Functional identities 1st Edition Matej Bresar

pdf download

https://ebookname.com/product/functional-identities-1st-edition-
matej-bresar/

Get Instant Ebook Downloads – Browse at https://ebookname.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Post Colonial Identities 1st Edition Chin Ce

https://ebookname.com/product/post-colonial-identities-1st-
edition-chin-ce/

Functional JavaScript Introducing Functional

Programming with Underscore js 1st Edition Michael
Fogus

https://ebookname.com/product/functional-javascript-introducing-
functional-programming-with-underscore-js-1st-edition-michael-
fogus/

Queer Identities Political Realities 1st Edition Bruce

Drushel

https://ebookname.com/product/queer-identities-political-
realities-1st-edition-bruce-drushel/

The Three Piece Suit and Modern Masculinity England

1550 1850 1St Edition Edition David Kuchta

https://ebookname.com/product/the-three-piece-suit-and-modern-
masculinity-england-1550-1850-1st-edition-edition-david-kuchta/
Icons 1. Auflage Edition Pavlovich Kondakov

https://ebookname.com/product/icons-1-auflage-edition-pavlovich-
kondakov/

Stalin s Usable Past A Critical Edition of the 1937

Short History of the USSR David Brandenberger

https://ebookname.com/product/stalin-s-usable-past-a-critical-
edition-of-the-1937-short-history-of-the-ussr-david-
brandenberger/

Understanding Operating Systems 7th edition Ann Mchoes

https://ebookname.com/product/understanding-operating-
systems-7th-edition-ann-mchoes/

Peace Culture and Violence 1st Edition Fuat Gursozlu

https://ebookname.com/product/peace-culture-and-violence-1st-
edition-fuat-gursozlu/

Implementing European Union Education and Training

Policy A Comparative Study of Issues in Four Member
States 1st Edition D. Phillips

https://ebookname.com/product/implementing-european-union-
education-and-training-policy-a-comparative-study-of-issues-in-
four-member-states-1st-edition-d-phillips/
Nitride Semiconductors Handbook on Materials and
Devices 1st Edition Pierre Ruterana

https://ebookname.com/product/nitride-semiconductors-handbook-on-
materials-and-devices-1st-edition-pierre-ruterana/
Frontiers in Mathematics

Advisory Editorial Board

Luigi Ambrosio (Scuola Normale Superiore, Pisa)

Leonid Bunimovich (Georgia Institute of Technology, Atlanta)
Benoît Perthame (Ecole Normale Supérieure, Paris)
Gennady Samorodnitsky (Cornell University, Rhodes Hall)
Igor Shparlinski (Macquarie University, New South Wales)
Wolfgang Sprössig (TU Bergakademie Freiberg)
Matej Brešar
Mikhail A. Chebotar
Wallace S. Martindale 3rd

Functional
Identities

Birkhäuser Verlag
Basel . Boston . Berlin
Authors:

Matej Brešar :DOODFH60DUWLQGDOHrd

Department of Mathematics Department of Mathematics
and Computer Science University of Massachusetts
FNM $PKHUVW0$
University of Maribor USA
2000 Maribor e-mail: [email protected]
Slovenia
e-mail: [email protected]

Mikhail A. Chebotar
Department of Mathematical Sciences
Kent State University
Kent, OH 44242
USA
e-mail: [email protected]

0DWKHPDWLFDO6XEMHFW&ODVVL¿FDWLRQ$15::
:%&%

/LEUDU\RI&RQJUHVV&RQWURO1XPEHU

Bibliographic information published by Die Deutsche Bibliothek

'LH'HXWVFKH%LEOLRWKHNOLVWVWKLVSXEOLFDWLRQLQWKH'HXWVFKH1DWLRQDOELEOLRJUD¿H
detailed bibliographic data is available in the Internet at <http://dnb.ddb.de>.

,6%1%LUNKlXVHU9HUODJ%DVHO%RVWRQ%HUOLQ

This work is subject to copyright. All rights are reserved, whether the whole or part
RIWKHPDWHULDOLVFRQFHUQHGVSHFL¿FDOO\WKHULJKWVRIWUDQVODWLRQUHSULQWLQJUHXVH
RILOOXVWUDWLRQVUHFLWDWLRQEURDGFDVWLQJUHSURGXFWLRQRQPLFUR¿OPVRULQRWKHUZD\V
and storage in data banks. For any kind of use permission of the copyright owner
must be obtained.

‹%LUNKlXVHU9HUODJ32%R[&+%DVHO6ZLW]HUODQG
3DUWRI6SULQJHU6FLHQFH%XVLQHVV0HGLD
&RYHUGHVLJQ%LUJLW%ORKPDQQ=ULFK6ZLW]HUODQG
3ULQWHGRQDFLGIUHHSDSHUSURGXFHGIURPFKORULQHIUHHSXOS7&)f
3ULQWHGLQ*HUPDQ\

,6%1   H,6%1

    ZZZELUNKDXVHUFK
To Konstantin I. Beidar (1951–2004)
Contents

Preface ix

I An Introductory Course 1
1 What is a Functional Identity? 3
1.1 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Simple Examples of Functional Identities . . . . . . . . . . . . . . 8
1.3 Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4 Finding Functional Identities in Different Areas . . . . . . . . . . . 22

2 The Strong Degree and the FI-Degree 29

2.1 The Strong Degree . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.2 Strongly d-Free Rings and the FI-Degree . . . . . . . . . . . . . . . 33
2.3 Strongly (t; d)-Free Rings . . . . . . . . . . . . . . . . . . . . . . . 37
2.4 s-deg(A) ≤ FI-deg(A) . . . . . . . . . . . . . . . . . . . . . . . . . 41

II The General Theory 47

3 Constructing d-Free Sets 49
3.1 Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2 d-Free Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Two Constructions of d-Free Sets . . . . . . . . . . . . . . . . . . . 58
3.4 (t; d)-Free Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.5 (∗; t; d)-Free Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

4 Functional Identities on d-Free Sets 87

4.1 Introducing the General Setting . . . . . . . . . . . . . . . . . . . . 87
4.2 d-Free Pairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3 Quasi-polynomials and Core Functions . . . . . . . . . . . . . . . 97
4.4 Remarks on “Nonlinear” Identities . . . . . . . . . . . . . . . . . . 105
viii Contents

5 Functional Identities in (Semi)prime Rings 111

5.1 The Fractional Degree . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.2 A List of d-Free Subsets of Prime Rings . . . . . . . . . . . . . . . 119
5.3 d-Freeness of Semiprime Rings . . . . . . . . . . . . . . . . . . . . 123
5.4 Commuting Maps on (Semi)prime Rings . . . . . . . . . . . . . . . 128
5.5 Generalized Functional Identities . . . . . . . . . . . . . . . . . . . 133

III Applications 145

6 Lie Maps and Related Topics 147
6.1 Lie Maps on Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.2 Lie Maps on Skew Elements . . . . . . . . . . . . . . . . . . . . . . 158
6.3 Lie Maps on Lie Ideals . . . . . . . . . . . . . . . . . . . . . . . . . 166
6.4 Jordan Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.5 f -Homomorphisms and f -Derivations . . . . . . . . . . . . . . . . 180

7 Linear Preserver Problems 189

7.1 Commutativity Preserving Maps . . . . . . . . . . . . . . . . . . . 190
7.2 Normality Preserving Maps . . . . . . . . . . . . . . . . . . . . . . 201
7.3 Zero Jordan Product Preserving Maps . . . . . . . . . . . . . . . . 210
7.4 Equal Product Preserving Maps . . . . . . . . . . . . . . . . . . . . 215

8 Further Applications to Lie Algebras 221

8.1 Lie-Admissible Algebras . . . . . . . . . . . . . . . . . . . . . . . . 221
8.2 Poisson Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
8.3 Maps Covariant Under the Action of Lie Algebras . . . . . . . . . 230

A Maximal Rings of Quotients 235

B The Orthogonal Completion 243

C Polynomial Identities 249

D Generalized Polynomial Identities 255

Bibliography 259

Index 271
Preface

A functional identity (FI) can be informally described as an identical relation

involving (arbitrary) elements in a ring together with (“unknown”) functions; more
precisely, elements are multiplied by values of functions. The goal of the general FI
theory is to determine the form of these functions, or, when this is not possible, to
determine the structure of the ring admitting the FI in question. This theory has
turned out to be a powerful tool for solving a variety of problems in different areas.
It is not always easy to recognize that the problem in question can be interpreted
through some FI; often this is the most intriguing part of the process. But once
one succeeds in discovering an FI that fits into the general theory, this abstract
theory then as a rule yields the desired conclusions at a high level of generality.
Among classical algebraic concepts, the one of a polynomial identity (PI)
seems to be, at least on the surface, the closest one to the concept of an FI. In
fact, a PI is formally just a very special example of an FI (where functions are
polynomials). However, the theory of PI’s has quite different goals than the theory
of FI’s. One could say, especially from the point of view of applications, that the
two theories are complementary to each other. Under some natural restrictions, PI
theory deals with rings that are close to algebras of low dimensions, while FI theory
gives definitive answers in algebras of sufficiently large or infinite dimensions.
FI theory is a relatively new one. Its roots lie in the Ph.D. Thesis of the
first author from 1990, which was followed in the ensuing years by a series of
papers in which he studied some basic FI’s, in particular those concerning the
so-called commuting maps, and also found first applications. The central figure in
establishing the foundations of the general theory was Konstantin I. Beidar. This
theory culminated around 2000 in the works by Beidar and the second author on
d-free sets. Over the last few years, the emphasis has been mainly on applications
of the general theory. A notable example are complete solutions of Herstein’s
conjectures on Lie homomorphisms and Lie derivations in associative rings (posed
in Herstein’s 1961 “AMS hour talk”). They were obtained in a series of papers by
Beidar and the authors of this book which ended in 2002. Practically all of the
advanced FI theory was used in these solutions. In fact, the main motivation for
developing this theory was searching for tools that would settle these conjectures.
So far no book has been written exclusively on FI’s. Commuting maps and
some of their applications were treated in the book Rings with generalized identities
x Preface

(Marcel Dekker, 1996) by Beidar, Martindale and Mikhalev, as well as in the book
Local multipliers of C*-algebras (Springer, 2003) by Ara and Mathieu, in both cases
in some rather special settings so that the overlap with the present book is very
small. Up until now the general FI theory has existed only in research journals.
We believe that the theory has reached maturity and that a monograph presenting
FI’s and their applications in a comprehensive and comprehensible manner shall
be useful for the interested mathematical community.
Through their applications FI’s have connections to different areas, say to Lie
algebras, Jordan and other nonassociative algebras, linear algebra, operator theory,
functional analysis, and mathematical physics. However, the basic setting of the
book is (noncommutative) ring theory. Including non-algebraic topics in the book
would make the exposition too diverse. Still, different parts of the book should be
of interest to mathematicians having different basic backgrounds. While writing
we kept in mind that the reader might not be a ring theorist. The prerequisites
needed to follow the exposition are therefore carefully explained. The basic ones are
surveyed at the very beginning of the book, while the somewhat more demanding
ones appear in four appendices at the end. They consider topics which are more
or less widely known and treated in other books. We concentrate on those aspects
that are important for this book. Some results in appendices are proved, and some
not; in any case we try to explain to the reader, not necessarily in a rigorous
manner, the background of the results and concepts that are treated. The book is
therefore fairly self-contained and suitable for self-study.
The book consists of three parts. Each of them has its own purpose. Part I is
an introductory one, Part II gives a full account of the general FI theory, and Part
III is devoted to applications of the general theory. Parts are divided into chapters
which are further divided into sections. Chapters end with comments about the
literature and the history.
Part I has two chapters. The purpose of Chapter 1 is to introduce the new-
comer into the subject and to explain in an informal manner, mainly through
examples, what this theory is all about. Chapter 2 contains results that are al-
ready of some interest in their own right, and they make it possible for one to use
FI theory at a basic level. The concept of d-freeness is introduced in its simplest
form, and the connections to the concept of the strong degree are discussed. The
results of Chapter 2 are only partially superseded in Part II; namely, the setting in
which they are obtained is somewhat different, and in a certain direction even more
general than the setting considered later. In principle, a graduate student should
be able to understand Part I without difficulties. One could use some sections as
a part of a course in ring theory.
Part II is the core of the book. The general theory has so far been exposed only
in numerous papers, which are often very technical and long. They depend upon
each other, so it is not always easy to read them. The goal of Part II is to extract
from these papers what is, in our opinion, most important and applicable, that is,
to show what this theory has to offer. The basic concept upon which everything is
based is that of a d-free set. Chapter 3 discusses basic properties and constructions
Preface xi

of these sets. Chapter 4 is devoted to the study of FI’s on d-free sets; the notions of
quasi-polynomials and core functions are of central importance in this context. The
purpose of Chapter 5 is to prove the existence of some important classes of d-free
sets in prime and semiprime rings, and also to consider some special FI’s in these
rings. The exposition in Part II is necessarily technical, and in spite of our efforts
to be as intelligible as possible, the reader studying the text in detail shall need
some patience. It is impossible to avoid long formulas and somewhat complicated
notation, so that “love at first sight” with this theory seems a bit unlikely; some
effort and time is needed in order to appreciate it and find it enjoyable. On the
other hand, the main results have clear statements and so hopefully Part II will
serve as a useful source of references.
The meaning of the results of Part II becomes evident in Part III. Appli-
cations are the main reason for existence of the general theory. We believe that
while this theory is already mature (at least in the direction in which it has been
mainly developed), the possibilities for its applications are still far from being
exploited. Our intention in Part III is therefore to show through several relevant
examples the way this theory can be used. We do not try to present the results
in utmost generality, but rather point out various areas where FI’s are applica-
ble and demonstrate the methods needed in applications. Chapter 6 deals with
Lie homomorphisms and other “nonassociative” maps. In particular, solutions of
Herstein’s Lie map conjectures are discussed. Chapter 7 considers some linear
preserver problems, particularly commutativity and normality preserving maps.
Chapter 8 discusses miscellaneous topics which all, however, belong to the Lie
algebra framework.
We are grateful to a number of colleagues for fruitful discussions on the
subject of the book. It would be impossible to list all whose influence is felt in our
writing. But we can not omit mentioning a man we can not thank enough and
without whom this book would not exist. Unfortunately he is no longer with us.
Konstantin (Kostia) I. Beidar passed away in March 2004. His impact on FI theory
was decisive. He wrote more than 20 papers on the subject. Some of them were
path-breaking. Kostia himself considered his fundamental results on FI’s as some
of his best mathematical achievements. It is just incredible that while being so
productive in FI theory, at the same time he was also heavily involved in the work
in many other fields. But his mathematical strength and knowledge was a kind
of a legend among his numerous coauthors and colleagues. Solving problems and
inventing new areas seemed so easy when working with Kostia. All three of us had
the privilege of cooperating with him on several projects. Our main mathematical
tie, however, was FI theory. The plan to write this book was made together with
him. In several of our frequent visits to Tainan in Taiwan, where he was working
after leaving Moscow State University in 1994, we discussed with him possible
concepts of the book. A small portion of the book, a draft of Part I and some
other preliminary notes, were written while he was still alive. Arriving at Part II,
technically the most entangled part of the book, it was somehow expected that
he would become involved in the actual writing. But his early and unexpected
xii Preface

death has forced us to continue without him. We miss him; as a friend and as
a mathematician. We hope that this book appropriately represents an important
part of Kostia’s mathematical legacy.

Maribor, Kent, Philadelphia Matej Brešar

February 2007 Mikhail A. Chebotar
Wallace S. Martindale 3rd
Chapter 1

What is a Functional Identity?

An exposition on a mathematical subject usually starts with basic definitions.

We feel, however, that it is more appropriate to introduce functional identities
through examples. We shall therefore present various simple examples of functional
identities, so that the reader may guess which conclusions could be derived when
facing these identities. These examples have been selected in order to illustrate
the general theory, and not all of them are of great importance in their own right.
Their consideration will be rather elementary; anyhow, many of the arguments
that we shall present here will be used, sometimes in a hidden way, in much more
general situations considered in further chapters. Examples will be followed by
some basic definitions and notation, but even these will be given in a somewhat
informal fashion. The last objective of this preliminary chapter is to point out
a few instances where functional identities appear naturally. That is, we wish to
indicate, without many details, why and where the theory of functional identities
is applicable. So, in summary, the goal of this chapter is to give an informal
introduction to functional identities which should be of help to a newcomer to the
subject.
At the very beginning, however, we will survey the ring-theoretic notions and
tools that will be needed throughout the book. They are very basic.

1.1 Prerequisites
Most readers are probably already familiar with the basic terminology in ring the-
ory. However, some parts of the book are supposed to be of interest not exclusively
to ring theorists. We shall therefore state and briefly comment on a few very fun-
damental definitions and results. Details can be found in many standard graduate
level texts on algebra. Simultaneously we shall also explain some conventions and
fix the notation that will be most frequently used. Thereby the reader can find
below all the information needed to follow Part I. Occasionally, however, we shall
4 Chapter 1. What is a Functional Identity?

also mention some facts about polynomial identities and generalized polynomial
identities in Part I. In such instances the reader may consult appendices C and
D or just continue reading without too much loss. A knowledge of the theory of
(generalized) polynomial identities should certainly help the reader to understand
Part I better, but it is not a necessity for following the essence of the exposition.
Parts II and III require some more prerequisites which are surveyed in appendices.
Throughout the text, by a ring (algebra) we shall mean an associative ring
(algebra over a commutative ring) possibly without unity and not necessarily
commutative. As a matter of fact, many of our results are meaningless in the
commutative setting: they are either trivially true or trivially false. When the
existence of a unity will be required, we shall speak about a unital ring (unital
algebra). We will also consider some nonassociative rings and algebras, but this will
always be pointed out (say, a Lie ring will never be referred to as just a ring). We
remark that a ring (algebra) is nonassociative if the associativity of multiplication
is not required, which does not necessarily mean that it does not hold. For most
of the time we shall denote rings and algebras by letters A, B, . . .. The centers of
rings will often play an important role, and even when not explicitly indicated we
shall regard rings as algebras over their centers.
Let us list the definitions of some fundamental types of rings. By a field we
mean a commutative division ring, while a division ring is a nonzero unital ring
all of whose nonzero elements are invertible. More generally, a ring A is said to be
simple if its only ideals are 0 and A, and ab = 0 for some a, b ∈ A. As usual, by
an ideal of a ring we shall always mean a two-sided ideal. A ring A is called (left)
primitive if it has a faithful simple (left) module M. The simplicity of a module
means that 0 and M are the only submodules of M and AM = 0, and faithfulness
means that aM = 0 for every nonzero a ∈ A. A common generalization of all types
mentioned so far is the concept of a prime ring: a ring A is called prime if the
product of two nonzero ideals of A is always nonzero. Equivalently, aAb = 0, where
a, b ∈ A, implies a = 0 or b = 0. Finally, a ring A containing no nonzero nilpotent
ideals is said to be semiprime. Equivalently, aAa = 0, where a ∈ A, implies a = 0.
Division, simple, primitive, prime, and semiprime algebras are defined analogously.
This is not a book about these types of rings; these definitions just provide
an appropriate setting for studying functional identities. So we give just a few
simple-minded examples to help a non-specialist to clarify these notions. Exam-
ples of fields are widely known. We assume the reader is familiar with the ring of
quaternions which is the simplest, the most famous and historically the first exam-
ple of a noncommutative division ring. It is well-known and easy to see that every
unital simple ring is primitive. If V is a vector space over a field F, then EndF (V),
the ring of all endomorphisms of V, is always primitive, and is simple only when V
is finite dimensional. In that case it is isomorphic to the matrix ring Mn (F) where
n is the dimension of V. This notation is standard: for any ring A, Mn (A) denotes
the ring of n × n matrices with entries from A. It turns out that a unital ring A is
simple, (left) primitive, prime, and semiprime, respectively, if and only if the same
1.1. Prerequisites 5

holds true for the ring Mn (A). The ring Z of integers is the simplest example of a
prime ring which is not primitive, and so, more generally, also Mn (Z) is prime but
not primitive. The direct product A1 × A2 of two nonzero prime rings A1 and A2
is semiprime but not prime. The ring of all upper triangular n × n matrices over
a field is not even semiprime since the strictly upper triangular matrices form a
nilpotent ideal.
There are many ways to construct new rings from a given ring A. Let us
mention a few that shall be of particular importance for us.
First of all, for any n in N, the set of all positive integers, we have the matrix
ring Mn (A). When A is unital, Mn (A) contains the so-called matrix units, that is,
matrices in which exactly one entry equals 1 and all other entries are 0. A matrix
unit with entry 1 in the position (i, j) will be denoted by eij .
By End(A) we denote the ring of endomorphisms of the additive group of A,
that is, the set of additive maps of A into itself which forms a ring under pointwise
addition and composition as multiplication. For any a, b ∈ A we define a two-sided
multiplication a Mb ∈ End(A) by a Mb (x) = axb. By M(A) we denote the set of all
elements in End(A) that can be written as finite sums of two-sided
multiplications
a Mb . A typical element f ∈ M(A) is thus of the form f : x → i ai xbi . Note that
M(A) is a subring of End(A). It will be called the multiplication ring of A.
The notation A M means that M is a left module over a ring A, and
EndA (M) denotes the ring of all A-module endomorphisms of M. Suppose that
the module M is simple. Since the kernel and the range of every endomorphism
f ∈ EndA (M) are clearly submodules of M, f is either 0 or it is invertible. That is
to say, EndA (M) is a division ring. This observation is known as Schur’s lemma. A
simple module M can therefore also be regarded as a vector space over a division
ring D = EndA (M). The simplicity of M implies that Ax = M for every nonzero
x ∈ M. That is to say, for every 0 = x ∈ M and y ∈ M there exists a ∈ A such
that ax = y. The celebrated Jacobson density theorem states that much more is
true: If x1 , . . . , xn , where n ∈ N is arbitrary, are elements in M which are linearly
independent over D, and y1 , . . . , yn are any elements in M, then there exist a ∈ A
such that axi = yi , i = 1, . . . , n. One usually refers to this property by saying that
A acts densely on the vector space M.
Let C be a commutative ring and X = {x1 , x2 , . . .} be a countable set. By
CX we denote the free algebra on the set X over C. The elements of CX are
polynomials in noncommuting variables from the set X, whose coefficients in C
commute with indeterminates. Their multiplication is defined in the obvious way
by juxtaposition. We remark here that the construction of the free algebra can be
made rigorous by various methods (e.g., forming the vector space over C with basis
the free semigroup on the set X or using the more high powered notion of tensor
products), but we will not elaborate further in this regard. Considering an element
f ∈ CX as a polynomial, the notions such as the degree of f or monomial of
f are self-explanatory. Let f = f (x1 , . . . , xm ); this notation indicates that no xi
with i > m occurs in f . Further, let A be a C-algebra, let R be a nonempty subset
6 Chapter 1. What is a Functional Identity?

of A, and let Rm = R× R×. . .× R denote the Cartesian product (m times). Then

f determines a function F : Rm → A given in the obvious way of substituting
r1 , . . . , rm for x1 , . . . , xm in f , where ri ∈ R, i = 1, . . . , m. Such a function F will
be called the polynomial function determined by f . One should carefully note the
distinction between a polynomial f and the polynomial function it determines,
e.g., if A is commutative, then f = x1 x2 − x2 x1 is a nonzero element in CX,
whereas F is the zero function on A × A.
For elements x and y in a ring A we shall write [x, y] = xy − yx and x ◦ y =
xy+yx. The element [x, y] is called the Lie product (or the commutator) of elements
x and y, and x ◦ y is called the Jordan product of x and y. In the case when 2A = 0
(i.e., 2a = 0 for every a ∈ A), the Lie and the Jordan product coincide, which
makes their treatment rather muddled. Therefore we shall usually restrict our
attention to rings which are 2-torsion free; we say that a ring (or just an additive
group, for that matter) A is n-torsion free, where n ∈ N, if na = 0 for every
nonzero a ∈ A. In case A is a prime ring, it is more customary to express its n-
torsion freeness through the notion of the characteristic (char(A)): A is n-torsion
free if and only if char(A) = n. We also remark that the characteristic of a prime
ring is always a prime number or 0.
Let us fix some more notation. For a nonempty subset T of a ring A we shall
write T  for the subring of A generated by T , and (T ) for the ideal of A generated
by T . The set C(T ) = {a ∈ A | [a, T ] = 0} is called the centralizer of T in A. The
center of A is of course the centralizer of A in A; however, instead of C(A) we will
usually denote it by Z(A). When T = {t}, we shall write C(t), t and (t) instead
of C({t}), {t} and ({t}). If R and S are additive subgroups of A, we denote by
RS, [R, S], and R ◦ S the additive subgroup of A generated by all rs, [r, s], and
r ◦ s, respectively, where r ∈ R and s ∈ S are arbitrary elements. Similarly we
define R1 R2 . . . Rn , [R1 , [R2 , [. . . , Rn ] . . .]] etc. A warning is necessary: sometimes
Rn will denote the additive subgroup generated by all r1 r2 . . . rn , ri ∈ R, but more
often it will denote the Cartesian product of n copies of R. But from the context
it should be clear in each case what we have in mind.
An additive subgroup L of A such that [L, L] ⊆ L is called a Lie subring of
a ring A. If an additive subgroup I of a Lie subring L satisfies [I, L] ⊆ I, then
I is said to be a Lie ideal of L. For example, [L, L] is a Lie ideal of L, and so
is Z(L) = C(L) ∩ L = {l ∈ L | [l, L] = 0}, the center of L. We shall be usually
interested in noncentral Lie ideals of L, i.e., those which are not contained in
Z(L). Of course, a Lie ideal of L is also a Lie subring of A. Jordan subrings and
their Jordan ideals are defined similarly. Important examples of Lie and Jordan
subrings are provided in rings with involution. An involution, usually denoted
by ∗, on a ring A is an antiautomorphism of A of period 1 or 2, i.e., ∗ satisfies
(x + y)∗ = x∗ + y ∗ , (xy)∗ = y ∗ x∗ and (x∗ )∗ = x for all x, y ∈ A. Note that the set
S(A) = {x ∈ A | x∗ = x} of symmetric elements of A is a Jordan subring of A,
and the set K(A) = {x ∈ A | x∗ = −x} of skew elements of A is a Lie subring of
A. When the context is clear we shall write S and K instead of S(A) and K(A),
1.1. Prerequisites 7

and similarly we shall write Z instead of Z(A).

Let B and A be rings and let L be a Lie subring of B. An additive map
α : L → A, x → xα , is called a Lie homomorphism if [x, y]α = [xα , y α ] for
all x, y ∈ L. A Jordan homomorphism from a Jordan subring of B into A is
defined analogously, i.e., it is an additive map that preserves the Jordan product.
There is another type of maps that will often appear. Let A be a ring and let
M be an (A, A)-bimodule. An additive map δ : A → M is called a derivation
if (xy)δ = xδ y + xy δ for all x, y ∈ A. For example, for any fixed a ∈ A the
map x → [a, x] is a derivation of A into itself. Such derivations are called inner
derivations. Further, an additive map δ : L → M, where L is a Lie subring of A
is called a Lie derivation if [x, y]δ = [xδ , y] + [x, y δ ] for all x, y ∈ L. The definition
of a Jordan derivation is self-explanatory.
There is a simple explanation for the terminology introduced above. Lie sub-
rings of associative rings are examples of Lie rings, and Jordan subrings are ex-
amples of Jordan rings. These are defined as follows. A nonassociative ring (resp.
algebra) L with multiplication (x, y) → {x, y} is called a Lie ring (resp. Lie alge-
bra) if {x, x} = 0 for all x ∈ L and {x, {y, z}} + {y, {z, x}} + {z, {x, y}} = 0 for
all x, y, z ∈ L. The latter identity is called the Jacobi identity. A nonassociative
ring (resp. algebra) J with multiplication (x, y) → x · y is called a Jordan ring
(resp. Jordan algebra) if x · y = y · x and (x2 · y) · x = x2 · (y · x) for all x, y ∈ J
(here, of course, x2 stands for x · x). One can verify that every associative ring can
be turned into a Lie ring (resp. Jordan ring) by defining a new product as a Lie
product [x, y] = xy − yx (resp. Jordan product x ◦ y = xy + yx). Actually, even in
an abstract Lie algebra it is more common to denote the product by [. , .], rather
than by {. , .}. However, we introduced this notation to avoid confusion, so [. , .]
will always denote the Lie product arising from an associative ring.
The reader should be familiar with the process of linearization. In a loose
manner this can be described as obtaining new relations from a given relation
by replacing, iteratively if necessary, an arbitrary element satisfying this relation
by the sum of two arbitrary elements. Let us explain this by an example. The
reader shall easily find out when the method presented in this example also works
in some related situations. Let G and H be additive groups and let B : G n =
G × G × . . . × G → H be an n-additive map, that is, a map additive in each
argument. The map x → B(x, x, . . . , x) will be called the trace of B. Suppose that
the trace of B is zero. We claim that in this case

B(xπ(1) , xπ(2) , . . . , xπ(n) ) = 0
π∈Sn

for all x1 . . . , xn ∈ G where Sn denotes the symmetric group of order n. One

usually says that the latter identity is obtained by linearizing B(x, x, . . . , x) = 0
for all x ∈ G. When n = 2, that is when B is a biadditive map, we get B(x1 , x2 ) +
B(x2 , x1 ) = B(x1 + x2 , x1 + x2 ) − B(x1 , x1 ) − B(x2 , x2 ) = 0 immediately. Now
8 Chapter 1. What is a Functional Identity?

consider the case when n = 3. Replacing x by x1 + x2 in B(x, x, x) = 0 we get

B(x1 , x1 , x2 ) + B(x1 , x2 , x1 ) + B(x2 , x1 , x1 )

+ B(x2 , x2 , x1 ) + B(x2 , x1 , x2 ) + B(x1 , x2 , x2 ) = 0

for all x1 , x2 ∈ G. Incidentally we remark that substituting x2 for −x2 in this

last identity yields B(x1 , x1 , x2 ) + B(x1 , x2 , x1 ) + B(x2 , x1 , x1 ) = 0 provided that
H

is 2-torsion free. In any case, however, replacing x2 by x2 + x3 we arrive at
π∈S3 B(xπ(1) , xπ(2) , xπ(3) ) = 0 for all x1 , x2 , x3 ∈ G, proving our claim. The idea
of this argument still works for any n > 3, but writing the proof in detail is rather
tedious and we omit it.

1.2 Simple Examples of Functional Identities

We shall use the abbreviation FI for a functional identity.
Let A be a ring and let E, F : A → A be functions such that

E(x)y + F (y)x = 0 for all x, y ∈ A. (1.1)

This is a very basic example of an FI. Let us stress that we do not assume any
further conditions on functions E and F besides that they satisfy (1.1). The FI the-
ory deals with set-theoretic functions satisfying certain identities. These functions
should be considered as “unknowns”.
A trivial possibility when (1.1) is fulfilled is when E = F = 0. Are there any
other possibilities? If A is commutative, then there certainly are: just take, for
example, E to be the identity map and F = −E. More generally, suppose that
A contains a nonzero central ideal I, i.e., an ideal contained in Z(A). Given any
c ∈ I we can define E and F by E(x) = −F (x) = cx and (1.1) is fulfilled. Now
let A be any ring. Note that (1.1) implies

(E(x)yz)w = −F (yz)xw = (E(xw)y)z = −(F (y)x)wz = E(x)ywz

for all x, y, z, w ∈ A. That is,

E(A)A[A, A] = 0.

If A is prime and noncommutative, it follows immediately that E = 0. In

that case also F = 0 by (1.1). Now suppose that A is semiprime and E = 0.
Let I = (E(A)), i.e., I is the ideal generated by the range of E. Note that
[I, A]A[I, A] = 0 and hence [I, A] = 0 by the semiprimeness of A. That is, I
is a central ideal of A. Finally, suppose that A = Mn (C), n ≥ 1, where C is a
commutative unital ring. If n ≥ 2, then it is easy to see that ([A, A]), the ideal
generated by [A, A], contains the identity matrix, which in turn implies that E = 0,
and hence also F = 0.
1.2. Simple Examples of Functional Identities 9

One way of looking at functional identities is that these are equations with
functions appearing as unknowns. Accordingly, we will often refer to functions
satisfying a certain FI as solutions of that FI. We shall say that E = F = 0 is the
standard solution of the FI (1.1). We have seen that the existence of a nonstandard
solution implies that the ring is very special. More precisely, we can summarize
our observations as follows:
Example 1.1. Let A be a semiprime (resp. prime) ring. Then there exists a non-
standard solution of (1.1) if and only if A contains a nonzero central ideal (resp.
A is commutative). Further, if A = Mn (C) where C is a commutative unital ring,
then there exists a nonstandard solution of (1.1) if and only if n = 1.
The next FI which we are going to consider is only slightly more complicated
than (1.1). Instead of requiring that the expression E(x)y + F (y)x is always zero
we assume that it is central, that is, we consider the FI

E(x)y + F (y)x ∈ Z = Z(A) for all x, y ∈ A. (1.2)

One might say that (1.2) does not look like an identity, but rather as an inclusion
of one set in another; anyway, we can rewrite (1.2) as [E(x)y + F (y)x, z] = 0 for
all x, y, z ∈ A which is certainly an identity.
As in the previous case, we define E = F = 0 to be the standard solution of
(1.2). However, nonstandard solutions of (1.2) exist not only in any commutative
ring C, but also in the ring A = M2 (C). Indeed, define E, F : A → A by E(x) =
F (x) = x − tr(x)1, where tr(x) denotes the trace of x and 1 is the identity matrix,
and check that this gives a nonstandard solution of (1.2). On the other hand, if
A = Mn (C) where n ≥ 3 and C is a commutative unital ring, then (1.2) has only
the standard solution. Indeed, set π(x, y) = E(x)y + F (y)x for x, y ∈ A. Our
assumption is that π(x, y) lies in Z. Therefore, for any x, y, t ∈ A the element
π(xt, y) − π(x, y)t = E(xt)y − E(x)yt lies in C(t). Therefore, commuting it with t
we get
−E(x)yt2 + (tE(x) + E(xt))yt − tE(xt)y = 0.
Let 1 ≤ i ≤ n, set t = e12 + e23 , y = ei1 in the last relation, and then multiply the
identity so obtained from the right by e3i . Hence we arrive at E(x)eii = 0 for all
x ∈ A, 1 ≤ i ≤ n. This clearly yields E = 0, and hence also F = 0. Now we can
state
Example 1.2. Let A = Mn (C) where C is a commutative unital ring. Then there
exists a nonstandard solution of (1.2) if and only if n ≤ 2.
How does one proceed from Examples 1.1 and 1.2? In the next step it seems
natural to seek FI’s where the ring M3 (C) would also play an exceptional role. Let
us first reveal that the example in M2 (C) was derived from the Cayley–Hamilton
theorem. The reader might be familiar with this theorem only for matrices over
(some) fields, so let us point out that “Every matrix is a zero of its character-
istic polynomial” is still true for matrices over any commutative ring C (see e.g.,
10 Chapter 1. What is a Functional Identity?

[187, p. 18]). Thus, for every matrix x ∈ A = Mn (C) we have

xn + α1 (x)xn−1 + α2 (x)xn−2 + . . . + αn−1 (x)x + αn (x)1 = 0
where αi (x) ∈ C, i = 1, . . . , n, are coefficients of the characteristic polynomial
det(λ1 − x). These coefficients can be expressed by Newton’s formulas, but for
our purposes it suffices to consider them in a less precise manner. Expanding the
characteristic polynomial we first see that α1 (x) = −tr(x) and so α1 : A → C is
an additive map (actually, even a C-linear map, but this is not important for our
immediate goal). Secondly, α2 (x) can be expressed as the sum of the terms each of
which is of the form ±xi1 i2 xj1 j2 , where xij ’s are entries of x. But then α2 : A → C
is the trace of a biadditive map ζ2 : A2 → C. Indeed, one can define ζ2 (x, y) in
the obvious way, as the sum of the terms of the form ±xi1 i2 yj1 j2 . Further, the
expression for α3 (x) consists of summands of the form ±xi1 i2 xj1 j2 xk1 k2 , and so
α3 : A → C is the trace of a 3-additive map ζ3 : A3 → C. In this way we find out
that the Cayley–Hamilton theorem implies that
xn + ζ1 (x)xn−1 + ζ2 (x, x)xn−2 + . . . + ζn−1 (x, x, . . . , x)x + ζn (x, . . . , x)1 = 0
for all x ∈ A, where each ζk : Ak → C is a k-additive map. The last term is a
scalar matrix (incidentally, it is equal to (−1)n det(x)1), and so we have
xn + ζ1 (x)xn−1 + ζ2 (x, x)xn−2 + . . . + ζn−1 (x, . . . , x)x ∈ Z
(note that the center Z of A is equal to C1). For n = 2 this reads as x2 +ζ1 (x)x ∈ Z.
A linearization gives rise to a nonstandard solution of (1.2) mentioned above. Now
let n = 3. Then a linearization gives
E(x, y)z + E(x, z)y + E(y, z)x ∈ Z
for all x, y, z ∈ A, where E : A2 → A is given by
E(x, y) = (x + ζ1 (x))y + (y + ζ1 (y))x + (ζ2 (x, y) + ζ2 (y, x))1. (1.3)
Note that E = 0 (after all, this also follows from Example 1.2).
The conclusion from the above paragraph can be interpreted as follows. The
FI
E(x, y)z + F (x, z)y + G(y, z)x ∈ Z for all x, y, z ∈ A (1.4)
has a nonstandard solution on A = M3 (C), namely, the one where E = F = G
is the function defined according to (1.3). By the standard solution of (1.4) we
of course mean that E = F = G = 0. It is clear from the above discussion that
nonstandard solutions of (1.4) also exist in Mn (C) for n = 1, 2.
Let us, for a while, treat (1.4) in any ring A. It should already be self-
explanatory that E, F, G : A2 → A are considered to be arbitrary functions. Set
π(x, y, z) = E(x, y)z + F (x, z)y + G(y, z)x ∈ Z. A direct computation shows that

π(xt, yt, z) − π(xt, y, z)t − π(x, yt, z)t + π(x, y, z)t2

= E(xt, yt)z − (E(xt, y) + E(x, yt))zt + E(x, y)zt2 .
1.2. Simple Examples of Functional Identities 11

Obviously, this element lies in C(t) for any t ∈ A. Therefore, commuting it with t
we arrive at

E(x, y)zt3 + E1 (x, y, t)zt2 + E2 (x, y, t)zt + E3 (x, y, t)z = 0

for all x, y, z, t ∈ A. Here, the Ei ’s are some new functions which can be eas-
ily expressed in terms of E; however, we omit this calculation since it is of no
importance for the proceeding discussion.
Assume now that A = Mn (C) with n ≥ 4 and set t = e12 + e23 + e34 and
z = ei1 in the last relation, and then multiply the identity so obtained from the
right by e4i . Then we get E(x, y)eii = 0 for all x, y ∈ A, 1 ≤ i ≤ n, so that E = 0.
In a similar fashion we show that F = 0 and G = 0. Thus, we have established
Example 1.3. Let A = Mn (C) where C is a commutative unital ring. Then there
exists a nonstandard solution of (1.4) if and only if n ≤ 3.
Let us mention that appropriate modifications of the above arguments give
similar results in rings quite different from the matrix rings. But to show this we
need to introduce some other concepts which will be done in subsequent chapters.
Now the reader can probably guess how the examples given so far could be
unified and extended. Nevertheless, we stop this line of investigation at this point
since we have no intention of proving general statements in this chapter. Instead
we consider a different kind of an extension of the basic FI (1.1). Let a be a fixed
nonzero element in a ring A and consider the identity

E(x)ya + F (y)xa = 0 for all x, y ∈ A. (1.5)

So, besides involving functions this identity also involves a fixed element of the
ring. Such identities will be called generalized functional identities. The reasons
for this terminology will become more clear later. Again we say that E = F = 0
is the standard solution of (1.5). In the case when a = 1, or equivalently when a
is invertible, (1.5) has already been treated in Example 1.1 and we have seen that
nonstandard solutions can exist only very exceptionally. In the case when a is not
invertible, or better when a is “far” from being invertible, this is no longer true.
For instance, for any n ∈ N there exists a nonstandard solution in A = Mn (C)
given by a = e11 , E(x) = −F (x) = x11 e11 . More generally, if a is an element in a
ring A such that
aAa = Za = 0, (1.6)

then one can verify that E(x) = −F (x) = axa gives a nonstandard solution
of (1.5). Now assume that A is a simple unital ring and let us prove a kind of a
converse to these examples. Assuming (1.5) we have, on the one hand, E(x)yaza =
−F (yaz)xa, and on the other hand, E(x)yaza = −F (y)xaza. Hence

F (yaz)xa = F (y)xaza for all x, y, z ∈ A. (1.7)

12 Chapter 1. What is a Functional Identity?

Suppose that a = 0 and
F (y0 ) = 0 for some y0 ∈ A. Since A is simple, there are
xi , yi ∈ A such that i xi F (y0 )yi = 1. Consequently, using (1.7) we get
 
xaza = xi (F (y0 )yi xaza) = xi F (y0 az)yi xa for all x, z ∈ A.
i i


This shows that for any z ∈ A there is c = i xi F (y0 az)yi ∈ A such that
xaza = cxa for all x ∈ A. This implies that cyxa = yxaza = ycxa for all x, y ∈ A,
that is, [c, A]Aa = 0 which clearly yields c ∈ Z. But then aza = ca. Thus we
proved aAa ⊆ Za. But then actually (1.6) holds true since Z, as the center of a
unital simple ring, is a field (indeed, just note that for every central element c the
set cA is an ideal).
Which are the simple unital rings A that contain elements a satisfying (1.6)?
As already indicated, Mn (C) with C a field is such an example: even any matrix unit
satisfies (1.6). In fact, matrix rings are the only examples. This is well-known; let
us show this by using the Jacobson density theorem. So assume that a ∈ A satisfies
(1.6). We claim that Aa is a minimal left ideal of A. This means that Aa does not
properly contain any nonzero left ideal of A. Indeed, assume that L is a nonzero
left ideal of A such that L ⊆ Aa. Then 0 = aL ⊆ aAa = Za, whence 0 = au = ca
for some u ∈ L, c ∈ Z. Thus a = c−1 au ∈ L and so L = Aa, proving our claim.
Now regard Aa as a left A-module. To each element x ∈ A we can associate a Z-
linear operator Lx : Aa → Aa given by Lx (ya) = xya. Since A is simple, the map
x → Lx is an embedding of the ring A into the ring EndZ (Aa). We may therefore
view A as a subring of EndZ (Aa). Since Aa is minimal as a left ideal of A it follows
that it is simple as an A-module. Pick f ∈ EndA (Aa). In view of (1.6) there is
b ∈ A such that aba = a, so that f (a) = f (aba) = abf (a) ⊆ aAa. Therefore,
f (a) = ca for some c ∈ Z and hence f (xa) = xf (a) = cxa for every x ∈ A. That
is, every f ∈ EndA (Aa) is a multiplication with an element from Z. Conversely,
every such multiplication is an element of EndA (Aa). In this way we can therefore
identify EndA (Aa) with Z. By the Jacobson density theorem the ring A acts
densely on the vector space Z (Aa). We claim that Z (Aa) is
finite dimensional.
Since A is simple and unital, there are xi , yi ∈ A such that i xi ayi = 1. Using
(1.6) it follows that  
Aa = ( xi ayi )Aa ⊆ Zxi a
i i
which proves our claim. Because of the dense action A must therefore contain all
linear operators on Aa. Accordingly, A is isomorphic to Mn (Z) where n is the
dimension of Aa over Z. We can now state
Example 1.4. Let A be a simple unital ring. Then there exists a ∈ A such that
(1.5) has a nonstandard solution if and only if A ∼
= Mn (Z) for some n ∈ N and
some field Z.
We continue with an FI whose standard solutions are not as trivial as in the
previous cases. Let F : A → A be such that
F (x)y + F (y)x = yF (x) + xF (y) for all x, y ∈ A. (1.8)
1.2. Simple Examples of Functional Identities 13

Obvious examples are the identity map and every map with its range in Z. Ac-
cordingly, every map of the form

F (x) = λx + µ(x), λ ∈ Z, µ : A → Z (1.9)

is a solution of (1.8), and these are the solutions that will be called standard in
the present case.
The FI (1.8) is closely related to the notion of a commuting map which will
be often considered in the sequel. A map F is said to to be commuting on a set
S ⊆ A if [F (x), x] = 0 for all x ∈ S. Incidentally we also mention a somewhat
more general notion of a map that is centralizing on S; by definition such a map
satisfies [F (x), x] ∈ Z for all x ∈ S. If F is additive and commuting on A, then
linearizing [F (x), x] = 0 we get (1.8).
The most convenient way to consider (1.8) is via the following concept: A bi-
additive map ∆ : A2 → A is called a biderivation if it is a derivation in each
argument, that is, for every y ∈ A the maps x → ∆(x, y) and x → ∆(y, x) are
derivations. For example, for any λ ∈ Z, (x, y) → λ[x, y] is a biderivation. We shall
call such maps inner biderivations. It is easy to construct non-inner biderivations
on commutative rings, say, take the ring of polynomials over a field and define
∆(f, g) = f  g  where f  is the formal derivative of the polynomial f . In noncom-
mutative rings, however, it happens quite often that all biderivations are inner. In
such rings (1.8) can have only a standard solution. Namely, (1.8) can be written
as [F (x), y] = [x, F (y)]. This implies that the map ∆(x, y) = [F (x), y] is (even an
inner) derivation in each argument, and so it is a biderivation. Therefore, by as-
sumption there exists λ ∈ Z such that [F (x), y] = λ[x, y] = [λx, y] for all x, y ∈ Z,
meaning that µ(x) = F (x) − λx lies in Z.
So let us consider an arbitrary biderivation ∆ on a ring A. Since ∆ is a
derivation in the first argument, we have

∆(xu, yv) = ∆(x, yv)u + x∆(u, yv),

and since it is also a derivation in the second argument it follows that

∆(xu, yv) = ∆(x, y)vu + y∆(x, v)u + x∆(u, y)v + xy∆(u, v).

On the other hand, first using the derivation law in the second and after that in
the first argument we get

∆(xu, yv) = ∆(xu, y)v + y∆(xu, v)

= ∆(x, y)uv + x∆(u, y)v + y∆(x, v)u + yx∆(u, v).

Comparing both relations we obtain ∆(x, y)[u, v] = [x, y]∆(u, v) for all x, y, u, v ∈
A. Replacing v by zv and using [u, zv] = [u, z]v + z[u, v], ∆(u, zv) = ∆(u, z)v +
z∆(u, v) we obtain
∆(x, y)z[u, v] = [x, y]z∆(u, v)
14 Chapter 1. What is a Functional Identity?

for all x, y, z, u, v ∈ A. This is the crucial identity. To illustrate its utility, assume
for example
that A is unital and ([A, A]) = A. Thus, there are zi , ui , vi , wi ∈ A
such that i zi [ui , vi ]wi = 1, hence
 
∆(x, y) = ∆(x, y)zi [ui , vi ]wi = [x, y]zi ∆(ui , vi )wi .
i i


That is, ∆(x, y) = [x, y]λ for all x, y ∈ A where λ = i zi ∆(ui , vi )wi ∈ A. We
claim that λ ∈ Z. Indeed, we have

[x, y]zλ + y[x, z]λ = [x, yz]λ = ∆(x, yz)

= ∆(x, y)z + y∆(x, z) = [x, y]λz + y[x, z]λ,

showing that [x, y][z, λ] = 0 for all x, y, z ∈ A. Replacing z by zw and using

[zw, λ] = [z, λ]w + z[w, λ] it follows that [A, A]A[λ, A] = 0, which in turn implies
λ ∈ Z in view of our assumption.
Example 1.5. Let A be a unital ring such that ([A, A]) = A. Then every bideriva-
tion on A is inner. Consequently, every solution of the FI (1.8) is a standard solu-
tion (1.9). In particular, every additive commuting map on A is of the form (1.9).
For instance, in a simple ring A, or in the ring A = Mn (C) with C a com-
mutative unital ring, we can now say that (1.8) has only standard solutions. Here
one does not exclude the case when A is commutative because every map on a
commutative ring is trivially of the form (1.9).
In our final example we consider a generalized FI

E(x)ya = axF (y) for all x, y ∈ A. (1.10)

Here of course a is a fixed element in a ring A and E, F : A → A. Suppose for

a moment that a is invertible. Setting y = a−1 in (1.10) we then get E(x) = axq
where q = F (a−1 ) ∈ A. This yields ax(F (y) − qya) = 0 and hence F (y) = qya. So
we see which solution we can naturally expect. We shall say

E(x) = axq, F (y) = qya, (1.11)

where q ∈ A, is a standard solution of (1.10). Let us modify the above argument to

show that (1.10) has only standard

solutions also under the milder assumption that
A is unital and (a)

= A. So let
xi , yi ∈ A. Then
i xi ayi = 1 hold true for some

we have E(x) = i E(x)xi ayi = i axF (xi )yi = axq where q = i F (xi )yi ∈ A.
As above this implies F (y) = qya, so we can state
Example 1.6. Let A be a unital ring. Suppose that a ∈ A is such that (a) = A.
Then every solution of (1.10) is a standard solution (1.11). In particular, in a
simple unital ring (1.10) has only standard solutions provided that a = 0.
Now consider (1.10) in the ring Z and let, for example, a = 2. Taking E and F
to be both identity maps we certainly get a solution of (1.10). However, there is no
1.3. Basic Concepts 15

q ∈ Z satisfying (1.11). On the other hand, if we change slightly the definition of a

standard solution by allowing that q lies in Q, then such q certainly exists: q = 12 .
Though trivial, this example already indicates that it is sometimes appropriate
to involve rings bigger than the original ring when studying functional identities.
In Parts II and III we shall indeed deal with such rings. In Part I, however, we
confine our attention to rings such that the presence of another bigger ring can be
avoided.

1.3 Basic Concepts

Examples from the preceding section indicate how one might investigate FI’s. One
first has to find all “obvious” solutions of a given FI, that is, the solutions which do
not depend on some structural properties of the ring but are merely consequences
of a formal calculation. Such solutions are called standard solutions. The eventual
existence of a nonstandard solution usually implies that the ring has a very special
structure. So, intuitively it should already be clear what functional identities are
about. But we still have to place these observations in some general framework.
Let us first explain what we formally mean by a (generalized) functional
identity. The definitions that we shall give are very general and admittedly will not
have much practical meaning in this book. We just want to point out what is the
“home” of the theory in the broadest sense. The meaning of the definition is thus
more or less informative, and besides, it just does not seem right in mathematics
to speak about notions without making precise what they are. However, if one is
unfamiliar with some of the notions that will be mentioned in the next paragraphs,
then one may simply ignore them since their understanding is not of essential
importance for the book.
Some of the readers may have already guessed that FI’s could be viewed
as generalizations of polynomial identities (PI’s).
Let us give a concrete hint.
A ring A satisfies a multilinear PI of degree 3 if π∈S3 nπ xπ(1) xπ(2) xπ(3) = 0 for
all x1 , x2 , x3 ∈ A, where the nπ ’s are integers, with at least one nπ equal to 1,
and S3 is the symmetric group of order 3. This can be written as E(x1 , x2 )x3 +
F (x1 , x3 )x2 +G(x2 , x3 )x1 = 0 where E(x1 , x2 ) = n1 x1 x2 +n(12) x2 x1 , and F and G
are defined analogously. But this is just a special case of the FI (1.4) where E, F, G
are supposed to be just entirely arbitrary functions. This example also illustrates
the earlier remark that nonstandard solutions usually imply that the ring is of a
special nature: here the standard solution should just be the zero functions and
PI-rings are of a special nature.
We shall define the concept of an FI by modifying appropriately the definition
of a PI. Let A be a ring, let R be a nonempty subset of A, and let Fi : Rm → A,
i = 1, . . . , n, be functions. Actually, we could assume more generally that the
Fi ’s map Rm into a ring containing A, or even into an (A, A)-bimodule, but let
us not bother about various possible generalizations right now. Next, let X =
{x1 , y1 , x2 , y2 , . . .} be a countable set, let ZX be the free algebra on X over Z,
16 Chapter 1. What is a Functional Identity?

and let
f = f (x1 , . . . , xm , y1 , . . . , yn ) ∈ ZX, m ≥ 1, n ≥ 0
be a polynomial such that at least one of its monomials of highest degree has
coefficient 1. We shall say that f is a functional identity on R with functions
F1 , . . . , Fn if
f (r1 , . . . , rm , F1 (r1 , . . . , rm ), . . . , Fn (r1 , . . . , rm )) = 0
for all r1 , . . . , rm ∈ R. In this case we shall also say that the functions F1 , . . . , Fn
are solutions of this functional identity.
For instance, the identity (1.4) is clearly equivalent to [E(x, y)z + F (x, z)y +
G(y, z)x, u] = 0 for all x, y, z, u ∈ A, and so it can now be described by saying
that
f (x1 , x2 , x3 , x4 , y1 , y2 , y3 ) = [y1 x1 + y2 x2 + y3 x3 , x4 ]
is an FI on A with functions F1 , F2 , F3 defined in the obvious way, that is, so
that Fi does not depend on the i-th variable (e.g., F1 (a1 , a2 , a3 ) = G(a2 , a3 ) etc.).
Similarly, the identity (1.8) can be expressed via the polynomial [y1 , x1 ] + [y2 , x2 ]
and functions F1 , F2 given by F1 (x1 , x2 ) = F (x2 ), F2 (x1 , x2 ) = F (x1 ), while saying
that a map F is commuting is the same as saying that [y1 , x1 ] is an FI on A with
the map F .
What is the connection between FI’s and PI’s? One could, of course, say
that a PI is just a special case of an FI with n = 0 (i.e., there are no functions)
in the definition of an FI. But this superficial remark says nothing about the
substantive connection between FI’s and PI’s such as the one hinted at in the
example on multilinear PI’s of degree 3. Strictly speaking the notion of a PI is not
quite comparable with that of an FI: a PI is an element of the free algebra whereas
the notion of an FI involves functions. But, this small technicality aside, any PI
can be shown to be in essence an FI as we now proceed to indicate. Let R be a
nonempty subset of a ring A and let g = g(x1 , . . . , xm ) ∈ ZX be a PI on R. We
write g = f1 x1 + . . . + fm xm , fi = fi (x1 , . . . , xm ), and let Fi be the polynomial
function Rm → A determined by fi , i = 1, . . . , m. Then f = y1 x1 + . . . + ym xm
with functions F1 , . . . , Fm is the appropriate FI on R. In this sense the PI g can
be viewed as the FI f . Of course there are other ways of decomposing g (e.g.,
g = x1 h1 + . . . + xm hm ) which could equally well have been used.
Similarly we define a generalized functional identity (GFI). Actually, the def-
inition is literally the same as that of an FI with one change: we require that
f , instead of belonging to ZX, is an element of the coproduct of A and ZX
(when n = 0, this definition basically reduces to the definition of a generalized
polynomial identity (GPI), cf. appendix D). For example, (1.10) can be expressed
as that f (x1 , x2 , y1 , y2 ) = y2 x2 a − ax1 y1 is a GFI on A with maps F1 , F2 given by
F1 (x1 , x2 ) = F (x2 ), F2 (x1 , x2 ) = F (x1 ).
Anyway, we shall use the phrases a “functional identity” and a “generalized
functional identity” rather informally in this book. Roughly speaking, those iden-
tities that in addition to functions also involve some fixed elements in a ring (such
Exploring the Variety of Random
Documents with Different Content
 Oranges, 304.
Time-piece, 85.
Signs in New York, 74.
Wagers, 184.

Curse of Scotland, 227.

Curse of innocent blood, 228.

Cuttle-fish, eyes of the, 290.

Customs, Hallow E'en, 232.

Cutting off the fiddler's head, 236.

Cutting timber by the moon, 213.

Cynocephalic apes, 278.

Dance, gipsy, 181.

Danger, animals forecasting, 282.

Dairus, tomb of, 106.

Day, first sale of the, 338.

Dead drunk for twopence, 66.

Body, odd way of discovering, 334.
Monkeys demanding their, 279.

Dean Swift's marriage ceremony, 67.

Death warrant of the Saviour, 45.

Death watch, 230.

Decapitation by the guillotine, 204.

Deepened damnation, 191.

Delicate machinery, 86.

Destiny, visions of, 342.

Deutsche luft, 143.

Devil regarded as a benefactor, 227.

Dials, ancient, 86.

Dialect rhyme, 26.

Diamonds, veneered, 141.

Diamond, queer place to secrete a, 222.

Diaries, 17.

Digestive powers, great, 289.

Digging for water, 346.

Dinner time, royal, 318.

Dinner, getting outside of his, 284.

Diogenes and his tub, 27.

Dipping the left shirt sleeve, 233.

Discovery of the body of Canute, 154.

Of glass, 134.
Of composition rollers, 133.
Disfigurement, fashionable, 187.

Divination by Bible and key, 342.

Divinity, queer evidence of, 182.

Divining rods, 316.

Doctor, African rain, 152.

Donne's apparition, 245.
Johnson's blunders, 22.

Dodo, story of the, 268.

Dogs, barking of, 274.

Charm against, 332.
Can they count? 279.

Donation to a fair, 69.

Don Quixote's sheep, 40.

Double-entendre, 39.

Double apparition, 241.

Dove, 253.
Double-headed, 254.
A harbinger of good news, 254.

Dragon, bogus, 68.

Drawing cabbages blindfolded, 233.

Dream, Lord Bacon's, 246.

Dreams, remedy for bad, 225.

Dresses of Queen Elizabeth, 160.

Drinking glass a yard long, 109.

Drop letter retort, 67.

Druid's seat, 110.

Drum made of human skin, 212.

Drunk for twopence, 66.

Duels fought by clergymen, 71.

Durability of bricks, 208.

Earliest clocks, 80.

Earliest traders, 197.

Earthquake, swallowed by an, 175.

Ear of birds not to be deceived, 272.

Ears of the elephant, 277.

Earwig, 289.

Easter box, 238.

Eating animals that have died a natural death, 193.

Clouds, 293.
Tea, 208.
For a wager, 184.

Eccentricity, Kant's, 44.

Eccentric will, 59.

Echoes, remarkable, 166.

Echo verse, 78.

Eden Hall, luck of, 101.

Edict against fiddlers, 58.

Edifices of the Polypi, 299.

Effect of a new nose, 215.

Efficacy in a mutilated saint, 216.

Egg mistaken for a pearl, 170.

Egg, penguin's solitary, 262.

Eggs, superstition about, 274.

Egyptian perfumes, 203.

Egyptian physicians, 169.

Electioneering in 1640, 62.

Elephants' ears, 277.

Elephants, tenacity of life in, 276.

Elephants, white, 276.

Eliot's Indian Bible, 12.

Elves, child-stealing, 350.

Embalmed in honey, 194.

Eminent men, foppery in, 159.

Epitaphs, curious, 110.

Escape from the guillotine, 204.

Essence of pearl, 134.

Esteem, inordinate self-, 76.

Etching upon glass, 136.

Eucharist, ordeal of the, 308.

Eve, St. Agnes, 234.

Evidences of divinity, queer, 182.

Executioner's sword, 101.

Expenses of Charles II., private, 53.

Exterminating vermin, 327.

Extraordinary muscular strength of the bat, 289.

Eyebrows, mole-skin, 189.

Eye, evil, in Spain, 320.

Eyes of the cuttle-fish, 290.

Fable of the pelican, 270.

Fabulous roc, 270.

Fair, donation to a, 69.

Fairy treasure, 220.

False conclusion, 52.

Famous astronomical clock, 81.

Famine, singular provision against, 283.

Farthing, half-penny and, 170.

Farthing, Queen Anne's, 177.

Fashionable disfigurement, 187.

Fatal day to the royal family of England, 57.

Feasts at coronations, 217.

Gluttony of the monks, 181.
Mammoth, 181.
Skeletons at, 183.

Feathers for the ladies, 176.

Feathers, house of hen's, 164.

Feat, manufacturing, 175.

Felling oaks, 246.

Few fish found at sea, 199.

Fiddlers, edict against, 58.

Fiddler's head, cutting off the, 236.

Fiddling to some purpose, 137.

Figures, acrobatic, 202.

Fillan's bell, St., 119.

Fineness of India muslins, 174.

Fine for insulting a king, 187.

Fire, ordeal by, 309.

First artificial limb, 198.

Brick house in Philadelphia, 54.
Butterfly, 350.
Hermits, 98.
Meerschaum pipe, 130.
Museums, 107.
New-laid egg, 274.
Opera, 198.
Oval lathe, 131.
Sale for the day, 338.

Fish, cuttle, 290.

Few found at sea, 199.
Fighting, 295.
Market at Scarborough, 198.
Mullet, 298.
Queer legend about, 296.

Fisherman's luck, 335.

Flea, jump of a, 291.

Fleas, remedy for, 256.

Flamingo, nest of the, 274.

Flower, passion, 311.

Flower pots, toppling, 62.

Fly, persistent, 292.

Following his nose, 133.

Foppery in eminent men, 159.

Food, yearly, of one man, 207.

Fog, crows lost in a, 267.

Fools, court, 155.

Force of imagination, 224.

Four-leaved clover, 303.

Freckles, May dew a remedy for, 231.

Freischutz, 168.

Friar Bacon's brazen head, 96.

Fungus, vegetable, 306.

Gall insect, 294.

Gambling for fingers, 160.

Gander, an old, 269.

Gander, led by a, 267.

Gems, 196.
Geranium, superstition about, 302.

German silver, 138.

Getting himself outside of his dinner, 284.

Giants, Chinese, 162.

Gifts, mystic, 327.

Gilt beards, 9.

Gipsy dance, 181.

Gipsy reticence, 170.

Glas, Bodach, 240.

Glass, discovery of, 134.

Drinking, a yard long, 109.
Etching upon, 136.

Glasses, hour, in coffins, 220.

Glastonbury thorn, 219.

Gloves, Anne Boleyn's, 179.

Gloves, winning a pair of, 237.

Gluttony of the monks, 181.

Gods, moving, 167.

Gold-headed canes for physicians, 207.

Gold letters, printed in, 10.

Golden tooth, 226.

Good tenant, 209.

Grace knives, 108.

Grandiloquent outbursts, 25.

Grave of Robin Hood, 118.

Great auk, 264.

Bell of Rouen, 119.
Digestive powers, 289.
Hero of the Bretons, 144.
Tom of Lincoln, 118.
Gretna Green marriages, 41.

Grimaldi's death, 244.

Grinning for a wager, 183.

Groaning boards, 212.

Guide posts, pious, 68.

Guillotine, decapitation by the, 204.

Guillotine, escape from the, 204.

Haarlem, sea-woman of, 150.

Hair cutting in Russia, 183.

Care of the beard, 214.
Human, 208.
Red, ancient antipathy to, 72.
Hallow E'en customs, 232.

Half-penny and farthing, 170.

Half-starved tramp, 137.

Halls, spacious, 171.

Hammock, bird, 272.

Handkerchiefs, love, 185.

Harrison's clock, 84.

Hat, color of, for cardinals, 173.

Hats, paradise of old, 205.

Head dresses, pagoda-shaped, 157.

Fiddler's, cutting off the, 236.
Friar Bacon's brazen, 96.
James IV. of Scotland, 153.
Orpheus, 104.
Of the bed, looking for the, 283.

Healing by the king, 232.

Hell-stones, 226.

Hens, can they count? 280.

Hens, cocks and, 266.

Hermits, first, 198.

Hero of the Bretons, 144.

High-sounding prologue, 34.

Hinder well-spout unlucky, 247.

Hindostan, trying land titles in, 163.

Hindoo oaths, 57.

Hindoo vow, singular, 231.

Holland, bell-ringing in, 122.

Honey, embalmed in, 194.

Hood, Robin, grave of, 118.

Honor, boots an object of, 193.

Honoring the lark, 252.

Horn of Oldenburg, 98.

Horn, unicorn, 319.

Horrible mode of assassination, 295.

Horses, Arabian, pedigree of, 284.

Curing vicious, 211.
Feeding one another, 286.
Witch-ridden, 340.

Horseshoes, 314.

Hour glasses in coffins, 220.

House, brick, first in Philadelphia, 54.

Carried on his head, 176.
Crickets, 330.
Leek, 307.
 Of hen's feathers, 164.

Houses, before numbered, 180.

Household rules in the sixteenth century, 56.

How the ancients rewarded inventors, 143.

How the Chinese select a pastor, 238.

How the prophesy of the destruction of Bath came about, 66.

How rats and mice use their tails, 280.

Huge copy of the Koran, 12.

Human hair, 208.

Human skin, drum made out of, 212.

Hundred families' lock, 188.

Hundred stomachs, 293.

Hundred years too soon, 55.

Hungary water, 141.

Hymn in the form of a cross, 49.

Illuminated manuscript Bible, 8.

Imprisoned during incubation, 270.

Imperishable prison literature, 16.

Imagination, force of, 224.

Immense value placed upon gems by the ancients, 196.

Immense zoological cabinet, 298.

Incredible liars, 223.

India, library of the king of, 19.

India, muslins, fineness of, 174.

Indian Bible, Eliot's, 12.

Inducements to subscribers, 34.

Infamous nankeen, 75.

Ingenuity, literary, 18.

Injury, adding insult to, 179.

Ink, river of, 200.

Innate appetite, 290.

Innocent blood, curse of, 228.

Inordinate self-esteem, 76.

Inscriptions on bells, 122.

In search of a looking glass, 70.

In search of a rhyme, 26.

Insects, phosphorescent, 292.

Instance of remarkable perseverance, 21.

Instance of sympathy, strange, 241.

Interesting manuscript Bible, 10.

Introducing a soul into a bird, 249.

Inventor, legend of, 228.

Inventors, how the ancients rewarded, 143.

Iron crown of Lombardy, 99.

Isabella color, 138.

Island, supposed to be a genuine, 18.

Isaiah, martyrdom of, 154.

Isthmus of Panama suggested three hundred years ago, 52.

Japan, children's day in, 195.

Java, leaf-butterfly of, 290.

Java, tree-planting in, 302.

Jericho, rose of, 303.

Jersey, bells of, 120.

Jew, wandering, 145.

John O'Gaunt's will, 58.

Judas Iscariot, legend of, 151.

Judy, Punch and, in 1669, 77.

Jumping jack, 185.

Jump of a flea, 291.

Kant's eccentricity, 44.

Kicked by a camel, 281.

Killing a robin, 254.

King, fine for insulting a, 187.

Fisher, 264.
Healing by the, 232.
Of India's library, 19.

King's cock-crower, 188.

King's health, a swallow drinks the, 258.

Kircher's speaking trumpet, 198.

Kissed while asleep, 237.

Kneeling statue of Atlas, 109.

Knives, grace, 108.

Knotted tails, cats with, 287.

Knot, true-lover's, 187.

Koran, huge copy of the, 12.

Ladies, feathers for the, 176.

Lantern, curious, 100.

Lark, honoring the, 252.

Lathe, oval, 131.

Latin Bible, magnificent, 10.

Lawyer, astonished, 70.

Law, curious, 60.

Lead pencils, no lead in, 177.

Leaf-butterfly of Java, 290.

Led by a gander, 267.

Legend, fish, 296.

Inventor, 228.
Judas Iscariot, 151.
Pontius Pilate, 149.
Strange, 229.
Wandering Jew, 145.

Lepers in England, treatment of, 237.

Letiche, 226.

Letters, gold, printed in, 10.

Letter M and the Napoleons, 161.

Letters, oddly addressed, 43.

Letter, superscription to a, 70.

Liars, incredible, 223.

Library, King of India's, 19.

Library, unique, 13.

Life in death, 225.

In the elephant, 276.
Time, washing but once in a, 317.

Light from potatoes, 178.

Lightning prints, 72.

Lightning, thunder and, 323.

Lilly's predictions, 36.

Lily, belladonna, 305.

Lily, veneration for the, 311.

Limb, first artificial, 198.

Lincoln, accidental rhymes of, 43.

Lincoln, Great Tom of, 118.

Lion, re-tailing a, 277.

Literary ingenuity, 18.

Locality, strange, for saying prayers, 169.

Lock, wonderful, 91.

Lock, hundred families', 188.

Lombardy, iron crown of, 99.

Long lost book recovered, 8.

Long psalm, value of a, 216.

Long-toed shoes, 209.

Looking back, 317.

Glass, in search of, 70.
For the head of the bed, 283.

Lord Bacon's dream, 246.

Lord Kingsborough's Mexico, 15.

Lotteries, 65.

Lottery, ancient, 64.

Lottery, cashing prizes, 63.

Lottery of women in India, 64.

Lost book, 13.

Lost in a fog, crows, 267.

Love birds, 271.

Charms, 314.
Handkerchiefs, 185.
Knot, true, 87.

Luck of Eden Hall, 101.

Of birthdays, 336.
Fisherman's, 335.
Lundyfoot's, 136.
Spitting for, 343.
Lucky find, 205.

Luminous human bodies, 178.

Luncheon, 210.

Luxury in 1652, 158.

Macduff's cross, 221.

Machinery, delicate, 86.

Magpie, 257.
Stoning a toad, 278.
Half a nest, 257.

Magnetic cures, 230.

Magic rain stone, 203.

Magnificent Latin Bible, 10.

Magnus, Albertus, recipes from, 74.

Mammoth bottle, 109.

Bell of Buddah, 119.
Feast, 181.
Pawnbroker's shop, 170.

Man carries his house on his head 176.

Man, yearly food of one, 207.

Manna marked with the number six, 324.

Manner of watchmen imitating the clock at Hernhuth, 55.

Manufacturing feat, 175.

Mark of the cross on the ass, 276.

Marriages, Gretna Green, 41.

Marriages in May unlucky, 344.

Martin Luther's tankard, 94.

Martyrs, blood of the, 338.

Martyrdom of Isaiah, 154.

May dew a cure for freckles, 231.

Mazarine Bible, 9.

Meal, one a day, 206.

Mecca, black stone of, 92.

Medallions only for the royal, 171.

Medical prescriptions, Chinese, 182.

Medicine, ancient practice of, 337.

Meerschaum pipe, first, 130.

Memnon, statue of, 103.

Memorandum, curious old, on the crucifixion of the Saviour, 38.

Menageries, phantom, 60.

Mexico, Lord Kingsborough's, 15.

Mezzotinting, 134.

Midsummer, rose at, 307.

Military salute, 75.

Minute mechanisms, 90.

Miniature time-piece, 88.

Missal, Bedford, 14.

Mixed moons, 338.

Mob wisdom, 173.

Moles (on the skin), 349.

Mole-skin eyebrows, 189.

Mole, voracity of the, 285.

Money, breaking a piece of, 314.

Monkish prayers, 180.

Monks, gluttony of the, 181.

Monks ordered to shave, 62.

Monkeys demanding their dead, 279.

Moons, mixed, 338.

Moon, cutting timber by the, 213.

Welcoming the new, 240.
Struck, 168.
Most curious book in the world, 7.

Motherly sacrifice of the gall insect, 294.

Mountain highway, 321.

Mouse ear, 305.

Moving gods, 167.

Mugwort, 305.

Mullet, colors in a dying, 298.

Mummies converted into paint, 174.

Museums, temples, the first, 107.

Muscles of the caterpillar, 292.

Muslins, fineness of India, 174.

Musical stones, 199.

Musical sand, 200.

Music, queer arctic, 174.

Mutilated Saint, efficacy in a, 216.

Mystic gifts, 327.

Name, simply on account of her, 240.

Nankeen, infamous, 75.

Napoleons and the letter M, 161.

Nature, copied from, 303.

Nebuchadnezzar's golden mask, 98.

Nest of the flamingo, 274.

Why the cuckoo builds no, 256.
Why the magpie builds but half a, 257.

Nettles, striking with, 236.

Newcastle, carrying coals to, 170.

New England primer, 14.

New moon, welcoming the, 240.

New nose, effect of a, 215.

Newspapers, vascillating, 22.

Nightingale, 252.

Night owls, 271.

Nile, crocodiles of the, 282.

No buttons but brass buttons, 73.

No lead in lead pencils, 177.

Nonius, opal of, 195.

Normandy treasures, 201.

Nose, following his, 133.

Noted anachronisms, 27.

Not buried at night, Sir John Moore, 28.

Not divine until smeared with red paint, 169.

Nothing new under the sun, 142.

Novel way of curing vicious horses, 211.

Number one, superstitious notion of, 323.

Seven, virtue in the, 325.
Six, manna marked with the, 324.

Oaks, felling, 246.

Oaths, Chinese, 172.

Oaths, Hindoo, 57.

Oddly addressed letters, 43.

Odd bill of repairs, 63.

Mode of revenge, 287.
Titles of old books, 29.
Way to discover a dead body, 334.

Odor of patchouli, 141.

Odors, tenacity of, 202.

Offending barbers, 77.

Oldenburg, Horn of, 98.

Oldest ballad, 40.

Old Grimaldi's death, 244.

Old pike, 297.

Old weather rhymes, 127.

Omens, cock-crowing, 319.

Picture, 245.
Poetry of, 329.
Reckless disregard of, 247.
Twelfth-night, 244.

One hundred years too soon, 55.

One meal a day, 206.

Onomancy, 326.

Opera, first, 197.

Opal of Nonius, 195.

Oranges, curious, 304.

Orchards, wassailing the, 235.

Ordeal in Africa, 308.

Of chewing rice, 309.
Of the cross, 307.
Of cold water, 309.
Of the Eucharist, 308.
By fire, 309.
By touch, 310.

Origin of attar of roses, 215.

Blue-tinted paper, 132.
Long-toed shoes, 209.
Orthography, singular specimen of, 34.

Orpheus, head of, 104.

Ostrich, 252.

Outburst, grandiloquent, 25.

Oval lathe, first, 131.

Owl, 259.

Owl, night, 271.

Owl's claws, 340.

Oyster dredging chaunt, 201.

Oysters growing on trees, 301.

Painters, blunders of, 23.

Paint, mummies converted into, 174.

Paint, not divine until smeared with, 169.

Pagoda-shaped head-dresses, 157.

Palindromes, 20.

Paper, blue-tinted, origin of, 132.

Paper, note, diminutive, 135.

Papers, watch, 78.

Parallel passages between Shakespeare and the Bible, 32.

Parallel, curious, 80.

Parisian scarlet, 139.

Paradise of old hats, 205.

Paradise, bird of, 259.

Passage through the Isthmus of Panama suggested three hundred years

ago, 52.

Passion flower, 311.

Passing bell, 121.

Pastor, how the Chinese secure a, 238.

Patagonian superstitions, 312.

Patchouli, odor of, 141.

Pawning Bibles and Waterloo medals, 212.

Pawnbroker's shop, mammoth, 170.

Pea crabs, 288.

Peacock at home, 268.

Peacocks' crests, 263.

Pearl, egg mistaken for a, 170.

Pearl, essence of, 134.

Pedigree of Arabian horses, 284.

Pelican, fable of the, 270.

Penguin breeding-grounds, 271.

Penguin's solitary egg, 262.

Peri, born of a, 214.

Perfumes, Egyptian, 293.

Perfumed butter, 194.

Perfume, saffron as a, 202.

Perforated stones, 328.

Persistent fly, 292.

Phantom menageries, 60.

Physicians, Egyptian, 169.

Gold-headed canes for, 207.
Symbol, 162.

Phœnix, 260.

Phosphorescent insects, 292.

Picnics centuries ago, 182.

Picture omens, 245.

Pigmies, 161.

Pike, an old, 297.

Pillory in Philadelphia, 54.

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookname.com