What is a Braid Group?

Daniel Glasscock, June 2012

These notes complement a talk given for the What is ... ? seminar at the Ohio State University.

Intro

The topic of braid groups fits nicely into this seminar. On the one hand, braids lend themselves immedi-
ately to nice and interesting pictures about which we can ask (and sometimes answer without too much
difficulty) interesting questions. On the other hand, braids connect to some deep and technical math;
indeed, just defining the geometric braid groups rigorously requires a good deal of topology.

I hope to convey in this talk a feeling of how braid groups work and why they are important. It is not
my intention to give lots of rigorous definitions and proofs, but instead to draw lots of pictures, raise
some interesting questions, and give some references in case you want to learn more.

Braids

A braid∗ on n strings is an object consisting of 2n points (n above and n below) and n strings such that

i. the beginning/ending points of the strings are (all of) the upper/lower points,
ii. the strings do not intersect,
iii. no string intersects any horizontal line more than once.

The following are braids on 3 strings:

We think of braids as lying in 3 dimensions; condition iii. is then that no string in the projection of the
braid onto the page (as we have drawn them) intersects any horizontal line more than once.

Two braids on the same number of strings are equivalent (≡) if the strings of one can be continuously
deformed – in the space strictly between the upper and lower points and without crossing – into the
strings of the other.

Denote by Bn be the set of all braids (up to equivalence) on n strings.

*These are actually braid diagrams. See the definitions at the end of this note for the textbook definition
of a geometric braid.
Braid groups

Two braids in Bn can be “added” (⊕) to yield a new braid by joining the bottom points of the first
braid to the top points of the second.

It is now a little exercise to check that (Bn , ⊕) forms a group. You may want to pause to check the
existence of inverses.

Let’s consider the first couple braid groups:

B1 This group consists of all braids on 1 string. It is the trivial group.

B2 The elements in this group are twists of two strings. By giving a twist in one direction the value
+1 and a twist in the other direction the value −1, it follows that B2 is isomorphic to the group
of integers under addition.

B3 This group is infinite and non-abelian (check that adding the two braids above in the opposite
order yields a different braid). By keeping one string fixed, we see that B3 contains several copies
of B2 .

To get a better idea about B3 , we might ask:

1. How does B3 relate to some groups we already know?

2. Are there braids with finite order? (Is B3 torsion free?)
3. Are there braids that commute with all other braids? (Does B3 have trivial center?)

4. Does B3 arise or act naturally? (What are the representations of B3 ?)

5. Is B3 even linear (a subgroup of GLm (C) for some m)?
6. What about homology/cohomology of B3 ?

Despite the seemingly innocent nature of these questions, most of them are hard to answer, especially
in the general case of Bn . The reader is encouraged to spend a few minutes pondering questions 2 and
3 especially. What does your intuition tell you?

1. To each braid in B3 we may associate a permutation of {1, 2, 3} by labeling the upper and lower
points with 1, 2, 3 and following the strings. For example, the very first braid drawn in these notes
corresponds to the permutation 1 7→ 3, 2 7→ 1, and 3 7→ 2. It is in this way that the braid groups
generalize the symmetric groups!

Let ϕ : B3 −→ S3 send a braid to its associated permutation. It is an exercise to check that ϕ is a

surjective group (anti-)homomorphism. Thus S3 ∼

= B3 ker ϕ. We recognize ker ϕ as the subgroup of
braids corresponding to the trivial permutation; these are called the pure braids.

2. Your gut may have told you that there are no braids which may be composed with themselves finitely
many times to yield the trivial braid. That intuition is correct, but proving such a result is not so
straightforward.

Just as S3 is generated by the transpositions (12), (23), we see that B3 is generated by a crossing of

strings 1 and 2 and a crossing of strings 2 and 3. More specifically, if σ1 is and σ2 is , then
B3 = hσ1 , σ2 i.

To ask what relations there are between σ1 and σ2 is to ask about the kernel of the homomorphism
ψ : F2 −→ B3 where F2 = ha, bi is the free group on two generators and ψ(a) = σ1 , ψ(b) = σ2 .

Emil Artin proved in 1925 that σ1 σ2 σ1 = σ2 σ1 σ2 (shown above) is the fundamental relationship
between σ1 and σ2 . In other words, B3 ∼
= ha, b | aba = babi.

One can deduce from the presentation ha, b | aba(bab)−1 i that B3 has no non-trivial elements of finite
order, answering question 2. (See propositions 5.17 and 5.18 in Combinatorial Group Theory by R.C.
Lyndon and P.E. Schupp.)

The general presentation of Bn is only slightly more complicated. If σi represents the (left over right)
crossing of strings i and i+1, then σ1 , σ2 , . . . , σn−1 generate Bn . Notice that σ1 and σ3 commute; indeed,
they involve a disjoint set of strings. One may guess, then, that Bn ∼ = ha1 , a2 , . . . , an−1 | ai ai+1 ai =
ai+1 ai ai+1 ∀ 1 ≤ i ≤ n − 2, aj ak = ak aj ∀ |j − k| ≥ 2i.

3. Are there any braids which commute with all other braids? The braid (σ1 σ2 )3 , drawn below, is one
such braid. It can be thought of as a full twist of the strings. The reader is strongly encouraged to give
a simple (non-algebraic) reason why this braid commutes with all others.

We may prove that this braid generates the center by utilizing a surprising connection between B3 and
P SL2 (Z). We will use two facts without proof: P SL2 (Z) ∼
= hv, w | v 2 , w3 i, and the center of P SL2 (Z)
is trivial.

The map ρ : B3 −→ P SL2 (Z) induced by ρ(σ1 ) = w−1 v and ρ(σ2 ) = v −1 w2 is a well defined homomor-
phism (check that σ1 σ2 σ1 (σ2 σ1 σ2 )−1 maps to theidentity). It is an exercise to see that ker ρ = h(σ1 σ2 )3 i.
Since Z(P SL2 (Z)) is trivial and P SL2 (Z) ∼ = B3 ker ρ, Z(B3 ) ⊆ ker ρ. Since h(σ1 σ2 )3 i ⊆ Z(B3 ), we see
3
that Z(B3 ) = h(σ1 σ2 ) i.

In general, Z(Bn ) = h(σ1 σ2 · · · σn−1 )n i is infinite cycle, generated by the full twist.
 Definitions

Group A group is a non-empty set G with a binary operation ⊕ : G × G −→ G satisfying the following
three properties:
i. (associativity) for all g, h, f ∈ G, (g ⊕ h) ⊕ f = g ⊕ (h ⊕ f ),
ii. (identity) there exists e ∈ G such that for all g ∈ G, g ⊕ e = e ⊕ g = g,
iii. (inverses) for all g ∈ G, there exists h ∈ G such that g ⊕ h = h ⊕ g = e.

Order An element g ∈ G has finite order if there exists n ∈ {1, 2, . . .} such that g ⊕ g ⊕ · · · ⊕ g = e.
| {z }
n

Torsion A group is torsion free if the only element with finite order is the identity.
Center The center of a group G, denoted Z(G), is the set of all elements c ∈ G such that for all g ∈ G,
c ⊕ g = g ⊕ c.
Braid Let I = [0, 1]. A geometric braid on n strings is a set B ⊆ R2 × I formed by n disjoint sets,
each of which is homeomorphic to I, such that the projection R2 × I −→ I maps each set home-
omorphically onto I and B ∩ (R2 × {0}) = {(1, 0, 0), (2, 0, 0), . . . , (n, 0, 0)}, B ∩ (R2 × {1}) =
{(1, 0, 1), (2, 0, 1), . . . , (n, 0, 1)}.
Bn The geometric braid group on n strings is the group of braids on n strings under composition.
Sn The symmetric group on n letters is the group of permutations of n distinct letters under compo-
sition. This may be realized concretely as the group of permutations of {1, 2, . . . , n}.
GLn (C) The general linear group of degree n over C is the group of all invertible n × n complex matrices
under multiplication.
P SL2 (Z) The special linear group of degree 2 over Z, denoted SL2 (Z), is the group of all 2 × 2 integer
matrices with determinant 1 under multiplication. The projective special linear group of degree 2
over Z is the factor group SL2 (Z) {±I} where I is the 2 × 2 identity matrix.

References

The father of the modern braid group(s) is Emil Artin. See any of the references below for references
to his original works. Consult [4–6] as general references. Braids surface in knot theory [2], physics [1],
and cryptography [3].

[1] John Baez. Braids, May 1992. http://math.ucr.edu/home/baez/braids.html.

[2] J. S. Birman and T. E. Brendle. Braids: A Survey. ArXiv Mathematics e-prints, September 2004.
[3] David Garber. Braid group cryptography. In Braids, volume 19 of Lect. Notes Ser. Inst. Math. Sci.
Natl. Univ. Singap., pages 329–403. World Sci. Publ., Hackensack, NJ, 2010.
[4] Juan González-Meneses. Basic results on braid groups. Ann. Math. Blaise Pascal, 18(1):15–59, 2011.

[5] Christian Kassel and Vladimir Turaev. Braid groups, volume 247 of Graduate Texts in Mathematics.
Springer, New York, 2008.
[6] Dale Rolfsen. Tutorial on the braid groups. In Braids, volume 19 of Lect. Notes Ser. Inst. Math. Sci.
Natl. Univ. Singap., pages 1–30. World Sci. Publ., Hackensack, NJ, 2010.