· 24,014 chars · 36 deps · depth 13 Reason: Proof of the complex bounded-maps toolkit (adjoints via real Riesz).
Proves the eight claims in turn: the operator norm is the least bound, the operations and norm inequalities follow pointwise, the real part of the inner product gives a real Hilbert structure, and adjoints are obtained from the real Riesz representation theorem via the identity h(u,v)=Re h(u,v)-i Re h(u,iv). Completeness is proved by pointwise limits with a uniform estimate, and the quadratic-form bound by polarisation and the parallelogram identity.
Proof
Throughout, subscripts on inner products and norms are omitted when the space is clear from the arguments, 0 also denotes the zero vector of each vector space, and x−y=x+(−y), where −y=(−1)y by claim 5 of Elementary Identities in a Vector Space. Elementary real order and arithmetic is used as licensed by The Real Numbers: Standing Notation and Background.
moreover i=−i (as i=0+1i), so ii=−(ii)=1 by condition 2 of The Complex Numbers. From Properties of Complex Conjugation and Modulus we use: conjugation respects sums and products, is involutive, and fixes exactly the real numbers, so a=a and −1=−1 (claim 1); z+z=2Rez (claim 2); ∣z∣=0 only if z=0 (claim 3); ∣zw∣=∣z∣∣w∣ (claim 4); Rez≤∣z∣ and −Rez≤∣z∣, so the absolute value of the real number Rez is at most ∣z∣ (claim 6); and ∣a∣=a when 0≤a, while ∣a∣ is a or −a for every real a (claim 8).
(P3) Separation. If x,y∈X satisfy ⟨x,v⟩=⟨y,v⟩ for every v∈X, then x=y. Indeed, by claims 1 and 2 of Elementary Properties of a Complex Inner Product and −1=−1 we get ⟨x−y,v⟩=⟨x,v⟩−⟨y,v⟩=0 for every v; taking v=x−y, claim 4 of that lemma gives x−y=0, hence x=y.
Claim 2 (Operations). First, for any linear maps S,T:V→W, R:W→U and c∈C, the maps S+T, cT and RT are linear: for u,u′∈V and λ∈C, linearity of S,T,R (Linear Map) and conditions 1, 2, 5 and 7 of Vector Space over a Field in W and U give (S+T)(u+u′)=(S+T)u+(S+T)u′, (S+T)(λu)=λSu+λTu=λ(S+T)u, (cT)(u+u′)=cTu+cTu′, (cT)(λu)=(cλ)Tu=(λc)Tu=λ(cTu), R(T(u+u′))=R(Tu)+R(Tu′) and R(T(λu))=λR(Tu).
Now let S,T∈L(V,W), R∈L(W,U), c∈C and v∈V. By (P2) and claim 1,
where the middle step of the third chain multiplies the bound for Tv by ∥R∥op≥0. The multipliers on the right are nonnegative real numbers, so they are bounds; hence S+T,cT∈L(V,W), RT∈L(V,U) and IV∈L(V) (IV is linear), and the second part of claim 1 gives ∥S+T∥op≤∥S∥op+∥T∥op, ∥cT∥op≤∣c∣∥T∥op, ∥RT∥op≤∥R∥op∥T∥op and ∥IV∥op≤1. For the reverse inequality ∣c∣∥T∥op≤∥cT∥op: if c=0, then ∣c∣=0 by claim 8 of Properties of Complex Conjugation and Modulus and 0≤∥cT∥op by claim 1. If c=0, then 0<∣c∣ by (P1), c−1(cT)=T by conditions 5 and 6 of Vector Space over a Field in W, and ∣c−1∣∣c∣=∣1∣=1 by (P1); the inequality just proved, applied to c−1 and cT, gives ∥T∥op≤∣c−1∣∥cT∥op, and multiplying by ∣c∣ gives the claim.
The zero map 0:V→W is linear and, since ∥0∥=0 by (P2), has the bound 0; so 0∈L(V,W). Two maps V→W are equal when they agree at every v∈V. Evaluating at each v, conditions 1, 2, 5, 6, 7 and 8 of Vector Space over a Field for L(V,W) with the operations above follow from the same conditions in W; condition 3 holds with the zero map, as Tv+0=Tv; and condition 4 holds with (−1)T∈L(V,W), as Tv+(−1)Tv=Tv+(−Tv)=0 by claim 5 of Elementary Identities in a Vector Space. So L(V,W) is a complex vector space with zero vector the zero map. Finally, for R,R′∈L(W,U), S,T∈L(V,W), c∈C and v∈V, linearity of R gives R(Sv+Tv)=R(Sv)+R(Tv) and R(cTv)=cR(Tv), while (R+R′)(Tv)=R(Tv)+R′(Tv) and (cR)(Tv)=cR(Tv) by definition; hence R(S+T)=RS+RT, R(cT)=c(RT), (R+R′)T=RT+R′T and (cR)T=c(RT).
Claim 3 (Underlying real structure). Since R⊆C with the same sums and products (condition 1 of The Complex Numbers) and the same identity 1 (claim 1 of Canonical Form and Arithmetic of Complex Numbers), conditions 1-8 of Vector Space over a Field for X with real scalars are instances of those with complex scalars, so X is a vector space over R. Let u,u′,v,w∈X, c∈C and a∈R. By the third hypothesis on h, the first two, and (P1),
Claim 5 (Existence of adjoints). Let V be a complex Hilbert space and T∈L(V,W). By claim 3, V with ⟨⋅,⋅⟩V,R=Re⟨⋅,⋅⟩V is a real Hilbert space with norm ∥⋅∥V. Fix w∈W and put ℓw(v)=Re⟨w,Tv⟩W for v∈V. For v,v′∈V and a∈R, linearity of T, conditions 2 and 3 of Complex Inner Product Space and (P1) give ℓw(v+v′)=ℓw(v)+ℓw(v′) and ℓw(av)=Re(a⟨w,Tv⟩)=aℓw(v); and by (P1), (P2) and claim 1,
Bounds. Let w∈W. The real number ∥T∗w∥2=⟨T∗w,T∗w⟩V=⟨w,T(T∗w)⟩W is nonnegative, so it equals its modulus by (P1), and by (P2) and claim 1
∥T∗w∥2≤∥w∥∥T(T∗w)∥≤∥w∥∥T∥op∥T∗w∥.
If T∗w=0, then ∥T∗w∥=0≤∥T∥op∥w∥; otherwise 0<∥T∗w∥ by (P2), and dividing gives ∥T∗w∥≤∥T∥op∥w∥. Hence ∥T∥op≥0 is a bound for T∗, so T∗∈L(W,V) and ∥T∗∥op≤∥T∥op by claim 1. Conversely, for v∈V the adjoint identity with w=Tv gives ∥Tv∥2=⟨Tv,Tv⟩W=⟨T∗(Tv),v⟩V, and in the same way ∥Tv∥2≤∥T∗(Tv)∥∥v∥≤∥T∗∥op∥Tv∥∥v∥, whence ∥Tv∥≤∥T∗∥op∥v∥ (trivially if Tv=0). By claim 1, ∥T∥op≤∥T∗∥op, so the two norms are equal.
the last number being real. Finally, the computation for RT uses only that the three spaces are complex inner product spaces; applied to T:V→W and to T∗:W→V (in the roles of T and R, with V in the role of U; the adjoint of T∗ is T), it shows that T∗T has the adjoint T∗T, so T∗T is self-adjoint by the equivalence just proved; and ⟨v,T∗Tv⟩=∥Tv∥2 is a nonnegative real number for every v, so T∗T is positive semi-definite.
Claim 7 (Completeness). Let W be a complex Hilbert space and (Tk) as stated. By claim 2 each Tk−Tl lies in L(V,W), and (Tk−Tl)v=Tkv−Tlv by claim 5 of Elementary Identities in a Vector Space, so by claim 1
T is linear. Let u,u′∈V and λ∈C. Given ε>0, choose N so large that ∥Tku−Tu∥<ε/2, ∥Tku′−Tu′∥<ε/2 and ∥Tku−Tu∥<ε/(∣λ∣+1) for k≥N. Then for k≥N, by (P2),
So (Tk(u+u′)) converges to Tu+Tu′ and (Tk(λu)) to λTu; since they converge to T(u+u′) and T(λu) by definition, Uniqueness of Limits in a Metric Space gives T(u+u′)=Tu+Tu′ and T(λu)=λTu.
Uniform estimate. Let ε>0 and choose N from the hypothesis for ε. Fix k≥N and v∈V, and let δ>0. Choose M with ∥Tlv−Tv∥<δ for l≥M, and put l=max(N,M). By (P2) and (†),
∥Tkv−Tv∥≤∥Tkv−Tlv∥+∥Tlv−Tv∥≤ε∥v∥+δ.
As δ>0 was arbitrary, Comparison of Real Numbers with Arbitrary Positive Slack §slack-above gives ∥(Tk−T)v∥=∥Tkv−Tv∥≤ε∥v∥. So ε is a bound for the linear map Tk−T (linear by claim 2), whence Tk−T∈L(V,W) and ∥Tk−T∥op≤ε for all k≥N by claim 1. In particular TN−T∈L(V,W), and T=TN+(−1)(TN−T) (evaluate at each v), so T∈L(V,W) by claim 2. Given ε>0, the above with ε/2 yields N with ∥Tk−T∥op−0=∥Tk−T∥op≤ε/2<ε for k≥N, so (∥Tk−T∥op) converges to 0 in the sense of Limit of a Sequence of Real Numbers.
Pointwise convergence for every such T. Let T∈L(V,W) with (∥Tk−T∥op) converging to 0, and let v∈V and ε>0. Choose N with ∥Tk−T∥op<ε/(∥v∥+1) for k≥N. Then claim 1 gives, for k≥N, d(Tkv,Tv)=∥(Tk−T)v∥≤∥Tk−T∥op∥v∥<ε, so (Tkv) converges to Tv in W (Convergent Sequence in a Metric Space).
Claim 8 (Quadratic-form bound). Let A and c be as stated; A is a linear operator and self-adjoint as in Self-Adjoint Operator. For x∈V, ⟨x,Ax⟩=⟨Ax,x⟩=⟨x,Ax⟩ by self-adjointness and condition 1 of Complex Inner Product Space, so ⟨x,Ax⟩ is real by (P1), and ⟨x,Ax⟩≤c∥x∥2 and −⟨x,Ax⟩≤c∥x∥2 because its modulus is ⟨x,Ax⟩ or −⟨x,Ax⟩ (P1).
With B=IV, adding, and (P2): ∥u+w∥2+∥u−w∥2=2∥u∥2+2∥w∥2. By the first paragraph applied to x=u+w and to x=u−w,
4Re⟨u,Aw⟩≤c∥u+w∥2+c∥u−w∥2=2c(∥u∥2+∥w∥2)(u,w∈V).
Now fix w∈V. If Aw=0, then ∥Aw∥=0≤c∥w∥. Otherwise w=0 (as A0=0 by (P2)), so 0<∥Aw∥ and 0<∥w∥; put s=∥w∥/∥Aw∥>0 and u=sAw. By claim 2 of Elementary Properties of a Complex Inner Product, (P1) and (P2), ⟨u,Aw⟩=s∥Aw∥2=∥w∥∥Aw∥, a real number, so Re⟨u,Aw⟩=∥w∥∥Aw∥, and ∥u∥=∣s∣∥Aw∥=∥w∥. The last inequality becomes 4∥w∥∥Aw∥≤4c∥w∥2, and dividing by 4∥w∥>0 gives ∥Aw∥≤c∥w∥. Hence the real number c≥0 is a bound for A, so A∈L(V), and ∥A∥op≤c by claim 1.