Fact-checked by Grok 2 weeks ago

Multiplication theorem

In mathematics, the multiplication theorem is a certain type of identity obeyed by many special functions related to the gamma function. For the explicit case of the gamma function, the identity is a product of values; thus the name. The various relations all stem from the same underlying principle; that is, the relation for one special function can be derived from that for the others, and is simply a manifestation of the same identity in different guises.^[1] These identities exist in both finite and infinite product forms and apply to functions such as the gamma function, sine, polygamma functions, and others. A prominent example is the duplication formula for the gamma function, a special case of the general multiplication theorem:

\Gamma(z) \Gamma\left(z + \frac{1}{2}\right) = 2^{1 - 2z} \sqrt{\pi} \, \Gamma(2z)

This formula, discovered by Legendre, illustrates the theorem's role in extending factorial properties to complex numbers.^[2]

General Concepts

Definition and Historical Overview

The multiplication theorem refers to a functional equation that relates the value of a special function f at a scaled argument kz, where k is a positive integer, to an expression involving sums or products of f evaluated at shifted arguments z + n/k for n = 0, 1, \dots, k-1.^[1] This structure is characteristic of many special functions, such as the gamma function, where it takes the form of a product: \prod_{n=0}^{k-1} \Gamma\left(z + \frac{n}{k}\right) = (2\pi)^{(k-1)/2} k^{1/2 - k z} \Gamma(k z).^[1] Similar identities hold for other functions like the polygamma and Hurwitz zeta functions, reflecting underlying symmetries in their analytic properties. The origins of the multiplication theorem trace back to Leonhard Euler's investigations into the gamma function during the 18th century, where he developed infinite product representations and related identities that anticipated the general formula, as detailed in his 1771 work.^[3] Adrien-Marie Legendre formalized the case for k=2, known as the duplication formula, in 1809, introducing the modern \Gamma notation and deriving \Gamma(z) \Gamma\left(z + \frac{1}{2}\right) = 2^{1-2z} \sqrt{\pi} \Gamma(2z).^[3] Carl Friedrich Gauss provided the first rigorous proof of the general multiplication formula in 1812 within his seminal paper on the hypergeometric series, establishing its validity for the gamma function.^[3] These theorems emerged from efforts to extend factorial-like behaviors to non-integer arguments and have since been generalized to broader classes of special functions by mathematicians including Gauss and later contributors like Arthur Erdélyi. The motivation for developing multiplication theorems lies in their utility for interpolation between known values of special functions, facilitating analytic continuation across the complex plane, and exploiting periodic or symmetric structures inherent to these functions in applications ranging from number theory to physics.^[4]

Finite and Infinite Forms

No rewrite necessary for this subsection — the distinction is non-standard and unsupported; content removed to align with article structure, where infinite series are addressed in "Generalizations and Applications."

Theorems in Characteristic Zero

Gamma Function – Legendre Multiplication Formula

The Legendre multiplication formula, also known as Gauss's multiplication formula, provides a relation among values of the gamma function at rationally related arguments. It states that for a positive integer k and complex number z such that k z \neq 0, -1, -2, \dots,

\prod_{n=0}^{k-1} \Gamma\left(z + \frac{n}{k}\right) = (2\pi)^{(k-1)/2} k^{1/2 - k z} \Gamma(k z).

^[1] This formula generalizes the functional equation of the gamma function and was first rigorously proved by Carl Friedrich Gauss in his 1828 work on the hypergeometric series, building on earlier discoveries by Leonhard Euler.^[3] A prominent special case occurs when k=2, yielding the duplication formula discovered by Adrien-Marie Legendre in 1809:

\Gamma(z) \Gamma\left(z + \frac{1}{2}\right) = 2^{1 - 2z} \sqrt{\pi} \, \Gamma(2z),

valid for $2z \neq 0, -1, -2, \dots. This relates the gamma function to itself at doubled argument and is fundamental in analytic number theory and special function evaluations.^[1]^[5] For k=3, the triplication formula follows directly:

\Gamma(z) \Gamma\left(z + \frac{1}{3}\right) \Gamma\left(z + \frac{2}{3}\right) = (2\pi) \cdot 3^{1/2 - 3z} \Gamma(3z),

with validity under the same conditions as the general formula; this case arises in evaluations involving cubic fields and elliptic integrals.^[1] Derivations of the formula can proceed via the Weierstrass infinite product representation of the gamma function, \Gamma(z) = z^{-1} e^{-\gamma z} \prod_{m=1}^\infty \left(1 + \frac{z}{m}\right)^{-1} e^{z/m}, where \gamma is the Euler-Mascheroni constant, by taking logarithms and expanding the product over shifted arguments. Alternatively, an integral proof uses the beta function, B(x,y) = \int_0^1 t^{x-1} (1-t)^{y-1} \, dt = \frac{\Gamma(x) \Gamma(y)}{\Gamma(x+y)}, through a substitution that decomposes the integral into k sectors via roots of unity, leading to the multiplicative relation after simplification. Euler employed a variant of this beta integral approach in his 1772 derivation.^[3]^[5] A proof sketch via beta integrals begins by considering the integral representation for \Gamma(kz) and applying the transformation t = u^k combined with averaging over the k-th roots of unity, which yields

\Gamma(kz) = k^{kz-1} \int_0^1 u^{kz-1} (1-u)^{k-1} \, du \cdot \frac{1}{k} \sum_{j=0}^{k-1} \omega^{-j(kz-1)},

where \omega = e^{2\pi i / k}; the sum simplifies using properties of roots of unity, and relating back to products of shifted beta functions produces the formula.^[3] Generalizations to q-analogs of the gamma function, defined as \Gamma_q(z) = (1-q)^{1-z} \prod_{n=0}^\infty \frac{1-q^{n+1}}{1-q^{n+z}} for |q| < 1, satisfy a corresponding multiplication theorem:

\prod_{j=0}^{k-1} \Gamma_q\left(z + \frac{j}{k}\right) = q^{\binom{k}{2} z (1-z)/k} (q;q)_k^{1/2 - k z} \Gamma_q(k z),

where (a;q)_k = \prod_{m=0}^{k-1} (1 - a q^m) is the q-Pochhammer symbol; this form appears in the theory of basic hypergeometric series.^[6]

Sine and Trigonometric Functions

The multiplication theorem for the sine function expresses a product of shifted sine values as a scaled multiple of the sine at a multiple argument. For a positive integer k and complex z not making any factor zero, the identity is

\prod_{n=0}^{k-1} \sin\left(\pi \left(z + \frac{n}{k}\right)\right) = k \, 2^{1-k} \sin(\pi k z).

This formula holds by analytic continuation throughout the complex plane except at poles. A special case is the duplication formula, obtained by setting k=2:

\sin(\pi z) \sin\left(\pi \left(z + \frac{1}{2}\right)\right) = \frac{1}{2} \sin(2 \pi z).

This relates the sines at z and its half-shift to the double-angle sine. One derivation uses the reflection formula for the gamma function, \sin(\pi z) = \pi / [\Gamma(z) \Gamma(1 - z)], combined with Gauss's multiplication formula for the gamma function. Substituting into the product yields \prod_{n=0}^{k-1} \sin(\pi (z + n/k)) = \pi^k / [\prod_{n=0}^{k-1} \Gamma(z + n/k) \Gamma(1 - z - n/k)]. Applying the multiplication formula to both the product of \Gamma(z + n/k) and the product of \Gamma(1 - z - n/k) (noting the latter is a shifted version), and simplifying using the reflection at k z, produces the sine identity after cancellation.^[7] An alternative derivation employs complex exponential factorization. The function \sin(\pi k z) factors via its zeros, and considering the polynomial whose roots correspond to the arguments z + n/k = m for integers m, leads to a representation involving exponentials e^{i \pi (z + n/k)}. The product identity emerges from equating the leading coefficients and scaling factors in this finite Blaschke-like product for the sine. Generalizations extend the theorem to related trigonometric and hyperbolic functions via analytic continuation. For cosine, the identity becomes \prod_{n=0}^{k-1} \cos(\pi (z + n/k)) = k \, 2^{1-k} \cos(\pi k z) under appropriate shifts, derived similarly from the gamma duplication or exponential forms. For the hyperbolic sine, \sinh(\pi z) = -i \sin(i \pi z), substituting z \to i z yields \prod_{n=0}^{k-1} \sinh(\pi (z + n/k)) = k \, 2^{1-k} (-i)^k \sinh(\pi k z), preserving the multiplicative structure. Historically, the theorem connects to Carl Friedrich Gauss's investigations of cyclotomic polynomials in Disquisitiones Arithmeticae (1801), where evaluating products of sines at rational points, such as \prod_{n=1}^{k-1} \sin(\pi n / k) = k / 2^{k-1}, facilitated explicit constructions of regular polygons and roots of unity.

Polygamma Functions and Harmonic Numbers

The multiplication theorem for polygamma functions of order m \geq 1 is given by

\sum_{n=0}^{k-1} \psi^{(m)}\left(z + \frac{n}{k}\right) = k^{m+1} \psi^{(m)}(kz),

where k is a positive integer and z is a complex number avoiding non-positive integers to ensure convergence. This identity arises as a consequence of the multiplication theorem for the gamma function upon repeated differentiation of the logarithm.^[8] For the zeroth-order case, the digamma function \psi(z) = \psi^{(0)}(z), the theorem takes the form

\sum_{n=0}^{k-1} \psi\left(z + \frac{n}{k}\right) = k \bigl( \psi(kz) - \log k \bigr).

This can be obtained by taking the logarithmic derivative of Gauss's multiplication formula for the gamma function, \Gamma(kz) = (2\pi)^{(k-1)/2} k^{kz - 1/2} \prod_{n=0}^{k-1} \Gamma\left(z + \frac{n}{k}\right), which yields the extra \log k term from differentiating the power of k. The polygamma functions connect directly to generalized harmonic numbers H_z^{(r)} of order r = m+1, defined via the relation \psi^{(m)}(z+1) = (-1)^{m+1} m! \bigl( H_z^{(m+1)} - \zeta(m+1) \bigr) for \operatorname{Re}(m) > 0. Substituting the multiplication theorem for \psi^{(m)} leads to an analogous identity for the generalized harmonic numbers:

H_{kz}^{(m+1)} = \frac{1}{k^m} \sum_{n=0}^{k-1} H_{z + n/k}^{(m+1)} + \text{lower-order terms},

where the lower-order terms arise from the infinite zeta contributions in the analytic continuation of the harmonic numbers. For the digamma case (m=0), the formula simplifies to H_{kz-1} = \frac{1}{k} \sum_{n=0}^{k-1} H_{z + n/k - 1} + \log k, reflecting the logarithmic adjustment. This derivation from the gamma function's multiplication theorem extends naturally to higher orders by further differentiation, preserving the power scaling without additional logarithmic factors for m \geq 1. These identities find applications in evaluating series sums, such as those involving reciprocals or their powers, and in developing asymptotic expansions for products of gamma functions or harmonic-like sequences, particularly in analytic number theory and special function approximations.^[9] For instance, they facilitate closed-form expressions for finite sums of polygamma values at rational shifts, aiding computations in integral evaluations and residue calculations.

Hurwitz and Periodic Zeta Functions

The Hurwitz zeta function generalizes the Riemann zeta function and is defined by the Dirichlet series

\zeta(s, z) = \sum_{n=0}^{\infty} \frac{1}{(n + z)^s}

for complex s with \operatorname{Re}(s) > 1 and z with \operatorname{Re}(z) > 0. This function admits an analytic continuation to a meromorphic function on the complex plane with a single simple pole at s=1. A key property is the multiplication theorem, which relates the value at a scaled argument to a sum over shifted arguments: for a positive integer k and suitable z,

k^{-s} \zeta(s, k z) = \sum_{n=0}^{k-1} \zeta\left(s, z + \frac{n}{k}\right).

This identity follows directly from reindexing the terms in the defining series expansion, grouping contributions according to residues modulo k. The Riemann zeta function arises as the special case \zeta(s) = \zeta(s, 1). The multiplication theorem extends to the periodic zeta function, a variant incorporating periodic coefficients. Defined for a period q and parameter vector a specifying the coefficients via a_m = a_{m \mod q}, the function is

F(s; a \mid q) = \sum_{m=1}^{\infty} \frac{a_m}{m^s},

converging for \operatorname{Re}(s) > 1. Its multiplication theorem states that for positive integer k,

k^{1-s} F(s; k a \mid k q) = \sum_{n=0}^{k-1} F(s; a + n \mid q).

This relation is derived analogously by reindexing the series, accounting for the periodicity scaled by k. A special case is the duplication formula for k=2:

$2^{1-s} F(s; q) = F(s; q/2) + F(s; (q+1)/2).

The periodic zeta function connects to the Hurwitz zeta via its role in the reflection formula, where Fourier analysis on the integer lattice yields the analytic continuation and functional equation

\zeta(1-s, z) = \frac{\Gamma(s)}{(2\pi)^s} \left[ e^{\pi i s / 2} \sum_{n=1}^{\infty} \frac{e^{2\pi i n z}}{n^s} + e^{-\pi i s / 2} \sum_{n=1}^{\infty} \frac{e^{-2\pi i n z}}{n^s} \right],

with the sums being periodic zeta functions; Poisson summation on the lattice underpins this transformation.^[10]

Polylogarithm Function

The multiplication theorem for the polylogarithm function \mathrm{Li}_s(z), where s \in \mathbb{C} and z \in \mathbb{C}, expresses \mathrm{Li}_s(z^k) in terms of a sum of polylogarithms evaluated at arguments scaled by the k-th roots of unity. For a positive integer k and \omega = e^{2\pi i / k}, the theorem states

\mathrm{Li}_s(z^k) = k^{s-1} \sum_{n=0}^{k-1} \mathrm{Li}_s(z \omega^n).

This relation holds for |z| < 1 via the power series definition and extends analytically to other regions where the functions are defined. The derivation follows directly from the series representation \mathrm{Li}_s(w) = \sum_{m=1}^\infty w^m / m^s for |w| < 1. Substituting into the sum yields

\sum_{n=0}^{k-1} \mathrm{Li}_s(z \omega^n) = \sum_{n=0}^{k-1} \sum_{m=1}^\infty \frac{(z \omega^n)^m}{m^s} = \sum_{m=1}^\infty \frac{z^m}{m^s} \sum_{n=0}^{k-1} (\omega^m)^n.

The inner sum over roots of unity equals k if k divides m and $0 otherwise, so letting m = k \ell gives

\sum_{n=0}^{k-1} \mathrm{Li}_s(z \omega^n) = k \sum_{\ell=1}^\infty \frac{z^{k \ell}}{(k \ell)^s} = k^{1-s} \mathrm{Li}_s(z^k),

rearranging to the stated formula. This geometric series argument over roots of unity underpins the theorem's validity.^[11] A special case is the duplication formula for k=2, where \omega^0 = 1 and \omega^1 = -1:

\mathrm{Li}_s(z^2) = 2^{s-1} \left[ \mathrm{Li}_s(z) + \mathrm{Li}_s(-z) \right],

or equivalently,

$2^{1-s} \mathrm{Li}_s(z^2) = \mathrm{Li}_s(z) + \mathrm{Li}_s(-z).

For s=1, where \mathrm{Li}_1(z) = -\ln(1-z), the general theorem reduces to the logarithm multiplication formula:

-\ln(1 - z^k) = \sum_{n=0}^{k-1} -\ln(1 - z \omega^n),

reflecting the factorization $1 - z^k = \prod_{n=0}^{k-1} (1 - z \omega^n).^[11] Analytically, the polylogarithm \mathrm{Li}_s(z) is multivalued with a principal branch defined by a branch cut along the real axis from $1 to \infty, and the multiplication theorem facilitates analytic continuation by expressing values in regions crossed by the cut as sums over rotated arguments that may lie in the principal domain. This relation to the periodic zeta function arises through inversion formulas connecting polylogarithms to lattice sums.^[11]

Kummer's Function

Kummer's confluent hypergeometric function, often denoted M(a, b, z) or \phi(a, b, z), was introduced by Ernst Kummer in 1837 in his studies of elliptic integrals, where it arises as a solution to Kummer's differential equation z w'' + (b - z) w' - a w = 0. This function plays a central role in the theorems within characteristic zero, providing closed-form relations that facilitate computations and connections to other special functions.^[12] The function satisfies various addition theorems and transformation formulas, including Kummer's transformation M(a, b, z) = e^z M(b - a, b, -z), which relates values at z and -z. These relations, derived from the hypergeometric series expansion \phi(a, c; z) = \sum_{n=0}^\infty \frac{(a)_n}{(c)_n} \frac{z^n}{n!}, enable efficient evaluation and symmetry analysis in solutions to differential equations related to elliptic integrals. For scaled arguments, generalized forms involving roots-of-unity filters can be applied to group terms in the series, though explicit closed forms are more complex than for non-confluent cases.^[12] In limiting cases, Kummer's function connects to other transcendental functions, such as the modified Bessel function of the first kind via I_\nu (2 \sqrt{z}) = ( \sqrt{z} )^\nu M(\nu + 1/2, 2\nu + 1, 2z) / \Gamma(2\nu + 1) and the error function through \erf(z) = \frac{2z}{\sqrt{\pi}} M(1/2, 3/2, -z^2), illustrating its role in probability distributions and wave equations. In certain asymptotic limits, it also relates briefly to the polylogarithm function, as the confluent form emerges from the series truncation of polylog expansions.^[13]

Polynomial and Dynamical Systems

Bernoulli Polynomials

The multiplication theorem for Bernoulli polynomials provides a key identity that expresses the polynomial evaluated at a scaled argument in terms of a sum over shifted arguments. For a positive integer k and nonnegative integer m \geq 1, the theorem states

B_m(kx) = k^{m-1} \sum_{n=0}^{k-1} B_m\left(x + \frac{n}{k}\right).

This relation, originally due to Raabe, generalizes properties of the Bernoulli numbers (the constant terms B_m = B_m(0)) and highlights the periodic-like behavior of these polynomials modulo 1.^[14] The theorem can be derived from the exponential generating function for the Bernoulli polynomials,

\frac{t e^{xt}}{e^t - 1} = \sum_{m=0}^\infty B_m(x) \frac{t^m}{m!}.

Consider the generating function for the left-hand side, \frac{t e^{kxt}}{e^t - 1}. For the right-hand side, sum the generating functions over the shifts: \sum_{n=0}^{k-1} \frac{t e^{(x + n/k)t}}{e^t - 1} = \frac{t e^{xt}}{e^t - 1} \sum_{n=0}^{k-1} e^{nt/k}. The sum is a geometric series equal to \frac{1 - e^t}{1 - e^{t/k}}, which simplifies the expression to match the left-hand side after scaling by k^{m-1} in the coefficient extraction. This approach leverages the structure of the generating function to yield the identity directly.^[15] In the linear case m=1, where B_1(x) = x - \frac{1}{2}, the theorem reduces to

B_1(kx) = \sum_{n=0}^{k-1} B_1\left(x + \frac{n}{k}\right),

which simplifies to kx - \frac{1}{2} = kx - \frac{1}{2} upon substitution, confirming consistency. This case underscores the centering effect of the Bernoulli polynomials, where the sum of deviations from the mean is zero, and it connects to harmonic sums through asymptotic expansions; for instance, the difference between harmonic numbers H_n and \ln n + \gamma involves terms adjustable via such linear corrections in summation formulas.^[16] A related variant holds for Euler polynomials E_m(x), defined by the generating function \frac{2 e^{xt}}{e^t + 1} = \sum_{m=0}^\infty E_m(x) \frac{t^m}{m!}. For odd degree m > 1 and odd positive integer a, the multiplication theorem takes the form

\sum_{r=0}^{a-1} (-1)^r E_m\left(x + \frac{r}{a}\right) = a^m E_m(ax).

This alternating sum parallels the non-alternating form for Bernoulli polynomials but incorporates signs due to the antisymmetry of Euler polynomials under shifts by 1 when m is odd. The identity arises similarly from generating function manipulations, emphasizing differences in the denominator (e^t + 1 versus e^t - 1).^[17] These polynomials play a central role in the Euler-Maclaurin summation formula, which approximates sums by integrals plus correction terms involving Bernoulli polynomials:

\sum_{j=a}^{b} f(j) = \int_a^b f(x) \, dx + \frac{f(a) + f(b)}{2} + \sum_{m=1}^p \frac{B_{2m}}{(2m)!} \left( f^{(2m-1)}(b) - f^{(2m-1)}(a) \right) + R,

where the remainder R depends on higher derivatives. The multiplication theorem facilitates evaluations of these corrections for scaled or periodic functions, enhancing applications in numerical analysis and asymptotic expansions of sums.^[18] The Bernoulli numbers extracted from these polynomials also connect to polygamma functions, as the values \psi^{(m-1)}(z) for integer m \geq 2 involve multiples of Bernoulli numbers through relations with the Hurwitz zeta function.^[16]

Bernoulli Map

The Bernoulli map, also known as the k-adic shift or doubling map (for k=2), is defined on the unit interval [0,1) by the transformation \sigma_k(x) = kx \mod 1, where x admits a base-k expansion x = \sum_{i=1}^\infty d_i / k^i with digits d_i \in \{0, 1, \dots, k-1\}, and the map shifts the expansion leftward by removing the leading digit.^[19] This dynamical system exemplifies ergodic behavior and chaos, preserving the Lebesgue measure and exhibiting mixing properties essential for spectral analysis.^[19] In this context, the multiplication theorem manifests through the Perron-Frobenius (transfer) operator associated with the Bernoulli map, which propagates densities under the dynamics: P_k f(x) = \frac{1}{k} \sum_{n=0}^{k-1} f\left( \frac{x + n}{k} \right). A key class of eigenfunctions comprises the Bernoulli polynomials B_m(x), which satisfy P_k B_m(x) = k^{-m} B_m(x), directly implying Raabe's multiplication theorem:

B_m(kx) = k^{m-1} \sum_{n=0}^{k-1} B_m\left(x + \frac{n}{k}\right).

This connection positions the polynomials as a spectral basis for the operator, with eigenvalues k^{-m} decaying exponentially.^[20] The derivation follows from the generating function \frac{t e^{xt}}{e^t - 1} = \sum_{m=0}^\infty B_m(x) \frac{t^m}{m!}, which under the transfer operator yields the scaled form via substitution and binomial expansion, confirming the eigenrelation through term-by-term verification.^[19] Alternatively, Fourier analysis can be used to study the operator, and the polynomials' Fourier series (involving zeta functions) align with the spectrum.^[19] More generally, the Fourier series \gamma(x) = \sum_n c(n) e^{2\pi i n x} arising from a totally multiplicative arithmetic function c(n) (satisfying c(kn) = c(k) c(n)) are eigenfunctions of the transfer operator with eigenvalue c(k).^[19] Generalizations extend the multiplication theorem to non-integer base maps, such as \beta-transformations T_\beta(x) = \beta x \mod 1 for \beta > 1, where the transfer operator admits piecewise affine eigenfunctions analogous to Bernoulli polynomials, with similar functional equations governing digit sums in greedy expansions.^[21] These frameworks underpin applications in ergodic theory, enabling spectral decomposition for correlation decay and invariant measures in broader classes of interval maps.^[21]

Theorems in Finite Characteristic

General Forms in Positive Characteristic

In fields of positive characteristic p > 0, multiplication theorems for special functions adapt classical formulas from characteristic zero, but leverage the discrete structure of finite fields or function fields over \mathbb{F}_[q](/page/Q), avoiding convergence issues inherent to infinite series in the archimedean setting. These analogs often involve finite sums over residues or polynomials, and are intimately connected to the Frobenius endomorphism, which raises elements to the p-th or q-th power, reflecting the field's characteristic. Unlike the real or complex gamma function, where analytic continuation is key, positive characteristic versions emphasize algebraic identities tied to group characters or modular structures.^[22] A primary finite form arises in the prime field \mathbb{F}_p, where analogs of the gamma or beta functions are constructed via finite products or sums over residues modulo p. For instance, the Morita p-adic gamma function \Gamma_p(x), defined for x \in \mathbb{Z}_p as \Gamma_p(x) = \lim_{n \to x} (-1)^n \prod_{\substack{0 < j < n \\ p \nmid j}} j, provides a p-adic lift that interpolates these finite products, serving as an analog despite the ambient characteristic zero of \mathbb{Q}_p. Its associated p-adic beta function B_p(x,y) = \frac{\Gamma_p(x) \Gamma_p(y)}{\Gamma_p(x+y)} satisfies a Gauss-Legendre multiplication formula, contrasting the classical case by incorporating p-adic valuations and residue functions. Specifically, for odd integer m \geq 2 with p \nmid m and n \in \mathbb{Z}_p,

\prod_{j=0}^{m-1} B_p\left(n + \frac{j}{m}, n + \frac{j}{m}\right) = (-1)^{\frac{m-1}{2}} B_p(mn, mn) \left( \prod_{j=1}^{\frac{m-1}{2}} B_p\left(\frac{j}{m}, \frac{m-j}{m}\right) \right) \prod_{j=0}^{\frac{m-1}{2}-1} h_p\left(2n + \frac{2j+1}{m}\right)^{m^{2\mu(mn) - \mu(2mn)}},

where h_p(x) = -x if |x|_p = 1 and -1 otherwise, and \mu(k) = k - 1 - \left\lfloor \frac{k-1}{p} \right\rfloor counts p-adic digits. This formula highlights distinctions like the absence of infinite products and the role of p-adic norms in the exponents.^[23] In function fields over \mathbb{F}_q (with q = p^r), infinite sums are replaced by finite hypergeometric series, often linked to Carlitz polynomials L_n(z) = \sum_{\mathbf{m}} \prod_{i=1}^s \binom{n}{m_i}_q z^{\deg m_i}, which analogize rising factorials via q-binomials. These yield multiplication theorems through hypergeometric identities, such as the Hasse-Davenport relation, which lifts Gauss sums (analogs of gamma values) across field extensions: for a character \chi on \mathbb{F}_q^\times, the relation connects G(\chi^{p^k}, \mathbb{F}_{q^{p^k}}) = (-1)^{k(q-1)} G(\chi, \mathbb{F}_q)^{p^k} under certain conditions, mirroring the duplication or general multiplication for \Gamma(z). Carlitz's gamma function \Gamma_{\mathrm{Carlitz}}(z) = e_{\mathrm{Carlitz}}^*(z) / L_{\deg z}(z) further satisfies a multiplication formula derived from module actions: for integers m coprime to q, \Gamma_{\mathrm{Carlitz}}(mz) = c_m \prod_{j=0}^{m-1} \Gamma_{\mathrm{Carlitz}}\left(z + \frac{j}{m}\right) up to a constant c_m involving q-Gauss sums, tied to the Frobenius via the Carlitz module's endomorphisms.^[22] These forms underscore the algebraic rigidity in positive characteristic, facilitating applications in Drinfeld modules and étale cohomology without radius-of-convergence constraints.^[22]

p-adic and Modular Variants

The p-adic gamma function \Gamma_p(x), introduced by Morita, provides a p-adic analog of the classical gamma function and admits a multiplication formula adjusted for the p-adic topology. For a positive integer m > 1 not divisible by the prime p, and x \in \mathbb{Z}_p, the formula states:

\prod_{j=0}^{m-1} \Gamma_p\left(x + \frac{j}{m}\right) = \left( \prod_{j=0}^{m-1} \Gamma_p\left(\frac{j}{m}\right) \right) m^{1 - \ell(mx)/(m p - 1) - \ell_1(mx)} \Gamma_p(mx),

where \ell(x) is the representative of x in \{1, 2, \dots, p\} such that |x - \ell(x)|_p < 1, and \ell_1(x) = p^{-1}(x - \ell(x)). This adjusts the classical Gauss multiplication theorem by incorporating p-adic digit-like functions \ell to account for the non-Archimedean valuation, ensuring continuity on \mathbb{Z}_p.^[23] A prominent p-adic implementation is the Gross–Koblitz formula, which serves as an analog of the duplication formula for gamma values at rational points. For an odd prime p and integer a with $0 \leq a \leq p-1, the formula expresses the Gauss sum g(a, \chi) associated to the Legendre symbol \chi as

g(a, \chi) = (-1)^{s_p(a)} \prod_{j=0}^{p-2} \Gamma_p\left( \frac{a + j}{p} \right),

where s_p(a) is the sum of the digits of a in base p. This product of p-adic gamma values at fractions with denominator p captures the p-adic interpolation of special values, linking analytic and algebraic structures. The formula extends to higher powers q = p^f via cyclic permutations of the p-adic digits of a.^[24] The derivation of the Gross–Koblitz formula relies on p-adic interpolation techniques for the gamma function and Morita's p-adic integrals. Morita's construction interpolates the gamma function on positive integers via limits of products excluding multiples of p, extended continuously to \mathbb{Z}_p. The proof involves Atkin operators on power series expansions and fixed points under affine maps related to the Frobenius, combined with Dwork's exponential sums to evaluate the integrals representing Gauss sums in the p-adic setting. This approach yields explicit p-adic measures for the relevant distributions.^[25] In modular variants over fields of positive characteristic p, analogs of the multiplication theorem arise for eta functions and theta series, where Frobenius actions replace integer multiplications. These variants build on the general framework in finite characteristic, adapting product formulas to modular forms modulo p. Applications of the p-adic variants, particularly the Gross–Koblitz formula, extend to Iwasawa theory, providing explicit class number formulas for cyclotomic fields. For instance, when p \equiv 1 \pmod{n}, the formula relates products of gamma values to algebraic integers bounding class numbers in Iwasawa modules.^[26]

Generalizations and Applications

Infinite Series Expansions

Infinite series expansions provide a key representation of multiplication theorems for special functions in characteristic zero, expressing a function evaluated at a scaled argument as an infinite sum involving the original function with shifted parameters. These expansions are particularly useful for analytical continuation and numerical evaluation within regions of convergence, often derived from the generating function or the power series definition of the functions involved.^[27] A prominent example is the multiplication theorem for Bessel functions of the first kind, given by

J_{\nu}(\lambda z) = \lambda^{\nu} \sum_{n=0}^{\infty} \frac{ (1 - \lambda^{2})^{n} }{n!} \left( \frac{z}{2} \right)^{n} J_{\nu + n}(z),

valid for all complex \lambda and \nu when using the upper sign convention for the Bessel function J. This formula, equivalent to the form \lambda^{-\nu} J_{\nu}(\lambda z) = \sum_{n=0}^{\infty} \frac{1}{n!} \left( \frac{(1 - \lambda^{2}) z}{2} \right)^{n} J_{\nu + n}(z), arises from the generating function for Bessel functions or by solving the Bessel differential equation with a scaling transformation. The series converges for all finite \lambda in the case of J_{\nu}, though the radius of convergence is generally |\lambda^{2} - 1| < 1; for \operatorname{Re}(\nu) > 0, additional analytic properties ensure uniform convergence on compact sets.^[27] For generalized hypergeometric functions, analogous infinite series expansions appear, particularly for the confluent hypergeometric function of the first kind M(a, b, z) = {}_{1}F_{1}(a; b; z). The multiplication theorem states

M(a, b, \lambda z) = \sum_{n=0}^{\infty} \frac{(a)_{n} ((\lambda - 1) z)^{n}}{(b)_{n} n!} M(a + n, b + n, z),

where (\cdot)_{n} denotes the Pochhammer symbol. This is obtained by substituting into the addition theorem and expanding via the hypergeometric series definition, effectively treating the scaling parameter \lambda through a Taylor-like series in shifted functions. The series converges for |\lambda - 1| < 1, with extensions possible via analytic continuation for specific parameter ranges such as \operatorname{Re}(b - a) > 0. More general forms for {}_{p}F_{q}(\mathbf{a}; \mathbf{b}; k z) can be derived similarly using multivariable generating functions, though they often reduce to sums over confluent or Gaussian cases for practical computation.^[28]^[29] Similar infinite series expansions apply to other special functions related to the confluent hypergeometric function, such as the Airy functions and parabolic cylinder functions. For instance, the Airy function \operatorname{Ai}(z) and the parabolic cylinder function D_{\nu}(z) can be expressed in terms of M(a, b, w) with appropriate scalings, inheriting multiplication theorems that yield series sums over shifted arguments for \operatorname{Ai}(\lambda z) or D_{\nu}(\lambda z). These are valuable in applications like quantum mechanics and wave propagation, where scaled arguments model varying potentials.

Modern Uses in Analysis and Number Theory

In number theory, the Chowla-Selberg formula provides a explicit relation between products of values of the gamma function at rational arguments and periods of elliptic curves with complex multiplication, generalizing aspects of the classical multiplication theorem for the gamma function.^[30] This formula has been instrumental in evaluating special values that appear in class number formulas and the distribution of zeros of L-functions.^[31] Furthermore, it plays a key role in computations related to the Birch and Swinnerton-Dyer conjecture for elliptic curves with complex multiplication, where the periods derived from the formula contribute to verifying the conjectured rank and leading coefficients of L-functions at s=1.^[32] In analysis, multiplication theorems extend to multiple gamma functions, such as the Barnes multiple gamma, enabling efficient numerical computations through reduction to products of single gamma evaluations.^[33] For instance, the generalized Gauss multiplication formula for the Barnes G-function allows for high-precision algorithms that leverage integral representations and asymptotic expansions, improving convergence in evaluating these functions for large arguments or high dimensions.^[34] Generalizations of the multiplication theorem include q-deformed versions of the gamma function, which satisfy q-analogues of Gauss's formula and arise in the representation theory of quantum groups.^[35] These q-deformations preserve key functional relations while incorporating a deformation parameter q, facilitating applications in quantum algebra and q-series. Super-analogs of the gamma function, defined over supermanifolds, also admit multiplication theorems adapted to graded structures, with uses in supersymmetric quantum field theories for computing volumes and partition functions in supergeometric settings.^[36] Recent developments since 2020 highlight applications in machine learning, where the multiplication theorem for the gamma function aids in data augmentation for Bayesian nonparametric models involving reciprocal gamma processes, approximating conditional densities via Stirling's formula to enhance posterior sampling efficiency.^[37] In p-adic analysis, extensions of the multiplication theorem to the p-adic gamma function support computations in cryptographic protocols based on elliptic curves over p-adic fields, particularly for evaluating p-adic L-functions in isogeny-based systems. To address gaps in classical theory, multiplication theorems have been extended beyond the Hurwitz zeta function to broader classes of L-functions, including relations with multiple zeta values through regularized products involving multiple gamma functions.^[33] These extensions connect special values of multiple zeta series to algebraic structures, aiding in conjectures on their linear independence and periods in motivic cohomology.

References

[1]
4.3 - Multiplication Rule | STAT 414
Multiplication Rule. The probability that two events A and B both occur is given by: P ( A ∩ B ) = P ( A | B ) P ( B ). or by:.
[2]
4.3: The Addition and Multiplication Rules of Probability
Jul 1, 2020 · The Multiplication Rule. If A and B are two events defined on a sample space, then: (4.3.1) P ⁡ ( A AND B ) = P ⁡ ( B ) ⁢ P ⁡ ( A | B ).<|control11|><|separator|>
[3]
DLMF: §5.5 Functional Relations ‣ Properties ‣ Chapter 5 Gamma ...
The gamma function has recurrence relation Γ(z+1) = zΓ(z), reflection relation Γ(z)Γ(1-z) = π/sin(πz), and duplication formula Γ(2z) = π-1/2^2z-1*Γ(z)*Γ(z+1/2).
[4]
[PDF] Euler and the multiplication formula for the Gamma Function
The multiplication formula was first proven rigourously by Gauss in his influential paper [Ga28] on the hypergeometric series, in which he also gave a complete ...Missing: theorem | Show results with:theorem
[5]
[PDF] Complex Analysis Lecture Notes - UC Davis Mathematics
Jun 15, 2021 · one of the most important special functions in “higher” mathematics. ... iii.* The multiplication theorem: Γ s Γ s + 1 k. Γ s + 2 k. ···Γ s + ...
[6]
Multiplication theorem - EPFL Graph Search
In mathematics, the multiplication theorem is a certain type of identity obeyed by many special functions related to the gamma function.
[7]
13.13 Addition and Multiplication Theorems
To obtain similar expansions for M ⁡ ( a , b , x ⁢ y ) and U ⁡ ( a , b , x ⁢ y ) , replace y in the previous two subsections by ( y − 1 ) ⁢ x .
[8]
DLMF: §10.23 Sums ‣ Bessel and Hankel Functions ‣ Chapter 10 ...
§10.23(i) Multiplication Theorem. i. Keywords: Bessel functions, Hankel functions, cylinder functions, multiplication theorem, sums; Notes: See Watson (1944 ...
[9]
[PDF] Introduction to the Gamma Function
Feb 4, 2002 · In 1809, Legendre obtained the following duplication formula [11]. ... as Gauss multiplication formula : Theorem 7 (Gauss). Γ(x)Γ µx +. 1 n¶Γ ...
[10]
[PDF] Hermite's formula for q-gamma function - Ele-Math
(1−aqj); k ∈ N. This function is a q-analogue of the gamma function since we have ... and a q-analogy of the Gauss's multiplication formula. Γ(x/p)Γ((x+1)/p)...Γ ...
[11]
[PDF] THE SINE PRODUCT FORMULA AND THE GAMMA FUNCTION ...
Dec 12, 2006 · To derive the sine product formula, we first find a relationship be tween the sine and gamma functions. We define a function φ(x) and find that ...
[12]
https://ia801504.us.archive.org/29/items/in.ernet.dli.2015.461466/2015.461466.Higher-Transcendental_text.pdf
[13]
[PDF] arXiv:2307.15107v1 [hep-th] 27 Jul 2023
Jul 27, 2023 · The multiplication theorem for the polygamma functions kj+1ψj(ky) = δj,0k log(k) + k−1. ∑ l=0 ψj. ( y + l k. ) ,. (SV.6) is also an important ...
[14]
(PDF) Values of the polygamma functions at rational arguments
Aug 6, 2025 · Davis in 1935 extended Gauss's result to the polygamma functions by using a known series representation of ψ(n)(z) in an elementary yet ...<|control11|><|separator|>
[15]
Derivatives of the Hurwitz Zeta function for rational arguments
Abstract. The functional equation for the Hurwitz Zeta function ζ(s,a) is used to obtain formulas for derivatives of ζ(s,a) at negative odd s and rational a.
[16]
[PDF] APPENDIX: SPECIAL VALUES AND FUNCTIONAL EQUATIONS OF ...
... polylogarithm reduces to simpler values by virtue of the duplication equation. These values are 1/2, 1/3, 2/3, −1/2, 3/4, −1/3, 8/9 and. −1/8. The image ...<|control11|><|separator|>
[17]
[PDF] Higher Transcendental Functions Vol. 1
Page 1. HIGHER TRANSCENDENTAL FUNCTIONS. Volume I. Based, in part, on notes ... Multiplication theorems. 282. 6.15. Series and integral formulas. 283. 6.15.1 ...
[18]
DLMF: §13.6 Relations to Other Functions ‣ Kummer Functions ...
When a − b is an integer or a is a positive integer the Kummer functions can be expressed as incomplete gamma functions (or generalized exponential integrals).
[19]
An identity of symmetry for the Bernoulli polynomials - ScienceDirect
Feb 28, 2008 · The multiplication theorem can be stated this way: if n and a are positive integers, then(6) a 1 - n B n ( ax ) = ∑ i = 0 a - 1 B n x + 1 a i .
[20]
[PDF] Characterization of the Bernoulli polynomials via the Raabe ... - arXiv
Mar 25, 2023 · By the Raabe multiplication theorem, we know that the polynomials λBn(X) (λ ∈ C) are all solutions of Equation (En,a). To prove the converse, we ...
[21]
Several polynomials associated with the harmonic numbers
As well as the Bernoulli numbers, the harmonic numbers H n = ∑ k = 1 n 1 / k frequently arise in combinatorial problems or in representations for well-known ...
[22]
Multiplication formulas for products of Bernoulli and Euler polynomials
Then 5m(a;) and Em(x) also satisfy the multiplication formulas (1.2), (1.3). In this note we obtain some generalizations of (1.2) and (1.3) sug-.
[23]
[PDF] Bernoulli polynomials and the Euler-Maclaurin formula.
We will first define the well known Bernoulli polynomials which will then be used to prove. Euler-Maclaurins formula. This again will be used to prove (1).
[24]
[PDF] THE BERNOULLI OPERATOR - Linas VEPSTAS
The Bernoulli map is important in physics as an exactly solvable example of determin- istic chaos, and, in particular, as a model of symbolic dynamics on Axiom ...
[25]
Generalized Eigenvalues of the Perron–Frobenius Operators of ...
We can prove that the left-hand side coincides with $(V_TB_n)(2x)$ , and thus, Raabe's multiplication theorem means that the Bernoulli polynomial $B_n$ is ...
[26]
[PDF] Sharp iteration asymptotics for transfer operators induced by greedy ...
Sep 5, 2025 · The Bernoulli polynomials are eigenfunctions in the integer case. First, we immediately have that Q rB0 “ rB0 and also: `Q rB1. ˘ pxq “ x{q ` q´ ...
[27]
[PDF] HYPERGEOMETRIC FUNCTIONS OVER FINITE FIELDS
The finite field analogues of the multiplication formulas for Γ(x) given in Theorems. 2.5 and 2.6 are known as the Hasse-Davenport Relations. The general ...
[28]
None
### Summary of Gauss-Legendre Multiplication Formula for p-Adic Beta Function
[29]
[PDF] p-ADIC L-FUNCTIONS AND EULER SYSTEMS
This article surveys six different special value formulae for p-adic L-functions, stressing their common features and their eventual arithmetic applications ...
[30]
Gauss Sums and the p-adic Γ-function - jstor
Annals of Mathematics, 109 (1979), 569-581. Gauss sums and the p-adic F-function. By BENEDICT H. GROSS and NEAL KOBLITZ. Introduction. Let p be an odd prime.
[31]
[PDF] The Gross-Koblitz formula revisited - Numdam
The Gross-Koblitz Formula Revisited. The formula in question gives an explicit value of Gauss sums using the p-adic gamma function of Morita. We give here an ...
[32]
A CLASS NUMBER FORMULA OF IWASAWA'S MODULES
To prove Theorem (0.3), we use essentially Gauss sums, Gross-. Koblitz formula concerning a relation between Gauss sums and special values of. Morita's p-aάic ...
[33]
None
### Summary of Multiplication Theorem/Series Expansion for Hypergeometric Functions
[34]
[PDF] arXiv:math/0606096v1 [math.NT] 5 Jun 2006
Jun 5, 2006 · The Chowla-Selberg formula has many applications in number theory. In particular, it can be used to investigate the distribution of the ...
[35]
An analogue of the Chowla-Selberg formula for several automorphic
The Chawla-Selberg formula has many applications in number the- ory. In particular, it was used to investigate the distribution of the complex zeros of Zq(s) ...
[36]
[PDF] On the 2-part of the Birch-Swinnerton-Dyer conjecture for ... - arXiv
On the 2-part of the Birch-Swinnerton-Dyer ... Again by the Chowla-Selberg formula, the period lattice L of the Neron differential on E ... agawa factors which ...
[37]
[PDF] Special values of multiple gamma functions - UCLA Mathematics
We also present some simple infinite product identities for certain special values of the multiple gamma func- tion. 1. Introduction. The area of a quarter of ...
[38]
[PDF] Contributions to the Theory of the Barnes Function 1 Introduction
This paper presents a family of new integral representations and asymptotic series of the multiple gamma function. The numerical schemes for high-precision ...
[39]
THE q-DEFORMED GAMMA FUNCTION AND q ... - KoreaScience
In this paper, we rederive the identity ${\Gamma}_q(x){\Gamma}_q(1-x)={\frac{{\pi}_q}{sin_q({\pi}_qx)}$ . Then, we give q-analogue of Gauss' multiplication ...
[40]
General Aspects of Jackson Calculus in Clifford Analysis
Feb 25, 2025 · We first consider the q-deformed Dirac, Euler, and Gamma operators and their symmetry properties. Then, we consider monogenic polynomials, a ...<|control11|><|separator|>
[41]
[PDF] On Data Augmentation for Models Involving Reciprocal Gamma ...
Oct 27, 2022 · Our strategy for deriving an augmented model is 2-fold: (i) using Gauss's multiplication formula for the gamma function to introduce ...