[ u(K). \]For the unknot,
\[ u(K)=0. \]For the trefoil,
\[ \boxed{u(K)=1}. \]Crossing Changes Are Not Reidemeister Moves
A crossing change switches which strand passes over and which passes under.
This can change the knot type.
Reidemeister moves do not change knot type.
Therefore crossing changes and Reidemeister moves play fundamentally different roles.
Orientation of a Knot
A knot can be given an orientation by choosing a direction around the loop.
This direction helps define signed crossings and several algebraic invariants.
Reversing the orientation does not necessarily change the underlying unoriented knot type.
Crossing Signs
In an oriented knot or link diagram, crossings can be assigned signs \(+1\) or \(-1\) according to a standard orientation convention.
Signed crossings are used in quantities such as writhe and linking number.
Writhe
The writhe of an oriented knot diagram is the sum of its crossing signs:
\[ w(D)=\sum_i \operatorname{sign}(c_i). \]Writhe depends on the diagram and is not itself a knot invariant under all Reidemeister moves.
In particular, Reidemeister I changes writhe.
What Is a Link?
A link is a collection of one or more disjoint embedded circles in three-dimensional space.
Formally, an \(n\)-component link can be modeled as an embedding of
\[ S^1\sqcup S^1\sqcup\cdots\sqcup S^1 \]with \(n\) copies of \(S^1\).
A knot is therefore a one-component link.
Components of a Link
Each closed loop in a link is called a component.
A two-component link has two disjoint circles.
A three-component link has three.
The number of components is preserved under link equivalence.
The Unlink
An unlink is a link whose components can be separated into disjoint unknotted circles.
The two-component unlink consists of two ordinary circles that are not linked together.
The Hopf Link
The simplest nontrivial two-component link is the Hopf link.
Its two components are linked once.
Neither component is individually knotted, but together they form a nontrivial link.
A Link Can Be Nontrivial Even If Every Component Is an Unknot
The Hopf link shows that knotting and linking are distinct ideas.
Each component can be an unknot while the collection as a whole cannot be separated.
Therefore component-by-component analysis is not sufficient to classify a link.
Linking Number
For an oriented two-component link, the linking number measures algebraically how the components wind around one another.
It is denoted
\[ \operatorname{lk}(L_1,L_2). \]One diagrammatic formula is
\[ \boxed{ \operatorname{lk}(L_1,L_2) = \frac12 \sum \operatorname{sign}(c_i) } \]where the sum runs over crossings involving one strand from each component.
Linking Number of the Hopf Link
With compatible orientations, the standard Hopf link has linking number
\[ \boxed{\operatorname{lk}=\pm1}. \]The sign depends on orientation choices.
Its absolute value is
\[ 1. \]Linking Number of the Two-Component Unlink
The two-component unlink has
\[ \boxed{\operatorname{lk}=0}. \]Therefore linking number distinguishes the Hopf link from the ordinary two-component unlink.
Linking Number Is Not a Complete Invariant
Two links can have the same linking number and still be topologically different.
In particular, linking number \(0\) does not necessarily mean that the components are completely unlinked.
More refined invariants may be required.
The Borromean Rings
The Borromean rings are a famous three-component link.
Any two components considered alone are unlinked.
Yet all three together form a nontrivial link.
This demonstrates that global linking behavior can be more subtle than pairwise linking.
Pairwise Information Can Miss Global Structure
In the Borromean rings, every pair has linking number
\[ 0. \]Nevertheless, the three-component link cannot be separated without cutting.
Thus pairwise invariants may fail to capture higher-order linking.
Knot Complements
Another way to study a knot is to remove it from the surrounding three-dimensional space.
The resulting space
\[ S^3\setminus K \]is called the knot complement.
Much of knot theory can be translated into the topology of these complements.
The Knot Group
The fundamental group of the knot complement,
\[ \pi_1(S^3\setminus K), \]is called the knot group.
It records information about loops moving around the knot without crossing it.
The knot group is a powerful algebraic invariant.
Why Complements Matter
The knot itself is only a circle topologically.
Every knot is homeomorphic to
\[ S^1 \]as an abstract space.
What distinguishes knots is how that circle is embedded in three-dimensional space.
The complement remembers that embedding.
Intrinsic Shape vs. Embedding
This distinction is essential.
As standalone spaces, the unknot and trefoil are both circles.
Therefore,
\[ K_{\text{unknot}}\cong K_{\text{trefoil}}\cong S^1 \]intrinsically.
They differ as embeddings in three-dimensional space.
Knot Theory Is a Study of Embeddings
Ordinary topology often asks whether two spaces are homeomorphic.
Knot theory asks a more refined question:
Are two embeddings of the same abstract circle equivalent inside the ambient space?
This makes knot theory a natural extension of the study of homeomorphism.
Prime Knots
A nontrivial knot is prime if it cannot be decomposed as a connected sum of two nontrivial knots.
The trefoil and figure-eight knot are prime.
Prime knots play a role analogous to prime numbers in arithmetic.
Connected Sum of Knots
Two knots \(K_1\) and \(K_2\) can be combined by cutting a small arc from each and joining the four resulting endpoints compatibly.
The result is the connected sum
\[ K_1\#K_2. \]The Unknot as an Identity
Under connected sum, the unknot behaves like an identity:
\[ \boxed{ K\#U\sim K } \]where \(U\) denotes the unknot.
Joining an unknot to a knot does not change its knot type.
Prime Decomposition
Every nontrivial knot can be decomposed into a connected sum of prime knots, and this decomposition is unique up to order and equivalence.
This is one of the major structural theorems of classical knot theory.
Mirror Images
Reflecting a knot produces its mirror image.
Some knots are equivalent to their mirror images.
Others are not.
This leads to the concept of chirality.
Chiral Knots
A knot is chiral if it is not equivalent to its mirror image.
The trefoil is chiral.
Therefore its left-handed and right-handed versions are distinct as oriented mirror types.
Amphichiral Knots
A knot is amphichiral if it is equivalent to its mirror image.
The figure-eight knot is a standard example.
Why Knot Tables Use Crossing Number
Knot tables are often organized by minimal crossing number.
The unknot has crossing number \(0\).
The trefoil is the unique prime knot with crossing number \(3\).
The figure-eight knot is the unique prime knot with crossing number \(4\).
Crossing Number Is Harder Than It Looks
A diagram with \(n\) crossings proves only that
\[ c(K)\le n. \]To prove
\[ c(K)=n, \]one must also show that no diagram with fewer crossings represents the same knot.
Minimality is the difficult part.
Knots and Topological Invariants
The general strategy of topology appears clearly in knot theory:
\[ \boxed{ \text{geometric object} \longrightarrow \text{topological invariant} \longrightarrow \text{classification evidence} }. \]Instead of trying every possible deformation, one searches for properties that all allowed deformations preserve.
Common Error: Thinking Every Crossing Is an Intersection
In a knot diagram, the crossing records one strand passing above another.
The strands are disjoint in three-dimensional space.
A genuine self-intersection would violate the embedding condition.
Common Error: Thinking Any Diagram with Crossings Is Knotted
The unknot can be drawn with many crossings.
Reidemeister moves may remove those crossings.
Therefore a complicated diagram does not automatically represent a nontrivial knot.
Common Error: Using Raw Crossing Count as an Invariant
Reidemeister I and II can change the number of crossings in a diagram.
Therefore the crossing count of one drawing is not invariant.
The minimal crossing number is the invariant.
Common Error: Allowing Strands to Pass Through One Another
If strands are allowed to pass through each other, every knot could be untangled.
Such a motion is not an allowed ambient isotopy.
Preserving embeddedness is essential.
Common Error: Confusing Knot Type with the Abstract Circle
Every knot is intrinsically homeomorphic to \(S^1\).
Knot type records how that circle sits inside three-dimensional space.
The embedding, not the abstract circle alone, contains the knotting information.
Common Error: Thinking Linking Number Completely Classifies Links
Linking number is useful but incomplete.
Different links can share the same linking number.
The Borromean rings show that even pairwise linking information can miss higher-order structure.
Common Error: Confusing Crossing Change with Reidemeister Move
Reidemeister moves preserve knot type.
A crossing change may alter knot type.
This distinction is central when studying unknotting number.
A Reliable Procedure for Comparing Knot Diagrams
- Verify that both drawings represent embedded closed loops.
- Look for obvious Reidemeister I, II, and III simplifications.
- Reduce unnecessary crossings where possible.
- Compare known invariants such as minimal crossing number or knot polynomials.
- If an invariant differs, conclude that the knots are not equivalent.
- If invariants agree, continue analysis; equality of one invariant may not prove equivalence.
A Reliable Procedure for Comparing Links
- Count the number of components.
- Determine whether individual components are knotted.
- Choose orientations when using signed invariants.
- Compute pairwise linking numbers when useful.
- Inspect for higher-order linking that pairwise data may miss.
- Use stronger invariants if simpler ones do not distinguish the links.
Example: Unknot vs. Trefoil
The unknot has minimal crossing number
\[ 0. \]The trefoil has minimal crossing number
\[ 3. \]Therefore,
\[ \boxed{\text{unknot}\not\sim\text{trefoil}}. \]Example: Trefoil vs. Figure-Eight Knot
The trefoil has
\[ c=3, \]while the figure-eight knot has
\[ c=4. \]Hence they are distinct knot types.
Example: Hopf Link vs. Two-Component Unlink
With suitable orientations, the Hopf link has
\[ |\operatorname{lk}|=1. \]The two-component unlink has
\[ \operatorname{lk}=0. \]Therefore they are not equivalent links.
Example: Why Component Count Matters
A trefoil is one component.
A Hopf link has two components.
Ambient isotopy cannot change the number of components.
Therefore a knot cannot be equivalent to a two-component link.
Practice 1: Knot Definition
What familiar topological space is embedded in three-dimensional space to model a knot?
Practice 2: Unknot
What is the unknot?
Practice 3: Reidemeister Moves
Name the three Reidemeister moves used to change knot diagrams without changing knot type.
Practice 4: Trefoil Crossing Number
What is the minimal crossing number of the trefoil?
Practice 5: Figure-Eight Crossing Number
What is the minimal crossing number of the figure-eight knot?
Practice 6: Link Definition
How does a link differ from a knot?
Practice 7: Linking Number
What is the absolute value of the linking number of a standard Hopf link?
Practice 8: Knot Invariant
If two knots have different values of a valid knot invariant, what can be concluded?
Practice 9: Knot Complement
What space is meant by the knot complement
\[ S^3\setminus K? \]Practice 10: Intrinsic vs. Embedded Structure
Why are the unknot and trefoil both homeomorphic to \(S^1\) as abstract spaces but still different knots?
Practice Answers
Practice 1
A knot is modeled by an embedded circle:
\[ \boxed{S^1}. \]Practice 2
The unknot is a knot equivalent by ambient isotopy to an ordinary round circle.
Practice 3
The three moves are:
- Reidemeister I;
- Reidemeister II;
- Reidemeister III.
Practice 4
\[ \boxed{3}. \]Practice 5
\[ \boxed{4}. \]Practice 6
A knot has one embedded circular component. A link may contain two or more disjoint embedded circular components.
Practice 7
\[ \boxed{1}. \]Practice 8
The knots cannot be equivalent:
\[ \boxed{I(K_1)\ne I(K_2)\Longrightarrow K_1\not\sim K_2}. \]Practice 9
It is the surrounding three-sphere with the knot removed.
Practice 10
Intrinsically, each knot is just a circle. The difference lies in how that circle is embedded in three-dimensional space. Knot theory studies the embedding, not merely the abstract topology of the loop.
What to Remember
- A knot is an embedded copy of \(S^1\) in three-dimensional space.
- Knot equivalence is modeled by ambient isotopy.
- Knot diagrams record over-under information at crossings.
- Reidemeister moves preserve knot type.
- The unknot is the trivial knot.
- The trefoil is the simplest nontrivial knot and has crossing number \(3\).
- The figure-eight knot has crossing number \(4\).
- A link is a collection of disjoint embedded circles.
- Linking number measures pairwise linking for oriented two-component links.
- Knot invariants convert topological structure into quantities or algebraic objects that can be compared.
- Every knot is intrinsically homeomorphic to \(S^1\); knot type records how that circle is embedded.
The core distinction is:
\[ \boxed{ \text{same abstract circle} \quad\ne\quad \text{same embedding in three-dimensional space} }. \]Knot theory studies the topology of that embedding.
Where We Go Next
Knots and links extend the study of homeomorphism into the topology of embeddings. The underlying loops may all be circles, yet the surrounding three-dimensional space records whether those circles are knotted or linked.
From here, topology can move toward fundamental groups, knot complements, homotopy, homology, surfaces, manifolds, and algebraic invariants. These tools make it possible to distinguish increasingly subtle forms of topological entanglement.