Alphabeta Math
ExampleConstruction: AI-adaptedVerification: AI-adaptedPipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Fibonacci fusion data and chiral convention

Example

Take the skeletal complex-linear category whose simple labels are 1,τ and whose objects are finite direct sums of them, including the empty direct sum 0. For objects A=⨁ss⊕ms and B=⨁ss⊕ns, set Hom⁡(A,B)=⨁sMat⁡ns×ms(C); composition is block-matrix multiplication, dagger is conjugate transpose, and the norm is the maximum of the operator norms of its blocks. Thus direct sums are orthogonal. Define tensor product on simple labels by the fusion rules 1⊗a=a⊗1=a and τ⊗τ=1⊕τ, and extend it distributively to direct sums. For each allowed simple fusion channel c⊂a⊗b, let iabc:c→a⊗b be the standard unit-vector isometric inclusion. Since the multiplicity space is Vabc=Hom⁡(a⊗b,c), its orthonormal basis vector is the adjoint projection (iabc)†:a⊗b→c; retain iabc itself for the coevaluation construction below. On simple-label blocks, tensor a pair of morphism matrices by their Kronecker product on multiplicity coordinates and by the identity on each simple fusion channel; extend over direct sums. This defines a bifunctor compatible with dagger. Use the strict unit gauge: every F or R coordinate involving 1 is the identity. Put

ϕ=1+52,u=ϕ−1,v=ϕ−1/2.

The only nontrivial associator blocks are

F1τττ=[1],Fττττ=(uvv−u),

in the intermediate-channel order (1,τ). In the pullback R convention of R-symbol as a channel exchange map, the exchange blocks are

R1ττ=x=e−4πi/5,Rτττ=y=e3πi/5.

Because these are pullback symbols, define the object-level braiding cτ,τ on the 1 and τ summands by x−1 and y−1, respectively. Every braiding with a unit input is the strict identity. These fixed chiral data define the Fibonacci unitary braided fusion category in this gauge; the verification checks every nontrivial pentagon and hexagon sector.

Facts & Assumptions

Given: the simple labels, fusion rules, strict unit gauge, and F/R blocks in the Example.

[F1]

Unitary fusion category: the dagger is conjugate-linear, tensor-compatible, and the associator and unit constraints must be unitary.

[F2]

Fusion multiplicity space: the multiplicity at a vertex is Nabc=dim⁡Vabc.

[F3]

Fusion-tree basis and dimension recursion: fusion-space dimensions are the finite sums/products of the vertex multiplicities over admissible internal labels.

[F4]

F-symbol as a recoupling matrix: F is the matrix between left and right fusion-tree bases induced by the associator.

[F5]

R-symbol as a channel exchange map: R is the channel-exchange matrix induced by the braiding in the specified pullback convention.

[F6]

Monoidal category: associators satisfy the pentagon axiom and the unitors satisfy the triangle axiom.

[F7]

Braiding: a braiding is a natural isomorphism satisfying both hexagon identities.

Verification

Given: the fusion, F, and R data in the Example; prove that they define a unitary braided fusion category.

1.1F1F4F5givenalgebra

The positive real choices give u2+u=1, v2=u and u2+v2=1, so the real symmetric block F satisfies F†F=F2=I. All scalar unit-gauge blocks are unitary. The phases x,y have modulus one and x=y2 exactly. To fix the root and sign without a numerical cosine approximation, y5=−1 and y≠−1 give y4−y3+y2−y+1=0. Dividing by y2 shows that z=y+y−1 satisfies z2−z−1=0. Since z=2cos⁡(3π/5)<0, it is (1−5)/2=−u. Thus y2+uy+1=0, with this chosen root, not its conjugate. The object-level braiding has eigenvalues x−1,y−1, so inverse-precomposition gives exactly the stated pullback eigenvalues x,y.

1.2F2F3givenalgebra

Formal fusion-tree labels can be counted before assuming coherence: for a fixed binary tree, enumerate each vertex's allowed outgoing simple label by the displayed fusion rules, and keep its unique channel coordinate. The recursion is mn+1(1)=mn(τ) and mn+1(τ)=mn(1)+mn(τ), with m1(1)=0,m1(τ)=1. Hence m3(1)=1,m3(τ)=2 and m4(1)=2,m4(τ)=3. These are direct-sum coordinate counts; once the category axioms are checked, they are exactly the orthonormal tree counts of [F3]. Zero-dimensional blocks are the unique empty maps.

2.1F4F6step 1.1step 1.2algebra

Pentagon types and unit cases. For four ordered inputs a,b,c,d, name the five trees T0=(((ab)c)d), T1=((a(bc))d), T2=(a((bc)d)), T3=(a(b(cd))), T4=((ab)(cd)). Their two nonroot edge-label coordinates are respectively (ab,abc), (bc,abc), (bc,bcd), (cd,bcd) and (ab,cd), ordered lexicographically with 1<τ. Let A1:T0→T1, A2:T1→T2, A3:T2→T3, A4:T0→T4 and A5:T4→T3 denote the inverse-precomposition matrices of the corresponding associators. Thus the typed pentagon is A3A2A1=A5A4. If an input is the unit, deleting it canonically identifies the endpoint tree-coordinate spaces. Every local move with that unit is identity; each route then contains the same associator on the remaining three inputs, or only identities if fewer than three nonunit inputs remain. Thus every unit-input case holds, including empty blocks. For four τ inputs with total 1, the bases of T0,T1,T2,T3 are (1,τ),(τ,τ), and that of T4 is (1,1),(τ,τ). Fixing the unchanged subtree in each move gives A1=A3=F and A2=A4=A5=I: the latter blocks use either a unit input or the scalar F1τττ=1. Therefore the two routes are F2=I and I.

2.2F2F3F5F7step 1.1step 1.2algebra

Hexagon unit and total-unit cases. In either hexagon, an input unit gives the canonical unit identification for its exchanges and every associator involving it. Deleting that strand reduces the two routes to the same exchange of the other two inputs. For three τ inputs and total 1, the only intermediate charge in either tree is τ, by step 1.2, so the associator is [1]. Exchanging τ with the fused pair has coordinate x, whereas the two successive simple exchanges have product y2. Both typed root hexagons therefore say x=y2, proved in step 1.1.

2.3F1F2F4step 1.1algebra

The unit object is self-dual. The 1 summand of τ⊗τ gives an isometric inclusion i:1→τ⊗τ; set coev⁡=ϕ i and ev⁡=ϕ i†. In either zig-zag composite for τ, the only intermediate channel is 1, so its scalar is ϕϕ Fττττ[1,1]=ϕu=1. The opposite zig-zag uses F−1=F from step 1.1 and has the same coefficient. Hence τ is self-dual; duality extends to finite direct sums, so the category is rigid.

3.1F3F4F6step 1.1step 2.1algebra

All total-τ pentagon entries. In the tree order of step 2.1, each of the five bases is (1,τ),(τ,1),(τ,τ). The moves A1,A3 fix the total of their three-leaf subtree, and mix precisely coordinates one and three; their other coordinate is the scalar total-unit block. The moves A2,A4 fix the pair subtree and mix precisely coordinates two and three. For A5, the fixed cd label is τ in source coordinates one and three; these map to target coordinates two and three by the columns of F. Source coordinate two has cd=1 and maps to target coordinate one with coefficient one. Consequently A1=A3=(u0v010v0−u),A2=A4=(1000uv0v−u),A5=(010u0vv0−u). Exact multiplication using u2+u=1 and v2=u gives A3A2A1=(0uvuu−uvv−uvu2)=A5A4. This exhibits every one of the nine entries, with their source and target bases fixed. Together with step 2.1 it covers both possible totals for every simple-input pentagon.

3.2F1F4F5F7step 1.1step 1.2step 2.2algebra

Both total-τ hexagons, with their types. In the intermediate-pair order (1,τ) put D=diag⁡(x,y) and E=diag⁡(1,y). Here D is a simple pair exchange in the tree where that pair is fused first; E exchanges one τ with the composite pair, whose summands are 1,τ, in the respective source and target composite-channel bases. It is not a braiding on two simple objects. The functor Φ(h)(f)=f∘h−1 satisfies Φ(kh)=Φ(k)Φ(h), so it preserves the displayed order of the typed root composites. For the first root hexagon [F7], the route through associator, composite exchange, associator has matrix FEF; the route through simple exchange, associator, simple exchange has matrix DFD. For the second root hexagon the inverse associators also have matrix F−1=F, giving the same two matrices. Thus both root equations are FEF=DFD, equivalently B2D=EF and DB2=FE, where B2=FDF. Expanding B2 gives entries u2x+uy, uv(x−y), uv(x−y), ux+u2y. The four entries of B2D=EF require exactly x(u2x+uy)=u,ux(x−y)=y,uy(x−y)=1,y(ux+u2y)=−uy. The last entry contains no extra factor u. From step 1.1, y3=u(1−y) and y4=y+u. Therefore the four left sides are respectively u(uy4+y3)=u(u2+u)=u, u(y4−y3)=u(1+u)y=y, u(y3−y2)=u(1+u)=1, and uy(y2+uy)=−uy. This proves all four entries. Transposing gives DB2=FE, since D,B2,E,F are symmetric. Together with step 2.2, all simple-input hexagon sectors hold exactly.

4.1F1F2F4F6F7step 1.1step 2.1step 3.1step 2.2step 3.2step 2.3∎

The matrix Hom spaces and fusion rules give a finite semisimple complex category with simple unit, orthogonal direct sums and the stated C∗ norms. The distributive tensor action on multiplicity matrices is a dagger-compatible bifunctor. On simple inputs choose the associator whose induced inverse-precomposition matrix is the specified F; this is well defined because its object-level block is (F−1)T=F in the only nonscalar case. Choose the object-level braiding eigenvalues x−1,y−1 as specified. Extend both maps by direct sums and identities on copy coordinates. This makes them natural for arbitrary block matrices, not just for scalar maps of simple objects. A pentagon or hexagon on arbitrary direct sums decomposes into the finite simple-input blocks already proved in steps 2.1–3.2; zero summands give empty blocks. The triangle is the identity in the strict unit gauge. Step 1.1 proves unitarity and step 2.3 proves rigidity. Hence these data define the claimed unitary braided fusion category on every finite direct sum. The finite block exhaustion is the full coherence proof; no numerical approximation, unproved category realization or choice principle is used.

Depends on

Used by

Dependency tree · two levels

19 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources