Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-22
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.

Gleason Kahane Zelazko

Statement

Let A be a complex unital Banach algebra (Unital Banach algebra) and let φ:AC be a complex linear map with φ(1)=1 which is nonzero on every invertible element: φ(a)0 whenever aA is invertible. Then φ is continuous and multiplicative:

φ(a)aandφ(ab)=φ(a)φ(b)(a,bA).

No commutativity of A is assumed, and the argument is choice-free. The hypothesis is that φ is nonzero on invertible elements, not that it is nonzero on A; together with φ(1)=1 it is the exact hypothesis used.

Facts & Assumptions

Given: A complex unital Banach algebra A and a complex-linear φ:AC with φ(1)=1 that is nonzero at every invertible element.

[F1]

A is complete, 1=1, the multiplication is associative and bilinear with xyxy, and 10 (Unital Banach algebra).

[F2]

If y<1 then 1y is invertible with inverse n0yn; hence λ1a is invertible whenever λ>a (Neumann series).

[F3]

A normed space is a Banach space if and only if every absolutely convergent series in it converges (Series criterion for Banach spaces).

[F4]
[F6]

If f is entire and zero-free with constants M>0, C0 satisfying f(z)MeCz for all z, then f(z)=f(0)eaz with a=f(0)/f(0) (Zero free entire function of exponential type is an exponential).

[F7]

Every bounded entire function is constant (Liouville's theorem: every bounded entire function is constant).

[F8]

Proof

technique · direct
1.1

φ(a)a for every a: otherwise put λ:=φ(a), so λ>a and λ1a=λ(1a/λ) is invertible by [F2], while φ(λ1a)=λφ(a)=0 by linearity and φ(1)=1; this contradicts the hypothesis that φ vanishes at no invertible element.

F1F2algebra
1.2

Put Ea(z)=n0znan/n!. Comparison with (za)n/n! and completeness prove convergence and Ea(z)eza. For SN(z)=k=0Nzkak/k!, the terms of total degree nN in SN(z)SN(z) sum to anznk=0n(1)nk/(k!(nk)!), namely 1 for n=0 and 0 otherwise by the finite binomial formula. Writing t=za, the remaining terms have norm at most n=N+12Ntnk=max(0,nN)min(n,N)1/(k!(nk)!)n>N(2t)n/n!0. Multiplication is continuous by submultiplicativity. Passing to the limit, and then exchanging the two factors in the same computation, gives Ea(z)Ea(z)=Ea(z)Ea(z)=1. No commutativity beyond the powers of the single element a is used.

F1F3F4F5algebra
1.3

The scalar analytic estimate used below is direct: if dnDrn/n! with D,r0, then for every R0, dnRnDerR< by comparison. The absolute-value coefficient series therefore converges at every real argument, and its radius, hence the complex radius, is infinite. By [F5] its sum is entire and its derivative at zero is d1. This includes D=0 and r=0, using the constant-term convention 00=1.

F4F5algebra
2.1

φ(Ea(z))=n0znφ(an)/n! for every z, by continuity of φ from [step 1.1] applied to the partial sums of the absolutely convergent series of [step 1.2]; consequently fa(z):=φ(Ea(z)) is an entire function of z, with fa(z)Ea(z)eza, fa(0)=φ(1)=1, derivative fa(0)=φ(a) by termwise differentiation, and fa is zero-free because Ea(z) is invertible by [step 1.2] and φ vanishes at no invertible element.

step 1.1step 1.2step 1.3F5algebra
3.1

By [F6] applied to the zero-free entire fa of [step 2.1] with M=1 and C=a: fa(z)=ezφ(a), that is, φ(Ea(z))=ezφ(a) for all zC and all aA.

step 2.1F6
4.1

Fix a,bA and put F(z,w):=φ(Ea(z)Eb(w)). For fixed w, the function zF(z,w)=nznφ(anEb(w))/n! is entire with F(z,w)ezaewb, the bound coming from [step 1.2] and [step 1.1]; F(0,w)=φ(Eb(w))=ewφ(b) by [step 3.1]; F(,w) is zero-free because Ea(z)Eb(w) is a product of invertibles [step 1.2]; and zF(0,w)=φ(aEb(w)) by termwise differentiation of this scalar power series, whose coefficient bound is anEb(w)/n!.

step 1.1step 1.2step 1.3step 3.1F5algebra
5.1

Applying [F6] to zF(z,w)/F(0,w) (zero-free, value 1 at 0, growth Mweaz with Mw:=e2bw) gives F(z,w)=F(0,w)ezc(w) for every z, where c(w):=zF(0,w)/F(0,w)=φ(aEb(w))ewφ(b).

step 1.1step 4.1F6F8algebra
6.1

The numerator φ(aEb(w))=n0wnφ(abn)/n! is entire by step 1.3, since its coefficients are bounded by abn/n!. The exponential factor in c(w)=φ(aEb(w))ewφ(b) is entire by the same estimate; the product is holomorphic by the product rule, obtained directly by splitting its difference quotient. Thus c is entire. Fix w. If c(w)0, set z=tc(w)/c(w), t>0. The identity in step 5.1 and [F8] give F(0,w)etc(w)eta+wb. Since F(0,w)=eRe(wφ(b)), this implies t(c(w)a)2wb. Divide by t and let t to obtain c(w)a. If c(w)=0 this bound is immediate.

step 1.1step 1.3step 4.1step 5.1F5F8algebra
7.1

By [F7] the bounded entire function c is constant. At zero, Eb(0)=1, so c(0)=φ(a) and c(w)=φ(a) for all w.

step 1.2step 5.1step 6.1F7algebra
8.1

Consequently F(z,w)=ewφ(b)ezφ(a) for all z,wC by [step 5.1] and [step 7.1].

step 5.1step 7.1algebra
9.1

Differentiate the scalar identity in step 8.1 with respect to z at zero, using the derivative established in step 4.1 and the exponential power series. It gives φ(aEb(w))=φ(a)ewφ(b). Differentiate this identity with respect to w at zero, using the numerator series in step 6.1 and [F5]. Its left derivative is φ(ab) and its right derivative is φ(a)φ(b). Thus φ(ab)=φ(a)φ(b) without invoking any double-series interchange.

step 4.1step 6.1step 8.1F4F5algebra
10.1

By [step 1.1] φ is bounded, hence continuous, and by [step 9.1] it is multiplicative; both assertions of the theorem are proved.

step 1.1step 9.1

Remarks

  • The hypothesis is used twice. It gives continuity through the spectrum argument [step 1.1] and zero-freeness of the functions fa and F(,w) in [steps 2.1 and 4.1]; no other invocation occurs.
  • The two-variable step is not a formal consequence of the one-variable step, which is why the function F and its w-dependent constant c(w) are introduced: the one-variable theorem applied for fixed w produces a constant that has to be shown independent of w.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

74 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