Volume 27 · Part Three · Chapter 10 of 23
Knots, Links, and the Braid Group
The mathematics the figure has been quietly borrowing.
The braid group
Take n strands running between two horizontal bars. A braid is a way of getting from the top to the bottom in which strands may cross but never turn back. Two braids compose by stacking, every braid has an inverse (undo the crossings in reverse), and Artin's presentation gives generators σ_i — strand i crossing over strand i+1 — with two relations: distant crossings commute, and the three-strand relation σ_i σ_{i+1} σ_i = σ_{i+1} σ_i σ_{i+1}.
That is a group, not a picture of one. The braid group on three or more strands is infinite and non-abelian: order of operations matters, which is the mathematical content of the sentence 'the history of the crossing is part of the state'.
Invariants
Closing a braid gives a knot or link, and the central problem is telling two apart when no deformation relates them. Invariants do this: quantities unchanged by smooth deformation. The Jones polynomial is the famous one. Its existence is what makes topology usable as physics rather than as description — an invariant is a number a measurement can, in principle, return.
Two honesties belong here. First, invariants are one-way tools: agreeing invariants do not prove two links are the same. Second, the informal use of 'topological' to mean 'robust' is not this. Robustness under disorder in a physical system needs its own argument, given in Chapter 11.
Hardness, and why it matters here
Braid groups are computationally awkward in an instructive way. The word problem — deciding whether two braid words describe the same braid — is solvable, and efficiently so; but the conjugacy problem is harder, and deciding whether two closed braids give the same link is harder still. Computing the Jones polynomial exactly is #P-hard, while approximating it is a natural problem for a quantum computer, which is the formal link between Chapter 10 and Chapter 12: the same structure that makes the invariant expensive to compute classically is the structure a topological quantum computer would exploit.
This matters to the volume because it marks a boundary the prose must respect. A structure can be exactly defined and still not be effectively decidable at the scale a physical claim needs. When a passage says a system's behaviour is determined by its topology, the honest follow-up question is whether anyone can compute the relevant invariant for a system of that size. Often no one can, and the sentence is then a statement about the existence of an explanation rather than the possession of one.
Equations borrowed
- Artin's presentation of the braid group B_n and its relations
- Markov's theorem relating braid closures to links
- The Jones polynomial as a link invariant
Validity band
Exact mathematics. Physical relevance requires a separate argument in each case that the physical configuration space really is the one the group describes.
Falsifier
None internally — these are theorems. The falsifiable claims are the physical ones that cite them.
Where this chapter is weakest
The chapter risks lending mathematical authority to physical passages that have not earned it. Its own guard is the sentence that must not be cut: a theorem about braids constrains a physical system only when someone has shown the system's configuration space is the braid group's.
The volume-wide audit of these weak points is collected in Where This Volume Is Weak.