Regular Functions of a Quaternionic

Variable 2nd Edition Graziano Gentili

Caterina Stoppato Daniele C Struppa
Visit to download the full and correct content document:
https://ebookmeta.com/product/regular-functions-of-a-quaternionic-variable-2nd-editio
n-graziano-gentili-caterina-stoppato-daniele-c-struppa/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

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/

Functions of One Complex Variable I 2nd Edition John B

Conway

https://ebookmeta.com/product/functions-of-one-complex-
variable-i-2nd-edition-john-b-conway/

Complete Solutions Manual for Calculus of a Single

Variable, Early Transcendental Functions 7th Edition
Ron Larson

https://ebookmeta.com/product/complete-solutions-manual-for-
calculus-of-a-single-variable-early-transcendental-functions-7th-
edition-ron-larson/

Calculus of one variable 2nd Edition M Thamban Nair

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/

Regular Extensions of Hermitian Operators A V Kuzhel S

A Kuzhel

https://ebookmeta.com/product/regular-extensions-of-hermitian-
operators-a-v-kuzhel-s-a-kuzhel/

C++17 Standard Library Quick Reference, 2nd Edition: A

Pocket Guide to Data Structures, Algorithms, and
Functions Peter Van Weert

https://ebookmeta.com/product/c17-standard-library-quick-
reference-2nd-edition-a-pocket-guide-to-data-structures-
algorithms-and-functions-peter-van-weert/

Extension of Holomorphic Functions 2nd Edition Marek

Jarnicki Peter Pflug

https://ebookmeta.com/product/extension-of-holomorphic-
functions-2nd-edition-marek-jarnicki-peter-pflug/

Statistical Analysis of Graph Structures in Random

Variable Networks V. A. Kalyagin

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)

ISSN 1439-7382 ISSN 2196-9922 (electronic)

Springer Monographs in Mathematics
ISBN 978-3-031-07530-8 ISBN 978-3-031-07531-5 (eBook)
https://doi.org/10.1007/978-3-031-07531-5

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

The theory of slice regular functions (originally Cullen-regular functions) was

born at George Mason University, in Virginia, where the first and third authors
collaborated for a 2-month period in the Fall of 2005. It was originated by
the desire to find a new class of quaternionic regular functions that included
polynomials and power series. The second author started working on this subject
in the Summer of 2006, and her doctoral thesis eventually became the skeleton for
this monograph. The theory of slice regular functions has rapidly developed, thanks
to a series of visits at Chapman University, in California and to the interest of many
mathematicians to whom we are greatly indebted.
We are very grateful to Fabrizio Colombo and Irene Sabadini, who immediately
realized that this theory could be applied to create a successful quaternionic
functional calculus. They also suggested the extension of these ideas to the case
of Clifford algebras and their impulse has greatly contributed to the development of
the theory.
We would like to thank Riccardo Ghiloni and Alessandro Perotti, who took an
active interest in these developments and introduced a new viewpoint on the theory
itself.
We warmly thank Cinzia Bisi, Alberto Damiano, Chiara Della Rocchetta, Giulia
Sarfatti, Irene Vignozzi, and Fabio Vlacci for their interest in the subject and for
their researches, which contributed to the expansion of the theory presented in this
book.
We should also express our gratitude to Michael (Misha) V. Shapiro and to María
Elena Luna–Elizarrarás for their discussions with us and especially to Misha for
his help in crafting a new introduction to this work, which better represents its
relationship with other lines of research in the quaternionic field.
Special thanks go to Simon Salamon for his role in an unexpected application
to the construction and classification of orthogonal complex structures in the
quaternionic space.
Last but not least, we want to express our gratitude to the institutions who
have supported us with the time needed for this work and in many cases have
granted travel or local living expenses to the three of us. We gratefully acknowledge

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.”

Orange, CA, USA Graziano Gentili

July 2012 Caterina Stoppato
Daniele C. Struppa
Preface to the Second Edition

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”.

Florence, Italy Graziano Gentili

Florence, Italy Caterina Stoppato
Orange, CA, USA Daniele C. Struppa
July 2022

ix
Introduction

This book presents a comprehensive treatment of a recent theory of quaternionic

functions, introduced in 2006 by two of the authors [141]. In this book, we will
develop this theory in a self-contained fashion and show in which sense it is another
important way to generalize the notion of holomorphy to the setting of quaternions.
Quaternions were introduced by Hamilton in 1843, adding a multiplicative
structure to R4 . In a modern notation, the vectors (x0 , x1 , x2 , x3 ) ∈ R4 are
represented as

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,

then extended to all elements of R4 in a bilinear fashion. This process defines a

(noncommutative) associative real algebra, usually denoted by H. In this algebra,
every nonzero element has a multiplicative inverse: after defining the conjugate of
an element q = x0 + x1 i + x2 j + x3 k ∈ H as

q̄ = x0 − x1 i − x2 j − x3 k

and its norm as

 
|q| = q q̄ = x02 + x12 + x22 + x32 ,

it is not difficult to prove that if q = 0 then

q −1 = |q|−2 q̄.

xi
xii Introduction

Hence, H is an associative division algebra (also called a skew field). Hamilton

arrived to his construction after attempting (and failing) to build a similar structure
on a three-dimensional vector space: only in 1877 Frobenius proved that R, C, H are
the only finite dimensional, associative real algebras with division. The construction
of H also shortly preceded the introduction of the modern vector notation, which
was later ostracized by Hamilton’s followers as a “perverted” version of the
quaternionic structure. The birth of quaternions and their algebraic properties are
well described in [108]. For general references in noncommutative algebra, see
[59, 210, 211, 255].
Since the beginning of last century, there have been many attempts to determine a
class of quaternion-valued functions of one quaternionic variable playing the same
role as the holomorphic functions of one complex variable. In the complex case,
a function f is called holomorphic if it admits complex derivative, i.e., a complex
number f  (z) such that, for h ∈ C,

f (z + h) − f (z)
lim = f  (z).
h→0 h

This is also sometimes expressed by saying that f is complex differentiable at z,

namely there exists a complex number f  (z) such that

f (z + h) − f (z) = f  (z) · h + o(h).

Equivalently, f is termed holomorphic if it is a solution of the Cauchy–Riemann

equation, i.e.,
 
∂f 1 ∂ ∂
:= +i f = 0.
∂ z̄ 2 ∂x ∂y

A third equivalent definition calls holomorphic a function that is complex analytic,

i.e., one that expands into complex Taylor series in a neighborhood of any point of
its domain of definition.
These concepts cease to be equivalent in the quaternionic context, not only for
lack of commutativity (from which distinct notions of right and left quotients arise).
For instance, asking for a function f : H → H to be quaternion-differentiable, i.e.,
imposing that, for h ∈ H,

lim h−1 [f (q + h) − f (q)]

h→0

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

functions of quaternionic variables is possible. It is indeed the case. The starting

point is observing that complex differentiability is inherently connected with the
Cauchy–Riemann operator, which, in turn, is related to one-dimensional directional
derivatives. Good ways to generalize complex differentiability to the quaternionic
case go through suitable generalizations of the Cauchy–Riemann system, which, in
turn, indicate the relevant directional derivatives. We will provide a quick overview,
but the interested reader is referred to [222], as well as [223–225, 228, 269] and
references therein.
The most successful analog of the Cauchy–Riemann equations was developed in
the 1920s through the efforts of various mathematicians such as Moisil, Teodorescu,
and finally Fueter, to whose work most current literature refers. In [122], Fueter
defined a quaternionic function to be regular if it solves the equation
 
∂f 1 ∂ ∂ ∂ ∂
= +i +j +k f ≡ 0.
∂ q̄ 4 ∂x0 ∂x1 ∂x2 ∂x3

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 (2) = idx2 ∧ dx3 + j dx3 ∧ dx1 + kdx1 ∧ dx2

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

d(σ (2) f ) = σ (3) · f  .

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

notion of differentiability in the quaternionic setting, compatible with the Cauchy–

Fueter operator, can be defined with a quaternionic increment along, so to speak, a
three-dimensional real manifold. This codimension-one real manifold is an analog
of the curves (codimension-one real manifolds) used in the complex plane.
Let us now turn to possible analogs of the third equivalent definition of
holomorphy, namely analyticity. Once again, in a naive analogy with the complex
case, one could require the existence at each q0 of a power series expansion

f (q) = Pn (q − q0 )
n∈N

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:

holomorphic in z1 and antiholomorphic in z2 ⇒ Fueter-regular

⇒ harmonic in x0 , x1 , x2 , x3 .

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

where Im(q) = x1 i + x2 j + x3 k denotes

 the imaginary part of the variable q =
x0 +x1 i+x2 j +x3 k and r = |Im(q)| = x12 + x22 + x32 . The number x0 is also called
the real part of q and denoted by Re(q). Gentili and Struppa restated and developed
Cullen’s definition in [141, 142] by means of the algebraic properties of H that we
Im(q)
are about to describe. For all q with Im(q) = 0, the normalized vector |Im(q)| is
an imaginary unit, i.e., a quaternion whose square is −1. Hence, every quaternion
q ∈/ R can be uniquely expressed as q = x + yI for some x, y ∈ R, y > 0 and
some I in the set of imaginary units

S = {q ∈ H : q 2 = −1} = {x1 i + x2 j + x3 k : x12 + x22 + x32 = 1},

xvi Introduction

which is a 2-sphere in the 3-space of purely imaginary quaternions. In other words

(denoting R = R · 1 ⊂ H), the algebra H is the union

H= (R + RI )
I ∈S

of the planes LI := R + RI , which are all isomorphic to C and intersect in the

real axis. This decomposition measures quite precisely the lack of commutativity
of H: two quaternions commute if, and only if, they belong to the same plane
LI ; this is always the case when one of them is real. To go back to Gentili
and Struppa’s definition, they originally called a function f Cullen-regular if it is
complex holomorphic when restricted to each plane LI , i.e., if it solves
 
1 ∂ ∂
+I f (x + yI ) = 0 (1)
2 ∂x ∂y

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

(with an ∈ H) define Cullen-regular functions on their sets of convergence, which

are Euclidean balls centered at the origin of H. Conversely, every Cullen-regular
function on such a ball admits a series expansion of this type. These two properties
yield that Cullen-regularity does not imply, nor is implied by, Fueter-regularity: the
function q → q 2 is an example of Cullen-regular function which is not harmonic,
whence not Fueter-regular nor holomorphic/antiholomorphic in z1 /z2 ; the function
x0 + x1 i + x2 j + x3 k → x0 + x1 i = z1 is holomorphic/antiholomorphic in
z1 /z2 , whence Fueter-regular, harmonic and quaternionic holomorphic, but not
Cullen-regular. In fact, over suitable quaternionic domains, the only quaternionic
functions that are both Fueter-regular and Cullen-regular are the constant functions
(see Corollary 3.6.2 in [236]).
In the original works [141, 142], the authors developed the foundational material
for Cullen-regular functions on balls centered at the origin. In particular, they proved
a first version of the Cauchy Formula and of some of its consequences (Cauchy’s
inequalities, Liouville Theorem, Morera Theorem, etc.). A fundamental technical
tool is the so-called Splitting Lemma:
Lemma 1 (Splitting) Let f be a Cullen-regular function defined on an open
connected set . Then for any I ∈ S and any J ∈ S with J ⊥ I , there exist
two holomorphic functions F, G :  ∩ LI → LI such that for every z = x + yI it
is

fI (z) = F (z) + G(z)J.

Introduction xvii

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

In Chap. 6 we present the regular analogs of many classical integral formulas

in complex analysis. We introduce versions of the Cauchy Theorem and of the
Morera Theorem. We present results such as the Cauchy Integral Formula, the
formula for derivatives, and the Pompeiu Formula, giving new proofs. We illustrate
the Cauchy Estimates, the Liouville Theorem, and an integral formula to compute
the coefficients of regular Laurent expansions. We conclude this chapter with an
argument principle for regular functions.
Chapter 7 presents the Maximum Modulus Principle for slice domains and sev-
eral important results valid over symmetric slice domains: the Minimum Modulus
Principle, the Open Mapping Theorem, and the study of the real parts of regular
functions. The chapter also contains Principles of Phragmén–Lindelöf type and an
Ehrenpreis–Malgrange Lemma for quaternionic polynomials.
In Chap. 8, we present series expansions valid in open subsets of H and a notion
of analyticity which is equivalent to regularity on axially symmetric slice domains.
This allows a detailed study of the real differentials of regular functions and of their
ranks.
Chapter 9 begins with the Schwarz Lemma, then it presents the regular trans-
formations of the quaternionic space, unit ball, and Riemann sphere. It overviews
rigidity results for regular functions on the unit ball, both of Cartan and of Burns–
Krantz type. It then presents the Borel–Carathéodory Theorem and the Bohr
Theorem for regular functions.
In Chap. 10, we overview several generalizations and applications of the theory.
Specifically, we describe the theory of regular functions on the space of octonions
and on the Clifford algebra R3 = Cl(0, 3), as well as the notion of slice monogenic
function with values in Rm = Cl(0, m). We also present the alternative approach
to regularity due to Ghiloni and Perotti, including the notion of slice function.
Chapter 11 is entirely devoted to the theory of regular functions on nonsymmetric
slice domains. The approach adopted here follows the lines of Ghiloni and Perotti’s
approach. Thus, the starting point is positioning the class of regular functions
on a slice domain within the larger class of locally slice functions. Then, it is
possible to endow both classes with an algebraic structure, which, in turn, allows
the study of the zero sets and the definition of quotients. For regular functions,
singularities are classified and studied. The Minimum Modulus Principle and the
Open Mapping Theorem are revisited in this new context, as well as the integral
representation formulas and the spherical series expansions. The case of spherical
Laurent expansions is special, as they are treated here for both symmetric and
nonsymmetric slice domains.
The final chapter, Chap. 12, overviews various applications of the theory of
regular functions. These include quaternionic functional calculus and function
spaces as well as the applications in differential geometry we already mentioned:
to the classification of Orthogonal Complex Structures on open dense subsets of R4
and to quaternionic manifolds. We also cite some applications in spatial kinematics,
namely to the study of Rational Rotation-Minimizing Frame curves and of motion
polynomials over dual quaternions.
Introduction xxi

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

1 Definitions and Basic Results . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1

1.1 Regular Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1
1.2 Affine Representation . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 5
1.3 Extension Results and Local Representation .. . . . . . . . . . . . . . . . . . . . 10
1.4 Algebraic Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 21
2 Regular Power Series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 29
2.1 The Distance σ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 31
2.2 Convergence of Power Series Centered at p . .. . . . . . . . . . . . . . . . . . . . 33
2.3 Series Expansion at p and Analyticity . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 36
3 Zeros . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 41
3.1 Basic Properties of the Zeros . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 41
3.2 Algebraic Properties of the Zeros . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 42
3.3 Topological Properties of the Zeros . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 45
3.4 On the Roots of Quaternions . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 46
3.5 Factorization of Polynomials . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 48
3.6 Multiplicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
3.7 Division Algorithm and Bezout Theorem . . . . .. . . . . . . . . . . . . . . . . . . . 60
3.8 Gröbner Bases for Quaternionic Polynomials . . . . . . . . . . . . . . . . . . . . 66
4 Infinite Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 73
4.1 Infinite Products of Quaternions . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 73
4.2 The Principal Branch of Quaternionic Logarithm . . . . . . . . . . . . . . . . 75
4.3 Infinite Products of Functions Defined on H . .. . . . . . . . . . . . . . . . . . . . 80
4.4 Convergence of an Infinite ∗ -Product . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 84
4.5 Convergence-Producing Regular Factors . . . . .. . . . . . . . . . . . . . . . . . . . 89
4.6 Weierstrass Factorization Theorem . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 93
5 Singularities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 99
5.1 Regular Reciprocal and Quotients . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 99
5.2 Laurent Series and Expansion . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 101
5.3 Classification of Singularities . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 106

xxiii
xxiv Contents

5.4 Poles and Quotients . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 108

5.5 Casorati–Weierstrass Theorem . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 112
6 Integral Representations .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 115
6.1 Cauchy Theorem and Morera Theorem . . . . . . .. . . . . . . . . . . . . . . . . . . . 115
6.2 Cauchy Integral Formula . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 116
6.3 Pompeiu Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 118
6.4 Derivatives Using the Cauchy Formula . . . . . . .. . . . . . . . . . . . . . . . . . . . 119
6.5 Coefficients of the Laurent Series Expansion .. . . . . . . . . . . . . . . . . . . . 121
6.6 Argument Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 122
7 Maximum Modulus Theorem and Applications . . . .. . . . . . . . . . . . . . . . . . . . 131
7.1 Maximum and Minimum Modulus . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 131
7.2 Open Mapping Theorem . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 133
7.3 Real Parts of Regular Functions . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 134
7.4 Phragmén–Lindelöf Principles . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 143
7.5 An Ehrenpreis–Malgrange Lemma . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 148
8 Spherical Series and Differential . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 157
8.1 Spherical Series and Expansions . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 157
8.2 Integral Formulas and Cauchy Estimates . . . . .. . . . . . . . . . . . . . . . . . . . 160
8.3 Symmetric Analyticity . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 163
8.4 Differentiating Regular Functions . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 165
8.5 Rank of the Differential . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 168
9 Fractional Transformations and the Unit Ball . . . . . .. . . . . . . . . . . . . . . . . . . . 175
9.1 Transformations of the Quaternionic Space . . .. . . . . . . . . . . . . . . . . . . . 175
9.2 Regular Fractional Transformations . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 176
9.3 Transformations of the Quaternionic Riemann Sphere . . . . . . . . . . . 180
9.4 Schwarz Lemma and Transformations of the Unit Ball . . . . . . . . . . 182
9.5 Rigidity and a Boundary Schwarz Lemma . . . .. . . . . . . . . . . . . . . . . . . . 185
9.6 Borel–Carathéodory Theorem . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 188
9.7 Bohr Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 192
10 Generalizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 197
10.1 Direct Generalizations to Algebras Other Than H . . . . . . . . . . . . . . . . 197
10.1.1 The Case of Octonions . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 197
10.1.2 The Case of R3 . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 200
10.1.3 The Slice Monogenic Case . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 204
10.2 Slice Functions Over H and Other Alternative Real
Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 206
10.3 An Alternative Approach to Slice Regularity . . . . . . . . . . . . . . . . . . . . . 209
11 Function Theory Over Non-symmetric Slice Domains . . . . . . . . . . . . . . . . . 213
11.1 Spherical Value and Derivative . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 213
11.2 Locally Slice Functions and their Algebraic Structure . . . . . . . . . . . 216
11.3 Zero Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 221
11.4 Quotients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 225
Contents xxv

11.5 Factorization of Zeros . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 230

11.6 Applications of Factorization . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 235
11.7 Semiregular Functions . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 242
11.8 Minimum Modulus Principle and Open Mapping
Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 245
11.9 Integral Representation . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 247
11.10 Spherical Series Expansions . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 250
12 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 257
12.1 Applications in Functional Analysis . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 257
12.1.1 Quaternionic Functional Calculus . . .. . . . . . . . . . . . . . . . . . . . 257
12.1.2 Some Quaternionic Function Spaces . . . . . . . . . . . . . . . . . . . . 259
12.2 Applications in Differential Geometry . . . . . . . .. . . . . . . . . . . . . . . . . . . . 260
12.2.1 Orthogonal Complex Structures Induced by
Regular Functions . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 260
12.2.2 A Direct Approach to Quaternionic Manifolds . . . . . . . . . 263
12.3 Applications in Spatial Kinematics . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 265
12.3.1 Rational Rotation-Minimizing Frame Curves . . . . . . . . . . . 266
12.3.2 Motion Polynomials over Dual Quaternions . . . . . . . . . . . . 267

References .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 269
Index . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 283
Chapter 1
Definitions and Basic Results

1.1 Regular Functions

Let  be a domain in the space of quaternions H, namely an open connected subset

of H = R + iR + j R + kR, and let

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.

Remark 1.2 It is useful to note that if f is regular in a domain  and if r is a real

number, then g(q) = f (q + r) is obviously regular on  − r. Note, however, that,
in general, the composition of two regular functions is not regular.
The relation between quaternionic regularity and complex holomorphy can be
clarified as follows. If U is an open subset of LI , a C 1 map φ : U → H is called
holomorphic if its real differential dφ commutes with the constant complex structure
induced by I (as defined in the forthcoming Sect. 12.2.1). Now, for each I ∈ S, let

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 1

G. Gentili et al., Regular Functions of a Quaternionic Variable, Springer
Monographs in Mathematics, https://doi.org/10.1007/978-3-031-07531-5_1
2 1 Definitions and Basic Results

us identify LI with C. Notice, moreover, that for all J ∈ S with J ⊥ I , the following
equality holds:

H = LI + LI J.

Lemma 1.3 (Splitting) Let f be a regular function defined on a domain . Then

for any I ∈ S and any J ∈ S with J ⊥ I , there exist two holomorphic functions
F, G : I → LI such that for every z = x + yI , it holds

fI (z) = F (z) + G(z)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

φ(z) = F (z) + G(z)J (1.2)

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

B(0, R) = {q ∈ H : |q| < R}

the Euclidean ball of radius R centered at 0 in H.

Theorem 1.6 (Abel’s Theorem) Let {an }n∈N be a sequence in H and let

1
R= . (1.3)
lim supn∈N |an |1/n

If R > 0, then the power series


q n an (1.4)
n∈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

on I . The slice derivative of f is the function f  = ∂c f :  → H defined by ∂I f

on I , for all I ∈ S.
The definition is well posed because, by direct computation, ∂I f = ∂J f in I ∩
J ⊆ R for any choice of I, J ∈ S. Furthermore, the fact that ∂I and ∂¯I commute
allows us to make the next remark.
Remark 1.9 For any regular function f :  → H, the slice derivative f  is regular
in .
Thus, it is possible to iterate the derivation process. Let us denote the nth slice
derivative as f (n) for each n ∈ N. We now come to the announced result.
Theorem 1.10 (Series Expansion) Let R > 0 and let f : B = B(0, R) → H be a
regular function. Then
 1 (n)
f (q) = qn f (0) (1.6)
n!
n∈N

for all q ∈ B. In particular, f ∈ C ∞ (B).

4 1 Definitions and Basic Results

Proof Fix I ∈ S and identify LI with C. Choose J ∈ S such that J ⊥ I : by the

Splitting Lemma 1.3, there exist holomorphic functions F, G : BI → LI such that
fI = F + GJ . Notice that, for all z ∈ BI ,

∂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

Now observe that the complex series

 1 ∂ nF
zn (0)
n! ∂zn
n∈N

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

as desired. Since I was arbitrarily chosen in S, the thesis follows. Finally, f ∈

C ∞ (B) because each summand q n n!1 f (n) (0) is clearly in C ∞ (B) and because the
convergence is uniform on compact sets. 

Propositions 1.6 and 1.10 are fundamental in the study of regular quaternionic
functions on balls B = B(0, R) centered at the origin of H. For instance, they allowed
the proof of an identity principle in [142], stating that if, for some I ∈ S, two regular
functions f, g : B → H coincide on a subset of BI having an accumulation point in
BI , then f = g in B. This principle does not hold for an arbitrarily chosen domain in
H, as shown by the next example.
Example 1.11 Let I ∈ S and let f : H \ R → H be defined as follows:

0 if q ∈ H \ LI
f (q) =
1 if q ∈ LI \ R

This function is clearly regular.

1.2 Affine Representation 5

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.

1.2 Affine Representation

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

Fig. 1.1 A view in

R + iR + j R of a symmetric
slice domain

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

f (x + yI ) = (J − K)−1 [Jf (x + yJ ) − Kf (x + yK)] +

(1.8)
+ I (J − K)−1 [f (x + yJ ) − f (x + yK)] .

The quaternion b = (J − K)−1 [Jf (x + yJ ) − Kf (x + yK)] and the quaternion

c = (J − K)−1 [f (x + yJ ) − f (x + yK)] do not depend on J, K but only on x, y.
Another random document with
no related content on Scribd:
muddy. Mr. Elephant knew a place where the bananas grew ripe and
thick. And they spent a pleasant day. On the way home Mr. Frog
hopped up close to Mr. Elephant, and he said in his high, squeaky
voice:
“Grandfather, I have no strength to walk. Let me get up on your
back.”
“Climb up, my grandson,” said Mr. Elephant.
 And just then they came toward home.

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:

“Reynard, Reynard, come out to the earth,

Or Drakestail’s life is of little worth.”

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.

“Ladder, ladder, come out to the wall,

 Drakestail does not wish to be broth at all,”

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.

“River, my sweetheart, put out this hot fire,

The flames that would cook me rise higher and higher,”

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 I was walking across London Bridge, and I

found a penny. So I bought a little kid. But the kid would not go. And I
saw by the moonlight it was long past midnight. It was time kid and I
were home an hour and a half ago.
Then I met a staff, and to the staff I said:
“Staff, staff, drive kid. I see by the moonlight it is long past
midnight. It is time kid and I were home an hour and a half ago.”
But the staff would not drive kid.
Then I met a hatchet, and to the hatchet I said:
“Hatchet, chop staff, staff will not drive kid. I see by the moonlight it
is long past midnight. It is time kid and I were home an hour and a
half ago.”
But the hatchet would not chop staff.
Then I met a torch, and to the torch I said:
“Torch, burn hatchet, hatchet will not chop staff, staff will not drive
kid. I see by the moonlight it is long past midnight. It is time kid and I
were home an hour and a half ago.”
But the torch would not burn the hatchet.
Then I met the wind, and to the wind I said:
“Wind, put out torch, torch will not burn hatchet, hatchet will not
chop staff, staff will not drive kid. I see by the moonlight it is long past
midnight. It is time kid and I were home an hour and a half ago.”
 But the wind would not put out the fire.
Then I met a tree, and to the tree I said:
“Tree, stop wind, wind will not put out torch, torch will not burn
staff, staff will not drive kid. I see by the moonlight it is long past
midnight. It is time kid and I were home an hour and a half ago.”
But the tree would not.
Then I met a wee mouse, and to the mouse I said:
“Mouse, gnaw tree, tree will not stop wind, wind will not put out
torch, torch will not burn staff, staff will not drive kid. I see by the
moonlight it is long past midnight. It is time kid and I were home an
hour and a half ago.”
Then the wee, wee mouse began to gnaw the tree, the tree began
to stop the wind, the wind began to put out the torch, the torch began
to burn the staff, the staff began to drive the kid, and the kid began to
go.
See by the moonlight it is almost midnight. But kid and I were
home an hour and a half ago.
 THE ROBIN’S
CHRISTMAS SONG.

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.”