Pipeline-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.

✓ 8 results · all verified · 8 also independently AI-judged

Every result on this page is machine-checked by a proof checker and read in full and owner-audited; the judge is an additional, independent cross-model AI review of the proofs; all 8 also cleared it.

Sl2 R Principal and Complementary Series

1 · Prerequisites

2 · Summary

This page studies the normalized principal series of $G = {S L}_{2} (R)$ using its minimal parabolic subgroup and compact picture. It fixes the parity parameter $ε$ , the inducing parameter $ν$ , and the half-modular normalization, then derives the one-dimensional $K = S O (2)$ types and their raising and lowering coefficients.

The algebraic, Hilbert-space, and smooth representations are irreducible away from the parity-compatible exceptional lattice. At the exceptional parameters, the ladder zeros determine the finite-dimensional factor, the two one-sided factors, their submodule and quotient orientations, and the odd-parity split at $ν = 0$ . The page also computes the standard intertwining operator on each $K$ -type and uses its continuation and invariant pairings to establish the unitary principal-series and spherical complementary-series ranges, parameter equivalence, and convergence of the spherical complementary series to the trivial representation. Detailed coordinate, ladder, and sign computations are collected on the companion examples page.

3 · Logical flowchart

4 · Definitions, theorems and proofs

DefinitionDefinition: Literature-sourcedProof: Not applicablejudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Iwasawa and minimal-parabolic data for SL2(R)

Definition

Assume the Axiom of Choice (The Axiom of Choice). Let $G = {S L}_{2} (R) = {g \in M_{2} (R) : \det g = 1}$ , the embedded matrix Lie group of General and special linear Lie groups, with the real Lie-group conventions of Lie group and Real and complex Lie groups. Set

$K = S O (2) = {k_{θ} = (\begin{array}{cc} \cos θ & \sin θ \\ - \sin θ & \cos θ \end{array}) : θ \in R / 2 π Z} .$

This is a closed embedded Lie subgroup (Orthogonal and special orthogonal Lie groups), isomorphic to the circle group $T = R / 2 π Z$ ; the isomorphism with the quotient $R / Z$ of The one-dimensional torus and its normalized Haar integral is $[θ]_{2 π} \mapsto [θ / (2 π)]_{1}$ . Write $d k$ for the normalized Haar probability measure on $K$ (Normalized Haar measure on a compact Lie group). Then $k_{θ + π} = - k_{θ}$ .

Let $A = {a_{t} = diag (e^{t / 2}, e^{- t / 2}) : t \in R}, N = {n_{x} = (\begin{array}{cc} 1 & x \\ 0 & 1 \end{array}) : x \in R}, M = {\pm I} .$ In fact $M = Z (G)$ : commuting with $a_{1}$ forces a central matrix to be diagonal, and commuting with $n_{1}$ forces its diagonal entries to agree; determinant one then gives precisely $I$ and $- I$ .

The coordinates $a_{t} \leftrightarrow t$ and $n_{x} \leftrightarrow x$ identify $A$ and $N$ with the additive group $R$ ; explicitly, $a_{t} a_{s} = a_{t + s}$ , $n_{x} n_{y} = n_{x + y}$ , and $a_{t} n_{x} a_{- t} = n_{e^{t} x}$ . The standard minimal parabolic subgroup is $P = M A N = {(\begin{array}{cc} u & b \\ 0 & u^{- 1} \end{array}) : u \in R^{\times}, b \in R},$ the closed stabilizer of the line $R (1, 0)$ in the standard action on $R^{2}$ : preservation of that line is exactly the vanishing of the lower-left entry. The factorization is unique: for the displayed matrix, set $m = sgn (u) I$ , $t = 2 \log ∣ u ∣$ , and $x = b / u$ ; then $p = m a_{t} n_{x}$ , and the diagonal and top-right entries force these same values from any such factorization. The displayed coordinates identify $P$ with $R^{\times} \times R$ , so its identity component is the $u > 0$ component $A N$ . A unipotent element has both eigenvalues equal to $1$ , forcing $u = 1$ ; hence the unipotent elements of $P$ are exactly $N$ , which is connected and normal. Thus $N$ is the unipotent radical. Directly, $K \cap P = {\pm I} = M$ .

Put $H = diag (1, - 1)$ , $e_{0} = (\begin{array}{cc} 0 & 1 \\ 0 & 0 \end{array})$ , $a = R H$ , and $n = R e_{0}$ . By General and special linear Lie groups, the Lie algebra of this real $G$ is the traceless real matrices; $H, e_{0}$ lie in it, and direct multiplication gives $[H, e_{0}] = 2 e_{0}$ , the same matrix relation used in The special linear Lie algebra sl_2. The explicit matrix exponential gives $a_{t} = \exp_{G} (t H / 2)$ and $n_{x} = \exp_{G} (x e_{0})$ , so their coordinate maps are one-parameter subgroups (Exponential map of a Lie group, One-parameter subgroup of a Lie group, One-parameter subgroups are exactly exponentials). In particular, $Ad (a_{t}) e_{0} = e^{t} e_{0}$ .

For $p \in P$ , define the parabolic modular character by $δ_{P} (p) = ∣ \det ⁣ (Ad (p) ∣_{n}) ∣ .$ Then $δ_{P} ∣_{M N} = 1$ and $δ_{P} (a_{t}) = e^{t}$ . To distinguish this character from the group modular function of Modular function of a locally compact group, write the latter as $Δ_{P}$ : with the convention $\int_{P} f (x p^{- 1}) d μ (x) = Δ_{P} (p) \int_{P} f d μ$ , one has $Δ_{P} = δ_{P}^{- 1}$ and hence $Δ_{P} (a_{t}) = e^{- t}$ . Indeed, on $A N$ the coordinates $a_{t} n_{x}$ have left Haar measure $d t d x$ ; right translation by $a_{- s}$ is $(t, x) \mapsto (t - s, e^{s} x)$ and scales the integral by $e^{- s}$ . Kerr denotes the parabolic modular character $δ_{P}$ by $Δ_{P}$ .

For $ε \in {0, 1}$ and $ν \in C$ , let $σ_{ε} (\pm I) = (\pm 1)^{ε}$ , extend it trivially over $A N$ , and define the character $e^{ν}$ to be trivial on $M N$ and satisfy $e^{ν} (a_{t}) = e^{ν t / 2}$ . Set $∣ α (a_{t}) ∣ = e^{t / 2}$ . The normalized inducing character is $σ_{ε} δ_{P}^{1 / 2} e^{ν}$ ; on $A$ it is $∣ α (a_{t}) ∣^{1 + ν} = e^{(1 + ν) t / 2} = δ_{P} (a_{t})^{1 / 2} e^{ν} (a_{t}) .$

All principal-series parameters on this page use the letter $ν$ with this normalization, and the half-modular shift is applied exactly once. AC is used to obtain $d k$ through Normalized Haar measure on a compact Lie group. A countable family of nonempty sets is a family to which AC applies, so AC supplies the ${A C}_{ω}$ hypothesis stated in The Axiom of Countable Choice ( ${A C}_{ω}$ ) and required by the cited exponential-map and one-parameter-subgroup results. The groups and coordinates above are otherwise explicit.

TheoremStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Iwasawa decomposition and Haar integration formula for SL2(R)

Statement

Assume the Axiom of Choice (The Axiom of Choice) and use the notation of Iwasawa and minimal-parabolic data for SL2(R).

(1) Diffeomorphism. Multiplication $K \times A \times N ⟶ G, (k, a, n) ⟼ k a n,$ is a diffeomorphism. For $g = (\begin{array}{cc} p & r \\ q & s \end{array}) \in G$ , put $a = \sqrt{p^{2} + q^{2}}, k = a^{- 1} (\begin{array}{cc} p & - q \\ q & p \end{array}), x = \frac{p r + q s}{a^{2}} .$ Then $a > 0$ , $k \in K$ , $a = e^{t / 2}$ for a unique $t \in R$ , and $g = k diag (a, a^{- 1}) n_{x}$ is the unique factorization with $k \in K$ , $diag (a, a^{- 1}) \in A$ , and $n_{x} \in N$ .

(2) Haar integral. With normalized Haar probability $d k$ on $K$ and Lebesgue measures $d t$ on $A ≅ R$ and $d x$ on $N ≅ R$ , $f ⟼ \int_{K} \int_{R} \int_{R} f (k a_{t} n_{x}) e^{t} d x d t d k$ is a left Haar integral on $G$ . In the opposite order the same measure is $\int_{K} \int_{R} \int_{R} f (n_{x} a_{t} k) e^{- t} d x d t d k .$

(3) Unimodularity. The group $G$ is unimodular; both displayed Haar integrals are right invariant as well.

Facts & Assumptions

Given: AC and the matrix group $G = {S L}_{2} (R)$ .

[F1]

$G, K, A, N$ , the matrices $a_{t}, n_{x}$ , and normalized $d k$ are fixed in Iwasawa and minimal-parabolic data for SL2(R). The displayed parametrization identifies $K$ with the circle $R / 2 π Z$ .

[F2]

Under $θ \mapsto [θ / (2 π)]$ , the Lebesgue fundamental-domain probability on $R / Z$ of The one-dimensional torus and its normalized Haar integral pulls back to the translation-invariant probability $d θ / (2 π)$ on $K$ ; uniqueness of normalized Haar probability identifies it with $d k$ (Normalized Haar measure on a compact Lie group).

[F3]

Every invertible real matrix has a unique QR factorization with an orthogonal factor and an upper-triangular factor with positive diagonal (Every invertible real or complex square matrix has a unique factorisation $A = Q R$ with $Q$ orthogonal or unitary and $R$ upper triangular with positive real diagonal).

[F4]

A smooth bijection with smooth inverse is a diffeomorphism; smoothness is checked in the matrix and angle coordinates ( $C^{r}$ and smooth maps between smooth manifolds, Diffeomorphisms and local diffeomorphisms of manifolds).

[F5]

$d t$ and $d x$ are Lebesgue measures; integration of a compactly supported smooth density changes by the absolute Jacobian under a diffeomorphism (Lebesgue measurable sets, the family $L (R^{n})$ , and the restricted set function $λ_{n}$ , Change of variables on oriented manifolds).

[F6]

A left Haar integral is a nonzero positive left-invariant functional; a left Haar measure is a nonzero Radon measure finite on compact sets (Left Haar integral and left Haar measure).

[F7]

A left Haar measure is also a right Haar measure exactly when $G$ is unimodular (Unimodular locally compact group).

[F8]

Under ${A C}_{ω}$ , a finite-valued positive smooth density on a second-countable smooth manifold defines a Radon measure finite on compact sets (Positive smooth densities give Radon volume).

[F9]

Under ${A C}_{ω}$ , Borel integration against this density measure agrees with its chart-density integral; for smooth compactly supported functions this is the smooth density integral (Measurable integration extends smooth density integration).

[F10]

A measurable transformation preserving a measure preserves integrals of nonnegative measurable and integrable functions (Measure-preserving transformations and systems, Integral invariance under measure-preserving maps).

[F11]

The Lebesgue integral is real- and complex-linear on $L^{1}$ (The Lebesgue integral is linear on $L^{1} (μ)$ ).

[A1]

AC supplies normalized Haar probability on $K$ and, by restriction to any countable family, the ${A C}_{ω}$ hypotheses for [F8] and [F9]. The matrix and Jacobian calculations make no other arbitrary choice (The Axiom of Choice, The Axiom of Countable Choice ( ${A C}_{ω}$ )).

Proof

technique · direct

1.1F1F3algebra

Write $g = Q R$ by [F3]. Since $\det g = 1$ and $R$ has positive diagonal, $\det Q = 1$ , so $Q \in K$ . Writing $a = R_{11} > 0$ and $R_{12} = a x$ , the determinant condition gives $R_{22} = a^{- 1}$ and $R = diag (a, a^{- 1}) n_{x}$ . The first column gives $a = \sqrt{p^{2} + q^{2}}$ and $Q = a^{- 1} (\begin{array}{cc} p & - q \\ q & p \end{array})$ ; the second gives $x = (p r + q s) / a^{2}$ . QR uniqueness proves the factorization unique.

2.1F4step 1.1algebra

Multiplication is smooth. Its inverse is given by the displayed formulas, with $a > 0$ because the first column of a determinant-one matrix is nonzero; $a$ , $a^{- 1}$ , $k$ , and $x$ therefore depend smoothly on $g$ . The coordinate $t = 2 \log a$ is smooth as well, so the multiplication map is a diffeomorphism.

3.1A1F1F2F5step 2.1algebra

For fixed $g_{0} \in G$ , write uniquely $g_{0} k_{θ} = k_{θ_{1}} a_{u (θ)} n_{v (θ)}$ . With $e_{1} = (1, 0)^{T}$ , put $w (θ) = g_{0} k_{θ} e_{1} = r (θ) k_{θ_{1}} e_{1}$ , where $r = e^{u / 2} > 0$ . Since $\det (k_{θ} e_{1}, (k_{θ} e_{1})^{'}) = - 1$ and $\det g_{0} = 1$ , differentiating this identity gives $- 1 = - r^{2} θ_{1}^{'}$ , hence $θ_{1}^{'} = e^{- u} > 0$ . Left translation sends $(k_{θ}, t, x)$ to $(k_{θ_{1}}, t + u, x + e^{- t} v)$ ; its $(t, x)$ -Jacobian is $1$ , so its full orientation-preserving Jacobian is $e^{- u}$ . The target density is $e^{t + u} d k d t d x$ , and its pullback is $e^{t + u} e^{- u} d k d t d x = e^{t} d k d t d x$ . Thus the smooth density is left invariant; the change-of-variables theorem applies to compactly supported smooth test densities.

4.1A1F6F8F9F10F11step 2.1step 3.1algebra

Let $η = e^{t} d k d t d x$ be the positive smooth density in the global $K \times R^{2}$ chart. By [F8] it defines a Radon Borel measure $μ_{η}$ finite on compact sets, and [F9] identifies its Borel integral with the displayed coordinate integral. Step 3.1 makes $μ_{η}$ left invariant. A nonnegative smooth bump supported in a nonempty coordinate box has positive integral because $η$ is positive, so $I (f) : = \int f d μ_{η}$ is nonzero. It is positive and real-linear by the Lebesgue integral properties, and [F10] gives left invariance of $I$ ; hence $I$ is a left Haar integral and $μ_{η}$ is a left Haar measure by [F6].

5.1F1F7F8step 2.1step 4.1algebra

Put $H = diag (1, - 1)$ , $e = (\begin{array}{cc} 0 & 1 \\ 0 & 0 \end{array})$ , $f = (\begin{array}{cc} 0 & 0 \\ 1 & 0 \end{array})$ , $J = (\begin{array}{cc} 0 & 1 \\ - 1 & 0 \end{array})$ , and $S = (\begin{array}{cc} 0 & 1 \\ 1 & 0 \end{array})$ . Direct conjugation gives $Ad (k_{θ}) J = J$ , $Ad (k_{θ}) H = \cos (2 θ) H - \sin (2 θ) S$ , and $Ad (k_{θ}) S = \sin (2 θ) H + \cos (2 θ) S$ , so its determinant is $1$ . Also $Ad (a_{t})$ has eigenvalues $1, e^{t}, e^{- t}$ on $(H, e, f)$ . On $(e, H, f)$ , $Ad (n_{x}) e = e$ , $Ad (n_{x}) H = H - 2 x e$ , and $Ad (n_{x}) f = f + x H - x^{2} e$ , so its determinant is $1$ . Since $Ad$ is multiplicative and $G = K A N$ , $\det Ad (g) = 1$ for every $g \in G$ . For a left-invariant density $η$ , comparing $d R_{g}$ with $d L_{g^{- 1}}$ at the identity gives $d L_{g^{- 1}} d R_{g} = Ad (g^{- 1})$ , hence $R_{g}^{*} η = ∣ \det Ad (g^{- 1}) ∣ η = η$ . Thus $μ_{η}$ is right invariant and $G$ is unimodular by [F7].

6.1A1F1F2F8F9F10step 4.1step 5.1algebra∎

Inversion pulls the right-invariant density $η$ back to a left-invariant density; its differential at the identity is $- I$ on the three-dimensional tangent space, whose absolute determinant is $1$ . Since a left-invariant density is determined by its value at the identity, inversion preserves $μ_{η}$ . Applying inversion invariance to the KAN integral and using $(k a_{t} n_{x})^{- 1} = n_{- x} a_{- t} k^{- 1}$ , then substituting $(k, t, x) \mapsto (k^{- 1}, - t, - x)$ , gives the same measure in NAK coordinates with density $e^{- t} d k d t d x$ ; here $d k = d θ / (2 π)$ is inversion invariant by [F2].

DefinitionDefinition: Literature-sourcedProof: Not applicablejudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

The normalized principal series I(epsilon, nu)

Definition

Assume the Axiom of Choice (The Axiom of Choice) and use the data of Iwasawa and minimal-parabolic data for SL2(R). For $ε \in {0, 1}$ and $ν \in C$ , extend $σ_{ε} (m) = (\pm 1)^{ε}$ on $m = \pm I$ and $e^{ν} (a_{t}) = e^{ν t / 2}$ to characters of $P = M A N$ that are trivial on the other factors. The normalized inducing character is $χ_{ε, ν} (p) = δ_{P} (p)^{1 / 2} σ_{ε} (m_{p}) e^{ν} (p) = ∣ α (p) ∣^{1 + ν} σ_{ε} (m_{p}), p = m_{p} a_{t} n_{x},$ where $∣ α (m_{p} a_{t} n_{x}) ∣ = e^{t / 2}$ and $δ_{P}$ is the parabolic modular character of the preceding definition.

The normalized smooth principal series $I_{ε, ν}$ is the space of smooth functions $φ : G \to C$ satisfying $φ (p g) = χ_{ε, ν} (p) φ (g) (p \in P, g \in G),$ with the right-translation action $(Π_{ν} (g_{0}) φ) (g) = φ (g g_{0}) .$ It preserves covariance since $(Π_{ν} (g_{0}) φ) (p g) = φ (p g g_{0}) = χ_{ε, ν} (p) φ (g g_{0})$ , and $(Π_{ν} (g_{1}) Π_{ν} (g_{2}) φ) (g) = φ (g g_{1} g_{2}) = (Π_{ν} (g_{1} g_{2}) φ) (g)$ . The inducing character $σ_{ε} e^{ν}$ is unitary exactly when $ν \in i R$ , since $∣ σ_{ε} (m) e^{ν} (a_{t}) ∣ = e^{(Re ν) t / 2}$ for all $t \in R$ ; $δ_{P}^{1 / 2}$ is the separate half-modular normalization.

The $K$ -finite subspace $I_{ε, ν}^{K}$ consists of those $φ$ for which the right $K$ -translates span a finite-dimensional space. It is the associated $({s l}_{2} (C), K)$ -module, but need not be stable under the full noncompact group $G$ ; the ambient smooth space above carries that action. This is Kerr's distinction between the smooth globalization and its Harish-Chandra module. Under inversion $F (g) = φ (g^{- 1})$ , write $B = P$ and let $t (b)$ be the signed top-left diagonal entry of $b \in B$ . Then $F (g b) = φ (b^{- 1} g^{- 1}) = χ_{ε, ν} (b^{- 1}) F (g) = ∣ t (b) ∣^{- 1 - ν} sgn (t (b))^{ε} F (g) .$ The left action $(g_{0} \cdot F) (g) = F (g_{0}^{- 1} g)$ corresponds under inversion to $Π_{ν} (g_{0})$ , so this is Etingof's $V^{ε} (s)$ model with $s = - ν$ .

The parity- $ε$ parameter lattice used below is $W_{ε} = {ν \in Z : ν \equiv ε + 1 (m o d 2)},$ so $W_{0}$ is the odd integers and $W_{1}$ is the even integers. AC is inherited through the Iwasawa data and the AC $_{ω}$ hypotheses of its exponential suppliers; this definition makes no additional choice.

TheoremStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

The compact picture of the SL2(R) principal series

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ , $ν \in C$ , and let $I_{ε, ν}$ be as in The normalized principal series I(epsilon, nu). Write $C_{ε}^{\infty} (K) = {f \in C^{\infty} (K) : f (k_{θ + π}) = (- 1)^{ε} f (k_{θ})},$ and define $L_{ε}^{2} (K)$ by the same parity condition in $L^{2} (K, d k)$ .

(1) Smooth compact picture. Restriction $φ \mapsto φ ∣_{K}$ is a linear isomorphism from the smooth covariant functions of $I_{ε, ν}$ onto $C_{ε}^{\infty} (K)$ . For any factorization $g = p k$ with $p = m_{p} a_{t} n_{x} \in P$ and $k \in K$ , its inverse is $φ (g) = ∣ α (p) ∣^{1 + ν} σ_{ε} (m_{p}) f (k) .$ The parity condition makes this independent of the $M = P \cap K$ ambiguity. The unique positive-diagonal factorization $g = a_{t} n_{x} k$ gives the canonical formula $φ (g) = e^{(1 + ν) t / 2} f (k)$ . This isomorphism intertwines right translation with the Iwasawa-cocycle action $(g_{0} \cdot f) (k) = ∣ α (p (k, g_{0})) ∣^{1 + ν} σ_{ε} (m_{p (k, g_{0})}) f (κ (k, g_{0})),$ where $k g_{0} = p (k, g_{0}) κ (k, g_{0})$ is the unique $A N \times K$ factorization; in this canonical factorization $p (k, g_{0}) \in A N$ and $m_{p (k, g_{0})} = I$ .

(2) Unitary case. If $ν \in i R$ , the inducing character $σ_{ε} e^{ν}$ is unitary. Restriction carries the normalized induced inner product to $⟨ f, h ⟩ = \int_{K} f (k) \overline{h (k)} d k$ on $L_{ε}^{2} (K)$ . The resulting representation is strongly continuous and unitary, and is equivalent to the right-covariant model of ${Ind}_{P}^{G} (σ_{ε} \otimes e^{ν})$ in Unitary induction from a closed subgroup. Under inversion and the rho half-density, the parameter remains $ν$ in that unitary model.

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in C$ , and the normalized principal series from part (1).

[F1]

The subgroups, characters, modular conventions, normalized Haar measure $d k$ , and coordinates on $K, A, N$ are fixed by Iwasawa and minimal-parabolic data for SL2(R).

[F2]

Multiplication gives unique smooth $K A N$ and $N A K$ coordinates; the Haar density in $K A N$ coordinates is $e^{t} d k d t d x$ , and $G$ is unimodular (Iwasawa decomposition and Haar integration formula for SL2(R)).

[F3]

The smooth model has left $P$ -covariance by $χ_{ε, ν} = δ_{P}^{1 / 2} σ_{ε} e^{ν}$ and right-translation action; $σ_{ε} e^{ν}$ is unitary for $ν \in i R$ (The normalized principal series I(epsilon, nu)).

[F4]

The unitary induction model uses continuous right- $P$ -covariant functions, the rho quotient norm, and its Hilbert completion (Continuous covariant model and measurable completion).

[F5]

A rho-function satisfies $ρ (x p) = Δ_{P} (p) Δ_{G} (p)^{- 1} ρ (x)$ (Rho-function for a closed subgroup).

[F6]

The group modular functions are continuous homomorphisms; their convention is fixed by Modular function of a locally compact group and The modular function is a continuous homomorphism.

[F7]

For closed $P$ , $G / P$ is locally compact Hausdorff, the quotient map is continuous, and subgroup averaging sends $C_{c} (G)$ to $C_{c} (G / P)$ (Compact lifts and averaging onto C_c(G/H)).

[F8]

A continuous image of a compact space is compact (A continuous image of a compact space is compact; a continuous real-valued map on a nonempty compact space attains a maximum and a minimum; and a continuous bijection from a compact space to a Hausdorff space is a homeomorphism).

[F9]

A positive functional on $C_{c} (G / P; R)$ is integration against a Radon measure (Positive functionals on C_c(X) are integration against a Radon measure).

[F10]

For fixed left Haar measures and a rho-function, the Weil formula gives a unique Radon quotient measure (Weil formula with a rho-function).

[F11]

If the inducing character is unitary, the completed covariant model with its cocycle action is a strongly continuous unitary representation (Unitary induction from a closed subgroup).

[F12]

The quotient Radon–Nikodym cocycle is $D_{g} (x P) = ρ (g^{- 1} x) / ρ (x)$ , and its square root corrects the left action (Continuous quotient translation cocycle).

[F13]

Trigonometric polynomials are finite linear combinations of the circle characters (Fourier coefficients and trigonometric polynomials on the torus).

[F14]

Trigonometric polynomials are uniformly dense in continuous functions on the circle (Trigonometric polynomials are uniformly dense in continuous functions on the torus).

[F15]

Continuous functions are dense in $L^{2}$ on the finite torus with normalized Haar measure (Continuous functions are dense in $L^{p}$ of finite tori and of bounded intervals).

[F16]

Normalized Haar measure on a compact Lie group is right-translation and inversion invariant (Normalized Haar measure on a compact Lie group).

[F17]

A left Haar measure is a nonzero left-invariant Borel Radon measure finite on compact sets, and Lebesgue measure on $R^{2}$ is Radon under ${A C}_{ω}$ (Left Haar integral and left Haar measure, Lebesgue measure is a Radon measure on R^n).

[A1]

The Axiom of Choice supplies the normalized Haar measure and is assumed by the Weil, cocycle, and unitary-induction suppliers (The Axiom of Choice).

[A2]

AC implies the countable-choice hypothesis used by the Fourier-density suppliers (The Axiom of Countable Choice ( ${A C}_{ω}$ )).

Proof

technique · direct

1.1F1F2algebra

By [F2], write each $g \in G$ uniquely as $g = n_{x} a_{t} k$ . Since $n_{x} a_{t} = a_{t} n_{e^{- t} x}$ , this is also the unique smooth factorization $g = a_{t} n_{e^{- t} x} k$ with $a_{t} n_{e^{- t} x} \in A N$ . Thus restriction to $K$ is injective on the induced function model, because covariance determines $φ$ on every $g = p k$ .

1.2A1F1F2F5F6F7F8F9F10F16F17A2algebra

By [F2], write uniquely $g = k a_{t} n_{x}$ and set $ρ (g) = e^{- t}$ . For $p_{0} = m a_{s} n_{y}$ , direct multiplication using centrality of $M$ gives $g p_{0} = k m a_{t + s} n_{e^{- s} x + y}$ ; hence $ρ (g p_{0}) = e^{- s} ρ (g) = Δ_{P} (p_{0}) Δ_{G} (p_{0})^{- 1} ρ (g)$ by [F1], [F2], and [F6]. Thus $ρ$ is a positive continuous rho-function by [F5], and $ρ (k) = 1$ for $k \in K$ . The continuous map $q : K \to G / P$ , $q (k) = k P$ , is onto because $G = K A N$ ; therefore $G / P$ is compact by [F7] and [F8]. Define $Λ (ψ) = \int_{K} ψ (k P) d k$ on $C_{c} (G / P; R)$ . This is positive and $Λ (1) = 1$ , so [F9] represents it by a Radon probability $\overset{ˉ}{d k}$ , the pushforward of $d k$ . Give $P = M \times A N$ left Haar measure $d p$ by normalized counting on $M$ and $d t d x$ on $A N$ ; left multiplication by $a_{s} n_{y}$ sends $(t, x)$ to $(s + t, x + e^{- t} y)$ with Jacobian $1$ , and $M$ is central. The Radon property follows from [F17] and finite disjoint union over $M$ . For $h \in C_{c} (G)$ , subgroup averaging [F7] makes $k P \mapsto \int_{P} h (k p) ρ (k p)^{- 1} d p$ a member of $C_{c} (G / P)$ . The Weil formula [F10], applied to $h / ρ$ , gives $\int_{G} h (g) d g = \int_{G / P} \int_{P} h (k p) ρ (k p)^{- 1} d p d μ_{ρ} (k P)$ . With $\overset{ˉ}{d k}$ instead, the right side is $\int_{K} \frac{1}{2} \sum_{m \in M} \int_{R^{2}} h (k m a_{s} n_{y}) e^{s} d y d s d k = \int_{K} \int_{R^{2}} h (k a_{s} n_{y}) e^{s} d y d s d k$ , since $d k$ is right- $M$ -invariant by [F16]. This is the KAN Haar integral [F2], so uniqueness in [F10] gives $\overset{ˉ}{d k} = μ_{ρ}$ .

1.3F13F14F15F16A2algebra

Identify $K$ with the circle by $k_{θ} \leftrightarrow [θ / (2 π)]$ . On continuous functions the projection $P_{ε} f (k) = \frac{1}{2} (f (k) + (- 1)^{ε} f (- k))$ has norm at most $1$ and range $C_{ε} (K)$ . By [F14], trigonometric polynomials approximate each continuous function uniformly; applying $P_{ε}$ gives parity trigonometric polynomials approximating each continuous parity function uniformly. By [F15], continuous functions are dense in $L^{2} (K)$ ; applying the same bounded projection, which is an $L^{2}$ contraction by [F16], shows continuous parity functions are dense in $L_{ε}^{2} (K)$ . Therefore parity trigonometric polynomials, which are $K$ -finite by [F13], are dense in $L_{ε}^{2} (K)$ .

2.1F1F2F3step 1.1algebra

If $φ \in I_{ε, ν}$ , its $M$ -covariance gives $f (- k) = (- 1)^{ε} f (k)$ for $f = φ ∣_{K}$ . Conversely, for $f \in C_{ε}^{\infty} (K)$ define $φ (a_{t} n_{x} k) = e^{(1 + ν) t / 2} f (k)$ in the unique $A N K$ coordinates. Left multiplication by $a_{s} n_{y}$ changes $t$ to $s + t$ and the $K$ factor remains fixed; left multiplication by $m \in M$ replaces $k$ by $m k$ and multiplies $f$ by $σ_{ε} (m)$ . Thus $φ (p g) = χ_{ε, ν} (p) φ (g)$ and $φ$ is smooth. For any factorization $g = p k$ with $p = m a_{t} n_{x}$ , centrality of $m$ gives the same canonical factorization $g = a_{t} n_{x} (m k)$ , so $e^{(1 + ν) t / 2} σ_{ε} (m) f (k) = e^{(1 + ν) t / 2} f (m k)$ ; this proves the inverse formula and shows its independence of the $M$ -ambiguity.

2.2F1F3F4F5F6F12step 1.2algebra

For $ν \in i R$ , define $U φ (x) = ρ (x)^{- 1 / 2} φ (x^{- 1})$ . By [F5] and [F6], $ρ (x p)^{- 1 / 2} = Δ_{P} (p)^{- 1 / 2} ρ (x)^{- 1 / 2}$ , since $Δ_{G} = 1$ . Also $χ_{ε, ν} = δ_{P}^{1 / 2} τ_{ν} = Δ_{P}^{- 1 / 2} τ_{ν}$ , so $φ (p^{- 1} x^{- 1}) = χ_{ε, ν} (p)^{- 1} φ (x^{- 1})$ cancels the modular factor and gives $U φ (x p) = τ_{ν} (p)^{- 1} U φ (x)$ . The resulting continuous section has compact quotient support because $G / P$ is compact by step 1.2. For $g_{0} \in G$ , direct substitution yields $U (Π_{ν} (g_{0}) φ) (x) = {(\frac{ρ (g_{0}^{- 1} x)}{ρ (x)})}^{1 / 2} U φ (g_{0}^{- 1} x) = D_{g_{0}} (x P)^{1 / 2} U φ (g_{0}^{- 1} x)$ by [F12]. Hence $U$ intertwines right translation on the left-covariant model with the cocycle-corrected left action on the right- $P$ -covariant model of ${Ind}_{P}^{G} (τ_{ν})$ ; the parameter remains $ν$ .

2.3F3F4F10F16step 1.2algebra

The quotient norm of $U φ$ is, by [F10] and step 1.2, $∥ U φ ∥^{2} = \int_{G / P} ∣ U φ (x) ∣^{2} d μ_{ρ} (x P) = \int_{K} ∣ U φ (k) ∣^{2} d k = \int_{K} ∣ φ (k^{- 1}) ∣^{2} d k$ . Inversion invariance of normalized Haar measure [F16] makes this $\int_{K} ∣ f (k) ∣^{2} d k$ , so restriction is isometric for the normalized induced inner product.

3.1F1F2F3step 2.1algebra

For $g_{0} \in G$ , factor $k g_{0} = a_{t (k, g_{0})} n_{x (k, g_{0})} κ (k, g_{0})$ by the canonical $A N K$ coordinates. Then $(Π_{ν} (g_{0}) φ) (k) = φ (k g_{0}) = e^{(1 + ν) t (k, g_{0}) / 2} f (κ (k, g_{0}))$ . In a general $P K$ factorization the same value is $∣ α (p) ∣^{1 + ν} σ_{ε} (m_{p}) f (κ)$ ; the parity relation makes this independent of the $M$ -choice. Since $- I$ is central, replacing $k$ by $- k$ leaves $t, x$ fixed and replaces $κ$ by $- κ$ , so the formula preserves the parity- $ε$ subspace. This is the stated Iwasawa-cocycle action, with $m_{p} = I$ in the canonical $A N K$ coordinates.

4.1A1F3F4F11step 1.3step 2.2step 2.3algebra∎

The coordinate construction in step 2.1, together with the inverse of $U$ from step 2.2, identifies continuous right- $P$ -covariant sections with continuous parity- $ε$ functions on $K$ . Step 2.3 identifies their quotient norm with the $L^{2} (K)$ norm, and step 1.3 shows the smooth parity functions are dense in $L_{ε}^{2} (K)$ ; thus the compact-picture isometry extends onto the completed induced Hilbert space. By [F11], ${Ind}_{P}^{G} (τ_{ν})$ is strongly continuous and unitary because $τ_{ν}$ is a continuous unitary character for $ν \in i R$ . The intertwining identity in step 2.2 transfers these properties to the compact-picture action.

LemmaStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

K-type decomposition of the SL2(R) principal series

Statement

Assume the Axiom of Choice (The Axiom of Choice). In the compact picture of The compact picture of the SL2(R) principal series, put $f_{n} (k_{θ}) = e^{i n θ}$ for $n \in Z$ . Then:

Here, a vector is $K$ -finite when the span of its right $K$ -orbit is finite-dimensional.

$f_{n} \in C_{ε}^{\infty} (K)$ exactly when $n \equiv ε (m o d 2)$ , and ${f_{n} : n \equiv ε (m o d 2)}$ is an orthonormal basis of $L_{ε}^{2} (K)$ ;
the right $K$ -action is $k_{ϕ} \cdot f_{n} = e^{i n ϕ} f_{n}$ , so every $K$ -type is one-dimensional, spanned by one such $f_{n}$ , and occurs with multiplicity one;
the algebraic direct sum $⨁_{n \equiv ε (2)} C f_{n}$ is exactly the space of $K$ -finite vectors and is dense in $L_{ε}^{2} (K)$ ; every closed $K$ -invariant subspace of $L_{ε}^{2} (K)$ is the closed span of the $f_{n}$ it contains.

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , and the compact picture from part (1).

[F1]

Restriction identifies the compact picture with parity- $ε$ functions on $K$ , and the right $K$ -action is translation (The compact picture of the SL2(R) principal series).

[F2]

The characters $e_{n} ([x]) = e^{2 π i n x}$ form an orthonormal basis of $L^{2} (R / Z)$ , and their symmetric Fourier sums converge in mean square (Fourier coefficients and trigonometric polynomials on the torus, The trigonometric system is complete in $L^{2}$ of the torus, Fourier series converge in mean square).

[F3]

The angle map $k_{θ} \mapsto [θ / (2 π)]$ identifies normalized Haar measure on $K$ with normalized Haar measure on the one-dimensional torus (The one-dimensional torus and its normalized Haar integral).

[F4]

A finite-dimensional subspace of a normed space is closed, including the zero subspace (A finite-dimensional normed subspace is closed).

[A1]

AC supplies normalized Haar measure and implies the AC $_{ω}$ hypothesis of the Fourier suppliers; this implication is the stated The Axiom of Countable Choice ( ${A C}_{ω}$ ) consequence of The Axiom of Choice. The finite parity and cyclic-average calculations make no further choice.

Proof

technique · direct

1.1F1F3A1algebra

Since $k_{θ + π} = - k_{θ}$ , $f_{n} (k_{θ + π}) = e^{i n π} f_{n} (k_{θ}) = (- 1)^{n} f_{n} (k_{θ})$ . This equals $(- 1)^{ε} f_{n} (k_{θ})$ exactly when $n \equiv ε (m o d 2)$ , proving both directions of the parity criterion. By [F3], for matching parities the inner product is $⟨ f_{n}, f_{m} ⟩ = \frac{1}{2 π} \int_{0}^{2 π} e^{i (n - m) θ} d θ = {\begin{cases} 1, & n = m, \\ 0, & n \neq m . \end{cases}$

2.1A1F2F3step 1.1algebra

The full Fourier basis in [F2], transported by [F3], is ${f_{n} : n \in Z}$ . If $f \in L_{ε}^{2} (K)$ , translation invariance of the coefficient integral by $1 / 2$ in the torus coordinate gives $\hat{f} (n) = (- 1)^{ε - n} \hat{f} (n)$ . Thus $\hat{f} (n) = 0$ when $n \equiv̸ ε (m o d 2)$ . The mean-square Fourier expansion of $f$ therefore uses only matching parity indices, proving the asserted orthonormal basis of $L_{ε}^{2} (K)$ .

2.2F1step 1.1algebra

For $k_{ϕ} \in K$ , right translation gives $(k_{ϕ} \cdot f_{n}) (k_{θ}) = f_{n} (k_{θ + ϕ}) = e^{i n ϕ} f_{n} (k_{θ})$ . The characters $ϕ \mapsto e^{i n ϕ}$ are distinct for distinct integers $n$ , so these are pairwise inequivalent one-dimensional $K$ -types.

3.1A1F1F2step 2.1algebra

Let $W \subseteq L_{ε}^{2} (K)$ be closed and $K$ -invariant, and let $f \in W$ . For $J \geq 0$ , set $s_{J} = \sum_{∣ m ∣ \leq J} \hat{f} (m) f_{m}$ , so $s_{J} \to f$ in $L^{2}$ by [F2]. For an integer $n$ , $J \geq ∣ n ∣$ , and $L > J + ∣ n ∣$ , define $Q_{n, L} h = \frac{1}{L} \sum_{j = 0}^{L - 1} e^{- 2 π i n j / L} (k_{2 π j / L} \cdot h)$ . Each $Q_{n, L} f$ belongs to $W$ , and $∥ Q_{n, L} ∥ \leq 1$ because it is an average of unitary operators with coefficients of modulus one. On $s_{J}$ , the finite geometric sum $\frac{1}{L} \sum_{j = 0}^{L - 1} e^{2 π i (m - n) j / L}$ is $1$ for $m = n$ and $0$ for all other $∣ m ∣ \leq J$ , since then $0 < ∣ m - n ∣ < L$ . Hence $Q_{n, L} s_{J} = \hat{f} (n) f_{n}$ . Taking $J \to \infty$ with $L = J + ∣ n ∣ + 1$ gives $Q_{n, L} f \to \hat{f} (n) f_{n}$ ; closedness of $W$ implies $f_{n} \in W$ whenever $\hat{f} (n) \neq 0$ . By the Fourier expansion from step 2.1, every $f \in W$ is the $L^{2}$ limit of finite sums of modes it contains. Therefore $W$ is their closed span.

4.1A1F2F4step 1.1step 2.1step 2.2step 3.1algebra∎

A finite sum of the $f_{n}$ has a finite-dimensional right $K$ -orbit span. Conversely, if $f$ is $K$ -finite in the local sense stated above, its orbit span $V$ is $K$ -invariant and closed by [F4]. Step 3.1 makes $V$ the closed span of the modes it contains. Since distinct $f_{n}$ are linearly independent by step 1.1, only finitely many can lie in $V$ ; hence $f$ is a finite sum of them. Thus the $K$ -finite vectors are exactly the algebraic direct sum of the parity-matching lines. By step 2.1 their $K$ -types each have multiplicity one, and that direct sum is dense in $L_{ε}^{2} (K)$ .

LemmaStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Derived action and raising/lowering formulas in the compact picture

Statement

Assume the Axiom of Choice (The Axiom of Choice). Use the compact-picture conventions of Iwasawa and minimal-parabolic data for SL2(R) and K-type decomposition of the SL2(R) principal series: $k_{θ} = (\begin{array}{cc} \cos θ & \sin θ \\ - \sin θ & \cos θ \end{array})$ and $f_{n} (k_{θ}) = e^{i n θ}$ . Put $J = (\begin{array}{cc} 0 & 1 \\ - 1 & 0 \end{array}), H = (\begin{array}{cc} 1 & 0 \\ 0 & - 1 \end{array}), S = (\begin{array}{cc} 0 & 1 \\ 1 & 0 \end{array}), W = - i J, E_{+} = \frac{1}{2} (H + i S), E_{-} = \frac{1}{2} (H - i S) .$ Thus $J, H, S$ are real Lie-algebra elements, $k_{θ} = \exp (θ J)$ , and $W, E_{+}, E_{-}$ are the explicitly normalized basis of ${s l}_{2} (C)$ with $[W, E_{\pm}] = \pm 2 E_{\pm}$ and $[E_{+}, E_{-}] = W$ . For a smooth compact-picture vector define $L_{X} f = \frac{d}{d t} ∣_{t = 0} Π_{ν} (\exp (t X)) f (X \in {s l}_{2} (R)),$ and extend this map complex-linearly to ${s l}_{2} (C)$ ; in particular $L_{W} = - i L_{J}$ and $L_{E_{\pm}} = (L_{H} \pm i L_{S}) / 2$ . No action of the real group on $\exp (t X)$ for nonreal $X$ is asserted.

These operators preserve smooth vectors and the $K$ -finite vectors of $I_{ε, ν}$ , satisfy $[L_{W}, L_{E_{\pm}}] = \pm 2 L_{E_{\pm}}, [L_{E_{+}}, L_{E_{-}}] = L_{W},$ and act on the $K$ -type basis by $L_{W} f_{n} = n f_{n}, L_{E_{\pm}} f_{n} = \frac{1 + ν \pm n}{2} f_{n \pm 2} .$ Consequently $L_{E_{-}} f_{n} = 0$ exactly when $ν = n - 1$ , and $L_{E_{+}} f_{n} = 0$ exactly when $ν = - (n + 1)$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in C$ , and a smooth vector in the compact picture of $I_{ε, ν}$ .

[F1]

Restriction to $K$ identifies the induced model with parity- $ε$ smooth functions and gives the cocycle action for right translation (The compact picture of the SL2(R) principal series).

[F2]

The Iwasawa theorem gives unique smooth KAN and NAK coordinates; the relation $a_{t} n_{x} = n_{e^{t} x} a_{t}$ converts the latter by a smooth coordinate change into unique smooth ANK coordinates (Iwasawa decomposition and Haar integration formula for SL2(R), Iwasawa and minimal-parabolic data for SL2(R)).

[F3]

The covariance factor in that action is $∣ α ∣^{1 + ν}$ , with the normalized half-modular character applied exactly once (The normalized principal series I(epsilon, nu)).

[F4]

The $K$ -finite vectors are the finite sums of $f_{n} (k_{θ}) = e^{i n θ}$ with $n \equiv ε (m o d 2)$ (K-type decomposition of the SL2(R) principal series).

[F5]

The bracket on the traceless matrix Lie algebra is the matrix commutator (The special linear Lie algebra sl_2); direct multiplication of the displayed matrices gives $[W, E_{\pm}] = \pm 2 E_{\pm}$ and $[E_{+}, E_{-}] = W$ .

[F6]

Real one-parameter subgroups are exponentials with initial velocity $X$ ; the real product and chain rules apply to the smooth matrix coordinates (One-parameter subgroup of a Lie group, One-parameter subgroups are exactly exponentials, Exponential map of a Lie group, Sums, scalar multiples, products and quotients: $(f + g)^{'} (c) = f^{'} (c) + g^{'} (c)$ , $(α f)^{'} (c) = α f^{'} (c)$ , $(f g)^{'} (c) = f^{'} (c) g (c) + f (c) g^{'} (c)$ , and $(f / g)^{'} (c) = (f^{'} (c) g (c) - f (c) g^{'} (c)) / g (c)^{2}$ when $g (c) \neq 0$ , The chain rule, in one line from Carathéodory: if $g$ is differentiable at $c$ and $f$ is differentiable at $g (c)$ , then $f \circ g$ is differentiable at $c$ with $(f \circ g)^{'} (c) = f^{'} (g (c)) g^{'} (c)$ ).

[F7]

For positive $r$ , the scalar factor $r^{1 + ν}$ means $\exp ((1 + ν) \log r)$ ; the complex exponential has derivative itself, so its derivative along a real smooth curve follows by applying the real chain rule to real and imaginary parts (The complex exponential is entire and its complex derivative is itself, The chain rule, in one line from Carathéodory: if $g$ is differentiable at $c$ and $f$ is differentiable at $g (c)$ , then $f \circ g$ is differentiable at $c$ with $(f \circ g)^{'} (c) = f^{'} (g (c)) g^{'} (c)$ ).

[F9]

The real logarithm is differentiable on $(0, \infty)$ with derivative $1 / r$ (The natural logarithm has derivative 1/x and equals the integral from 1 to x of 1/t).

[F8]

The circle matrices satisfy $k_{θ} k_{ϕ} = k_{θ + ϕ}$ , and $\frac{d}{d θ} k_{θ} ∣_{θ = 0} = J$ , by the sine/cosine addition and derivative formulas (The addition formulas for sine and cosine, The derivatives of sine and cosine are cosine and minus sine).

[A1]

AC supplies the normalized Haar inputs of the compact-picture and K-type suppliers; it implies AC $_{ω}$ for the Fourier and one-parameter-subgroup/exponential suppliers (The Axiom of Countable Choice ( ${A C}_{ω}$ )). No vector is selected when defining $L_{X}$ (The Axiom of Choice, Iwasawa and minimal-parabolic data for SL2(R)).

Proof

technique · differentiate the actual right-translation cocycle along real matrix directions

1.1F1F2F6F8A1algebra

By [F8], $θ \mapsto k_{θ}$ is a one-parameter subgroup with initial velocity $J$ ; uniqueness in [F6] gives $k_{θ} = \exp (θ J)$ . Fix $θ$ and $X \in {J, H, S}$ , and write $k_{θ} \exp (t X) = a_{u (t)} n_{x (t)} k_{ϕ (t)}$ in the unique smooth $A N K$ coordinates near $t = 0$ , with $u (0) = x (0) = 0$ and $ϕ (0) = θ$ . If the bottom row is $v (t)$ , then $v (t) = e^{- u (t) / 2} (- \sin ϕ (t), \cos ϕ (t))$ , so $∣ α ∣ = ∥ v ∥^{- 1}$ , $(\log ∣ α ∣)^{'} = - v \cdot v^{'} / ∥ v ∥^{2}$ , and $ϕ^{'} = \det (v, v^{'}) / ∥ v ∥^{2}$ . By [F6], $\frac{d}{d t} \exp (t X) ∣_{t = 0} = X$ , hence $v (0) = (- \sin θ, \cos θ)$ and $v^{'} (0) = (- \sin θ, \cos θ) X$ . For $X = H$ , this gives $v \cdot v^{'} = - \cos 2 θ$ and $\det (v, v^{'}) = \sin 2 θ$ ; for $X = S$ , it gives $v \cdot v^{'} = - \sin 2 θ$ and $\det (v, v^{'}) = - \cos 2 θ$ ; for $X = J$ , it gives $v \cdot v^{'} = 0$ and $\det (v, v^{'}) = 1$ . Thus the pairs $((\log ∣ α ∣)^{'}, ϕ^{'})$ for $J, H, S$ are respectively $(0, 1)$ , $(\cos 2 θ, \sin 2 θ)$ , and $(\sin 2 θ, - \cos 2 θ)$ . The positive diagonal fixes the local $M$ factor as $I$ .

2.1F1F3F6F7F9step 1.1

Differentiating the compact-picture factor $∣ α ∣^{1 + ν} f (k_{ϕ})$ using step 1.1 gives $L_{J} = \partial_{θ}$ , $L_{H} = (1 + ν) \cos 2 θ + \sin 2 θ \partial_{θ}$ , and $L_{S} = (1 + ν) \sin 2 θ - \cos 2 θ \partial_{θ}$ . Indeed, [F7] and [F9] give $\frac{d}{d t} ∣ α ∣^{1 + ν} = (1 + ν) ∣ α ∣^{1 + ν} (\log ∣ α ∣)^{'}$ ; at $t = 0$ the factor is $1$ . The compact-coordinate derivative contributes $ϕ^{'} (0) \partial_{θ}$ .

3.1F4step 2.1algebra

Complex-linear extension gives $L_{W} = - i D$ and $L_{E_{\pm}} = \frac{1}{2} e^{\pm 2 i θ} ((1 + ν) \mp i D)$ , where $D = \partial_{θ}$ . Applying these to $f_{n} = e^{i n θ}$ yields $L_{W} f_{n} = n f_{n}$ and $L_{E_{\pm}} f_{n} = \frac{1 + ν \pm n}{2} f_{n \pm 2}$ ; the coefficients preserve the parity lattice and shift each Fourier mode by one allowed $K$ -type.

4.1F4F5step 3.1algebra∎

The displayed operators are differential operators with smooth periodic coefficients, so they preserve smooth vectors, and step 3.1 shows they preserve finite Fourier sums. Using $[D, e^{\pm 2 i θ}] = \pm 2 i e^{\pm 2 i θ}$ gives $[L_{W}, L_{E_{\pm}}] = \pm 2 L_{E_{\pm}}$ . Writing $a = 1 + ν$ and $P_{\pm} = L_{E_{\pm}}$ , the product rule gives $P_{+} P_{-} = \frac{1}{4} ((a - 2 - i D) (a + i D)) = \frac{1}{4} (a (a - 2) + D^{2} - 2 i D)$ and $P_{-} P_{+} = \frac{1}{4} ((a - 2 + i D) (a - i D)) = \frac{1}{4} (a (a - 2) + D^{2} + 2 i D)$ , hence $[L_{E_{+}}, L_{E_{-}}] = - i D = L_{W}$ on every smooth vector. Finally, the coefficient of $f_{n - 2}$ in $L_{E_{-}} f_{n}$ vanishes iff $1 + ν - n = 0$ , i.e. $ν = n - 1$ , and the coefficient of $f_{n + 2}$ in $L_{E_{+}} f_{n}$ vanishes iff $1 + ν + n = 0$ , i.e. $ν = - (n + 1)$ .

Remarks

Kerr's formulas (2.5)–(2.6) use the same matrices $W, E_{\pm}$ and parameter normalization as this item, so the ladder coefficients and vanishing loci match directly. Kerr leaves the coordinate derivation as an exercise; this item supplies it from the bottom row of $k_{θ} \exp (t X)$ . Kowalski's Lemma 7.4.9, printed pp. 298–299, computes one real derived direction and leaves the other two to the reader; the local calculation above verifies all three. Etingof's formulas (4)–(5) use abstract $e, f, h$ and a separately normalized weight basis on $P^{\pm} (s)$ . In the convention $h = W$ , $e = E_{+}$ , $f = E_{-}$ , the local ladder coefficients give $f e f_{n} = (ν^{2} - (n + 1)^{2}) f_{n} / 4$ , so $f e + (h + 1)^{2} / 4$ acts by $ν^{2} / 4$ . This matches Etingof's Casimir scalar $s^{2} / 4$ when $s = \pm ν$ ; it is a central-character check, not a coefficient-by-coefficient identification. The displayed formulas above are derived locally.

LemmaStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

K-finite vectors detect nonzero closed invariant subspaces

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ and $ν \in C$ , and set $H = L_{ε}^{2} (K)$ with normalized Haar measure. The smooth compact-picture action of The compact picture of the SL2(R) principal series extends uniquely to a strongly continuous representation on $H$ ; unitarity is not asserted when $ν \notin i R$ . With this action:

the $K$ -finite vectors of $H$ are smooth for the action and are stable under the derived action, so $I_{ε, ν}^{K}$ is a $(g, K)$ -submodule of the smooth vectors;
every nonzero closed subspace $W \subseteq H$ invariant under the $K$ -action contains a nonzero $K$ -finite vector: for $n \equiv ε (m o d 2)$ , let $P_{n}$ be the isotypic projection onto the $K$ -type $C f_{n}$ of K-type decomposition of the SL2(R) principal series. Then $P_{n} (W) \subseteq W$ for every such $n$ , and $v = \sum_{n \equiv ε (2)} P_{n} v$ for every $v \in H$ , so $v \neq 0$ forces $P_{n} v \neq 0$ for some $n$ ;
consequently, for every closed $G$ -invariant subspace $W$ , the space $W \cap I_{ε, ν}^{K}$ is a $(g, K)$ -submodule of $I_{ε, ν}^{K}$ , and $W \neq {0}$ if and only if $W \cap I_{ε, ν}^{K} \neq {0}$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in C$ , the compact-picture Hilbert space $H = L_{ε}^{2} (K)$ , and a closed subspace $W \subseteq H$ when specified.

[F1]

The compact-picture action is $(Π_{ν} (g) f) (k) = ∣ α (p (k, g)) ∣^{1 + ν} f (κ (k, g))$ , where $k g = p (k, g) κ (k, g)$ is the unique $A N \times K$ factorization; both the cocycle and its coordinates are smooth in $(k, g)$ . Its restriction to $K$ is right translation. The normalized inducing character fixes the positive-base complex-power convention and gives ${∣ ∣ α ∣^{1 + ν} ∣}^{2} = ∣ α ∣^{2 + 2 Re ν}$ (The compact picture of the SL2(R) principal series, The normalized principal series I(epsilon, nu)).

[F2]

The unique Iwasawa coordinates are smooth (Iwasawa decomposition and Haar integration formula for SL2(R)); the subgroup $K = {k_{θ} : θ \in R / 2 π Z}$ is the compact circle (Iwasawa and minimal-parabolic data for SL2(R)).

[F3]

The vectors $f_{n} (k_{θ}) = e^{i n θ}$ with $n \equiv ε (m o d 2)$ form an orthonormal basis of $H$ , have $K$ -character $e^{i n ϕ}$ , and their finite linear combinations are exactly the $K$ -finite vectors (K-type decomposition of the SL2(R) principal series).

[F4]

The derived action has $L_{W} f_{n} = n f_{n}$ and $L_{E_{\pm}} f_{n} = (1 + ν \pm n) f_{n \pm 2} / 2$ , preserving finite Fourier sums (Derived action and raising/lowering formulas in the compact picture).

[F5]

For a strongly continuous unitary $K$ -representation and one-dimensional character $σ_{n} (k_{ϕ}) = e^{i n ϕ}$ , the compact-group definition constructs the bounded Bochner averaging map $P_{n} v = \int_{K} \overline{σ_{n} (k)} Π (k) v d k$ ; the integral is a norm limit of finite linear combinations of its range values (Compact-group isotypic projection).

[F6]

Under $k_{θ} \leftrightarrow [θ / (2 π)] \in R / Z$ , normalized Haar measure is $d k = d θ / (2 π)$ : the torus integral is normalized translation-invariant Lebesgue measure, and its pushforward is the unique normalized Haar measure on compact $K$ (Iwasawa and minimal-parabolic data for SL2(R), The one-dimensional torus and its normalized Haar integral, Normalized Haar measure on a compact Lie group).

[F7]

An orientation-preserving diffeomorphism of the compact oriented circle changes a top-form integral by its positive angular Jacobian (Change of variables on oriented manifolds).

[F8]

A continuous real-valued function on compact metric $K$ is bounded and attains its maximum (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value).

[A1]

AC supplies normalized Haar measure on $K$ for the compact-picture and Fourier arguments. AC implies AC $_{ω}$ through The Axiom of Countable Choice ( ${A C}_{ω}$ ), used by the Fourier approximation in [F3], the Bochner-integral construction in [F5], and the one-parameter/exponential input in [F4]. The witness $v \in W ∖ {0}$ in step 6.1 follows from the stated nonzero hypothesis and uses no choice axiom (The Axiom of Choice).

Proof

technique · extend the smooth cocycle by a Jacobian bound, then use its compact-group Fourier averages and closedness

1.1A1F1F2F6F7algebra

Fix $g \in G$ and write $k_{θ} g = a_{t (θ)} n_{x (θ)} k_{ψ (θ)}$ in the unique smooth $A N K$ coordinates. In a local lift of the angle $ψ$ , the bottom row is $v (θ) = e^{- t (θ) / 2} (- \sin ψ (θ), \cos ψ (θ))$ . It is also $(- \sin θ, \cos θ) g$ , so $\det (v, v^{'}) = \det ⁣ (\begin{array}{cc} - \sin θ & \cos θ \\ - \cos θ & - \sin θ \end{array}) \det g = 1$ . The first expression gives $\det (v, v^{'}) = e^{- t} ψ^{'}$ , hence $ψ^{'} = e^{t} = ∣ α (p (k_{θ}, g)) ∣^{2} > 0$ . Thus $κ (\cdot, g)$ is an orientation-preserving circle diffeomorphism; uniqueness of the $A N K$ factorization gives inverse $κ (\cdot, g^{- 1})$ . By [F6], $d k = d θ / (2 π)$ , and the compactly supported top-form change of variables [F7] applies because $K$ is compact; it gives Haar Jacobian $d k^{'} = ∣ α (p (k, g)) ∣^{2} d k$ .

2.1F1F3F8step 1.1algebra

For smooth $f$ , [F1] and step 1.1 give $∥ Π_{ν} (g) f ∥_{2}^{2} = \int_{K} ∣ α (p (k, g)) ∣^{2 + 2 Re ν} ∣ f (κ (k, g)) ∣^{2} d k = \int_{K} ∣ α (p (κ^{- 1} (k^{'}), g)) ∣^{2 Re ν} ∣ f (k^{'}) ∣^{2} d k^{'} \leq C_{g} ∥ f ∥_{2}^{2}$ , where $C_{g} : = \max_{k \in K} ∣ α (p (k, g)) ∣^{2 Re ν} < \infty$ by [F8]; this is the same supremum because $κ (\cdot, g)$ is a diffeomorphism. Thus each smooth operator extends uniquely to a bounded operator on $H$ , since smooth Fourier sums are dense by [F3].

3.1F1F2F3step 2.1algebra

The bounded extensions satisfy the group law because the original smooth action does and smooth Fourier sums are dense by [F3]. The coefficient $∣ α (p (k, g)) ∣^{2 Re ν}$ is jointly continuous in $(k, g)$ ; compactness of $K$ and a finite subcover near any fixed $g_{0}$ give a local uniform bound $M$ for the operator norms. For smooth $f$ , [F1] gives a jointly smooth function $Π_{ν} (g) f (k)$ ; compactness of $K$ makes every parameter derivative continuous uniformly in $k$ , so its orbit map is smooth into $H$ . For arbitrary $h \in H$ , approximate by a smooth Fourier sum $f$ and use $∥ Π_{ν} (g) h - Π_{ν} (g_{0}) h ∥_{2} \leq M ∥ h - f ∥_{2} + ∥ Π_{ν} (g) f - Π_{ν} (g_{0}) f ∥_{2} + ∥ Π_{ν} (g_{0}) (f - h) ∥_{2}$ near $g_{0}$ ; this proves strong continuity. For $ν \notin i R$ no unitarity of the $G$ -action is asserted.

4.1F1F3F4step 3.1algebra

By [F1] and [F3], $Π_{ν} (k_{ϕ}) f_{n} = e^{i n ϕ} f_{n}$ . Since these vectors are an orthonormal basis, the $K$ -action extends to a unitary action on $H$ ; it is strongly continuous by step 3.1. Every $K$ -finite vector is a finite Fourier sum by [F3]. Joint smoothness in [F1] and compactness of $K$ show that its orbit map is $C^{\infty}$ as an $H$ -valued map. The formulas in [F4] preserve finite Fourier sums, and the $K$ -action does too; hence $I_{ε, ν}^{K}$ is a $(g, K)$ -submodule of the smooth vectors.

5.1A1F3F5step 4.1algebra

For $n \equiv ε (m o d 2)$ , let $σ_{n} (k_{ϕ}) = e^{i n ϕ}$ . By step 4.1 the representation of $K$ on $H$ is strongly continuous and unitary, so [F5] defines $P_{n}$ . For each basis vector $f_{m}$ , [F3] gives $P_{n} f_{m} = (\int_{K} e^{i (m - n) ϕ} d k) f_{m} = ⟨ f_{m}, f_{n} ⟩ f_{m} = δ_{n m} f_{m}$ . Boundedness of $P_{n}$ and the orthonormal-basis expansion in [F3] therefore give $P_{n} v = ⟨ v, f_{n} ⟩ f_{n}$ and $v = \sum_{n \equiv ε (2)} P_{n} v$ in Hilbert norm.

6.1A1F3F5step 5.1

If $W$ is $K$ -invariant and $v \in W$ , every value $\overline{σ_{n} (k)} Π (k) v$ in the integrand of [F5] belongs to $W$ . The simple-function approximants to its Bochner integral are finite linear combinations of such values, and their norm limit lies in $W$ because $W$ is closed. Thus $P_{n} v \in W$ for every allowed $n$ . If $W \neq {0}$ , take any $v \in W ∖ {0}$ ; the expansion in step 5.1 has a nonzero term, so $P_{n} v \neq 0$ for some $n$ , and this vector is $K$ -finite. For $W = {0}$ every projection is zero. This proves detection for closed $K$ -invariant subspaces.

7.1F1F4step 6.1algebra∎

If $W$ is closed and $G$ -invariant, then it is $K$ -invariant, so step 6.1 proves $W \neq {0}$ iff $W \cap I_{ε, ν}^{K} \neq {0}$ , including $W = {0}$ where both sides are false. For $w \in W \cap I_{ε, ν}^{K}$ and real $X \in g$ , the difference quotients $(Π_{ν} (\exp (t X)) w - w) / t$ lie in $W$ ; their limit $L_{X} w$ also lies in $W$ because $W$ is closed, and is $K$ -finite by [F4]. The $K$ -action preserves the intersection, so it is a $(g, K)$ -submodule.

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Generic irreducibility and the exceptional parameter lattice

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ , $ν \in C$ and $W_{ε} = {ν \in Z : ν \equiv ε + 1 (2)}$ . Write $g = {s l}_{2} (C)$ for the complexified Lie algebra acting on $I_{ε, ν}^{K}$ .

(a) Generic irreducibility. If $ν \notin W_{ε}$ , then $I_{ε, ν}^{K}$ is an algebraically irreducible $(g, K)$ -module. The compact-picture representation on $L_{ε}^{2} (K)$ is irreducible in the Hilbert-space sense, and the smooth induced representation $I_{ε, ν} ≅ C_{ε}^{\infty} (K)$ is topologically irreducible in its usual $C^{\infty}$ compact-picture topology (there is no nonzero proper closed $G$ -invariant subspace).

(b) Exceptional parameters. Let $n \in W_{ε}$ , $n \geq 1$ . Then $I_{ε, n}^{K}$ has composition length $3$ with composition factors $L_{n - 1} (finite-dimensional, K-types n - 1, n - 3, \dots, - (n - 1)),$ $M_{n + 1}^{-} (K-types n + 1, n + 3, \dots), M_{- (n + 1)}^{+} (K-types - (n + 1), - (n + 3), \dots) .$ Here $M_{n + 1}^{-}$ and $M_{- (n + 1)}^{+}$ denote the irreducible one-sided modules with these respective K-type strings and inherited ladder action. For $ν = n$ the finite-dimensional factor $L_{n - 1}$ is the unique irreducible quotient and $M_{n + 1}^{-} \oplus M_{- (n + 1)}^{+}$ is the unique maximal proper submodule; for $ν = - n$ the roles are reversed ( $L_{n - 1}$ is the unique irreducible submodule and $M_{n + 1}^{-} \oplus M_{- (n + 1)}^{+}$ is the quotient). For $ε = 1$ and $ν = 0$ one has the direct sum $I_{1, 0}^{K} ≅ M_{1}^{-} \oplus M_{- 1}^{+}$ of the two limits of discrete series.

The quadratic Casimir element of The quadratic Casimir element acts on $I_{ε, ν}^{K}$ by the scalar $\frac{1}{8} (ν^{2} - 1)$ , and at $ν = n$ on the factor $L_{n - 1}$ by $\frac{1}{8} (n^{2} - 1)$ . For ${s l}_{2}$ the center of $U (g)$ is generated by this Casimir, so the central character exists and determines $ν$ up to sign.

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in C$ , the smooth compact-picture model, and $H = L_{ε}^{2} (K)$ .

[F1]

The smooth compact-picture action is $(Π_{ν} (g) f) (k) = ∣ α (p (k, g)) ∣^{1 + ν} f (κ (k, g))$ , and its restriction to $K$ is right translation (The compact picture of the SL2(R) principal series, The normalized principal series I(epsilon, nu)).

[F2]

The modes $f_{r} (k_{θ}) = e^{i r θ}$ , $r \equiv ε (m o d 2)$ , form an orthonormal basis of $H$ ; their finite span is $I_{ε, ν}^{K}$ and is dense in $H$ (K-type decomposition of the SL2(R) principal series).

[F3]

The derived action is $L_{W} f_{r} = r f_{r}$ , $L_{E_{+}} f_{r} = (1 + ν + r) f_{r + 2} / 2$ , and $L_{E_{-}} f_{r} = (1 + ν - r) f_{r - 2} / 2$ (Derived action and raising/lowering formulas in the compact picture).

[F4]

The smooth compact-picture action extends to a strongly continuous representation on $H$ for every $ν \in C$ ; every nonzero closed $G$ -invariant subspace of $H$ contains a nonzero $K$ -finite vector, and its $K$ -finite intersection is a $(g, K)$ -submodule (K-finite vectors detect nonzero closed invariant subspaces).

[F5]

Under $k_{2 π t} \leftrightarrow [t] \in T = R / Z$ , normalized Haar measure on $K$ is the normalized torus integral on $[0, 1)$ (Iwasawa and minimal-parabolic data for SL2(R), The one-dimensional torus and its normalized Haar integral, Normalized Haar measure on a compact Lie group).

[F6]

For a one-periodic integrable $g$ , $\hat{g} (r) = \int_{0}^{1} g (t) e^{- 2 π i r t} d t$ , $S_{N} g = \sum_{∣ r ∣ \leq N} \hat{g} (r) e^{2 π i r t}$ , and $σ_{N} g = (N + 1)^{- 1} \sum_{j = 0}^{N} S_{j} g$ (Period-one Fourier coefficients, partial sums, and convolution on the torus, Cesaro and Abel means of a Fourier series).

[F7]

If $g$ is continuous and one-periodic, its Fejér means satisfy $\sup_{t} ∣ σ_{N} g (t) - g (t) ∣ \to 0$ (Fejer means converge uniformly for continuous periodic functions).

[F8]

Repeated integration by parts gives $\hat{g^{(q)}} (r) = (2 π i r)^{q} \hat{g} (r)$ for smooth periodic $g$ : apply the real formula to real and imaginary parts, whose continuous derivatives are Riemann integrable, and use equality of bounded Riemann and Lebesgue integrals (If $u, v$ are differentiable on $[a, b]$ with $u^{'}, v^{'}$ integrable, then $\int_{a}^{b} u v^{'} = u (b) v (b) - u (a) v (a) - \int_{a}^{b} u^{'} v$ , A continuous function on $[a, b]$ is Riemann integrable, by Heine-Cantor and Riemann's criterion, A bounded Riemann integrable function on a closed bounded interval is Lebesgue measurable and has the same integral).

[F9]

The closed box $[0, 1]^{2}$ is compact in its Euclidean metric, so every continuous function on it is uniformly continuous (Heine-Borel in $R^{n}$ : with the Euclidean metric a subset of $R^{n}$ is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line, Heine-Cantor: a continuous map from a compact metric space to any metric space is uniformly continuous).

[F10]

The matrix basis $W, E_{+}, E_{-}$ from the derived-action supplier has brackets $[W, E_{\pm}] = \pm 2 E_{\pm}$ , $[E_{+}, E_{-}] = W$ (The special linear Lie algebra sl_2, Derived action and raising/lowering formulas in the compact picture).

[F11]

For semisimple $g$ , the quadratic Casimir is the sum formed from Killing-dual bases and is central; the shifted Harish-Chandra map is the algebra isomorphism ${HC}_{ρ} (z) (λ) = pr (z) (λ - ρ)$ onto $S (h)^{W}$ (The Killing form of a semisimple Lie algebra, The quadratic Casimir element, The quadratic Casimir element is central, The Harish-Chandra projection, Harish-Chandra isomorphism for the center).

[F12]

Every finite-dimensional simple $g$ -module is the unique simple highest-weight module for its dominant integral highest weight (Finite-dimensional simple modules are classified by dominant highest weights).

[F13]

A central character is a unital algebra homomorphism $χ : Z (U (g)) \to C$ such that each central element acts by its scalar value under $χ$ (Central character of a Lie algebra module).

[F14]

A Cartan subalgebra is nilpotent and self-normalizing, its roots and root spaces are the nonzero eigenspaces of the adjoint action and form a reduced crystallographic root system, each root has its Killing-dual vector, and a root reflection acts by $s_{α} (λ) = λ - λ (α^{\lor}) α$ (Cartan subalgebra, Root and root space, The root set is a reduced crystallographic root system, The Killing-dual vector attached to a root, Root reflections and the Weyl group action).

[F15]

For a finite-dimensional Lie algebra, $B (X, Y) = tr ({ad}_{X} {ad}_{Y})$ ; over a characteristic-zero field, the algebra is semisimple exactly when this Killing form is nondegenerate (Killing form, Cartan's semisimplicity criterion).

[F16]

For the chosen rank-one simple root, the fundamental weight satisfies $ω (α^{\lor}) = 1$ (Fundamental weights for a chosen simple root system).

[A1]

AC supplies normalized Haar probability on $K$ , is the premise of the $K$ -finite detection result [F4], the root-system and Killing-dual-vector inputs [F14], and the Harish-Chandra isomorphism [F11], and implies the AC $_{ω}$ hypothesis used for the period-one Fourier coefficients and Fejér means [F6]–[F7]. No vector or weight is selected by an additional choice principle (The Axiom of Countable Choice ( ${A C}_{ω}$ ), The Axiom of Choice).

Proof

technique · establish the rank-one Lie data, connect the K-weight graph, identify exceptional boundary cuts, and compute the central scalar

1.1A1F10F14F15F16algebra

In the basis of [F10], ${ad}_{W} = diag (0, 2, - 2)$ and ${ad}_{E_{+}} {ad}_{E_{-}}$ has diagonal entries $2, 2, 0$ . The reversed composition has diagonal entries $2, 0, 2$ by the same brackets; each product with one $W$ and one $E_{\pm}$ , or with two equal $E_{\pm}$ , is off-diagonal. Thus the Killing matrix is $(\begin{array}{ccc} 8 & 0 & 0 \\ 0 & 0 & 4 \\ 0 & 4 & 0 \end{array})$ with determinant $- 128$ , so $g$ is semisimple by [F15]. Put $h = C W$ . It is abelian, hence nilpotent; for $X = a W + b E_{+} + c E_{-}$ the brackets in [F10] give $[X, W] = - 2 b E_{+} + 2 c E_{-}$ , so $N_{g} (h) = h$ and [F14] makes $h$ a Cartan subalgebra. Its root spaces are $C E_{\pm}$ with roots $\pm α$ , where $α (W) = 2$ ; by [F14] they form the reduced root system of type $A_{1}$ . Choose $α$ positive. Since $B (W, W) = 8$ , the Killing-dual vector is $H_{α} = W / 4$ and $α (H_{α}) = 1 / 2$ , so $α^{\lor} = W$ . The fundamental weight has $ω (W) = 1$ by [F16], hence $α = 2 ω$ and $ρ = α / 2 = ω$ . The root reflection in [F14] sends $t = λ (W)$ to $- t$ .

1.2F1F2algebra

Let $M$ be an algebraic $(g, K)$ -submodule and let $0 \neq v = \sum_{r \in S} a_{r} f_{r} \in M$ , where $S$ is its finite support and every $a_{r} \neq 0$ . If $S$ has more than one element, put $q = 1 + \sum_{{r, s} \subset S, r \neq s} ∣ r - s ∣$ ; then $q > ∣ r - s ∣$ for all distinct $r, s \in S$ , so the eigenvalues $e^{2 π i r / q}$ of $k_{2 π / q}$ on these modes are distinct. Lagrange interpolation in $Π (k_{2 π / q})$ extracts each $a_{r} f_{r}$ as a finite linear combination of $K$ -translates of $v$ , hence $f_{r} \in M$ ; for singleton support this is immediate. Thus every nonzero algebraic submodule contains a K-type.

1.3A1F5F6F7F8algebra

Write $g (t) = f (k_{2 π t})$ for a smooth parity- $ε$ function. Then $g (t + \frac{1}{2}) = (- 1)^{ε} g (t)$ , so $\hat{g} (r) = 0$ unless $r \equiv ε (m o d 2)$ . By [F8], $\hat{g^{(q)}} (r) = (2 π i r)^{q} \hat{g} (r)$ for every derivative order $q$ , since the endpoint terms vanish by periodicity. Differentiating the finite Fourier sums in $σ_{N} g = (N + 1)^{- 1} \sum_{j = 0}^{N} S_{j} g$ gives $(σ_{N} g)^{(q)} = σ_{N} (g^{(q)})$ . Each $σ_{N} g$ is a finite sum of the allowed K-types, and [F7] applied to every $g^{(q)}$ shows $σ_{N} g \to g$ in every $C^{q}$ seminorm.

2.1F3step 1.2algebra

If $ν \notin W_{ε}$ , neither coefficient in [F3] vanishes for an allowed weight $r$ : either zero would force $ν = - r - 1$ or $ν = r - 1$ , an integer congruent to $ε + 1$ modulo $2$ . Repeated raising and lowering therefore connects every allowed weight to every other. By step 1.2 every nonzero algebraic submodule contains one weight, hence all of them, proving algebraic irreducibility.

2.2F3F12F16step 1.1step 1.2algebra

Let $n \in W_{ε}$ , $n \geq 1$ , and set $ν = n$ . Put $T^{+} = ⨁_{j \geq 0} C f_{n + 1 + 2 j}$ , $T^{-} = ⨁_{j \geq 0} C f_{- (n + 1 + 2 j)}$ , and $F = ⨁_{j = 0}^{n - 1} C f_{n - 1 - 2 j}$ . The zeros $L_{E_{-}} f_{n + 1} = 0$ and $L_{E_{+}} f_{- n - 1} = 0$ make $T^{\pm}$ submodules; their internal arrows are nonzero, so each is simple by the weight-extraction argument of step 1.2. In the quotient by $T^{+} \oplus T^{-}$ , the finite chain $F$ has nonzero internal arrows, and the class of $f_{n - 1}$ is a highest-weight vector because its $E_{+}$ image lies in $T^{+}$ . Its highest weight is $(n - 1) ω$ , which is dominant integral since $n \geq 1$ and [F16]. The same interpolation as in step 1.2 extracts a weight from any nonzero submodule of this quotient; the internal arrows connect every weight of $F$ , so $F$ is simple and [F12] identifies it with $L ((n - 1) ω) = L_{n - 1}$ in the notation of the Statement. By step 1.2, any submodule not contained in the two tails has a weight in $F$ ; its internal arrows reach all of $F$ , and the arrows out of $f_{n - 1}$ and $f_{1 - n}$ have coefficient $n \neq 0$ , so it then contains both tails. Consequently $T^{+} \oplus T^{-}$ is the unique maximal proper submodule and the finite quotient is the unique irreducible quotient. The filtration $0 < T^{+} < T^{+} \oplus T^{-} < I_{ε, n}^{K}$ has three nonzero simple factors, so the composition length is $3$ . Here $T^{+} ≅ M_{n + 1}^{-}$ is the lowest-weight string and $T^{-} ≅ M_{- (n + 1)}^{+}$ the highest-weight string.

2.3F3step 1.2algebra

If $ε = 1$ and $ν = 0$ , then $L_{E_{-}} f_{1} = 0$ and $L_{E_{+}} f_{- 1} = 0$ . The positive odd chain $⨁_{j \geq 0} C f_{1 + 2 j}$ and negative odd chain $⨁_{j \geq 0} C f_{- (1 + 2 j)}$ are invariant; every internal arrow is nonzero, so each is simple by step 1.2. They have disjoint K-types and together contain every odd K-type, hence $I_{1, 0}^{K} = M_{1}^{-} \oplus M_{- 1}^{+}$ .

3.1A1F2F4step 2.1

Assume $ν \notin W_{ε}$ and let $W \neq {0}$ be a closed $G$ -invariant subspace of $H$ . By [F4], $W \cap I_{ε, ν}^{K}$ is a nonzero algebraic $(g, K)$ -submodule. Step 2.1 makes this intersection all of $I_{ε, ν}^{K}$ ; its finite Fourier sums are dense in $H$ by [F2], so closedness gives $W = H$ .

3.2A1F1F3F5F7F9step 2.1step 1.3

Assume $ν \notin W_{ε}$ , and let $V \neq {0}$ be a closed $G$ -invariant subspace of $C_{ε}^{\infty} (K)$ . Take $0 \neq f \in V$ and write $g (t) = f (k_{2 π t})$ . For an allowed $r$ , form $P_{r} g (t) = \int_{0}^{1} e^{- 2 π i r ϕ} g (t + ϕ) d ϕ$ . Riemann sums lie in $V$ by $K$ -invariance and converge in every $C^{q}$ seminorm: each $\partial_{t}^{q} (e^{- 2 π i r ϕ} g (t + ϕ))$ is uniformly continuous on the compact angle square by [F9], so the Riemann-sum error tends uniformly to zero in $t$ . Periodicity and $s = t + ϕ$ give $P_{r} g (t) = \hat{g} (r) e^{2 π i r t}$ , hence $P_{r} g = \hat{g} (r) f_{r} \in V$ . Some allowed $\hat{g} (r)$ is nonzero, since otherwise every Fejér mean of $g$ would vanish and [F7] would force $g = 0$ . Thus $V$ contains a K-type. For each real Lie algebra element, joint smoothness of the compact-picture cocycle implies that the derived difference quotients converge together with every angular derivative, hence in $C^{\infty}$ ; closedness puts the derived vector in $V$ , and complex-linear combinations give the raising and lowering operators in [F3]. Step 2.1's connected weight graph therefore puts every K-type in $V$ , and step 1.3 plus closedness yields $V = C_{ε}^{\infty} (K)$ .

3.3F3F12F16step 1.1step 1.2step 2.2algebra

For $ν = - n$ , the same finite chain $F$ is a submodule: its outward boundary arrows vanish, and its internal arrows are nonzero. Its highest-weight vector is $f_{n - 1}$ , killed by $E_{+}$ , with highest weight $(n - 1) ω$ , which is dominant integral since $n \geq 1$ and [F16]. By the same interpolation and internal-arrow argument as in step 2.2, it is simple, so [F12] gives $F ≅ L ((n - 1) ω) = L_{n - 1}$ . Modulo $F$ , the positive and negative tails are separate submodules, because the crossing arrows land in $F$ and vanish in the quotient. Each has one-dimensional weight spaces and, by [F3], nonzero arrows in both directions between every adjacent pair of tail weights; only the inward boundary arrow vanishes in the quotient. The same interpolation as in step 1.2 extracts a weight from any nonzero submodule, and these internal raising and lowering arrows generate the whole tail, so both are simple. In either parameter, a one-sided string with fixed lowest (or highest) weight is unique up to rescaling its successive weight vectors: normalize each nonzero outward arrow to $1$ , then $[E_{+}, E_{-}] = W$ recursively fixes the inward arrows from the boundary condition. Thus the quotient factors are again $M_{n + 1}^{-}$ and $M_{- (n + 1)}^{+}$ . Any irreducible submodule not contained in $F$ has a tail weight by step 1.2; repeated nonzero inward arrows first reach a tail boundary, whose inward arrow at $ν = - n$ is nonzero into $F$ . Its intersection with the simple submodule $F$ is then all of $F$ , forcing the irreducible submodule to equal $F$ . Hence $F$ is the unique irreducible submodule. If ${\tilde{T}}^{+}$ is the preimage of the positive quotient tail, the filtration $0 < F < {\tilde{T}}^{+} < I_{ε, - n}^{K}$ has three nonzero simple factors, so the composition length is $3$ .

3.4F3F11step 1.1step 2.2algebra

By step 1.1, the Killing-dual basis gives $C = \frac{1}{8} W^{2} + \frac{1}{4} (E_{+} E_{-} + E_{-} E_{+}) = \frac{1}{8} W^{2} - \frac{1}{4} W + \frac{1}{2} E_{+} E_{-}$ using [F11]. Formula [F3] yields $L_{E_{+}} L_{E_{-}} f_{r} = \frac{1}{4} (ν^{2} - (r - 1)^{2}) f_{r}$ , and substituting $L_{W} f_{r} = r f_{r}$ gives $C f_{r} = \frac{1}{8} (ν^{2} - 1) f_{r}$ for every weight. Therefore $C$ acts by this scalar on $I_{ε, ν}^{K}$ and on its subquotient $L_{n - 1}$ at $ν = n$ .

4.1A1F11F13F14step 1.1step 3.4algebra∎

Use the rank-one Cartan, root, coroot, positive-system and Weyl data from step 1.1. In PBW order $E_{-} < W < E_{+}$ , $E_{+} E_{-} = E_{-} E_{+} + W$ , so the Harish-Chandra projection of the Casimir is $pr (C) = W^{2} / 8 + W / 4$ ; hence the shifted polynomial is ${HC}_{ρ} (C) (t) = (t^{2} - 1) / 8$ . This generates $S (h)^{W} = C [t^{2}]$ , so [F11] implies $Z (U (g)) = C [C]$ . Since $C$ acts by $λ_{ν} = (ν^{2} - 1) / 8$ , every $p (C)$ acts by $p (λ_{ν})$ and defines the central character of [F13]; two such characters are equal exactly when $ν^{2} = (ν^{'})^{2}$ , that is, when $ν^{'} = \pm ν$ .

Remarks

Kerr's §2 formula (2.6), Examples 2.6–2.7 and classification paragraph (printed pp. 10–12) cross-check the ladder coefficients, the odd $ν = 0$ splitting, and the orientation of the finite and one-sided factors. Etingof's §9.1 formulas (4)–(5) and short exact sequences (printed pp. 48–49) cross-check the generic lattice and factor strings after the parameter dictionary $s = - ν$ ; its basis normalization is separate, so it is not used for the local arrow coefficients. Kowalski's §7.4 Proposition 7.4.3(2) (statement printed p. 294, proof pp. 297–301) is only a unitary-character cross-check and does not establish the complex-parameter claim. All irreducibility and Casimir arguments above are proved locally.

DefinitionDefinition: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

The standard intertwining operator A(nu)

Definition

Assume AC and let $ε \in {0, 1}$ and $ν \in C$ with $Re ν > 0$ . Use the smooth model of The normalized principal series I(epsilon, nu), the compact picture of The compact picture of the SL2(R) principal series, and the K-type basis of K-type decomposition of the SL2(R) principal series. Put $w = (\begin{array}{cc} 0 & - 1 \\ 1 & 0 \end{array}) = k_{- π / 2} \in K .$ The standard intertwining integral is $(A (ν) φ) (g) = \int_{R} φ (w n_{u} g) d u,$ where $d u$ is Lebesgue measure in the $N$ coordinate. For positive real bases in complex powers use $x^{z} : = \exp (z \log x)$ with the real logarithm.

The proof below shows that the integral is absolutely convergent for every smooth $φ$ and $g \in G$ , and defines a linear operator from the smooth model $I_{ε, ν}$ to $I_{ε, - ν}$ . It commutes with right translation, maps K-finite vectors to K-finite vectors, and in the compact picture is diagonal on the parity- $ε$ K-types: $A (ν) f_{n} = c_{n} (ν) f_{n} (n \equiv ε (m o d 2)) .$ For $ε = 0$ , the base K-type is $f_{0} = 1$ and its eigenvalue is $c_{0} (ν) = \int_{R} (1 + u^{2})^{- (1 + ν) / 2} d u,$ which the proof identifies with $B (1 / 2, ν / 2)$ ; it is positive for real $ν > 0$ . For $ε = 1$ , $f_{0}$ is not a K-type, and $c_{0} (ν)$ denotes this same scalar function, not an eigenvalue. Its meromorphic continuation is established by K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing ↗. The full family of continuous K-diagonal maps on $C_{ε}^{\infty} (K)$ has the meromorphic continuation proved in Meromorphic continuation and intertwining identity for A(nu) ↗: a common local scalar factor clears its poles, with holomorphy in every smooth seminorm. The defining integral itself is only asserted on $Re ν > 0$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $Re ν > 0$ , and a smooth left- $P$ -covariant function $φ \in I_{ε, ν}$ .

[F1]

The model has covariance $φ (p g) = χ_{ε, ν} (p) φ (g)$ , with $χ_{ε, ν} (m a_{s} n_{x}) = σ_{ε} (m) e^{(1 + ν) s / 2}$ for $m = \pm I$ , and right action $(Π_{ν} (g_{0}) φ) (g) = φ (g g_{0})$ (Iwasawa and minimal-parabolic data for SL2(R), The normalized principal series I(epsilon, nu)).

[F2]

If $y = p k$ with $p = m a_{s} n_{x}$ and $k \in K$ , then $φ (y) = e^{(1 + ν) s / 2} σ_{ε} (m) φ (k)$ ; the Euclidean norm $r$ of the bottom row of $y$ is $e^{- s / 2}$ , so $∣ φ (y) ∣ \leq ∥ φ ∣_{K} ∥_{\infty} r^{- 1 - Re ν}$ by $∣ e^{z} ∣ = e^{Re z}$ (The compact picture of the SL2(R) principal series, $\exp (x + i y) = e^{x} (\cos y + i \sin y)$ , $∣ \exp (x + i y) ∣ = e^{x}$ , and $e^{i π} + 1 = 0$ ).

[F3]

In the compact picture, the parity-matching functions $f_{n} (k_{θ}) = e^{i n θ}$ are precisely the one-dimensional K-types (K-type decomposition of the SL2(R) principal series).

[F4]

The nonnegative integral is monotone and homogeneous; nonnegative improper Riemann integrals on a half-line agree with their Lebesgue integrals; positive-base real powers have the stated derivatives (Monotonicity and nonnegative homogeneity of the nonnegative integral, A nonnegative improper Riemann integral on a half-line agrees with the Lebesgue integral, Continuity and derivatives of positive-base real powers).

[F5]

A C1 diffeomorphism obeys change of variables for nonnegative measurable functions and for L1 complex functions. Nonnegative integrals are additive over disjoint measurable pieces, integrals on null sets vanish, and a one-dimensional box has its length as Lebesgue measure (A C^1 diffeomorphism satisfies the change-of-variables formula for nonnegative Lebesgue measurable functions, A C^1 diffeomorphism satisfies the change-of-variables formula for L^1 functions, Additivity of the nonnegative Lebesgue integral, A nonnegative integral over a null set vanishes, A box in $R^{n}$ with parameters $a_{i} \leq b_{i}$ is Lebesgue measurable of measure $\prod_{i < n} (b_{i} - a_{i})$ , whichever of its faces are included).

[F6]

Differentiation under the integral sign applies to a common integrable majorant; dominated convergence gives continuity of the resulting parameter integrals; the complex Lebesgue integral is linear on L1 (Differentiation under the integral sign, Dominated convergence, The Lebesgue integral is linear on $L^{1} (μ)$ ).

[F7]

Products in a Lie group and smooth functions between smooth manifolds are smooth, with the chain rule for coordinate derivatives (Lie group, $C^{r}$ and smooth maps between smooth manifolds, The chain rule for total derivatives: $D (g \circ f) (a) = D g (f (a)) \circ D f (a)$ ).

[F8]

A continuous real-valued function on a compact metric space is bounded (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value); continuous scalar functions on Euclidean spaces are Borel measurable (Continuous functions on Euclidean spaces are Borel measurable).

[F9]

AC supplies the normalized Haar probabilities used by the compact-picture and K-type suppliers and implies the countable-choice hypotheses of the half-line integral and Lebesgue change-of-variables results (Iwasawa and minimal-parabolic data for SL2(R), The Axiom of Choice, The Axiom of Countable Choice ( ${A C}_{ω}$ )).

[F10]

For $Re p, Re q > 0$ , the Euler Beta integral is $B (p, q) = \int_{0}^{1} t^{p - 1} (1 - t)^{q - 1} d t$ (Euler's Beta function on the right half-planes, The Beta-Gamma identity).

Verification

technique · establish a uniform compact-parameter decay estimate, then verify smoothness, covariance, and the K-type claims directly

1.1F2F4F5F8F9algebra

Let $C$ be a compact coordinate box in $G$ , write $g = (\begin{array}{cc} a & b \\ c & d \end{array})$ , and set $B = \max_{g \in C} ∥ (a, b) ∥$ and $B_{1} = \max_{g \in C} ∥ (c, d) ∥$ ; these are finite and positive by [F8] and $\det g = 1$ . For any two real rows $x, y$ , direct expansion gives $∥ x ∥^{2} ∥ y ∥^{2} - \det (x, y)^{2} = (x \cdot y)^{2} \geq 0$ . Thus the rows of $g$ give $1 \leq ∥ (a, b) ∥ ∥ (c, d) ∥$ and hence $∥ (c, d) ∥ \geq B^{- 1}$ . The rows of $w n_{u} g$ are $(- c, - d)$ and $(a + u c, b + u d)$ , so its determinant also gives $∥ (a + u c, b + u d) ∥ \geq B_{1}^{- 1}$ . Set $R : = \max (1, 2 B^{2})$ . If $∣ u ∣ \geq R$ , the triangle inequality gives $∥ (a + u c, b + u d) ∥ \geq ∣ u ∣ / (2 B) \geq (1 + ∣ u ∣) / (4 B)$ ; for $∣ u ∣ \leq R$ the determinant bound gives $∥ (a + u c, b + u d) ∥ \geq (1 + ∣ u ∣) / (B_{1} (1 + R))$ . Thus for $c_{C} : = \min ((4 B)^{- 1}, (B_{1} (1 + R))^{- 1}) > 0$ , the bottom-row norm $r (u, g)$ is at least $c_{C} (1 + ∣ u ∣)$ throughout $C$ . The compact-picture formula [F2] therefore gives $∣ φ (w n_{u} g) ∣ \leq ∥ φ ∣_{K} ∥_{\infty} c_{C}^{- 1 - σ} (1 + ∣ u ∣)^{- 1 - σ}, σ = Re ν > 0.$ The majorant is integrable: for $T > 0$ , the antiderivative supplied by [F4] gives $\int_{0}^{T} (1 + u)^{- 1 - σ} d u = (1 - (1 + T)^{- σ}) / σ \to 1 / σ$ ; reflection $u \mapsto - u$ in [F5] gives the same finite integral on $(- \infty, 0)$ , and [F5] combines the two pieces. Taking $C$ to contain any fixed $g$ proves absolute convergence there.

2.1F6F8F9step 1.1

For each fixed $g$ , step 1.1 makes $u \mapsto φ (w n_{u} g)$ integrable; the integrand is continuous, hence measurable by [F8]. Thus the displayed formula defines a value for every $g$ . Applying [F6] to the integrands for $φ_{1}, φ_{2}$ and their linear combination, which are all integrable by step 1.1, shows $A (ν)$ is complex-linear.

2.2F6F7F8F9step 1.1

In a smooth coordinate chart $g = γ (z)$ and on any compact sub-box, the map $(u, z) \mapsto φ (w n_{u} γ (z))$ is smooth by [F7]. Every coordinate derivative is a finite sum of right derivatives of $φ$ at $w n_{u} γ (z)$ with smooth coefficients bounded on that sub-box. Each such right derivative remains left- $P$ -covariant with character $χ_{ε, ν}$ , and its restriction to compact $K$ is bounded; therefore the estimate of step 1.1 gives one integrable majorant $C_{α} (1 + ∣ u ∣)^{- 1 - σ}$ for each coordinate derivative $\partial_{z}^{α}$ , uniformly on the sub-box. These derivatives are continuous in $u$ and hence measurable by [F8]. Applying the differentiation-under-the-integral theorem [F6] successively to the coordinates gives $\partial_{z}^{α} (A (ν) φ) (γ (z)) = \int_{R} \partial_{z}^{α} [φ (w n_{u} γ (z))] d u$ ; dominated convergence [F6] makes each such derivative continuous in $z$ . All coordinate derivatives therefore exist and are continuous, so $A (ν) φ$ is smooth.

3.1F1F5F9step 1.1step 2.2algebra

For $n_{x} \in N$ , $n_{u} n_{x} = n_{u + x}$ , so translation change of variables [F5] gives $A (ν) φ (n_{x} g) = A (ν) φ (g)$ . For $a_{s} \in A$ , use $n_{u} a_{s} = a_{s} n_{e^{- s} u}$ and $w a_{s} = a_{- s} w$ ; covariance contributes $e^{- (1 + ν) s / 2}$ , and the change of variables for integrable complex functions $v = e^{- s} u$ contributes $e^{s}$ . Hence $A (ν) φ (a_{s} g) = e^{(1 - ν) s / 2} A (ν) φ (g) .$ For $m \in M = {\pm I}$ , centrality gives $w n_{u} m g = m w n_{u} g$ , so $A (ν) φ (m g) = σ_{ε} (m) A (ν) φ (g)$ . These are exactly the $P = M A N$ covariance rules for $I_{ε, - ν}$ . For every $g_{0} \in G$ , $A (ν) Π_{ν} (g_{0}) φ (g) = \int_{R} φ (w n_{u} g g_{0}) d u = (Π_{- ν} (g_{0}) A (ν) φ) (g),$ so the operator intertwines right translations. Together with step 2.2 this proves that its target is the smooth model $I_{ε, - ν}$ .

4.1F3step 3.1algebra

The compact picture identifies source and target with the same parity space and its K-types with the one-dimensional lines $C f_{n}$ . Since step 3.1 intertwines every right translation, $A (ν)$ maps each finite-dimensional right- $K$ orbit span into a finite-dimensional right- $K$ orbit span. If $f_{n}$ is a K-type vector, its image has the same right- $K$ character; the corresponding target character space is exactly $C f_{n}$ by [F3]. Hence $A (ν) f_{n} = c_{n} (ν) f_{n}$ for a scalar $c_{n} (ν)$ and every allowed $n$ , proving the K-finite-target and diagonalization assertions.

5.1F2F4F5F8F9F10step 1.1step 4.1∎

For $ε = 0$ , the compact vector $f_{0} = 1$ extends by [F2]. At $g = I$ , the bottom row of $w n_{u}$ is $(1, u)$ , so its norm is $r = \sqrt{1 + u^{2}}$ . The positive-real-log convention for complex powers gives $φ (w n_{u}) = r^{- 1 - ν} = (1 + u^{2})^{- (1 + ν) / 2}$ and $c_{0} (ν) = (A (ν) f_{0}) (I) = \int_{R} (1 + u^{2})^{- (1 + ν) / 2} d u .$ For real $ν > 0$ the integrand is positive and on $[0, 1]$ is at least $2^{- (1 + ν) / 2}$ ; that interval has measure $1$ by [F5], so $c_{0} (ν) > 0$ by monotonicity [F4]. For complex $Re ν > 0$ , the integrand is even and absolutely integrable by step 1.1. Splitting off the null endpoint and reflecting the negative half-line by [F5] gives twice its integral on $(0, \infty)$ . Under the C1 diffeomorphism $t = u^{2} / (1 + u^{2}) : (0, \infty) \to (0, 1)$ , one has $d t / d u = 2 u / (1 + u^{2})^{2}$ and $t^{- 1 / 2} (1 - t)^{ν / 2 - 1} \frac{d t}{d u} = 2 (1 + u^{2})^{- (1 + ν) / 2} .$ The change-of-variables formula for integrable complex functions [F5] therefore yields $c_{0} (ν) = \int_{0}^{1} t^{- 1 / 2} (1 - t)^{ν / 2 - 1} d t = B (1 / 2, ν / 2),$ as claimed by [F10]. For $ε = 1$ , the same integral defines the formal scalar $c_{0} (ν)$ but is not an eigenvalue because $f_{0}$ is not an allowed K-type; its scalar continuation is proved by K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing ↗, while the full smooth-operator continuation is supplied by Meromorphic continuation and intertwining identity for A(nu) ↗.

Remarks

Continuation discharge: Meromorphic continuation and intertwining identity for A(nu) ↗, Proof 1.1–5.1, proves the common simple-pole set, uniform polynomial multiplier bounds for both signs of the K-index, locally convergent operator power series in every smooth seminorm, agreement with this integral on its initial half-plane, and the full group-intertwining identity. Its actual base eigenvalue is $c_{0}$ in even parity and $c_{1}$ in odd parity; the formal odd-parity scalar $c_{0}$ is not used as that base. The normalized common-pole extensions and exceptional kernels are established there and in K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing ↗.
Source convention: Kerr's Exercise 2.8 uses $w_{K} = (\begin{array}{cc} 0 & 1 \\ - 1 & 0 \end{array}) = - w$ . Since $- I \in M$ , that source integral differs from this item's chosen $w = k_{- π / 2}$ by the factor $σ_{ε} (- I) = (- 1)^{ε}$ . The local formulas use the explicit $w$ fixed in the Definition; transfer of any source eigenvalue normalization must include that factor.
Kerr, Exercise 2.8(i)–(iii), printed p. 12, asks the reader to verify right intertwining, target covariance, and nonvanishing; it does not provide those proofs or meromorphic continuation. Kowalski, Proposition 7.4.3(3) (statement p. 294, discussion pp. 301–302) and Exercise 7.4.12 (p. 302), classify equivalence and leave construction of an inverse-character intertwiner as an exercise. Etingof, §9.1–9.2, printed pp. 48–50, gives the algebraic $P^{\pm} (s) ≅ P^{\pm} (- s)$ equivalence only in the irreducible regime and the right- $P$ model with parameter $s = - ν$ . These are motivation and convention checks, not substitutes for the local proof or its verified suppliers.

LemmaStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ , let $ν \in C$ with $Re ν > 0$ , and let $c_{r} (ν)$ be the eigenvalue of $A (ν)$ on the $K$ -type $f_{r}$ for each $r \equiv ε (m o d 2)$ in The standard intertwining operator A(nu). Then, for every such $r$ , $(r + 1 + ν) c_{r + 2} (ν) = (r + 1 - ν) c_{r} (ν), c_{- r} (ν) = (- 1)^{r} c_{r} (ν) .$ When $r + 1 + ν \neq 0$ , the first identity is equivalently $c_{r + 2} (ν) = c_{r} (ν) \frac{r + 1 - ν}{r + 1 + ν};$ at a zero denominator the cross-multiplied identity is the meaning of the recurrence. The scalar eigenvalues have the closed form $c_{r} (ν) = (- i)^{r} \sqrt{π} Γ (ν / 2) Γ ((ν + 1) / 2) \frac{1}{Γ ((ν + r + 1) / 2)} \frac{1}{Γ ((ν - r + 1) / 2)},$ where reciprocal Gamma is understood as its entire continuation, so this formula is valid for $Re ν > 0$ and its right side gives the meromorphic continuation of each scalar eigenvalue. Define the two base scalars $b_{0} (ν) = \int_{R} (1 + u^{2})^{- (1 + ν) / 2} d u, b_{1} (ν) = \int_{R} (u - i) (1 + u^{2})^{- (2 + ν) / 2} d u .$ Then $b_{0} = \sqrt{π} Γ (ν / 2) / Γ ((ν + 1) / 2)$ and $b_{1} = - i \sqrt{π} Γ ((ν + 1) / 2) / Γ ((ν + 2) / 2)$ . For $ε = 0$ , $c_{0} = b_{0}$ is the base eigenvalue and $b_{1}$ is only a formal odd scalar; for $ε = 1$ , $c_{1} = b_{1}$ is the base eigenvalue and $b_{0}$ is only a formal even scalar.

For $m \in W_{ε}$ , $m \geq 1$ , the exceptional zero sets are exact: at $ν = m$ , $c_{r} (m) = 0$ exactly for allowed $r$ with $∣ r ∣ \geq m + 1$ ; at $ν = - m$ , $c_{r} (- m) = 0$ exactly for allowed $r$ with $∣ r ∣ \leq m - 1$ . In particular $c_{m - 1} (m) \neq 0$ and $c_{m + 1} (- m)$ is finite and nonzero.

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $Re ν > 0$ , and the smooth compact-picture principal series.

[F1]

For $Re ν > 0$ , the defining integral for $A (ν)$ is absolutely convergent, smooth, covariant for $I_{ε, - ν}$ , right- $G$ intertwining, and diagonal on the allowed $K$ -types. These initial-half-plane claims are verified in steps 1.1–4.1 of The standard intertwining operator A(nu); this proof uses only those claims. The formal even scalar and its continuation needed when $ε = 1$ are established below. The separate meromorphic continuation of the full operator family in that Definition is not assumed here.

[F2]

The compact-picture action has $K$ -types $C f_{r}$ with $f_{r} (k_{θ}) = e^{i r θ}$ and $r \equiv ε (m o d 2)$ (The compact picture of the SL2(R) principal series, K-type decomposition of the SL2(R) principal series). The exceptional lattice is $W_{0} = 2 Z + 1$ and $W_{1} = 2 Z$ (The normalized principal series I(epsilon, nu)).

[F3]

The complex-linear derived action satisfies $L_{E_{\pm}} f_{r} = (1 + ν \pm r) f_{r \pm 2} / 2$ and preserves smooth vectors (Derived action and raising/lowering formulas in the compact picture).

[F4]

For $Re p, Re q > 0$ , the Beta and Gamma integrals satisfy $B (p, q) = Γ (p) Γ (q) / Γ (p + q)$ (Euler's Beta function on the right half-planes, Euler's Gamma function on the right half-plane, The Beta-Gamma identity), and $Γ (1 / 2) = \sqrt{π}$ (The value of Gamma at one half).

[F5]

Gamma has meromorphic continuation with simple poles of nonzero residue at the nonpositive integers, no zeros, and reciprocal Gamma is entire with simple zeros exactly there. Its functional equation holds meromorphically (Meromorphic continuation of Gamma, Gamma has no zeros).

[F6]

If $T : U \to V$ is a $C^{1}$ diffeomorphism of open Euclidean sets and $h \in L^{1} (V)$ , then $\int_{V} h = \int_{U} (h \circ T) ∣ \det D T ∣$ (A C^1 diffeomorphism satisfies the change-of-variables formula for L^1 functions).

[F7]

$L (R)$ is the Lebesgue sigma-algebra, and a real or complex function is in $L^{1}$ when it is measurable and its absolute value has finite integral (Lebesgue measurable sets, the family $L (R^{n})$ , and the restricted set function $λ_{n}$ , Integrable real and complex functions, and their integrals).

[F8]

Measurable restrictions of nonnegative functions are measurable, dominated nonnegative functions are integrable when the majorant is, and the integral over a measurable set is the integral of the indicator restriction (Integral over a measurable subset, Closure properties of measurable functions used by the integral, Monotonicity and nonnegative homogeneity of the nonnegative integral).

[F9]

A nonnegative integrable function has integral zero over a null set, and every singleton in $R$ is null by the degenerate interval case (A nonnegative integral over a null set vanishes, A box in $R^{n}$ with parameters $a_{i} \leq b_{i}$ is Lebesgue measurable of measure $\prod_{i < n} (b_{i} - a_{i})$ , whichever of its faces are included).

[F10]

The Lebesgue integral is complex-linear on $L^{1}$ (The Lebesgue integral is linear on $L^{1} (μ)$ ).

[A1]

The item assumes full AC. It implies ${A C}_{ω}$ through The Axiom of Countable Choice ( ${A C}_{ω}$ ), which supplies the countable-choice hypothesis in [F6] and the singleton-null supplier in [F7]. The algebraic and reflection calculations make no further choices (The Axiom of Choice).

Proof

technique · evaluate the integral on each $K$-type, differentiate its intertwining identity, and determine all scalars from one base index in each parity

1.1F1F2F7algebra

Put $R (u) = \sqrt{1 + u^{2}}$ and $q_{r} (u) = (u - i)^{r} R (u)^{- 1 - ν - r}$ . At $g = I$ , the bottom row of $w n_{u}$ is $(1, u)$ , so the compact coordinate satisfies $e^{i θ_{u}} = (u - i) / R (u)$ . Thus the defining integral gives $c_{r} (ν) = \int_{R} q_{r} (u) d u$ . Also $b_{0} = \int_{R} q_{0}$ and $b_{1} = \int_{R} q_{1} (u) d u$ . For every integer $r$ , $∣ q_{r} (u) ∣ = R (u)^{- 1 - Re ν}$ ; this equals the absolute value of the defining integrand for any allowed mode, so it is integrable by [F1]. It also shows both formal base integrands are integrable when their parity is not allowed.

1.2F1F2F3algebra

Differentiate the right- $G$ intertwining identity in [F1] along real Lie-algebra directions and extend complex-linearly as in [F3]. Applying $A (ν) L_{E_{+}}^{ν} = L_{E_{+}}^{- ν} A (ν)$ to $f_{r}$ gives $(r + 1 + ν) c_{r + 2} (ν) = (r + 1 - ν) c_{r} (ν)$ . This cross-multiplied identity holds even when $r + 1 + ν = 0$ ; division gives the ratio form only when that denominator is nonzero.

1.3F1F4F6F7F8F9F10A1algebra

The even function $q_{0}$ is in $L^{1} (R)$ . For measurable $E$ , write $\int_{E} q_{0} = \int_{R} q_{0} χ_{E}$ ; split real and imaginary parts into positive and negative parts, use [F8] for measurability of their restrictions, and use $∣ q_{0} χ_{E} ∣ \leq ∣ q_{0} ∣$ and [F7, F8] for integrability. The singleton ${0}$ is null by the degenerate interval case of [F9]; its complex integral is zero by applying the nonnegative null-integral result in [F9] to the four restricted parts and recombining by [F10]. The three indicators of $(- \infty, 0)$ , ${0}$ , and $(0, \infty)$ sum to $1$ , so [F10] splits the full integral into these subset integrals. Reflection $u \mapsto - u$ in [F6] equates the two open half-line integrals, giving $b_{0} = 2 \int_{0}^{\infty} q_{0} (u) d u$ . The inverse change of variables for $t = u^{2} / (1 + u^{2})$ is $u = \sqrt{t / (1 - t)}$ with derivative $1 / (2 \sqrt{t} (1 - t)^{3 / 2})$ . Applying [F6] to this inverse diffeomorphism and the $L^{1}$ function $q_{0}$ gives $b_{0} = B (1 / 2, ν / 2)$ ; these Beta parameters have positive real parts because $Re ν > 0$ . By [F4], this equals $\sqrt{π} Γ (ν / 2) / Γ ((ν + 1) / 2)$ .

2.1F1F4F6F7F8F9F10step 1.1step 1.3algebra

Put $h_{1} (u) = (1 + u^{2})^{- 1 - ν / 2}$ . Both $h_{1}$ and $u h_{1}$ are in $L^{1}$ : pointwise $∣ h_{1} (u) ∣ \leq ∣ (u - i) h_{1} (u) ∣$ and $∣ u h_{1} (u) ∣ \leq ∣ (u - i) h_{1} (u) ∣$ , and the latter is the absolutely integrable formal base integrand by step 1.1. Reflection in [F6] sends $u h_{1} (u)$ to its negative, so its full integral is zero. The same indicator splitting, reflection and null-singleton argument as in step 1.3 give $\int_{R} h_{1} = 2 \int_{0}^{\infty} h_{1}$ . The inverse substitution from step 1.3 now gives $\int_{R} h_{1} = B (1 / 2, (ν + 1) / 2)$ ; these Beta parameters have positive real parts because $Re ν > 0$ . By linearity and [F4], $b_{1} = \int_{R} (u - i) h_{1} (u) d u = - i \sqrt{π} Γ ((ν + 1) / 2) / Γ ((ν + 2) / 2)$ . Thus $c_{0} = b_{0}$ when $ε = 0$ and $c_{1} = b_{1}$ when $ε = 1$ ; the other scalar is only a formal base integral.

2.2F6step 1.1algebra

The density formula gives $q_{- r} (u) = (- 1)^{r} q_{r} (- u)$ for every integer $r$ . Since both sides are integrable by step 1.1, the $C^{1}$ reflection $u \mapsto - u$ in [F6] yields $c_{- r} (ν) = (- 1)^{r} c_{r} (ν)$ directly, with no division by a recurrence coefficient.

3.1F4F5step 1.2step 1.3step 2.1step 2.2

Define ${\tilde{c}}_{r} (ν)$ by the Gamma formula in the Statement and set $x = (ν + r + 1) / 2$ , $y = (ν - r + 1) / 2$ . The Gamma functional equation gives ${\tilde{c}}_{r + 2} = - (y - 1) x^{- 1} {\tilde{c}}_{r} = \frac{r + 1 - ν}{r + 1 + ν} {\tilde{c}}_{r}$ wherever this ratio is defined, so $(r + 1 + ν) {\tilde{c}}_{r + 2} = (r + 1 - ν) {\tilde{c}}_{r}$ holds meromorphically, including at zero denominators. Swapping the denominator factors and using $(- i)^{- 2 r} = (- 1)^{r}$ gives ${\tilde{c}}_{- r} = (- 1)^{r} {\tilde{c}}_{r}$ . The substitutions in steps 1.3 and 2.1 give ${\tilde{c}}_{0} = b_{0} = c_{0}$ for even parity and ${\tilde{c}}_{1} = b_{1} = c_{1}$ for odd parity. For $r \geq 0$ , $r + 1 + ν \neq 0$ on $Re ν > 0$ , so the recurrence determines every nonnegative allowed index from that base; step 2.2 determines the negative indices. Hence $c_{r} = {\tilde{c}}_{r}$ on the initial half-plane, and the Gamma expression supplies the meromorphic continuation of each scalar without asserting convergence of the original integral outside that half-plane. At $ν = 0$ , for even $r$ the denominator arguments are half-integers, so the numerator pole remains; for odd $r$ , exactly one denominator argument is a nonpositive integer, whose reciprocal zero cancels the numerator pole and leaves a finite nonzero value.

4.1F5step 3.1algebra∎

Let $m \in W_{ε}$ be positive. Then $m$ and every allowed $r$ have opposite parity, so all four denominator arguments at $ν = \pm m$ are integers. At $ν = m$ , both arguments are positive for $∣ r ∣ \leq m - 1$ , making the Gamma quotient finite and nonzero; for $∣ r ∣ \geq m + 1$ , exactly one argument is nonpositive, so its reciprocal-Gamma factor vanishes and the numerator is finite. At $ν = - m$ , exactly one numerator Gamma factor has a simple pole and the other is finite and nonzero. If $∣ r ∣ \leq m - 1$ , both denominator arguments are nonpositive integers, so their two simple reciprocal-Gamma zeros leave a zero after multiplication by the single numerator pole. If $∣ r ∣ \geq m + 1$ , exactly one denominator argument is a nonpositive integer and the other is positive; its simple reciprocal-Gamma zero cancels the numerator pole and leaves a finite nonzero value. In particular, the denominator arguments at $r = m - 1, ν = m$ are $m, 1$ , and at $r = m + 1, ν = - m$ they are $1, - m$ , proving the two stated boundary values. These cases prove both directions of the exact zero-set assertions.

Remarks

Kerr's formulas (2.5)–(2.6) use the same compact-picture ladder normalization and give the same derived-action coefficients. His Exercise 2.8 asks for the intertwiner properties and K-type computation without supplying a solution; the integral eigenvalues and their continuation are derived above. Kerr's Weyl matrix is $- w$ relative to the $w$ fixed in The standard intertwining operator A(nu), so its integral eigenvalues differ by $(- 1)^{ε}$ in parity $ε$ .

Etingof's §9.1 formulas (4)–(5) use an abstractly normalized $({s l}_{2}, K)$ basis, and §9.2 uses a right- $P$ -covariant model with left $G$ -action. In that model $F (g b) = ∣ t (b) ∣^{s - 1} σ_{ε} (t (b)) F (g)$ , while inversion of the present left- $P$ model gives exponent $- 1 - ν$ and hence $s = - ν$ . This is a convention and ladder check; it does not supply the integral eigenvalue constants proved here.

LemmaStatement: Literature-sourcedProof: AI-adaptedjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

The invariant pairing between opposite principal-series parameters

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ and $ν \in C$ . The sesquilinear pairing $⟨ f, h ⟩ = \int_{K} f (k) \overline{h (k)} d k$ between the smooth compact-picture spaces $C_{ε}^{\infty} (K)$ of $I_{ε, - \overset{ˉ}{ν}}$ and $I_{ε, ν}$ is $G$ -invariant: $(Π_{- \overset{ˉ}{ν}} (g) f, Π_{ν} (g) h) = (f, h) (f, h \in C_{ε}^{\infty} (K), g \in G) .$ For real $ν$ it pairs $I_{ε, - ν}$ with $I_{ε, ν}$ ; on the $K$ -type basis $f_{n}$ of K-type decomposition of the SL2(R) principal series, $⟨ f_{m}, f_{n} ⟩ = δ_{m n}$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in C$ , and smooth compact-picture vectors $f, h$ .

[F1]

The compact-picture action is $(Π_{λ} (g) u) (k) = ∣ α (p (k, g)) ∣^{1 + λ} u (κ (k, g))$ for $u \in C_{ε}^{\infty} (K)$ , where $k g = p (k, g) κ (k, g)$ is the canonical $A N \times K$ factorization. Here the $A N$ factor has trivial $M$ -character (The compact picture of the SL2(R) principal series).

[F2]

The $N A K$ coordinates are smooth global coordinates; using $a_{t} n_{x} = n_{e^{t} x} a_{t}$ converts them by a smooth coordinate change into the unique smooth $A N K$ coordinates (Iwasawa decomposition and Haar integration formula for SL2(R), Iwasawa and minimal-parabolic data for SL2(R)).

[F3]

The model parameter and its normalized inducing character are fixed by The normalized principal series I(epsilon, nu).

[F4]

The parity basis is $f_{n} (k_{θ}) = e^{i n θ}$ for $n \equiv ε (m o d 2)$ , and is orthonormal for normalized Haar measure on $K$ (K-type decomposition of the SL2(R) principal series).

[F5]

Under $k_{θ} \mapsto [θ / (2 π)] \in R / Z$ , normalized Haar probability on $K$ is $d k = d θ / (2 π)$ : the torus definition gives this normalized translation-invariant probability, and uniqueness of normalized Haar probability identifies its pullback with $d k$ (Iwasawa and minimal-parabolic data for SL2(R), The one-dimensional torus and its normalized Haar integral, Normalized Haar measure on a compact Lie group).

[F6]

An orientation-preserving diffeomorphism of the circle preserves the integral of a smooth top form; every smooth top form on compact $K$ has compact support (Change of variables on oriented manifolds).

[F7]

The complex exponential obeys $e^{z + w} = e^{z} e^{w}$ , $∣ e^{z} ∣ = e^{Re z}$ , and $\overline{e^{z}} = e^{\overset{ˉ}{z}}$ ( $\exp (z + w) = \exp z \exp w$ , and the complex exponential extends the real exponential, $\exp (x + i y) = e^{x} (\cos y + i \sin y)$ , $∣ \exp (x + i y) ∣ = e^{x}$ , and $e^{i π} + 1 = 0$ ).

[F8]

The product of smooth functions is smooth by the coordinatewise product rule, and complex conjugation is the coordinate map $(x, y) \mapsto (x, - y)$ (Sums, scalar multiples, products and quotients: $(f + g)^{'} (c) = f^{'} (c) + g^{'} (c)$ , $(α f)^{'} (c) = α f^{'} (c)$ , $(f g)^{'} (c) = f^{'} (c) g (c) + f (c) g^{'} (c)$ , and $(f / g)^{'} (c) = (f^{'} (c) g (c) - f (c) g^{'} (c)) / g (c)^{2}$ when $g (c) \neq 0$ , Real and imaginary parts, complex conjugation, and modulus).

[F9]

The integral of a complex function is defined by integrating its real and imaginary parts and combining the two real integrals (Integrable real and complex functions, and their integrals).

[F10]

A continuous real-valued function is bounded on compact $K$ ; since $d k (K) = 1$ , a bounded measurable function on $K$ is integrable by monotonicity (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value, Monotonicity and nonnegative homogeneity of the nonnegative integral, Normalized Haar measure on a compact Lie group).

[A1]

AC supplies normalized Haar probability on compact $K$ and implies the countable-choice hypothesis for the torus measure; no vector is selected in this proof (The Axiom of Choice, The Axiom of Countable Choice ( ${A C}_{ω}$ ), Normalized Haar measure on a compact Lie group).

Proof

technique · direct Jacobian calculation in the compact coordinate

1.1F1F2F5F6F8F9A1algebra

For fixed $g = (\begin{array}{cc} a & b \\ c & d \end{array}) \in G$ , write $k_{θ} g = a_{t (θ)} n_{x (θ)} k_{ψ (θ)}$ in the unique smooth $A N K$ coordinates [F2], using the compact-picture notation [F1]. The bottom row of $k_{θ} g$ is $v (θ) = (- a \sin θ + c \cos θ, - b \sin θ + d \cos θ) = e^{- t (θ) / 2} (- \sin ψ (θ), \cos ψ (θ))$ , so its norm is $e^{- t / 2} = ∣ α (p (k_{θ}, g)) ∣^{- 1}$ . Direct differentiation gives $\det (v, v^{'}) = a d - b c = 1$ , while $\det ((- \sin ψ, \cos ψ), (- \cos ψ, - \sin ψ)) = 1$ ; therefore $e^{- t (θ)} ψ^{'} (θ) = 1$ , or $ψ^{'} (θ) = e^{t (θ)} = ∣ α (p (k_{θ}, g)) ∣^{2} > 0$ . If $k g = p (k, g) κ (k, g)$ with $p (k, g) \in A N$ , then $κ (k, g) g^{- 1} = p (k, g)^{- 1} k$ ; uniqueness of the same coordinates gives $κ (κ (k, g), g^{- 1}) = k$ , and the reversed identity gives the inverse map, so $κ (\cdot, g)$ is an orientation-preserving diffeomorphism. By [F5], $d k = d θ / (2 π)$ . For a smooth complex $F$ , [F8] makes $Re F$ and $Im F$ smooth; apply [F6] separately to the compactly supported real top forms $(Re F) d θ / (2 π)$ and $(Im F) d θ / (2 π)$ and combine by [F9]. This yields $\int_{K} ∣ α (p (k, g)) ∣^{2} F (κ (k, g)) d k = \int_{K} F (k) d k$ .

2.1F1F3F7F9F10step 1.1algebra

By [F1] and [F3], $Π_{- \overset{ˉ}{ν}} (g) f$ contributes $e^{(1 - \overset{ˉ}{ν}) t / 2} f (κ)$ , while the conjugate of $Π_{ν} (g) h$ contributes $e^{(1 + \overset{ˉ}{ν}) t / 2} \overline{h (κ)}$ by [F7]; their product is $∣ α ∣^{2} f (κ) \overline{h (κ)}$ . Apply step 1.1 with $F = f \overline{h}$ to obtain $⟨ Π_{- \overset{ˉ}{ν}} (g) f, Π_{ν} (g) h ⟩ = \int_{K} f (k) \overline{h (k)} d k$ . The compact Haar probability and smoothness make these integrals finite by [F9]–[F10].

3.1F4F5F9F10algebra∎

For $m, n \equiv ε (m o d 2)$ , [F4]–[F5] give $⟨ f_{m}, f_{n} ⟩ = (2 π)^{- 1} \int_{0}^{2 π} e^{i (m - n) θ} d θ$ , which equals $1$ when $m = n$ and $0$ otherwise by direct integration. For real $ν$ , $- \overset{ˉ}{ν} = - ν$ , giving the stated opposite-parameter pairing.

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Meromorphic continuation and intertwining identity for A(nu)

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ , and let $A (ν)$ and its eigenvalues $c_{n} (ν)$ be as in The standard intertwining operator A(nu) and K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing.

(Meromorphic continuation.) Each eigenvalue $c_{n} (ν)$ extends to a meromorphic function on $C$ , with no zeros outside the exceptional lattice $W_{ε}$ ; the family $A (ν)$ has a unique meromorphic continuation in $ν$ as a family of continuous $K$ -diagonal maps on the smooth compact-picture spaces.
(Intertwining identity.) For every $ν$ at which the continuation is regular, $A (ν) Π_{ν} (g) = Π_{- ν} (g) A (ν)$ on smooth compact-picture vectors (and therefore on the algebraic $K$ -finite core where its derived action is defined). If $ν \notin W_{ε}$ and both $A (ν)$ and $A (- ν)$ are regular, then $A (- ν) A (ν)$ is the scalar $c_{n_{0}} (ν) c_{n_{0}} (- ν)$ on every $K$ -type of parity $ε$ , where $n_{0} = 0$ for $ε = 0$ and $n_{0} = 1$ for $ε = 1$ . At these parameters, $\hat{A} (ν) : = A (ν) / c_{n_{0}} (ν)$ satisfies $\hat{A} (- ν) \hat{A} (ν) = i d$ and ${\hat{c}}_{n} (- ν) {\hat{c}}_{n} (ν) = 1$ , where ${\hat{c}}_{n} (ν) = c_{n} (ν) / c_{n_{0}} (ν)$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , the normalized smooth principal-series models, and the standard integral $A (ν)$ on its initial half-plane $Re ν > 0$ .

[F1]

For $Re ν > 0$ , $A (ν)$ is the absolutely convergent integral over $N$ of The standard intertwining operator A(nu); its right-translation action is the smooth compact-picture action The normalized principal series I(epsilon, nu), The compact picture of the SL2(R) principal series. The full-family continuation left open in the Definition is not assumed here; it is proved below.

[F2]

For every $r \equiv ε (m o d 2)$ , the eigenvalue has the meromorphic Gamma expression, cross-multiplied recurrence, and symmetry $c_{- r} (ν) = (- 1)^{r} c_{r} (ν)$ in K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing. Gamma has simple poles at the nonpositive integers, no zeros, and reciprocal Gamma is entire with simple zeros exactly at those poles, as used in that supplier.

[F3]

The compact picture identifies smooth vectors with $C_{ε}^{\infty} (K)$ , whose K-types are $f_{r} (k_{θ}) = e^{i r θ}$ for $r \equiv ε (m o d 2)$ ; in this angle coordinate normalized Haar measure is $d k = d θ / (2 π)$ (The compact picture of the SL2(R) principal series, K-type decomposition of the SL2(R) principal series, Iwasawa and minimal-parabolic data for SL2(R), The one-dimensional torus and its normalized Haar integral).

[F4]

AC supplies countable choice, which is the countable-choice premise of complex integration by parts and the Cauchy integral estimates (AC supplies the countable and dependent choices used in Banach integration, The Axiom of Choice).

[F5]

The usual modulus makes $C$ a complex normed space (Real and complex scalar conventions for normed spaces, Conjugation is an involutive real-field automorphism, $z \overline{z} = ∣ z ∣^{2}$ , and modulus is definite, multiplicative, and subadditive); its norm metric is $d (z, w) = ∣ z - w ∣$ (The Euclidean metric, convergence, Cauchy sequences, and continuity on the complex plane), which is complete (The complex plane is complete, and convergence is equivalent to convergence of real and imaginary parts), so $C$ is a complex Banach space by Banach space. The Banach-valued Cauchy theorem therefore gives Taylor expansions and coefficient estimates for scalar holomorphic functions on discs (Cauchy integral formula and Cauchy estimates for Banach-valued holomorphic functions).

[F6]

Complex integration by parts on a period interval gives rapid decay of Fourier coefficients of a smooth periodic function (Complex integration by parts on intervals and decaying lines). The Weierstrass M-test, uniform derivative-limit theorem, and convergence of $\sum_{m \geq 1} m^{- 2}$ justify uniform convergence and termwise angular differentiation (Weierstrass M-test for complex-valued function series, If continuously differentiable functions converge at one point and their derivatives converge uniformly on a closed interval, then the functions converge uniformly to a differentiable function whose derivative is the derivative limit, For rational $p > 0$ , $\sum 1 / k^{p}$ converges iff $p > 1$ ).

[F7]

The smooth compact-picture action is differentiable in the real Lie-algebra directions with the displayed ladder operators, and the ordinary real product and chain rules apply (Derived action and raising/lowering formulas in the compact picture, Sums, scalar multiples, products and quotients: $(f + g)^{'} (c) = f^{'} (c) + g^{'} (c)$ , $(α f)^{'} (c) = α f^{'} (c)$ , $(f g)^{'} (c) = f^{'} (c) g (c) + f (c) g^{'} (c)$ , and $(f / g)^{'} (c) = (f^{'} (c) g (c) - f (c) g^{'} (c)) / g (c)^{2}$ when $g (c) \neq 0$ , The chain rule, in one line from Carathéodory: if $g$ is differentiable at $c$ and $f$ is differentiable at $g (c)$ , then $f \circ g$ is differentiable at $c$ with $(f \circ g)^{'} (c) = f^{'} (g (c)) g^{'} (c)$ ).

[F8]

Every group element has a $K A N$ factorization; the factors are generated by the one-parameter subgroups from $J$ , $H / 2$ , and $E_{0} = (S + J) / 2$ (Iwasawa decomposition and Haar integration formula for SL2(R), Iwasawa and minimal-parabolic data for SL2(R)).

[F9]

Two holomorphic functions on a connected complex domain that agree on a nonempty open set agree throughout the domain (Identity theorem for holomorphic functions).

[F10]

A real-valued function continuous on an interval and differentiable with zero derivative at every interior point is constant there (A function continuous on an interval $I$ whose derivative vanishes at every interior point of $I$ is constant on $I$ ; consequently two such functions with the same derivative differ by a constant).

[F11]

Holomorphic functions are continuous (Complex differentiability at a point implies continuity there), and a continuous real-valued function on a compact metric space is bounded and attains its maximum (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value).

[A1]

AC supplies normalized Haar probability on $K$ for the compact Fourier coefficients and implies the countable-choice premises of [F4]–[F6]; no additional choice is used (The Axiom of Choice, Iwasawa and minimal-parabolic data for SL2(R), [F4]).

Proof

technique · continue the K-type multipliers with uniform polynomial bounds, then integrate the resulting derived intertwining identity along the real one-parameter subgroups

1.1F2algebra

For $r \equiv ε (m o d 2)$ , write the Gamma expression from [F2] as $c_{r} (ν) = (- i)^{r} \sqrt{π} Γ (ν / 2) Γ ((ν + 1) / 2) / (Γ ((ν + r + 1) / 2) Γ ((ν - r + 1) / 2))$ . Its numerator has no zeros and only simple poles at nonpositive integers. The denominator contributes zeros only when $ν = - r - 1 - 2 j$ or $ν = r - 1 - 2 j$ for some $j \geq 0$ , and each such parameter belongs to $W_{ε}$ ; hence $c_{r}$ has no zeros outside that lattice. At a possible numerator pole $ν = - m$ , if $m \equiv ε (m o d 2)$ the denominator arguments are half-integers and do not cancel the simple pole; if $m \equiv̸ ε (m o d 2)$ , the two denominator arguments are integers with at least one nonpositive, so a reciprocal-Gamma zero cancels the numerator pole. Thus all scalar poles are simple and lie in the locally finite set $P_{ε} = {- m : m \in N_{0}, m \equiv ε (m o d 2)}$ .

2.1F2F11step 1.1algebra

For each center $ν_{*} \in C$ , choose $ρ > 0$ so that the closed disc $D = \overline{D (ν_{*}, ρ)}$ has boundary disjoint from $P_{ε}$ and its intersection with $P_{ε}$ is either empty or ${ν_{*}}$ ; this is possible because $P_{ε}$ is locally finite. Put $h (ν) = 1$ if $ν_{*} \notin P_{ε}$ and $h (ν) = ν - ν_{*}$ if $ν_{*} \in P_{ε}$ . Every $h c_{r}$ is holomorphic on a neighbourhood of $D$ by step 1.1. Let $M = ∣ ν_{*} ∣ + ρ$ , so $∣ ν ∣ \leq M$ on $D$ by the triangle inequality for the complex modulus. Choose a nonnegative $r_{0}$ in the parity lattice so large that $r + 1 \geq 2 M$ for every $r \geq r_{0}$ , and choose an integer $L \geq \max (1, 4 M)$ . The cross-multiplied recurrence in [F2], with $h$ multiplied into both sides, gives $∣ h (ν) c_{r + 2} (ν) ∣ \leq \frac{r + 1 + M}{r + 1 - M} ∣ h (ν) c_{r} (ν) ∣ \leq (1 + \frac{L}{r + 1}) ∣ h (ν) c_{r} (ν) ∣$ for $ν \in D$ and $r \geq r_{0}$ ; the denominator here is the scalar $∣ r + 1 + ν ∣ \geq r + 1 - M > 0$ , and no division by $c_{r} (ν)$ is used. Since $h c_{r_{0}}$ is bounded on $D$ , iteration bounds $∣ h (ν) c_{r} (ν) ∣$ by a constant times $\prod_{k = 1}^{N} (1 + L / k) = (\binom{N + L}{L}) \leq (N + L)^{L}$ , where $N$ is the number of recurrence steps from $r_{0}$ to $r$ ; hence $∣ h (ν) c_{r} (ν) ∣ \leq C (1 + ∣ r ∣)^{L}$ uniformly. The finitely many smaller indices are bounded by holomorphy, and negative indices obey the same estimate by the symmetry in [F2].

3.1F3F4F6F11A1step 2.1

Using the angular Haar normalization in [F3], put $\hat{f} (r) = (2 π)^{- 1} \int_{0}^{2 π} f (k_{θ}) e^{- i r θ} d θ$ . For every positive integer $N$ , periodicity removes the boundary terms in [F6], giving $∣ \hat{f} (r) ∣ \leq p_{N} (f) ∣ r ∣^{- N}$ for $r \neq 0$ , where $p_{N} (f) = \max_{0 \leq j \leq N} ∥ \partial_{θ}^{j} f ∥_{\infty}$ . For each derivative order $q$ , take $N = q + 2$ ; integration by parts gives $(i r)^{q} \hat{f} (r) = \hat{\partial_{θ}^{q} f} (r)$ and the differentiated Fourier terms are bounded by $p_{N} (f) ∣ r ∣^{- 2}$ . The M-test and p-series fact [F6], applied separately to the positive and negative parity tails, give uniform convergence of the input Fourier series and every derivative series. The uniform derivative-limit theorem applied successively to real and imaginary parts shows these limits are the derivatives of the sum; their Fourier coefficients match those of $f$ and its derivatives, so completeness in [F3] and continuity identify them with those functions. Hence Fourier partial sums converge to $f$ in every $C^{q}$ seminorm. Next, for each angular derivative order $q$ , choose $N = L + q + 2$ in step 2.1; the differentiated terms of $\sum_{r} h (ν) c_{r} (ν) \hat{f} (r) e^{i r θ}$ are bounded by $C p_{N} (f) (1 + ∣ r ∣)^{- 2}$ . Applying the M-test separately to the positive and negative parity tails gives uniform convergence on the circle, locally uniformly in $ν$ , for this multiplier series and every angular derivative; applying the uniform derivative-limit theorem to real and imaginary parts shows its sum is smooth and the derivatives are the termwise derivatives. In particular each regular $A (ν)$ defined by this series is a continuous linear map $C_{ε}^{\infty} (K) \to C_{ε}^{\infty} (K)$ , with each output seminorm bounded by a constant times one input seminorm.

4.1F3F5F11step 2.1step 3.1

The same polynomial bound makes the meromorphic family holomorphic in the smooth topology after multiplication by $h$ . Choose concentric parameter discs $\overline{D (ν_{*}, R)} \subset D (ν_{*}, S)$ with $0 < R < S$ and closed radius- $S$ disc contained in the neighbourhood from step 2.1. Cauchy's coefficient estimate [F5] gives Taylor coefficients $a_{r, j}$ of $h c_{r}$ at $ν_{*}$ with $∣ a_{r, j} ∣ \leq C (1 + ∣ r ∣)^{L} S^{- j}$ . For each $j$ , the Fourier multiplier $T_{j} f = \sum_{r} a_{r, j} \hat{f} (r) e^{i r θ}$ is smooth by the argument of step 3.1 and satisfies $p_{q} (T_{j} f) \leq C_{q} S^{- j} p_{L + q + 2} (f)$ . Therefore, on every smaller closed parameter disc of radius $ρ < S$ , $\sum_{j \geq 0} T_{j} f (ν - ν_{*})^{j}$ converges in every $C^{q}$ seminorm, uniformly for $f$ in bounded subsets of $C_{ε}^{\infty} (K)$ . Its Fourier coefficients are the holomorphic values $(h c_{r}) (ν) \hat{f} (r)$ ; for regular $ν$ , completeness in [F3] makes the sum equal to $h (ν) A (ν) f$ , and its value at $ν_{*}$ is the removable extension when $ν_{*} \in P_{ε}$ (and the ordinary value otherwise). This is the required local power-series definition of a meromorphic family of continuous K-diagonal maps on $C_{ε}^{\infty} (K)$ .

4.2F1F2F3F9step 3.1

On $Re ν > 0$ , the operator from [F1] agrees with the Fourier multiplier in step 3.1: they agree on finite Fourier sums by [F2], and for a general smooth $f$ its Fourier partial sums converge uniformly by step 3.1 while the compact-picture integral satisfies $∥ A (ν) (f - f_{N}) ∥_{\infty} \leq C_{ν} ∥ f - f_{N} ∥_{\infty}$ , since for $k \in K$ the bottom-row norm in the integral is $\sqrt{1 + u^{2}}$ and $(1 + u^{2})^{- (1 + Re ν) / 2}$ is integrable. Thus the series is a continuation of the stated integral. Any other meromorphic K-diagonal continuation has on each K-type a scalar meromorphic coefficient agreeing with $c_{r}$ on this open half-plane; clearing local pole factors and applying [F9] on connected parameter discs makes the coefficients equal as meromorphic functions. The two continuous operators then agree on finite Fourier sums and, by the density from step 3.1, on all of $C_{ε}^{\infty} (K)$ , proving uniqueness.

4.3F2F3F7step 3.1

The recurrence in [F2], interpreted cross-multiplied at zero denominators, gives $A (ν) L_{E_{+}}^{ν} f_{r} = L_{E_{+}}^{- ν} A (ν) f_{r}$ on every allowed K-type; the same recurrence at index $r - 2$ gives the $E_{-}$ identity, and K-diagonality gives the $W$ identity. These are meromorphic scalar equalities, so they hold at every regular parameter. By linearity they hold for the real generators $J, H, S$ on finite Fourier sums; their operators are continuous on $C^{\infty}$ by [F7], and step 3.1 gives density of Fourier sums in that topology, so the derived intertwining identities hold on all smooth vectors.

5.1F7F8F10step 3.1step 4.3algebra

Fix a real generator $X \in {J, H, (S + J) / 2}$ and a smooth $f$ . The smooth orbit maps, continuity of $A (ν)$ from step 3.1, and the real chain rule [F7] make $F (t) = Π_{- ν} (\exp (- t X)) A (ν) Π_{ν} (\exp (t X)) f$ continuous and differentiable in every smooth seminorm. Its derivative is $Π_{- ν} (\exp (- t X)) (- L_{X}^{- ν} A (ν) + A (ν) L_{X}^{ν}) Π_{ν} (\exp (t X)) f = 0$ by step 4.3. Evaluation at each $k_{θ}$ gives a complex-valued function of $t$ whose real and imaginary parts satisfy [F10], so $F (t) = F (0)$ pointwise. Therefore $A (ν) Π_{ν} (\exp (t X)) = Π_{- ν} (\exp (t X)) A (ν)$ . Each element of $G$ is $k a n$ by [F8], with $K, A, N$ generated by these one-parameter subgroups, so the asserted $G$ -intertwining identity follows by multiplying the three factor identities.

6.1F2step 1.1step 3.1algebra∎

Suppose $ν \notin W_{ε}$ and both $A (ν)$ and $A (- ν)$ are regular. Then $r + 1 + ν$ and $r + 1 - ν$ are nonzero for every $r \equiv ε (m o d 2)$ . Applying the recurrence at $ν$ and $- ν$ gives $c_{r + 2} (ν) c_{r + 2} (- ν) = c_{r} (ν) c_{r} (- ν)$ ; the symmetry $c_{- r} = (- 1)^{r} c_{r}$ handles negative indices. Hence the product is independent of $r$ and equals $c_{n_{0}} (ν) c_{n_{0}} (- ν)$ , with $n_{0} = 0$ in even parity and $n_{0} = 1$ in odd parity. Continuity and density from step 3.1 extend this K-type calculation to the composite operator on smooth vectors. The base eigenvalues are finite and nonzero at these regular parameters by step 1.1, so dividing each factor by its base eigenvalue gives $\hat{A} (- ν) \hat{A} (ν) = i d$ and ${\hat{c}}_{r} (- ν) {\hat{c}}_{r} (ν) = 1$ .

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Unitarity of the unitary principal series

Statement

Assume the Axiom of Choice (The Axiom of Choice). For every $ε \in {0, 1}$ and every $ν \in i R$ the compact-picture action of $G$ on $L_{ε}^{2} (K)$ preserves the $L^{2}$ inner product and is strongly continuous; hence $I_{ε, ν}$ is a strongly continuous unitary representation of $G$ with $K$ -types $f_{n}$ , $n \equiv ε (2)$ , of multiplicity one. At $ν = 0$ the spherical representation $I_{0, 0}$ is irreducible (unitary spherical principal series). In odd parity the $K$ -finite core is $I_{1, 0}^{K} ≅ M_{1}^{-} \oplus M_{- 1}^{+}$ , and the Hilbert representation is the orthogonal direct sum of the irreducible unitary completions of these two limit-of-discrete-series modules.

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , $ν \in i R$ , and the compact-picture Hilbert space $L_{ε}^{2} (K)$ .

[F1]

For imaginary $ν$ , restriction to $K$ identifies the smooth compact picture with an isometric subspace of the right-covariant unitary-induction model by inversion and the $ρ^{- 1 / 2}$ half-density; the completed action is strongly continuous and unitary (The compact picture of the SL2(R) principal series(2)).

[F2]

The functions $f_{n} (k_{θ}) = e^{i n θ}$ , $n \equiv ε (m o d 2)$ , are an orthonormal basis of $L_{ε}^{2} (K)$ , and their finite spans are exactly the K-finite vectors (K-type decomposition of the SL2(R) principal series).

[F3]

At $ν = 0$ , the spherical compact-picture representation is irreducible; in odd parity its K-finite module splits into the positive chain $M_{1}^{-}$ with K-types $1, 3, 5, \dots$ and the negative chain $M_{- 1}^{+}$ with K-types $- 1, - 3, - 5, \dots$ (Generic irreducibility and the exceptional parameter lattice).

[F4]

In the compact picture at $ν = 0$ , $(Π_{0} (g) f) (k_{θ}) = ∣ α (p (k_{θ}, g)) ∣ f (k_{κ_{g} (θ)})$ , where $k_{θ} g = a_{t} n_{x} k_{κ_{g} (θ)}$ is the canonical $A N \times K$ factorization (Iwasawa and minimal-parabolic data for SL2(R), The compact picture of the SL2(R) principal series).

[F5]

Complex modulus is multiplicative and subadditive, so $∣ q z ∣ = ∣ q ∣ ∣ z ∣$ and $∣ 1 + q z ∣ \geq 1 - ∣ q z ∣ \geq 1 - ∣ q ∣$ (Conjugation is an involutive real-field automorphism, $z \overline{z} = ∣ z ∣^{2}$ , and modulus is definite, multiplicative, and subadditive). For $q \in C$ with $∣ q ∣ < 1$ and $∣ z ∣ \leq 1$ , the finite identity $(1 + q z)^{- 1} = \sum_{m = 0}^{N} (- q z)^{m} + (- q z)^{N + 1} / (1 + q z)$ is algebraic, and its remainder is bounded by $∣ q ∣^{N + 1} / (1 - ∣ q ∣)$ . The real geometric series with ratio $∣ q ∣$ converges by For $∣ r ∣ < 1$ , $\sum_{k \geq 0} r^{k} = 1 / (1 - r)$ , and for $∣ r ∣ \geq 1$ the series diverges, so these partial sums converge uniformly on the closed disk and have supremum at most $(1 - ∣ q ∣)^{- 1}$ . For every positive integer $j$ , the $j$ -fold powers of the partial sums are polynomials in $z$ with only nonnegative powers and converge uniformly to $(1 + q z)^{- j}$ : use $∣ u^{j} - v^{j} ∣ \leq j M^{j - 1} ∣ u - v ∣$ when $∣ u ∣, ∣ v ∣ \leq M = (1 - ∣ q ∣)^{- 1}$ (algebra).

[F6]

A nonzero closed $G$ -invariant subspace of this Hilbert model contains a nonzero K-finite vector, and its K-finite intersection is a $(g, K)$ -submodule (K-finite vectors detect nonzero closed invariant subspaces).

[F7]

A strongly continuous unitary representation is a homomorphism into unitary operators whose orbit maps are norm-continuous (Strongly continuous unitary representations, invariant linear subspaces and intertwiners).

[A1]

AC supplies the normalized Haar probability on $K$ and is inherited through the unitary compact-picture construction; no additional choice is used (The Axiom of Choice, [F1]).

Proof

technique · transfer unitarity through the exact compact-picture model, then identify the odd zero-parameter summands by their Hardy-type Fourier spaces

1.1F1F2F7A1

For every $ν \in i R$ , [F1] gives the strongly continuous unitary compact-picture action on $L_{ε}^{2} (K)$ ; the inversion and $ρ^{- 1 / 2}$ map in that supplier is the model equivalence, so no left/right covariance convention is silently identified. By [F2], the K-types are the one-dimensional mutually orthogonal lines $C f_{n}$ , $n \equiv ε (m o d 2)$ , and their finite span is dense. By [F7] this is a strongly continuous unitary representation with K-type multiplicity one.

1.2F3algebra

At $ε = 0$ , $W_{0}$ consists of odd integers, so $0 \notin W_{0}$ . Part (a) of [F3] therefore says that the Hilbert compact-picture representation $I_{0, 0}$ is irreducible.

1.3F2construct

At $ε = 1$ and $ν = 0$ , set $z = - e^{2 i θ}$ . The odd positive Fourier polynomials have the form $f (k_{θ}) = e^{i θ} F (z)$ with $F$ a polynomial; their closure $H_{+}$ is the closed span of $f_{1}, f_{3}, f_{5}, \dots$ . The negative odd Fourier polynomials have the form $\overline{f (k_{θ})}$ for such positive polynomials; let their closure be $H_{-}$ . By [F2], $H_{+}$ and $H_{-}$ are orthogonal and $L_{1}^{2} (K) = H_{+} \oplus H_{-}$ .

1.4F4F5algebra

Put $C = 2^{- 1 / 2} (\begin{array}{cc} 1 & i \\ 1 & - i \end{array})$ and let $v_{θ} = (- \sin θ, \cos θ)^{T}$ . Direct multiplication gives $C v_{θ} = (i e^{i θ}, - i e^{- i θ})^{T} / \sqrt{2}$ , so the ratio of its coordinates is $z = - e^{2 i θ}$ . For $g \in G$ , direct multiplication using $\det g = 1$ gives $C g^{T} C^{- 1} = (\begin{array}{cc} a & b \\ \overline{b} & \overline{a} \end{array})$ with $∣ a ∣^{2} - ∣ b ∣^{2} = 1$ . The row action $v_{θ}^{T} g$ therefore sends $z$ to $Φ_{g} (z) = (a z + b) / (\overline{b} z + \overline{a})$ ; its derivative is $Φ_{g}^{'} (z) = (\overline{b} z + \overline{a})^{- 2}$ . Since $∣ a ∣^{2} - ∣ b ∣^{2} = 1$ and both moduli are nonnegative, $∣ a ∣ > ∣ b ∣$ , so the denominator has no zero on the closed disk.

2.1F4F5step 1.4algebra

On the boundary, $z = - e^{2 i θ}$ and $Φ_{g} (z) = - e^{2 i κ_{g} (θ)}$ . Differentiating in $θ$ gives $∣ κ_{g}^{'} (θ) ∣ = ∣ Φ_{g}^{'} (z) ∣ = ∣ \overline{b} z + \overline{a} ∣^{- 2}$ . To determine the sign, write the bottom row of $k_{θ} g$ as the column $u (θ) = g^{T} v_{θ}$ , where $v_{θ} = (- \sin θ, \cos θ)^{T}$ . For $k_{θ} g = a_{t (θ)} n_{x (θ)} k_{κ_{g} (θ)}$ one has $u = e^{- t / 2} v_{κ_{g} (θ)}$ . Since $\det g = 1$ and $\det (v_{θ}, v_{θ}^{'}) = 1$ , we have $\det (u, u^{'}) = 1$ ; the factorized expression gives $\det (u, u^{'}) = e^{- t} κ_{g}^{'}$ . Therefore $κ_{g}^{'} = e^{t} > 0$ , and the boundary derivative identity gives $κ_{g}^{'} = ∣ Φ_{g}^{'} (z) ∣ = ∣ \overline{b} z + \overline{a} ∣^{- 2} = ∣ α (p (k_{θ}, g)) ∣^{2}$ . On $∣ z ∣ = 1$ , $a z + b = z \overline{(\overline{b} z + \overline{a})}$ , so $e^{2 i (κ_{g} (θ) - θ)} = \overline{d} / d$ for $d = \overline{b} z + \overline{a}$ and $e^{i κ_{g} (θ)} = \pm e^{i θ} \overline{d} / ∣ d ∣$ with a constant sign on the circle. Using $∣ α (p (k_{θ}, g)) ∣ = e^{t / 2} = ∣ d ∣^{- 1}$ , if $f (k_{θ}) = e^{i θ} F (z)$ then [F4] gives $(Π_{0} (g) f) (k_{θ}) = \pm e^{i θ} (\overline{b} z + \overline{a})^{- 1} F (Φ_{g} (z))$ .

3.1F1F4F5step 2.1step 1.3

For polynomial $F$ , each term of $(\overline{b} z + \overline{a})^{- 1} F (Φ_{g} (z))$ is a polynomial divided by a positive integer power of $\overline{a} + \overline{b} z$ . Since $∣ \overline{b} / \overline{a} ∣ < 1$ , [F5] expands each reciprocal power uniformly on $∣ z ∣ \leq 1$ as a series with only nonnegative powers of $z$ . Thus $Π_{0} (g)$ maps positive odd Fourier polynomials into $H_{+}$ . The action is unitary by [F1], so approximation by these polynomials and closedness give $Π_{0} (g) H_{+} \subseteq H_{+}$ ; applying the same argument to $g^{- 1}$ gives equality. At $ν = 0$ the cocycle and the odd inducing sign are real, so complex conjugation commutes with $Π_{0} (g)$ ; consequently $H_{-}$ is also invariant.

4.1F2F3F6step 1.3step 3.1∎

By [F3], the K-finite parts of $H_{+}$ and $H_{-}$ are exactly $M_{1}^{-}$ and $M_{- 1}^{+}$ . Each is algebraically irreducible by the chain argument in [F3]. If a closed invariant subspace of either summand is nonzero, [F6] puts a nonzero K-finite vector in it; irreducibility then gives the whole corresponding chain, which is dense in that summand. Hence both invariant Hilbert summands are irreducible unitary limits of discrete series, and their orthogonal sum is $I_{1, 0}$ .

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Unitarity of the complementary series

Statement

Assume the Axiom of Choice (The Axiom of Choice) and let $ν \in R$ . In spherical parity define $a_{0} (ν) = 1, a_{\pm 2 j} (ν) = \prod_{l = 1}^{j} \frac{2 l - 1 - ν}{2 l - 1 + ν} (j \geq 1) .$ For every real $ν \notin {- 1, - 3, - 5, \dots}$ , the Fourier form $B_{ν} (f, h) = \sum_{n \in 2 Z} a_{n} (ν) \hat{f} (n) \overline{\hat{h} (n)}$ is a finite continuous Hermitian $G$ -invariant form on $C_{0}^{\infty} (K)$ . It agrees with $⟨ A (ν) f, h ⟩_{0} / c_{0} (ν)$ wherever that quotient is defined and with its regular meromorphic continuation at the common scalar poles, including $ν = 0$ . The full spherical K-finite module $I_{0, ν}^{K}$ admits a positive-definite invariant Hermitian form (in the $(g, K)$ sense) if and only if $∣ ν ∣ < 1$ . For $0 < ∣ ν ∣ < 1$ , $B_{ν}$ is positive definite and its Hilbert completion is the irreducible strongly continuous unitary spherical complementary series; at $ν = 0$ , $B_{0}$ is the usual $L^{2}$ form and the representation is the spherical unitary principal series. At regular real $ν$ with $∣ ν ∣ > 1$ , $a_{0}$ and $a_{2}$ have opposite signs, so $B_{ν}$ is indefinite. At negative odd integers the normalized weights have poles, but the invariant-form recurrence still rules out a positive-definite form on the full K-finite spherical module.

At $ν = 1$ , the regular normalized form has $a_{0} = 1$ and every nonzero even weight zero. At $ν = - 1$ , the rescaled limit $q_{n} = \lim_{ν \to - 1^{+}} (1 + ν) a_{n} (ν), q_{0} = 0, q_{\pm 2 j} = 2 j (j \geq 1)$ is a different nonzero degenerate $G$ -invariant form. The $ν = 1$ form detects the trivial quotient; the $ν = - 1$ limiting form has radical $C f_{0}$ , the trivial submodule. No signature claim is made for other exceptional rescaled forms.

For odd parity and real $ν \neq 0$ , if $B$ is an invariant Hermitian form and $b_{n} = B (f_{n}, f_{n})$ , the recurrence gives $b_{1} = - b_{- 1}$ , so no positive-definite invariant form exists on the full odd K-finite module and there is no nonspherical complementary series. At a nonexceptional real $ν$ , every nonzero invariant Hermitian form on that odd module is nondegenerate and indefinite. At positive even $ν = 2 m > 0$ , its tail weights $∣ n ∣ \geq 2 m + 1$ vanish; at negative even $ν = - 2 m < 0$ , its central weights $∣ n ∣ \leq 2 m - 1$ vanish, so every such form is degenerate at these nonzero exceptional parameters. At odd $ν = 0$ , the K-finite module splits into the two unitary limits of discrete series as in Unitarity of the unitary principal series.

Facts & Assumptions

Given: AC, the normalized principal-series models for real $ν$ , their compact-picture smooth vectors, and the parity parameter $ε \in {0, 1}$ .

[F1]

The compact-picture action is $(Π_{ν} (g) f) (k) = ∣ α (p (k, g)) ∣^{1 + ν} f (κ (k, g))$ in the canonical $A N \times K$ factorization and is smooth in the group and compact variables (The compact picture of the SL2(R) principal series, The normalized principal series I(epsilon, nu)).

[F2]

The Fourier vectors $f_{n} (k_{θ}) = e^{i n θ}$ , $n \equiv ε (m o d 2)$ , form an orthonormal basis of $L_{ε}^{2} (K)$ and their finite spans are the K-finite core (K-type decomposition of the SL2(R) principal series, Fourier coefficients and trigonometric polynomials on the torus).

[F3]

For real $ν$ , $⟨ u, h ⟩_{0} = \int_{K} u (k) \overline{h (k)} d k$ is a G-invariant pairing between $I_{ε, - ν}$ and $I_{ε, ν}$ (The invariant pairing between opposite principal-series parameters).

[F4]

The standard integral defines $A (ν)$ for $Re ν > 0$ ; its meromorphic family is continuous and K-diagonal on smooth vectors and intertwines $Π_{ν}$ with $Π_{- ν}$ wherever regular (The standard intertwining operator A(nu), Meromorphic continuation and intertwining identity for A(nu)). The quotient by $c_{0} (ν)$ at common poles is proved locally in Step 2.2.

[F5]

The eigenvalues satisfy $(n + 1 + ν) c_{n + 2} = (n + 1 - ν) c_{n}$ , $c_{- n} = (- 1)^{n} c_{n}$ , and the displayed Gamma formula with its exact exceptional zeros and poles (K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing).

[F6]

The real derived action is $L_{E_{+}} f_{n} = (1 + ν + n) f_{n + 2} / 2$ and $L_{E_{-}} f_{n} = (1 + ν - n) f_{n - 2} / 2$ . For an invariant Hermitian form on the K-finite module, infinitesimal invariance makes real generators skew-adjoint; since $E_{-} = \overline{E_{+}}$ , sesquilinearity gives $B (L_{E_{+}} u, v) = - B (u, L_{E_{-}} v)$ (Derived action and raising/lowering formulas in the compact picture).

[F7]

At $ν = 0$ , the spherical module is irreducible and the odd K-finite module splits into the two chains $M_{1}^{-}$ and $M_{- 1}^{+}$ ; at $ν = 1$ the trivial representation is the finite-dimensional quotient, and at $ν = - 1$ it is the submodule $C f_{0}$ (Generic irreducibility and the exceptional parameter lattice).

[F8]

The compact-picture action at imaginary parameter is a strongly continuous unitary representation in the sense of Strongly continuous unitary representations, invariant linear subspaces and intertwiners; at $ν = 0$ the odd representation is the direct sum of the two unitary limits of discrete series (Unitarity of the unitary principal series).

[F9]

Complex integration by parts on a period interval bounds Fourier coefficients of a smooth function by $C_{N} ∣ n ∣^{- N}$ , and $\sum_{n \geq 1} n^{- p}$ converges for rational $p > 1$ (Complex integration by parts on intervals and decaying lines, For rational $p > 0$ , $\sum 1 / k^{p}$ converges iff $p > 1$ ).

[F10]

AC supplies countable choice for the complex integration-by-parts supplier (AC supplies the countable and dependent choices used in Banach integration, The Axiom of Choice).

[F11]

The compact-picture action and its parameter depend continuously in each $C^{r}$ seminorm; the product and chain rules compute derivatives of its multiplier and composed argument (The compact picture of the SL2(R) principal series, Sums, scalar multiples, products and quotients: $(f + g)^{'} (c) = f^{'} (c) + g^{'} (c)$ , $(α f)^{'} (c) = α f^{'} (c)$ , $(f g)^{'} (c) = f^{'} (c) g (c) + f (c) g^{'} (c)$ , and $(f / g)^{'} (c) = (f^{'} (c) g (c) - f (c) g^{'} (c)) / g (c)^{2}$ when $g (c) \neq 0$ , The chain rule, in one line from Carathéodory: if $g$ is differentiable at $c$ and $f$ is differentiable at $g (c)$ , then $f \circ g$ is differentiable at $c$ with $(f \circ g)^{'} (c) = f^{'} (g (c)) g^{'} (c)$ ).

[F12]

For fixed $g$ and smooth $f$ , the function $ν \mapsto p_{N} (Π_{ν} (g) f)$ is continuous on a compact real parameter interval by [F11], hence bounded there by the extreme-value theorem (A continuous real-valued function on a nonempty compact metric space is bounded and attains a greatest and a least value). Thus the Fourier-decay bounds in [F9] can be chosen uniformly on such an interval.

[A1]

AC supplies normalized Haar probability on $K$ through the compact-picture/Fourier suppliers and implies the countable-choice hypothesis in [F9] through [F10]. The proof makes no further selections (The Axiom of Choice, [F10]).

Proof

technique · solve the invariant-form recurrence, prove the smooth multiplier bounds, and pass the resulting invariant forms to their Hilbert completions and endpoint limits

1.1F5algebra

In spherical parity put $a_{0} (ν) = 1$ . The Gamma formula in [F5], using $Γ (z + 1) = z Γ (z)$ , gives the cross-multiplied recurrence as a meromorphic identity for every $ν$ , so it may be used even at its zero denominators. For $ν \notin {- 1, - 3, \dots}$ , successive use at $n = 0, 2, \dots, 2 j - 2$ gives $a_{2 j} = \prod_{l = 1}^{j} (2 l - 1 - ν) / (2 l - 1 + ν)$ . The meromorphic symmetry $c_{- n} = (- 1)^{n} c_{n}$ gives $a_{- 2 j} = a_{2 j}$ . At a positive odd $ν = 2 m + 1$ with $m \geq 0$ , the numerator $2 l - 1 - ν$ vanishes at $l = m + 1$ , so every weight with $j \geq m + 1$ is zero; a denominator vanishes exactly at a negative odd integer. Thus the displayed weights are finite precisely on the real parameter set in the Statement.

1.2F6algebra

Let $B$ be any positive-definite invariant Hermitian form on the full spherical K-finite module, and put $b_{n} = B (f_{n}, f_{n}) > 0$ . K-invariance makes distinct K-types orthogonal. By the infinitesimal invariance relation [F6], $B (L_{E_{+}} f_{0}, f_{2}) = - B (f_{0}, L_{E_{-}} f_{2})$ , so the ladder formulas give $(1 + ν) b_{2} = (1 - ν) b_{0}$ . If $ν = - 1$ , the left side is zero and the right side positive; if $ν < - 1$ , the left coefficient is negative and the right positive; if $ν = 1$ , the left is positive while the right is zero; and if $ν > 1$ , the right side is negative. Hence such a form can exist only when $∣ ν ∣ < 1$ .

1.3F5F6algebra

In odd parity, any invariant Hermitian form on the K-finite module is K-diagonal with real weights $b_{n} = B (f_{n}, f_{n})$ . Applying [F6] to each adjacent pair gives $(1 + ν + n) b_{n + 2} = (n + 1 - ν) b_{n}$ ; at $n = - 1$ this is $ν b_{1} = - ν b_{- 1}$ . For real $ν \neq 0$ , a positive-definite form would have $b_{1}, b_{- 1} > 0$ , contradicting this identity. If $ν$ is nonexceptional and the form is nonzero, K-diagonality gives some nonzero $b_{n}$ ; no ladder coefficient vanishes, so the recurrence propagates this to $b_{1} \neq 0$ and then to every odd weight. Since $b_{- 1} = - b_{1}$ , the form is nondegenerate and indefinite.

1.4F7F8F5

At $ε = 1$ , $ν = 0$ , the K-finite odd module splits into the two unitary limit chains by [F8]; the recurrence at $n = - 1$ is $0 = 0$ and leaves their invariant weights independent. This is the separate zero-parameter unitary splitting, not an odd complementary series.

2.1F2F9F10A1step 1.1algebra

Fix a real $ν \notin {- 1, - 3, \dots}$ . For all sufficiently large $l$ , $∣ (2 l - 1 - ν) / (2 l - 1 + ν) ∣ \leq 1 + C / l$ for a constant depending on $ν$ ; bound the finitely many earlier factors separately. Choose an integer $M \geq C$ ; then $\prod_{l = 1}^{j} (1 + M / l) = (\binom{j + M}{M}) \leq (j + M)^{M}$ , so $∣ a_{n} (ν) ∣ \leq C^{'} (1 + ∣ n ∣)^{M}$ . If an initial factor is zero the subsequent weights are zero and the same estimate holds. By [F9], $∣ \hat{f} (n) ∣ \leq C_{N} p_{N} (f) ∣ n ∣^{- N}$ for $n \neq 0$ . Choose $N$ with $2 N - M > 1$ ; the p-series bound then gives $∣ B_{ν} (f, h) ∣ \leq C p_{N} (f) p_{N} (h)$ , so the Fourier form is finite and continuous. Its weights are real and symmetric, hence it is Hermitian.

2.2F2F3F4F5F11step 1.1

Define $R_{ν}$ as the smooth K-diagonal multiplier with coefficients $a_{n} (ν)$ . If $ν$ is not a common scalar pole and $c_{0} (ν) \neq 0$ , it equals $A (ν) / c_{0} (ν)$ ; the smooth continuity follows from [F4]. At $ν_{0} = - 2 m$ , $m \geq 0$ , the Gamma formula in [F5] has a simple pole with nonzero residue in its numerator, while both denominator arguments for every even $n$ are half-integers and finite. Thus all $c_{n}$ , including $c_{0}$ , have the same simple pole and $(ν - ν_{0}) c_{0} (ν)$ is holomorphic and nonzero at $ν_{0}$ . In the local Laurent expansion of the meromorphic K-diagonal family [F4], every coefficient below degree $- 1$ has zero multiplier on each K-type because each scalar $c_{n}$ has at most a simple pole by [F5]. Such a continuous K-diagonal coefficient sends every smooth vector to a smooth function with all Fourier coefficients zero, hence is zero by completeness in [F2]. Therefore $(ν - ν_{0}) A (ν)$ is holomorphic in the smooth operator topology, and dividing by $(ν - ν_{0}) c_{0} (ν)$ defines the continuous extension of $R_{ν}$ there; its K-type multipliers are exactly those in step 1.1. At every real parameter under consideration, [F4] gives the intertwining identity after division when $c_{0} \neq 0$ ; at the common poles pass to the limit from neighboring regular parameters, using [F11] for continuity of the compact-picture action in $ν$ . Therefore $R_{ν}$ intertwines $Π_{ν}$ with $Π_{- ν}$ for every $ν$ in the stated real domain. By [F2], $B_{ν} (f, h) = ⟨ R_{ν} f, h ⟩_{0}$ ; [F3] and the intertwining identity give $B_{ν} (Π_{ν} (g) f, Π_{ν} (g) h) = B_{ν} (f, h)$ for every $g \in G$ .

2.3F2step 1.1algebra

If $∣ ν ∣ < 1$ , every numerator and denominator in the spherical weight product is positive, so all $a_{n} (ν) > 0$ . Fourier completeness [F2] then makes $B_{ν}$ positive definite; at $ν = 0$ all weights equal one and $B_{0}$ is the ordinary $L^{2}$ inner product.

2.4F5step 1.1algebra

At a regular real parameter with $∣ ν ∣ > 1$ , the displayed normalized form has $a_{0} = 1$ and $a_{2} = (1 - ν) / (1 + ν) < 0$ , so it is indefinite. Negative odd parameters are excluded from this assertion because the normalized weights have poles there; the recurrence argument in step 1.2 still rules out any positive-definite full-module form at those parameters.

2.5F6step 1.3algebra

Let $ν = 2 m > 0$ . The recurrence from [F6] at $n = 2 m - 1$ forces $b_{2 m + 1} = 0$ and then all upper tail weights vanish; at $n = - 2 m - 1$ it forces $b_{- 2 m - 1} = 0$ and then all lower tail weights vanish. If $ν = - 2 m < 0$ , the same two recurrence equations force $b_{2 m - 1} = b_{- 2 m + 1} = 0$ , propagating through every central odd weight $∣ n ∣ \leq 2 m - 1$ . Thus every invariant form at either nonzero even exceptional parameter is degenerate. These equations force zeros, not signs on the remaining chains, so no blanket indefiniteness conclusion is asserted.

3.1F2F7step 2.2

At $ν = 1$ , the product in step 1.1 has $a_{0} = 1$ and $a_{\pm 2 j} = 0$ for every $j \geq 1$ . Hence $B_{1} (f, h) = \hat{f} (0) \overline{\hat{h} (0)}$ . It is a finite nonzero degenerate invariant form by steps 2.1–2.2. Its radical in $C_{0}^{\infty} (K)$ is ${f : \hat{f} (0) = 0}$ , whose K-finite part is the algebraic span of the nonzero even K-types; the quotient is one-dimensional, and [F7] identifies its K-finite quotient with the trivial module $L_{0}$ .

3.2F7F9F12step 2.2algebra

For $- 1 < ν \leq 0$ and $j \geq 1$ , rewrite $(1 + ν) a_{2 j} (ν) = (1 - ν) \prod_{l = 2}^{j} (2 l - 1 - ν) / (2 l - 1 + ν)$ . Each factor in the product is at most $l / (l - 1)$ , so $0 < (1 + ν) a_{2 j} (ν) \leq 2 j$ . The $j = 0$ weight $(1 + ν) a_{0}$ tends to zero, while for $j \geq 1$ the product tends to $2 j$ as $ν \to - 1^{+}$ . By [F9] and this bound, the rescaled Fourier forms converge on smooth vectors to the finite nonzero form with weights $b_{0} = 0$ and $b_{\pm 2 j} = 2 j$ . To pass invariance to the rescaled limit, [F12] bounds the Fourier-decay seminorms of $Π_{ν} (g) f$ and $Π_{ν} (g) h$ uniformly for $ν \in [- 1, 0]$ ; hence the same summable majorant applies to both sides of the invariance identity. The limit is invariant, its radical is exactly $C f_{0}$ , and [F7] identifies that line as the trivial submodule.

4.1F1F2F6F9step 2.1step 2.3algebra∎

Suppose $0 < ∣ ν ∣ < 1$ . The polynomial bound in step 2.1 and Fourier decay in [F9] give a finite $r$ and $C$ with $∥ f ∥_{B_{ν}} \leq C p_{r} (f)$ for every smooth $f$ . The smooth compact action is continuous in each $C^{r}$ seminorm by [F1, F11], so every smooth vector has a continuous orbit in the $B_{ν}$ norm. Invariance makes $Π_{ν} (g)$ an isometry with inverse $Π_{ν} (g^{- 1})$ ; it extends to a unitary on the completion, and density plus the isometry bound extends strong continuity to every completed vector. For irreducibility, the completion is the weighted $ℓ^{2}$ completion of the even Fourier modes. If $W$ is a nonzero closed invariant subspace, choose $ξ = \sum_{n \in 2 Z} x_{n} f_{n} \neq 0$ in $W$ and an $n$ with $x_{n} \neq 0$ . For integers $N > ∣ n ∣$ , the finite average $Q_{N} ξ = N^{- 1} \sum_{j = 0}^{N - 1} e^{- 2 π i n j / N} Π_{ν} (k_{2 π j / N}) ξ$ lies in $W$ and retains exactly the modes $r \equiv n (m o d N)$ . As $N \to \infty$ , all retained modes other than $n$ lie in the weighted $ℓ^{2}$ tail $∣ r ∣ \geq N - ∣ n ∣$ , so $Q_{N} ξ \to x_{n} f_{n}$ . Hence $f_{n} \in W$ . Smooth difference quotients in the $B_{ν}$ norm put both ladder images in $W$ ; for $∣ ν ∣ < 1$ all even ladder coefficients are nonzero, so $W$ contains every even K-type. Their finite span is dense by construction of the completion, giving $W$ equal to the full Hilbert space.

TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

Parameter-sign equivalence and its exceptional failures for SL2(R)

Statement

Assume the Axiom of Choice (The Axiom of Choice). Let $ε \in {0, 1}$ , $ν \notin W_{ε}$ (The normalized principal series I(epsilon, nu)), and $n_{0} = 0$ for $ε = 0$ and $n_{0} = 1$ for $ε = 1$ . The base-normalized meromorphic intertwiner $\hat{A} (ν) : = A (ν) / c_{n_{0}} (ν)$ extends through every common scalar pole of numerator and denominator to a continuous $K$ -diagonal isomorphism $I_{ε, ν}^{K} \to I_{ε, - ν}^{K}$ with inverse $\hat{A} (- ν)$ ; hence the smooth compact-picture representations $I_{ε, ν}$ and $I_{ε, - ν}$ are continuously equivalent. At every nonexceptional parameter where the unnormalized $A (ν)$ is regular, $c_{n_{0}} (ν)$ is finite and nonzero, so $A (ν)$ is a nonzero scalar multiple of $\hat{A} (ν)$ and is itself an isomorphism.

If $ν \in i R ∖ W_{ε}$ , then $\hat{A} (ν)$ is a unitary intertwiner and $∣ {\hat{c}}_{r} (ν) ∣ = 1$ for every allowed parity- $ε$ $K$ -type; hence the unitary principal series at $ν$ and $- ν$ are unitarily equivalent. Complex conjugation of the compact picture is also an anti-unitary intertwiner between them. For $ε = 0$ and nonzero imaginary $ν$ , this is the usual normalization $A (ν) / c_{0} (ν)$ .

At an exceptional parameter $ν \in W_{ε}$ , $ν \neq 0$ , write $n = ∣ ν ∣$ . The unnormalized $A (ν)$ is regular but not injective: by K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing, at $ν = n$ it kills the two tail submodules with $∣ r ∣ \geq n + 1$ , while at $ν = - n$ it kills the finite-dimensional submodule with $∣ r ∣ \leq n - 1$ . Moreover, $I_{ε, n}$ and $I_{ε, - n}$ are not isomorphic as $(g, K)$ -modules, and therefore are not continuously equivalent as smooth $G$ -representations: the former has irreducible tail submodules, whereas the latter has the finite-dimensional submodule as its unique irreducible submodule, and their $K$ -type supports are disjoint. At $ε = 1, ν = 0$ , the two representations coincide and the odd K-finite module splits into the two limit chains by Generic irreducibility and the exceptional parameter lattice(b). Thus $I_{ε, ν} ≅ I_{ε, - ν}$ exactly when $ν \notin W_{ε} ∖ {0}$ .

Facts & Assumptions

Given: AC, $ε \in {0, 1}$ , the smooth principal-series models $I_{ε, ν}$ , and the normalized parameter lattice $W_{ε}$ .

[F1]

The integral $A (ν)$ and its compact-picture K-type eigenvalues $c_{r} (ν)$ are defined on an initial half-plane (The standard intertwining operator A(nu)). The meromorphic family is continuous K-diagonal on smooth vectors, its proof step 1.1 locates the common simple scalar poles in $P_{ε}$ , and it intertwines $Π_{ν}$ with $Π_{- ν}$ at every regular parameter (Meromorphic continuation and intertwining identity for A(nu)).

[F2]

The recurrence $(r + 1 + ν) c_{r + 2} (ν) = (r + 1 - ν) c_{r} (ν)$ , the symmetry $c_{- r} (ν) = (- 1)^{r} c_{r} (ν)$ , and the exact zeros at $ν = \pm n \in W_{ε}$ hold for each allowed $r$ . The Gamma formula has simple poles with nonzero residues, no zeros, and reciprocal-Gamma zeros that cancel the common numerator poles outside $P_{ε}$ (K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing).

[F3]

For $n \in W_{ε}$ , $n \geq 1$ , the finite-dimensional $L_{n - 1}$ is the unique irreducible quotient at $ν = n$ and the unique irreducible submodule at $ν = - n$ ; its K-types are $- (n - 1), - (n - 3), \dots, n - 1$ . At $(ε, ν) = (1, 0)$ the odd module splits into the two limit chains (Generic irreducibility and the exceptional parameter lattice).

[F4]

The compact-picture action is $(Π_{ν} (g) f) (k) = ∣ α (p (k, g)) ∣^{1 + ν} f (κ (k, g))$ in the canonical $A N \times K$ factorization (The compact picture of the SL2(R) principal series, The normalized principal series I(epsilon, nu)). For $ν \in i R$ this is a strongly continuous unitary representation (Unitarity of the unitary principal series).

[F5]

On $L_{ε}^{2} (K)$ the vectors $f_{r} (k_{θ}) = e^{i r θ}$ , $r \equiv ε (m o d 2)$ , form a complete orthonormal basis; the finite Fourier sums are precisely the K-finite core (K-type decomposition of the SL2(R) principal series).

[F6]

On the K-type $f_{r}$ , the derived operators are $L_{W} f_{r} = r f_{r}$ , $L_{E_{+}} f_{r} = (1 + ν + r) f_{r + 2} / 2$ , and $L_{E_{-}} f_{r} = (1 + ν - r) f_{r - 2} / 2$ . Differentiating a continuous smooth group intertwining identity along real one-parameter subgroups, then complexifying, gives a $(g, K)$ -module map (Derived action and raising/lowering formulas in the compact picture).

[F7]

A unitary representation is strongly continuous, and an equivalence of unitary representations is a unitary intertwiner (Strongly continuous unitary representations, invariant linear subspaces and intertwiners).

[A1]

AC supplies the normalized Haar probability $d k = d θ / (2 π)$ and the Hilbert compact-picture realization through [F4]–[F5], and is stated explicitly at the beginning. The proof makes no additional choices (The Axiom of Choice).

Proof

technique · normalize away the common scalar poles, use the exact K-type recurrence for the inverse, and compare the exceptional composition-series orientations separately

1.1F1F2algebra

Let $P_{ε} = {- m : m \in Z_{\geq 0}, m \equiv ε (m o d 2)}$ be the common scalar pole set from [F1]. It is disjoint from $W_{ε}$ . For $ν \notin W_{ε} \cup P_{ε}$ , [F2] shows that $c_{n_{0}} (ν)$ is finite and nonzero, so $\hat{A} (ν) = A (ν) / c_{n_{0}} (ν)$ is a continuous K-diagonal map. At $ν_{0} = - m \in P_{ε}$ , the Gamma formula in [F2] has one numerator Gamma factor with a simple pole of nonzero residue, while both denominator arguments are half-integers and hence finite and nonzero for every allowed $r$ . Thus every $c_{r}$ , including $c_{n_{0}}$ , has the same simple pole and $h (ν) c_{n_{0}} (ν)$ is holomorphic and nonzero at $ν_{0}$ , where $h (ν) = ν - ν_{0}$ . By [F1], $h A (ν)$ is holomorphic in the smooth operator topology there. Therefore $\hat{A} = (h A) / (h c_{n_{0}})$ extends continuously through that pole as a K-diagonal map.

1.2F1F2algebra

Let $ν \in W_{ε} ∖ {0}$ and write $n = ∣ ν ∣ \geq 1$ . The pole set $P_{ε}$ has parity $ε$ and is disjoint from $W_{ε}$ , so $A (ν)$ is regular. If $ν = n$ , [F2] gives $c_{r} (n) = 0$ on every allowed type with $∣ r ∣ \geq n + 1$ ; if $ν = - n$ , it gives $c_{r} (- n) = 0$ for $∣ r ∣ \leq n - 1$ . In either case a nonzero K-type is killed, so $A (ν)$ is not injective.

2.1F2step 1.1algebra

For every allowed $r$ , normalize the recurrence in [F2] to obtain ${\hat{c}}_{r + 2} (ν) = {\hat{c}}_{r} (ν) (r + 1 - ν) / (r + 1 + ν)$ and ${\hat{c}}_{- r} (ν) = (- 1)^{r} {\hat{c}}_{r} (ν)$ , with ${\hat{c}}_{n_{0}} = 1$ . If $ν \notin W_{ε}$ , neither $r + 1 + ν$ nor $r + 1 - ν$ vanishes for any allowed $r$ , since $r + 1$ has the opposite parity. Thus every multiplier is finite and nonzero, including at $P_{ε}$ by step 1.1.

3.1A1F1F2F5step 2.1

Applying the recurrence at $ν$ and $- ν$ gives ${\hat{c}}_{r + 2} (ν) {\hat{c}}_{r + 2} (- ν) = {\hat{c}}_{r} (ν) {\hat{c}}_{r} (- ν)$ ; the base value is $1$ and the symmetry handles negative indices. Therefore $\hat{A} (- ν) \hat{A} (ν)$ is the identity on every K-type. Both operators are continuous and K-diagonal by step 1.1; for smooth $f$ , every Fourier coefficient of $(\hat{A} (- ν) \hat{A} (ν) - I) f$ is zero, so [F5] makes that vector zero in $L^{2}$ . A smooth function that is zero almost everywhere is zero everywhere, since a nonzero value would by continuity give a nonempty open arc on which its modulus is bounded below, and every open arc has positive normalized Haar measure. The same argument applies in the reverse order, proving that the maps are continuous inverses on $C_{ε}^{\infty} (K)$ .

4.1F1F2F4step 1.1step 3.1

For $ν \notin W_{ε} \cup P_{ε}$ , [F1] gives the G-intertwining identity for the regular operator $A (ν)$ , and [F2] gives $0 \neq c_{n_{0}} (ν)$ ; division yields the identity for $\hat{A} (ν)$ . At a pole $ν_{0} \in P_{ε}$ , take regular $ν_{j} \to ν_{0}$ outside $W_{ε} \cup P_{ε} \cup (- P_{ε})$ ; step 1.1 gives convergence of $\hat{A} (ν_{j})$ in the smooth-operator topology. For fixed $g$ , the compact formula in [F4] depends continuously on $ν$ in every smooth seminorm, because $\log ∣ α (p (k, g)) ∣$ is smooth and bounded on compact $K$ . Passing to the limit in the intertwining identity gives it at $ν_{0}$ as well. Consequently the inverse maps of step 3.1 give $I_{ε, ν} ≅ I_{ε, - ν}$ for every $ν \notin W_{ε}$ . When in addition the unnormalized $A (ν)$ is regular, [F2] gives finite nonzero $c_{n_{0}} (ν)$ , so $A (ν) = c_{n_{0}} (ν) \hat{A} (ν)$ is an isomorphism.

5.1A1F2F4F5F7step 2.1step 4.1

Suppose $ν \in i R ∖ W_{ε}$ . Every recurrence factor $(r + 1 - ν) / (r + 1 + ν)$ has modulus one, and the base multiplier is ${\hat{c}}_{n_{0}} = 1$ ; the symmetry gives $∣ {\hat{c}}_{r} (ν) ∣ = 1$ on every allowed K-type. Thus the Fourier multiplier preserves the $L^{2}$ norm on finite Fourier sums and its inverse multiplier does too; by density [F5] it extends to a unitary operator on $L_{ε}^{2} (K)$ . Step 4.1 makes it a smooth G-intertwiner, so continuity of the two unitary actions and density extend the intertwining identity to all of $L_{ε}^{2} (K)$ . Separately, for $C f = \overline{f}$ , the compact-picture formula [F4] gives $C Π_{ν} (g) = Π_{- ν} (g) C$ because $∣ α (p (k, g)) ∣$ is positive real and $\overline{ν} = - ν$ ; thus $C$ is an anti-unitary intertwiner.

6.1F3F5F6step 1.2∎

Let $n \geq 1$ lie in $W_{ε}$ . In $I_{ε, n}^{K}$ put $T^{+} = ⨁_{j \geq 0} C f_{n + 1 + 2 j}$ . By [F6], $L_{E_{-}} f_{n + 1} = 0$ , while $E_{+}$ preserves these weights and both arrows are nonzero between every adjacent pair above the boundary; $K$ also preserves $T^{+}$ . Hence $T^{+}$ is a nonzero $(g, K)$ -submodule. It is simple: for any nonzero submodule choose a nonzero finite Fourier sum; a $k_{θ} \in K$ can be chosen with distinct eigenvalues on its finite set of K-types, since only finitely many angles cause a collision, and a polynomial in its action isolates one nonzero K-type. The nonzero ladder arrows then generate all of $T^{+}$ . If a $(g, K)$ -module isomorphism $T$ existed from $I_{ε, n}^{K}$ to $I_{ε, - n}^{K}$ , it would send $T^{+}$ to an irreducible submodule at $- n$ , which [F3] says must be $L_{n - 1}$ . But $L_{n - 1}$ has only K-types $- (n - 1), - (n - 3), \dots, n - 1$ , disjoint from the support of $T^{+}$ , contradicting K-equivariance. Thus the K-finite modules are not isomorphic. Any continuous G-equivalence of the smooth compact-picture representations restricts to a K-finite module isomorphism: continuity lets one differentiate $T Π_{n} (g) = Π_{- n} (g) T$ along real one-parameter subgroups, while K-commutation preserves K-finiteness. Therefore no continuous G-equivalence exists; exchanging $n$ and $- n$ gives the same conclusion in the reverse direction. The odd case $ν = 0$ is excluded: there $ν \in W_{1}$ and [F3] gives a direct sum of the two limit chains.

CorollaryStatement: Literature-sourcedProof: AI-adaptedprecheck passjudge pass (gpt-6.1-sol)audited 2026-10-08Open item page →

The spherical complementary series converge to the trivial representation

Statement

Assume the Axiom of Choice (The Axiom of Choice). For $0 < ν < 1$ let $φ_{ν} (g) = B_{ν} (Π_{ν} (g) f_{0}, f_{0})$ be the spherical function of the spherical complementary series $I_{0, ν}$ (the matrix coefficient of the $K$ -fixed vector $f_{0} = 1$ with respect to the normalized invariant form of Unitarity of the complementary series). Then $φ_{ν} (g) = \int_{K} ∣ α (p (k, g)) ∣^{1 - ν} d k,$ so $φ_{ν} \to 1$ uniformly on every compact subset of $G$ as $ν ↑ 1$ . Consequently, for any sequence $0 < ν_{j} < 1$ increasing to $1$ , the trivial representation is weakly contained in ${⨁^{^}}_{j} I_{0, ν_{j}}$ , and the classes of the spherical complementary series converge to the trivial class in the Fell topology of The Fell topology on the unitary dual as $ν ↑ 1$ (equivalently, by $I_{0, ν} ≅ I_{0, - ν}$ , as $∣ ν ∣ ↑ 1$ ).

Facts & Assumptions

Given: AC, $G = {S L}_{2} (R)$ , a real parameter $0 < ν < 1$ , the smooth compact-picture representation $I_{0, ν}$ , its normalized positive form $B_{ν}$ , and the constant vector $f_{0} = 1$ .

[F1]

In the compact picture, $(Π_{ν} (g) f) (k) = ∣ α (p (k, g)) ∣^{1 + ν} f (κ (k, g))$ , where $k g = p (k, g) κ (k, g)$ is the canonical positive-diagonal Iwasawa factorization and the factors depend continuously on $(k, g)$ (The compact picture of the SL2(R) principal series, The normalized principal series I(epsilon, nu), Iwasawa and minimal-parabolic data for SL2(R)).

[F2]

The constant vector $f_{0}$ is $K$ -fixed, has norm one for $B_{ν}$ , and the even Fourier vectors form the complete K-type basis (K-type decomposition of the SL2(R) principal series, Unitarity of the complementary series).

[F3]

The normalized form is $B_{ν} (f, h) = ⟨ R_{ν} f, h ⟩_{0}$ , where $R_{ν} = A (ν) / c_{0} (ν)$ is the smooth normalized intertwiner and $⟨ u, h ⟩_{0} = \int_{K} u (k) \overline{h (k)} d k$ is the invariant pairing between $I_{0, - ν}$ and $I_{0, ν}$ (Unitarity of the complementary series, The standard intertwining operator A(nu), The invariant pairing between opposite principal-series parameters).

[F4]

The normalized intertwiner is continuous, satisfies $R_{ν} Π_{ν} (g) = Π_{- ν} (g) R_{ν}$ , and has $R_{ν} f_{0} = f_{0}$ ; its spherical K-type multipliers are $a_{n} (ν)$ (Meromorphic continuation and intertwining identity for A(nu), Unitarity of the complementary series).

[F5]

For $0 < ν < 1$ , the completion in $B_{ν}$ is a strongly continuous irreducible unitary representation. The smooth Fourier vectors are dense in its weighted Hilbert completion (Unitarity of the complementary series, K-type decomposition of the SL2(R) principal series).

[F6]

A basic Fell neighborhood tests finitely many diagonal coefficients of one representation on a compact set, each uniformly within a positive tolerance of a finite sum of diagonal coefficients of the candidate representation. Every diagonal coefficient of the trivial representation is a nonnegative constant (The Fell topology on the unitary dual, The unitary dual of a locally compact group, Matrix coefficient of a unitary representation).

[F7]

For a strongly continuous unitary representation, the trivial representation is weakly contained exactly when the representation has almost invariant unit vectors uniformly on each compact subset; the Hilbert direct sum acts componentwise and remains strongly continuous (Weak containment of the trivial representation and almost invariant vectors, Weak containment of unitary representations, Hilbert direct sums of unitary representations).

[F8]

Outside the exceptional lattice, the base-normalized smooth intertwiner $R_{ν}$ has inverse $R_{- ν}$ (Parameter-sign equivalence and its exceptional failures for SL2(R)).

[F9]

The product of two compact topological spaces is compact (A product of finitely many compact spaces is compact in the product topology).

[F10]

The complex exponential has derivative itself and restricts to the real exponential; the real chain rule therefore gives $(e^{a t})^{'} = a e^{a t}$ for real $a$ . The real mean value theorem applies on every nondegenerate closed interval for this smooth function (The complex exponential is entire and its complex derivative is itself, $\exp (z + w) = \exp z \exp w$ , and the complex exponential extends the real exponential, The chain rule, in one line from Carathéodory: if $g$ is differentiable at $c$ and $f$ is differentiable at $g (c)$ , then $f \circ g$ is differentiable at $c$ with $(f \circ g)^{'} (c) = f^{'} (g (c)) g^{'} (c)$ , The mean value theorem, as the case $g (x) = x$ of Cauchy's: for $f$ continuous on $[a, b]$ with $a < b$ and differentiable on $(a, b)$ there is $c \in (a, b)$ with $f (b) - f (a) = f^{'} (c) (b - a)$ ).

[A1]

AC supplies the normalized Haar probability $d k$ through the compact-picture construction and is the declared hypothesis of the unitary-completion, Hilbert-direct-sum, Fell-topology and weak-containment suppliers. The sequence $(ν_{j})$ is given, and no further selection is made (The Axiom of Choice).

Proof

technique · identify the normalized spherical coefficient through the opposite-parameter intertwiner, prove compact-uniform convergence directly, and apply the coefficient and almost-invariant-vector definitions

1.1F1F2F3F4

Put $R_{ν} = A (ν) / c_{0} (ν)$ on the smooth compact picture. By [F3]–[F4], $B_{ν} (f, h) = ⟨ R_{ν} f, h ⟩_{0}$ , $R_{ν} f_{0} = f_{0}$ , and $R_{ν} Π_{ν} (g) = Π_{- ν} (g) R_{ν}$ . Therefore $φ_{ν} (g) = ⟨ Π_{- ν} (g) f_{0}, f_{0} ⟩_{0}$ . Substituting the compact-picture action at parameter $- ν$ and $f_{0} = 1$ gives $φ_{ν} (g) = \int_{K} ∣ α (p (k, g)) ∣^{1 - ν} d k$ , with the exponent $1 - ν$ coming from the intertwiner rather than the original $1 + ν$ action.

1.2F4F5F8algebra

For $0 < ν < 1$ , the normalized intertwiner has $R_{ν} f_{n} = a_{n} (ν) f_{n}$ , and the displayed product weights satisfy $a_{n} (- ν) = a_{n} (ν)^{- 1}$ for every even $n$ . Consequently, on finite Fourier sums, $B_{- ν} (R_{ν} f, R_{ν} h) = B_{ν} (f, h)$ . The finite Fourier sums are dense in both Hilbert completions by [F5]; the inverse $R_{- ν}$ from [F8] gives surjectivity. Thus $R_{ν}$ extends to a unitary operator between the completions. The smooth intertwining identity [F4] extends to the completions by density, since both representation operators and the extended intertwiner are bounded. Hence the two unitary representations are equivalent, so their classes for parameters $ν$ and $- ν$ agree. Thus parameter-sign equivalence reduces the limit as $∣ ν ∣ ↑ 1$ to the positive-parameter limit.

2.1F1F9F10step 1.1A1algebra

Let $Q \subseteq G$ be compact. If $Q = \emptyset$ the assertion is vacuous, so assume $Q \neq \emptyset$ and set $d (k, g) = ∣ α (p (k, g)) ∣$ . By [F1], $d$ is positive and continuous on $K \times Q$ . The Iwasawa data make $K$ a circle, hence compact; [F9] makes $K \times Q$ compact. For each $(k, g)$ continuity gives a neighborhood on which $d > d (k, g) / 2 > 0$ and a neighborhood on which $d < d (k, g) + 1$ . Finite subcovers of these two covers give constants $0 < m \leq d \leq M < \infty$ on $K \times Q$ . Thus $∣ \log d ∣ \leq C : = \max (∣ \log m ∣, ∣ \log M ∣)$ there. For $0 < ν < 1$ and $∣ x ∣ \leq C$ , the function $t \mapsto e^{(1 - ν) t}$ is continuously differentiable on the closed interval with endpoints $0, x$ by [F10]. If $x = 0$ its increment is zero; otherwise the mean value theorem in [F10] and the derivative bound $(1 - ν) e^{(1 - ν) t} \leq (1 - ν) e^{C}$ on that interval give $∣ e^{(1 - ν) x} - 1 ∣ \leq (1 - ν) C e^{C}$ . This gives $\sup_{g \in Q} ∣ φ_{ν} (g) - 1 ∣ \leq (1 - ν) C e^{C} \to 0$ as $ν ↑ 1$ , using that $d k$ is normalized Haar probability.

3.1F2F5F6step 2.1

A basic Fell neighborhood of the trivial class tests a finite list of its diagonal coefficients on a compact set $Q$ , each to a common tolerance $ϵ > 0$ . Each coefficient is a constant $c_{i} \geq 0$ ; if the list is empty or all $c_{i} = 0$ , the zero approximants suffice. Otherwise put $C = \max_{i} c_{i} > 0$ . For each $c_{i} > 0$ , the vector $\sqrt{c_{i}} f_{0}$ in $I_{0, ν}$ gives the diagonal coefficient $c_{i} φ_{ν}$ ; for $c_{i} = 0$ , use the zero vector. By step 2.1, choose $ν$ sufficiently close to $1$ that $\sup_{g \in Q} ∣ φ_{ν} (g) - 1 ∣ < ϵ / C$ . Then every tested coefficient is within $ϵ$ of its constant on $Q$ . Hence every basic Fell neighborhood of the trivial class contains $[I_{0, ν}]$ for all sufficiently large $ν < 1$ , proving $[I_{0, ν}] \to [1_{G}]$ .

3.2F2F5F7step 2.1

Let $0 < ν_{j} < 1$ increase to $1$ , and form the strongly continuous Hilbert direct sum $Π = {⨁^{^}}_{j} Π_{ν_{j}}$ . If $ξ_{j}$ is the unit vector $f_{0}$ in its $j$ th summand, then $\sup_{g \in Q} ∥ Π (g) ξ_{j} - ξ_{j} ∥^{2} = \sup_{g \in Q} 2 (1 - Re φ_{ν_{j}} (g)) \to 0$ for each compact $Q$ by step 2.1. Hence $Π$ has almost invariant unit vectors and $1_{G} ≺ Π$ by [F7]. Each fixed summand has no nonzero invariant vector: it is irreducible by [F5] and has infinitely many K-types by [F2], so it is not the one-dimensional trivial representation. Since the direct sum acts componentwise, it too has no nonzero invariant vector. This is a sequence-level weak-containment conclusion; no such assertion is made for an individual fixed $I_{0, ν}$ .

4.1A1F7∎

AC is already declared and used for normalized Haar measure, the unitary completions, and the direct-sum weak-containment suppliers; the coefficient limits and the two displayed estimates are choice-free.

5 · Examples, counterexamples and false statements

None yet.

Sl2 R Principal and Complementary Series

1 · Prerequisites

2 · Summary

3 · Logical flowchart

4 · Definitions, theorems and proofs

Iwasawa and minimal-parabolic data for SL2(R)

Definition

Iwasawa decomposition and Haar integration formula for SL2(R)

Statement

Facts & Assumptions

Proof

The normalized principal series I(epsilon, nu)

Definition

The compact picture of the SL2(R) principal series

Statement

Facts & Assumptions

Proof

K-type decomposition of the SL2(R) principal series

Statement

Facts & Assumptions

Proof

Derived action and raising/lowering formulas in the compact picture

Statement

Facts & Assumptions

Proof

Remarks

K-finite vectors detect nonzero closed invariant subspaces

Statement

Facts & Assumptions

Proof

Generic irreducibility and the exceptional parameter lattice

Statement

Facts & Assumptions

Proof

Remarks

The standard intertwining operator A(nu)

Definition

Facts & Assumptions

Verification

Remarks

K-type eigenvalues of A(nu): recurrence, closed form and nonvanishing

Statement

Facts & Assumptions

Proof

Remarks

The invariant pairing between opposite principal-series parameters

Statement

Facts & Assumptions

Proof

Meromorphic continuation and intertwining identity for A(nu)

Statement

Facts & Assumptions

Proof

Unitarity of the unitary principal series

Statement

Facts & Assumptions

Proof

Unitarity of the complementary series

Statement

Facts & Assumptions

Proof

Parameter-sign equivalence and its exceptional failures for SL2(R)

Statement

Facts & Assumptions

Proof

The spherical complementary series converge to the trivial representation

Statement

Facts & Assumptions

Proof

5 · Examples, counterexamples and false statements

Sources