Professional Documents
Culture Documents
Regular Functions of A Quaternionic Variable 2nd Edition Graziano Gentili Caterina Stoppato Daniele C Struppa
Regular Functions of A Quaternionic Variable 2nd Edition Graziano Gentili Caterina Stoppato Daniele C Struppa
Regular Functions of A Quaternionic Variable 2nd Edition Graziano Gentili Caterina Stoppato Daniele C Struppa
https://ebookmeta.com/product/regular-functions-of-a-
quaternionic-variable-2nd-edition-graziano-gentili-caterina-
stoppato-daniele-c-struppa-2/
https://ebookmeta.com/product/functions-of-one-complex-
variable-i-2nd-edition-john-b-conway/
https://ebookmeta.com/product/complete-solutions-manual-for-
calculus-of-a-single-variable-early-transcendental-functions-7th-
edition-ron-larson/
https://ebookmeta.com/product/calculus-of-one-variable-2nd-
edition-m-thamban-nair/
Historic Churches of New Mexico Today Graziano
https://ebookmeta.com/product/historic-churches-of-new-mexico-
today-graziano/
https://ebookmeta.com/product/regular-extensions-of-hermitian-
operators-a-v-kuzhel-s-a-kuzhel/
https://ebookmeta.com/product/c17-standard-library-quick-
reference-2nd-edition-a-pocket-guide-to-data-structures-
algorithms-and-functions-peter-van-weert/
https://ebookmeta.com/product/extension-of-holomorphic-
functions-2nd-edition-marek-jarnicki-peter-pflug/
https://ebookmeta.com/product/statistical-analysis-of-graph-
structures-in-random-variable-networks-v-a-kalyagin/
Springer Monographs in Mathematics
Graziano Gentili
Caterina Stoppato
Daniele C. Struppa
Regular
Functions of a
Quaternionic
Variable
Second Edition
Springer Monographs in Mathematics
Editor-in-Chief
Minhyong Kim, School of Mathematics, Korea Institute for Advanced Study, Seoul,
South Korea
International Centre for Mathematical Sciences, Edinburgh, UK
Katrin Wendland, School of Mathematics, Trinity College Dublin, Dublin, Ireland
Series Editors
Sheldon Axler, Department of Mathematics, San Francisco State University, San
Francisco, CA, USA
Mark Braverman, Department of Mathematics, Princeton University, Princeton, NY,
USA
Maria Chudnovsky, Department of Mathematics, Princeton University, Princeton,
NY, USA
Tadahisa Funaki, Department of Mathematics, University of Tokyo, Tokyo, Japan
Isabelle Gallagher, Département de Mathématiques et Applications, Ecole Normale
Supérieure, Paris, France
Sinan Güntürk, Courant Institute of Mathematical Sciences, New York University,
New York, NY, USA
Claude Le Bris, CERMICS, Ecole des Ponts ParisTech, Marne la Vallée, France
Pascal Massart, Département de Mathématiques, Université de Paris-Sud, Orsay,
France
Alberto A. Pinto, Department of Mathematics, University of Porto, Porto, Portugal
Gabriella Pinzari, Department of Mathematics, University of Padova, Padova, Italy
Ken Ribet, Department of Mathematics, University of California, Berkeley, CA,
USA
René Schilling, Institute for Mathematical Stochastics, Technical University Dres-
den, Dresden, Germany
Panagiotis Souganidis, Department of Mathematics, University of Chicago,
Chicago, IL, USA
Endre Süli, Mathematical Institute, University of Oxford, Oxford, UK
Shmuel Weinberger, Department of Mathematics, University of Chicago, Chicago,
IL, USA
Boris Zilber, Mathematical Institute, University of Oxford, Oxford, UK
This series publishes advanced monographs giving well-written presentations of the
“state-of-the-art” in fields of mathematical research that have acquired the maturity
needed for such a treatment. They are sufficiently self-contained to be accessible to
more than just the intimate specialists of the subject, and sufficiently comprehensive
to remain valuable references for many years. Besides the current state of knowledge
in its field, an SMM volume should ideally describe its relevance to and interaction
with neighbouring fields of mathematics, and give pointers to future directions of
research.
Graziano Gentili • Caterina Stoppato •
Daniele C. Struppa
Regular Functions
of a Quaternionic Variable
Graziano Gentili Caterina Stoppato
Dept of Math and Computer Science Dept of Math and Computer Science
University of Florence University of Florence
Florence, Italy Florence, Italy
Daniele C. Struppa
Donald Bren Presidential Chair in Math
Chapman University
Orange, CA, USA
This work was supported by Istituto Nazionale di Alta Matematica “Francesco Severi” (http://dx.
doi.org/10.13039/100009112), Ministero dell’Università e della Ricerca (http://dx.doi.org/10.13039/
501100021856)
Mathematics Subject Classification: 30G35, 30B10, 30C15, 30C80, 30D10, 30D30, 30E20
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2013, 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
The first author dedicates this work to Luisa,
and to Alessandro and Lorenzo.
The second author dedicates this work to
Arturo, who shares her passion for
mathematics, and to Lisa and Federico, her
role models (and greatest supporters).
The third author dedicates this work to his
princesses, Arianna and Athena, and to
Queen Lisa.
Preface
vii
viii Preface
the support of: Chapman University, where most of the work has been done;
George Mason University; Università degli Studi di Firenze; Università degli Studi
di Milano; GNSAGA of the Istituto Nazionale di Alta Matematica “F. Severi”;
European Social Fund; Regione Lombardia; MIUR—Italian Ministry of University
and Research—via the projects PRIN “Proprietà geometriche delle varietà reali
e complesse,” PRIN “Geometria Differenziale e Analisi Globale,” and FIRB
“Geometria Differenziale Complessa e Dinamica Olomorfa.”
Nine years after the first edition of the present monograph, the theory of slice regular
functions has vastly developed. It has also found new exciting applications.
A significant novelty is a new branch of the theory of slice regular functions,
namely function theory over slice domains that are not symmetric. Slice regular
functions on general slice domains exhibit new phenomena that were not visible
in the symmetric case. This prompted us to update the monograph, whose first
edition focused on symmetric slice domains. About the new topic, we added
Chap. 11, which in turn required updates to Sect. 1.3 and to Chap. 10. We are
grateful to Xinyuan Dou and Guangbin Ren, who found a flaw in the formulation
of former Theorem 1.24: we added extra hypotheses to the corresponding current
Theorem 1.32, a change which did not affect any other section of the monograph
but motivated the aforementioned study of regularity in the nonsymmetric case.
We took this chance to make minor improvements throughout the monograph.
The Bibliographic Notes and the references have been updated, to the best of our
knowledge of the literature. Finally, in this second edition, the overview of old and
new applications of the theory of slice regular functions forms a whole chapter,
namely Chap. 12. We are grateful to all collaborators involved in these applications.
We gratefully acknowledge partial support from Istituto Nazionale di Alta
Matematica, through GNSAGA and through the INdAM project “Hypercomplex
function theory and its applications”, as well as from Ministero dell’Istruzione,
dell’Università e della Ricerca, through the projects Finanziamento Premiale FOE
2014 “Splines for accUrate NumeRics: adaptive models for Simulation Environ-
ments” and PRIN 2017 “Real and complex manifolds: topology, geometry and
holomorphic dynamics”.
ix
Introduction
q = x0 + x1 i + x2 j + x3 k,
where {1, i, j, k} denotes the standard basis. Multiplication is defined on the basis
by the formulas
i 2 = j 2 = k 2 = −1,
ij = −j i = k, j k = −kj = i, ki = −ik = j,
q̄ = x0 − x1 i − x2 j − x3 k
q −1 = |q|−2 q̄.
xi
xii Introduction
f (z + h) − f (z)
lim = f (z).
h→0 h
exists for all q ∈ H, implies that f is an affine function of the form f (q) = qa + b
for some a, b ∈ H. This is shown in detail, for instance, in [222, 279]. Thus,
the naive approach to quaternionic differentiability is inadequate to work as an
analog of the theory of holomorphic functions of a complex variable. The reader
is referred again to [222] for a historical discussion and appropriate references.
One may then wonder whether a less naive approach to differentiability for
Introduction xiii
It turns out that the operator defined by ∂∂q̄ is a very good analog for the Cauchy–
Riemann operator. In other words, the theory of its zero-solutions enjoys many of the
key properties of the theory of holomorphic functions. The theory of Fueter-regular
functions is well established and includes, for instance, a Cauchy integral theorem
and a Cauchy kernel (for a first introduction, see [279]). We note, in particular,
that pointwise product of two Fueter-regular functions is not Fueter-regular, but
that a multiplication preserving regularity can be defined in terms of the Cauchy–
Kowalewski extension (see [52, 257, 258, 270]). This theory also has significant
applications to physics and engineering and it is the object of a vast literature.
Without any pretense of completeness, we mention the monographs [198, 199, 209].
Going back for a moment to the discussion in [222], it is possible to define
a notion of hyperdifferentiability (distinct from the naive notion of quaternionic
differentiability) by means of the following real differential forms with quaternionic
coefficients:
σ (3) = dx1 ∧ dx2 ∧ dx3 − idx0 ∧ dx2 ∧ dx3 + j dx0 ∧ dx1 ∧ dx3 − kdx0 ∧ dx1 ∧ dx2 ,
the latter being the analog, in the present setting, of the differential form dz =
dx + idy used in the complex setting. Now, a quaternion-valued function f on
an open subset of H, of class C1 in its four real variables x0 , x1 , x2 , x3 , is called
hyperdifferentiable if, for each q in its domain, there exists a quaternion f (q), so
that
Just as in the complex case, it turns out that the operator f → f can be expressed in
terms of the conjugate of the Cauchy–Fueter operator (see [222] for details). Thus, a
xiv Introduction
where Pn (q) is a finite sum of monomials of the type a0 qa1 . . . an−1 qan . It turns
out, however, that this condition is too weak: it is equivalent to analyticity in the
four real variables x0 , x1 , x2 , x3 because
1
x0 = (q − iqi − j qj − kqk),
4
1
x1 = (q − iqi + j qj + kqk),
4i
1
x2 = (q + iqi − j qj + kqk),
4j
1
x3 = (q + iqi + j qj − kqk).
4k
Taylor and Laurent expansions are available for Fueter-regular functions in a very
specific sense, discovered already in [123]. The first step is considering the three real
variables x1 , x2 , x3 from R3 = {q ∈ H | x0 = 0} and extending them in a Fueter-
regular fashion to the whole space H. If we set e1 = i, e2 = j, e3 = k, then the
unique Fueter-regular extension of xl to H is the Fueter variable ζl = xl − x0 el . The
Fueter variables do the trick: every Fueter-regular function expands into a series
involving powers of ζ1 , ζ2 , ζ3 . The story on Taylor expansions for Fueter-regular
functions does not end here: new bases can be introduced, whose structures lead to
a better understanding of the set of all Fueter-regular functions. This is done, for
instance, in [6]: the setting there is more general, but the last section shows how to
deal with the quaternionic case.
Overall, the theory of holomorphic functions can indeed be nicely general-
ized to the quaternionic case, with the Cauchy–Riemann system replaced by the
Cauchy–Fueter system, the notion of complex differentiability by the notion of
hyperdifferentiability, and the traditional Taylor series in the complex variable z by
the Taylor series in the three Fueter variables ζ1 , ζ2 , ζ3 . As a consequence, over
many years, the theory of Fueter-regularity has been developed and generalized
in many directions. These studies include a fairly well-developed study in several
quaternionic variables [71], as well as a theory of Clifford-valued regular functions
[52].
Introduction xv
Despite the richness of the theory of Fueter, some of its features motivated the
search for an alternative definition of regularity: on the one hand, the class of Fueter-
regular functions is quite large; on the other hand, it does not include the (one-sided)
polynomials in the variable q. Indeed, if we set z1 = x0 + ix1 , z2 = x2 + ix3, then
the following (proper) implications hold:
More details can be found in [280]. But powers q n are not harmonic in x0 , x1 , x2 , x3
when n > 1. Even in case n = 1, we find that the identity function id : H → H is
harmonic in x0 , x1 , x2 , x3 but not Fueter-regular because ∂∂q̄ id ≡ − 12 . Quaternionic
polynomials are included in the class of quaternionic holomorphic functions,
defined by Fueter himself in [121] as solutions of the equation
∂
f (q) = 0.
∂ q̄
Here, denotes the Laplacian in the four real variables x0 , x1 , x2 , x3 . This study
has been generalized by Laville and Ramadanoff in [212]. Notice, however, that
the class of quaternionic holomorphic functions is extremely large, as it includes
the whole class of harmonic functions of four real variables. (Restrictions of)
quaternionic polynomials are also contained in the class of functions of the reduced
variable x0 + x1 i + x2 j studied by Leutwiler in [213].
A different definition of regularity in the quaternionic context was given in [92]
by Cullen, who considered solutions of the equation
∂ Im(q) ∂
+ f (q) = 0
∂x0 r ∂r
for x, y ∈ R and for all I ∈ S (more details will be provided in Chap. 1). It turned
out that all polynomials and power series of the form
q n an
n∈N
This lemma is, e.g., instrumental to prove the Identity Principle and the Maximum
Modulus Principle for Cullen-regular functions by means of the corresponding
classical (complex) results. Proving more refined results for Cullen-regular func-
tions on open balls centered at the origin required new ideas. This was the case
for the study of the zero sets of Cullen-regular functions [135, 143, 151], which
required the natural generalization of the classical product of polynomials over a
noncommutative ring to the so-called ∗-product of regular power series, and for
the study of the poles of the quaternionic analogs of meromorphic functions [273].
It was also the case for the following results: versions of the Minimum Modulus
Principle and the Open Mapping Theorem [136], as well as [149] some results in
the spirit of the celebrated Cartan Fixed Point Theorems [120] and Burns–Krantz
Theorem [53]. A comprehensive survey of the state of the art up to this point can be
found in [152].
A turning point in the theory of these functions came when it became apparent
that in order to avoid pathological phenomena (e.g., regular functions which are
not even continuous, see Example 1.11), it was necessary to carefully choose the
domains of definition of regular functions. The first ideas on the features of these
domains appeared in [62, 136], which uncovered several aspects of the theory where
axial symmetry with respect to the real axis R is an important feature of the domain,
and [273], which proved a version of the Identity Principle valid on each domain
that intersects the real axis and whose intersection with LI is connected for every
I ∈ S. The subsequent paper [78] named such domains slice domains. In [74], the
authors discovered that if is an axially symmetric slice domain, if f : → H is
Cullen-regular, and if x, y ∈ R are such that x + yS ⊂ , then the Representation
Formula
1 JI
f (x + yJ ) = [f (x + yI ) + f (x − yI )] + [f (x − yI ) − f (x + yI )]
2 2
holds true for all I, J ∈ S. This formula has very deep implications, including
the fact that every holomorphic function φ : ∩ LI → H extends to a unique
Cullen-regular function → H. Consequently, from the works [74, 78] on, we and
our coauthors began to refer to Cullen-regular functions as slice regular functions,
s-regular functions, or simply regular functions (the expression we will adopt
henceforth in the present monograph). The new step in the theory also allowed the
definition of the ∗-product between regular functions, which extends the definition
valid for regular power series. Furthermore, a new and significantly different
Cauchy Formula was discovered in [78], which, along with the Representation
Formula, quickly became the central tool for a series of far-reaching extensions
of known results. On symmetric slice domains, it was immediately possible to
prove a Pompeiu Formula [82] and to extend [137] all properties of the zero set
of regular functions, as well as the Open Mapping Theorem, known in the case of
balls centered at the origin. A vast amount of work has been devoted to symmetric
slice domains in subsequent years. This produced a full-fledged theory of regular
functions in this context, which was the main focus of the first edition of the present
xviii Introduction
monograph [154]. Significant exceptions, where symmetry was not assumed, are:
the study [138] of power series centered at nonreal quaternions and their domains of
convergence, which led to a notion of weak analyticity that is equivalent to regularity
on general quaternionic domains; and the Maximum Modulus Principle, valid on all
slice domains in the version proven in [137]. We point out that a stronger notion of
analyticity was introduced and proven equivalent to regularity over symmetric slice
domains in [276].
For several years, regularity has not been investigated on general (nonsymmetric)
slice domains. Indeed, it was believed that every regular function on a slice domain
would extend to a unique regular function on the so-called symmetric completion
of , namely the smallest axially symmetric subset of H that contains . This
extension phenomenon was stated in [74] and even in the first edition of this
monograph [154], as Theorem 1.24. In recent years, Dou and Ren [104] provided a
counterexample to the original formulation of the theorem. Its current formulation,
proved in [139] and stated in this second edition of the monograph as Theorem 1.32,
includes additional hypotheses on the domain. This novelty motivated the study
of regularity on general slice domains, addressed in [139, 140]. New exciting
phenomena appeared, which were not visible in the symmetric case. The current
edition of the monograph describes them in detail. Further work in this direction
includes [105, 106].
In parallel with the development of the theory of regular functions over the
algebra of quaternions, the cases of other algebras have been addressed. Gentili
and Struppa generalized the definition of regular function to the octonions in [145]
and to the Clifford algebra Cl(0, 3) in [144] (both theories are surveyed in [152]).
The zeros of octonionic regular functions have been studied in [167]. The case
of functions defined on Rm+1 with values in the Clifford algebra Cl(0, m) has
been considered by Colombo, Sabadini, and Struppa in [75, 80], where the notion
of slice monogeneity was introduced. The theory of slice monogenic functions
has vastly developed in subsequent years and is surveyed in the monograph [81],
where slice monogenic and slice regular functions are called slice hyperholomorphic
functions. In [166], Ghiloni and Perotti have made a significant step forward: they
generalized the definition to other (finite dimensional) alternative real algebras and
endowed all the aforementioned theories (including the quaternionic one) with new
working tools. One of these tools is the notion of slice function, which encodes the
Representation Formula mentioned earlier in this Introduction without taking any
further (continuity, smoothness, holomorphy etc.) assumption. The study of regular
functions over alternative real algebras is fairly developed now, especially in the
realm of geometric function theory.
Every time mathematicians define a new object, two questions arise naturally.
The first is whether the new object leads to an interesting theory. One aim of this
book is showing that the notion of regularity is indeed a very stable notion that yields
a rich theory for functions of a quaternionic variable. At the same time, the book
highlights some crucial differences between the complex and the quaternionic case.
The second question is whether the new theory, in addition to its intrinsic value,
can also contribute to the solution of some outstanding problem. An answer to this
Introduction xix
question, in the present setting, can be found by looking at the various applications
the theory of regular functions has already had. It has been applied in analysis, in
the construction of a new functional calculus in a noncommutative setting, starting
with [73, 76, 77], and more generally in the study of function spaces and operator
theory. It has found several applications in differential geometry: in the construction
and classification of orthogonal complex structures on dense open sets in H, starting
with [155]; and in a direct approach to quaternionic manifolds [43, 157–159]. It has
been applied in spatial kinematics: e.g., in the characterization of Rational Rotation-
Minimizing Frame curves [118] and in the study of motion polynomials over the
algebra of dual quaternions [162]. It is also involved in the construction of a theory
of coherent states in quaternionic quantum mechanics [282].
We now describe in detail the contents of every chapter.
Chapter 1 explains the definition of regularity. It presents the basic results of the
theory: the existence of power series expansions in all Euclidean balls centered at
the origin, the Identity Principle, and a Representation Formula that is fundamental
in the study of regular functions on axially symmetric slice domains. The same
formula yields some interesting extension phenomena. Finally, we show that regular
functions on a symmetric slice domain form a ring with respect to addition and to
an appropriately defined regular multiplication.
Chapter 2 explores the possibility of series expansions at points other than the
origin. Power series expansions exist at all points of the domain of definition, but
their sets of convergence are balls with respect to a non-Euclidean distance σ .
This leads to the notion of σ -analyticity, which is equivalent to regularity on any
quaternionic domain.
In Chap. 3, we treat the zero sets of regular functions on symmetric slice domains,
which exhibit interesting algebraic and topological properties. For instance, the
zero set of a regular function consists of isolated points or isolated 2-spheres of a
special type. We present the study of nth roots of quaternions and factorizations
of regular polynomials, explaining their relations with the zeros and comparing
different notions of multiplicity. Finally, we overview the Bezout Theorem and the
construction of Gröbner bases for quaternionic polynomials.
Chapter 4 contains the Weierstrass Factorization Theorem for quaternionic entire
functions and all the tools that are necessary for its proof: the study of infinite
products of regular functions and of their convergence, based on the use of the
principal branch of the quaternionic logarithm. The formulation of this theorem
fully reflects the peculiarities of the structure of the zero sets in the quaternionic
setting.
Chapter 5 is devoted to the classification of the singularities of regular functions
on symmetric slice domains. We first present the construction of the ring of quotients
of regular functions. We then study regular Laurent series and expansions, which
are particularly interesting when centered at a point p other than the origin. Laurent
expansions allow us to classify the singularities as removable, essential, or as poles.
Poles are studied by means of regular quotients, while for essential singularities we
present a version of the Casorati–Weierstrass Theorem.
xx Introduction
To conclude, we point out that the bibliographical notes that appear at the end
of each chapter are essentially meant to collect the references to material due to
our nearest collaborators and us. Instead, references to production by other authors
are mostly disseminated in the text. There are a couple of exceptions, namely our
overviews in Chaps. 10 and 12, where all references are disseminated in the text.
Contents
xxiii
xxiv Contents
References .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 269
Index . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 283
Chapter 1
Definitions and Basic Results
S = {q ∈ H : q 2 = −1}
denote the 2-sphere of quaternionic imaginary units. We define the notion of regular
function as follows.
Definition 1.1 Let f be a quaternion-valued function defined on a domain . For
each I ∈ S, let I = ∩ LI and let fI = f|I be the restriction of f to I . The
restriction f I is called holomorphic if it has continuous partial derivatives and
1 ∂ ∂
∂¯I f (x + yI ) = +I fI (x + yI ) ≡ 0. (1.1)
2 ∂x ∂y
The function f is called regular if, for all I ∈ S, the restriction f I is holomorphic.
us identify LI with C. Notice, moreover, that for all J ∈ S with J ⊥ I , the following
equality holds:
H = LI + LI J.
The previous Lemma can be reformulated in a way that will be useful in the
sequel.
Lemma 1.4 Let I ∈ S, let U be open in LI , and let φ : U → H. The function φ
is holomorphic if, and only if, for all J ∈ S with J ⊥ I , there exist complex-valued
holomorphic functions of one complex variable F, G : I → LI such that
for all z ∈ I .
Let us now review the first examples and the basic properties of regular functions.
Example 1.5 The identity function q → q is regular in H. The same holds for any
polynomial function of the type q → a0 + qa1 + . . . + q n an , al ∈ H for all l.
This class of examples extends as follows. For all R ∈ (0, +∞], let us denote by
1
R= . (1.3)
lim supn∈N |an |1/n
converges absolutely and uniformly on compact sets in B(0, R). Moreover, its sum
defines a regular function on B(0, R).
1.1 Regular Functions 3
The proof, based on the root test, is completely analogous to that of the complex
Abel theorem.
Example 1.7 Setting
qn
exp(q) =
n!
n∈N
for all q ∈ H defines the (quaternionic) exponential function exp : H → H. For each
I ∈ S, if we identify LI with C, then the restriction expI coincides with the complex
exponential function. For this reason, we will also use the notation eq for exp(q).
For the same reason, exp(s + I t) = es (cos(t) + I sin(t)) for all s, t ∈ R, I ∈ S.
The equality exp(a + b) = exp(a) exp(b) holds if the quaternions a and b
belong to the same complex plane LI but not for general a, b ∈ H: of course,
there cannot exist a homomorphism between the commutative group (H, +) and the
noncommutative group (H \ {0}, ·).
Conversely, all regular functions on B(0, R) can be expressed as power series. In
order to prove this, we first introduce an appropriate notion of derivative.
Definition 1.8 Let be a domain and let f : → H be a regular function. For
each I ∈ S, the I-derivative of f is defined as
1 ∂ ∂
∂I f (x + yI ) = −I fI (x + yI ) (1.5)
2 ∂x ∂y
∂F ∂G
f (z) = ∂I f (z) = (z) + (z)J
∂z ∂z
and, similarly,
∂ nF ∂ nG
f (n) (z) = n
(z) + (z)J.
∂z ∂zn
converges to F(z) for z ∈ BI (absolutely and uniformly on its compact subsets). The
same can be proved for G, so that for all z ∈ BI
1 ∂ nF 1 ∂ nG
f (z) = F (z) + G(z)J = zn (0) + zn (0)J
n! ∂zn n! ∂zn
n∈N n∈N
1 (n)
= zn f (0)
n!
n∈N
The previous example proves that if the domain is not carefully chosen, then a
regular function f : → H does not even need to be continuous. It is possible to
prevent such pathologies by imposing further conditions on the domain .
Definition 1.12 Let be a domain in H. Then is called a slice domain if it
intersects the real axis and if, for all I ∈ S, its intersection I with the complex
plane LI is connected.
The identity principle holds true on all slice domains.
Theorem 1.13 (Identity Principle) Let f, g be regular functions on a slice domain
. If, for some I ∈ S, the functions f and g coincide on a subset of I having an
accumulation point in I , then f = g in .
Proof The restrictions fI , gI are holomorphic functions. Under our hypotheses, f I
and gI must coincide in I . In particular, f must coincide with g in ∩ R. For all
K ∈ S, the intersection ∩ R is a subset of K that has an accumulation point in
K . Thus, fK = gK in K for all K ∈ S, and we conclude that f = g in
= K .
K∈S
Notice that, in the proof of Theorem 1.13, both properties that define slice
domains are essential: the fact that ∩ R = ∅ and the connectedness of I for
all I ∈ S.
In the next section, we will present a symmetry condition on the domain that
guarantees continuity and differentiability for regular functions → H.
We now present a very peculiar property of regular functions. Let us start with the
following property of quaternionic powers, which is a direct consequence of the
(complex) binomial theorem.
Remark 1.14 For each x, y ∈ R, there exist sequences {αn }n∈N , {βn }n∈N ⊂ R such
that (x + yI )n = αn + βn I for all I ∈ S.
As a consequence, the following formula holds for a regular function f (q) =
na :
n∈N q n
f (x + yI ) = αn an + I βn a n .
n∈N n∈N
6 1 Definitions and Basic Results
This formula has a nice geometric interpretation: the restriction of f to the sphere
x + yS = {x + yI : I ∈ S}
is affine in the imaginary unit I, that is, there exist b, c ∈ H such that
f (x + yI ) = b + I c (1.7)
for all I ∈ S. This is not only true for power series but for all regular functions on
the slice domains that have the following property.
Definition 1.15 A set T ⊆ H is called axially symmetric if, for all points x + yI ∈
T with x, y ∈ R and I ∈ S, the set T contains the whole sphere x + yS.
Since no confusion can arise, we will refer to such a set as symmetric, tout court.
An example of symmetric slice domain is portrayed in Fig. 1.1. We are now ready
for the announced result.
Theorem 1.16 (Representation Formula) Let f be a regular function on a sym-
metric slice domain and let x + yS ⊂ . For all I, J, K ∈ S with J = K, we have
He put his trunk down for a ladder, and Mr. Frog climbed up. They
had not gone very far when Mr. Frog hopped up close to Mr.
Elephant’s ear, and he said:
“I am going to fall, grandfather. Give me some small cords from
the roadside that I may bind your mouth, and hold myself upon your
back.”
“I will, grandson,” said Mr. Elephant.
So Mr. Elephant stripped some small cords from a birch tree by
the roadside, and handed them to Mr. Frog. Then Mr. Frog bound Mr.
Elephant’s mouth, and they went on a little farther. It was not long,
though, before Mr. Frog spoke again to Mr. Elephant.
“Grandfather,” he said, “find me a small, green twig that I may fan
the mosquitoes from your ears.”
“I will, grandson,” said Mr. Elephant, so he broke a small, green
twig from the birch tree, and reached it up to Mr. Frog; and just then
they came toward home.
“See Mr. Elephant,” cried Mr. Hare.
“See Mr. Elephant,” cried Mr. Tiger.
“See Mr. Elephant,” cried Mr. Lion and all the others, “Mr. Elephant
is Mr. Frog’s horse.”
Mr. Elephant turned himself about, and he saw Mr. Frog on his
back, holding the reins and the whip.
“Why, so I am, grandson,” said Mr. Elephant.
Then Mr. Frog jumped down to the ground, and he laughed and he
laughed until he nearly split his coat, because he had played a trick
on Mr. Elephant.
HOW DRAKESTAIL
WENT TO THE
KING.
Once upon a time there was a wee little duck, with a very long tail,
so he was called Drakestail. Now, Drakestail had some money of his
very, very own, and the king asked if he might take it. So Drakestail
loaned all his money to the king.
But the king kept Drakestail’s money for a year and a day, and still
he did not send it back. Drakestail said he would go to the king and
bring back the money himself.
So off he started, one very fine morning, for the king’s house. The
sun was shining on the ponds, and Drakestail waddled along in the
middle of the road, feeling very fine. As he traveled, he met a fox,
and the fox said, “Where do you go this fine morning, Friend
Drakestail?”
“To the king,” said Drakestail, “for he owes me money.”
“I will travel along with you,” said the fox.
“Ah,” said Drakestail, “your four legs would soon tire. Come along
with me this way,” and he opened his wee little bill very wide, and
down his wee little throat went the fox.
Then Drakestail traveled on a little farther. As he went he came to
a ladder lying beside the road.
“Where do you go this fine morning, Friend Drakestail?” asked the
ladder.
“To the king,” said Drakestail, “for he owes me money.”
“I will travel along with you,” said the ladder.
“Your wooden legs would soon tire,” said Drakestail. “Come along
with me this way,” and he opened his little bill very wide, and down
his wee little throat went the ladder.
Then Drakestail traveled on a bit farther until he came to his friend,
the river, lying and glistening in the sunshine.
“Where do you go this fine morning, Friend Drakestail?” asked the
river.
“To the king, for he owes me money,” said Drakestail.
“I will travel with you,” said the river.
“You would soon tire if you ran so far, my friend,” said Drakestail.
“Come along with me this way.” He opened his wee bill very wide,
and down his wee little throat went the little river.
Then Drakestail traveled and traveled until he came to the king’s
house. Now Drakestail thought that the king would meet him at the
gate, so he called out very loudly:
“Honk! Honk! Drakestail waits at the gate.”
But the king did not come out to meet him. Who should appear at
the gate but the king’s cook, and the cook took Drakestail by his two
little legs and flung him into the poultry yard. The other fowls, who
were ill-bred birds, ran up to Drakestail and bit him, and jeered at his
large tail. It would have gone very badly with Drakestail, but he
called to his friend, the fox:
So the fox came out, and he ate up all the ill-bred fowls in the
king’s poultry yard. But still Drakestail was badly off. He heard the
king’s cook putting the broth pot over the fire.
he cried. So the ladder came out and leaned against the wall, and
Drakestail climbed over in safety. But the king’s cook saw Drakestail
and set out after him. He caught poor Drakestail and clapped him
into the broth pot, and hung him over the fire.
cried Drakestail. So the river put out the fire with a great noise and
sputtering, which the king heard. And the king came running to the
kitchen.
“Good morning to you, King,” said Drakestail, hopping out of the
broth pot, and making a very low bow, “are you through with my
money, which you have kept for a year and a day?”
“That I am, Drakestail,” said the king. “You shall have it at once.”
So the king gave Drakestail the money that he owed him, and
Drakestail waddled home again to tell of all his travels.
THE GREEDY CAT.
Once upon a time there lived a cat and a mouse, and they thought
they would ask each other to dinner, turn and turn about. First it was
the cat’s turn to ask the mouse, and he set his table and invited her,
but he did not have much to eat; only a dry crust of bread and some
water. But the mouse, who was very polite, ate it and thanked the
cat.
When it was the mouse’s turn to give a dinner, she spread a fine
feast, platters of fish, and saucers of milk, and joints of meat. Then
she baked a large cake with sugar on the top for the cat, and for
herself she made a very tiny cake with no frosting.
The cat came to the mouse’s dinner, and he ate the fish and the
meat, and lapped the milk, and ate the cake. Then he looked around
in a greedy way, and he said:
“What a very light dinner. Have you nothing more in the house to
eat, mouse?”
“Here is my cake,” said the mouse, who was not at all greedy.
So the cat ate the mouse’s cake, and then he looked about again
in a greedy way, and he said:
“Have you anything more to eat, mouse?”
“Nothing, kind sir,” said the mouse, “unless you eat me.”
She thought the cat would never be so greedy as that, but he
opened his mouth wide, and down his throat went the mouse.
Then the greedy cat walked out of the mouse’s house and down
the road, swinging his tail, for he felt very fine.
On his way he met an old woman. Now the old woman had been
peeping in at the window, and she had seen what that greedy cat
had done.
“You greedy cat,” she said, “to eat your friend, the mouse.”
“Greedy, indeed,” said the cat, “I have a mind to eat you.”
Then he opened his mouth very wide, and down his throat went
the old woman.
Then on down the road went the cat, swinging his tail, and feeling
finer than ever. As he went he met an old man taking his load of
apples to market. The old man was beating his donkey to make it go
faster.
“Scat, scat, pussy,” said the man, “my donkey will tread on you.”
“Tread on me, indeed,” said the cat, shaking his fat sides, “I have
eaten my friend the mouse, I have eaten an old woman. What is to
hinder my eating you?”
So the greedy cat opened his mouth very wide, and down his
throat went the man and his donkey.
Then he walked along in the middle of the road again. After a
while he spied a great cloud of dust, and he heard a great tramping
of feet. It was the king riding in his chariot, and behind him marched
all his soldiers and his elephants.
“Scat, scat, pussy,” said the king, “my elephants might step on
you.”
“Step on me, indeed,” said the cat, “I have eaten my friend the
mouse, I have eaten an old woman, I have eaten an old man and a
donkey. What is to hinder my eating a king and a few elephants?”
So the cat opened his mouth wide, and down his throat went the
king and the soldiers and all the elephants.
Then the cat started on again, but more slowly. He was really not
hungry any more. As he traveled he met two land crabs, scuttling
along in the dust.
“Scat, scat, pussy,” squeaked the crabs.
“I have eaten my friend the mouse,” said the cat, “I have eaten an
old woman, and a man and a donkey, and a king and all his soldiers
and all his elephants. What is to hinder my eating you, too?”
Then the cat opened his mouth wide, and down his throat went the
two crabs.
But the crabs began to look about them there in the dark. There
were the soldiers trying to form in fours, but there was not room. The
elephants were stepping on each other’s toes. The old woman was
scolding, and in a corner sat the poor little mouse, her paws and
ears all drooping.
“We must go to work,” said the crabs.
Then they began snipping and snipping with their sharp little
claws. Soon there was a hole large enough, and they crept out.
Then out came the king and his soldiers and all his elephants. Out
came the old woman scolding her cat. Out came the man and his
donkey. Last of all, out came the little mouse with one little cake
under her arm, for one cake was all that she had wanted.
But the greedy cat had to spend all the rest of the day sewing up
the hole in his coat.
THE THREE BILLY
GOATS GRUFF.
Once upon a time there were three Billy Goats, and one was a
very large Goat, and one was a middle-sized Goat, and one was a
tiny Goat, but the three had the very same name, which was Gruff.
One morning the three Billy Goats started away from home, for
they had decided to go far, far to a hillside where there was a
quantity of green grass, and they might eat of it and make
themselves fat.
Now, on the way to the hillside there ran a brook, and over the
brook was a bridge, and under the bridge lived a Troll with eyes as
large as saucers, and a nose as long as a poker. And this Troll was
fond of eating Billy Goats.
First of all came the youngest Billy Goat Gruff to cross over the
bridge. Trip trap, trip trap, his little feet pattered upon the boards.
“Who is that tripping over my bridge?” called up the Troll in a surly
voice.
“Oh, it’s only I, the tiniest Billy Goat Gruff, going over to the hillside
to make myself fat,” the Goat called back in a wee small voice.
“I am going to gobble you up, Billy Goat Gruff,” said the Troll.
“Oh, no, pray do not take me,” said the tiniest Billy Goat Gruff; “I
am too little, that I am. Wait until the second Billy Goat Gruff comes
along. He is ever so much bigger than I.”
“Well, be off with you,” said the Troll.
Then came the middle-sized Billy Goat Gruff, to cross the bridge.
Trip trap, trip trap, his middle-sized feet pattered upon the boards.
“Who is that tripping over my bridge?” called up the Troll.
“Oh, it’s only I, the middle-sized Billy Goat Gruff, going over to the
hillside to make myself fat,” the Goat called back in a middle-sized
voice.
“I am coming to gobble you up, Billy Goat Gruff,” said the Troll.
“Oh, no, pray do not take me,” said the middle-sized Billy Goat
Gruff; “I am a little larger than the tiniest Billy Goat, but I am not large
enough to make a mouthful for you. Of that I am quite sure.”
“Well, be off with you,” said the Troll.
Then, last of all, came the great Billy Goat Gruff, to cross over the
bridge.
Trip trap, trip trap, his great feet tramped across the boards.
“Who is that tramping over my bridge?” called up the Troll.
“It is I, the great Billy Goat Gruff, going over to the hillside to make
myself fat,” the Goat called back in a great voice.
“I am coming to gobble you up, Billy Goat Gruff,” said the Troll.
“Come along,” said the great Billy Goat Gruff.
So the Troll, whose eyes were as large as saucers and his nose
as long as a poker, came hurrying up to the top of the bridge,—but,
ah, this is what happened to him.
The Goat tossed the Troll so high with his horns.
There on the bridge stood the great Billy Goat Gruff with his feet
firmly planted on the boards and his head lowered, and as soon as
the Troll came near—rush, scamper—the Goat tossed the Troll so
high with his horns that no one has ever seen a Troll under a bridge
from that day to this.
Then the great Billy Goat Gruff went on to the hillside, and the
three Billy Goats ate, and ate, and made themselves so fat that they
could scarcely walk home again.
THE HOBYAHS.
Once upon a time there lived a little old man and a little old woman
in a house all made of hemp stalks. And they had a little dog named
Turpie who always barked when any one came near the house.
One night when the little old man and the little old woman were
fast asleep, creep, creep, through the woods came the Hobyahs,
skipping along on the tips of their toes.
“Tear down the hemp stalks. Eat up the little old man, and carry
away the little old woman,” cried the Hobyahs.
Then little dog Turpie ran out, barking loudly, and he frightened the
Hobyahs so that they ran away home again. But the little old man
woke from his dreams, and he said:
“Little dog Turpie barks so loudly that I can neither slumber nor
sleep. In the morning I will take off his tail.”
So when it came morning, the little old man took off little dog
Turpie’s tail to cure him of barking.
The second night along came the Hobyahs, creep, creep through
the woods, skipping along on the tips of their toes, and they cried:
“Tear down the hemp stalks. Eat the little old man, and carry away
the little old woman.”
Then the little dog Turpie ran out again, barking so loudly that he
frightened the Hobyahs, and they ran away home again.
But the little old man tossed in his sleep, and he said:
“Little dog Turpie barks so loudly that I can neither slumber nor
sleep. In the morning I will take off his legs.”
So when it came morning, the little old man took off Turpie’s legs
to cure him of barking.
The third night the Hobyahs came again, skipping along on the
tips of their toes, and they called out:
“Tear down the hemp stalks. Eat up the little old man, and carry
away the little old woman.”
Then little dog Turpie barked very loudly, and he frightened the
Hobyahs so that they ran away home again.
But the little old man heard Turpie, and he sat up in bed, and he
said:
“Little dog Turpie barks so loudly that I can neither slumber nor
sleep. In the morning I will take off his head.”
So when it came morning, the little old man took off Turpie’s head,
and then Turpie could not bark any more.
That night the Hobyahs came again, skip- ping along on the tips of
their toes, and they called out:
“Tear down the hemp stalks. Eat the little old man, and carry off
the little old woman.”
Now, since little dog Turpie could not bark any more, there was no
one to frighten the Hobyahs away. They tore down the hemp stalks,
they took the little old woman away in their bag, but the little old man
they could not get, for he hid himself under the bed.
Then the Hobyahs hung the bag which held the little old woman up
in their house, and they poked it with their fingers, and they cried:
“Look you! Look you!”
But when it came daylight, they went to sleep, for Hobyahs, you
know, sleep all day.
The little old man was very sorry when he found that the little old
woman was gone. He knew then what a good little dog Turpie had
been to guard the house at night, so he brought Turpie’s tail, and his
legs, and his head, and gave them back to him again.
Then Turpie went sniffing and snuffing along to find the little old
woman, and soon came to the Hobyahs’ house. He heard the little
old woman crying in the bag, and he saw that the Hobyahs were all
fast asleep. So he went inside.
Then he cut open the bag with his sharp teeth, and the little old
woman hopped out and ran home; but Turpie got inside the bag to
hide.
When it came night, the Hobyahs woke up, and they went to the
bag, and they poked it with their long fingers, crying:
“Look you! Look you!”
But out of the bag jumped little dog Turpie, and he ate every one
of the Hobyahs. And that is why there are not any Hobyahs now.
THE KID WHO
WOULD NOT GO.
Once upon a time there was an old gray Pussy and she was down
by the waterside when the trees and ground were white with snow.
And there she saw a wee, wee Robin Redbreast hopping upon a
branch, so Pussy said to him:
“Where are you going, Robin Redbreast, this frosty Yuletide
weather?”
Then the wee, wee Robin said to the Pussy, “I am going to the
King to sing him a song this good Yule morning.”
And the gray Pussy replied, “Go not yet. Come here, Robin
Redbreast, and I will let you see the bonny white necklace I wear
around my neck.”
But the wee, wee Robin said, “No, no, gray Pussy. You may show
the bonny white necklace that you wear around your neck to the little
mice, but not to me.”