PDF Differential Geometry of Manifolds Second Edition Lovett Ebook Full Chapter

Differential geometry of manifolds
Second Edition Lovett

Visit to download the full and correct content document:
https://textbookfull.com/product/differential-geometry-of-manifolds-second-edition-lov
ett/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Differential geometry of manifolds Second Edition

Lovett Stephen T
https://textbookfull.com/product/differential-geometry-of-
manifolds-second-edition-lovett-stephen-t/
Lectures on the Geometry of Manifolds 3rd Edition Liviu

I. Nicolaescu
https://textbookfull.com/product/lectures-on-the-geometry-of-
manifolds-3rd-edition-liviu-i-nicolaescu/
From Differential Geometry to Non commutative Geometry

and Topology Neculai S. Teleman
https://textbookfull.com/product/from-differential-geometry-to-
non-commutative-geometry-and-topology-neculai-s-teleman/
Introduction to Riemannian manifolds Corrected version

of second edition Lee
https://textbookfull.com/product/introduction-to-riemannian-
manifolds-corrected-version-of-second-edition-lee/
Topics in Modern Differential Geometry 1st Edition
Stefan Haesen
https://textbookfull.com/product/topics-in-modern-differential-
geometry-1st-edition-stefan-haesen/
Differential geometry and Lie groups A computational

perspective Gallier J.
https://textbookfull.com/product/differential-geometry-and-lie-
groups-a-computational-perspective-gallier-j/
Differential Geometry Connections Curvature and

Characteristic Classes 1st Edition Loring W. Tu
https://textbookfull.com/product/differential-geometry-
connections-curvature-and-characteristic-classes-1st-edition-
loring-w-tu/
Differential Geometry Connections Curvature and

Characteristic Classes 1st Edition Loring W. Tu
https://textbookfull.com/product/differential-geometry-
connections-curvature-and-characteristic-classes-1st-edition-
loring-w-tu-2/
Elementary differential equations Second Edition

Roberts
https://textbookfull.com/product/elementary-differential-
equations-second-edition-roberts/
Differential Geometry
of Manifolds
Second Edition
Textbooks in Mathematics
Series editors:
Al Boggess and Ken Rosen
CRYPTOGRAPHY: THEORY AND PRACTICE, FOURTH EDITION

Douglas R. Stinson and Maura B. Paterson
GRAPH THEORY AND ITS APPLICATIONS, THIRD EDITION
Jonathan L. Gross, Jay Yellen and Mark Anderson
COMPLEX VARIABLES: A PHYSICAL APPROACH WITH APPLICATIONS, SECOND EDITION
Steven G. Krantz
GAME THEORY: A MODELING APPROACH
Richard Alan Gillman and David Housman
FORMAL METHODS IN COMPUTER SCIENCE
Jiacun Wang and William Tepfenhart
AN ELEMENTARY TRANSITION TO ABSTRACT MATHEMATICS
Gove Effinger and Gary L. Mullen
ORDINARY DIFFERENTIAL EQUATIONS: AN INTRODUCTION TO THE FUNDAMENTALS, SECOND
EDITION
Kenneth B. Howell
SPHERICAL GEOMETRY AND ITS APPLICATIONS
Marshall A. Whittlesey
COMPUTATIONAL PARTIAL DIFFERENTIAL PARTIAL EQUATIONS USING MATLAB®, SECOND EDITION
Jichun Li and Yi-Tung Chen
AN INTRODUCTION TO MATHEMATICAL PROOFS
Nicholas A. Loehr
DIFFERENTIAL GEOMETRY OF MANIFOLDS, SECOND EDITION
Stephen T. Lovett
MATHEMATICAL MODELING WITH EXCEL
Brian Albright and William P. Fox
THE SHAPE OF SPACE, THIRD EDITION
Jeffrey R. Weeks
CHROMATIC GRAPH THEORY, SECOND EDITION
Gary Chartrand and Ping Zhang
PARTIAL DIFFERENTIAL EQUATIONS: ANALYTICAL METHODS AND APPLICATIONS
Victor Henner, Tatyana Belozerova, and Alexander Nepomnyashchy
Advanced Problem Solving Using Maple: Applied Mathematics, Operation Research, Business Analytics, and
Decision Analysis
William P. Fox and William C. Bauldry
DIFFERENTIAL EQUATIONS: A MODERN APPROACH WITH WAVELETS
Steven G. Krantz
https://www.crcpress.com/Textbooks-in-Mathematics/book-series/CANDHTEXBOOMTH
Differential Geometry
of Manifolds
Second Edition
Stephen Lovett
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2020 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
Printed on acid-free paper
International Standard Book Number-13: 978-0-367-18046-1 (Hardback)
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish
reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the
consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in
this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright
material has not been acknowledged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any
form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.
copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400.
CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been
granted a photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identifica-
tion and explanation without intent to infringe.
Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com
and the CRC Press Web site at

http://www.crcpress.com
Contents
Preface vii
Acknowledgements xiii
1 Analysis of Multivariable Functions 1

1.1 Functions from Rn to Rm . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Continuity, Limits, and Differentiability . . . . . . . . . . . . . . . . 9
1.3 Differentiation Rules; Functions of Class C r . . . . . . . . . . . . . . 20
1.4 Inverse and Implicit Function Theorems . . . . . . . . . . . . . . . . 28
2 Variable Frames 39
2.1 Frames Associated to Coordinate Systems . . . . . . . . . . . . . . . 39
2.2 Frames Associated to Trajectories . . . . . . . . . . . . . . . . . . . 49
2.3 Variable Frames and Matrix Functions . . . . . . . . . . . . . . . . . 57
3 Differentiable Manifolds 65
3.1 Definitions and Examples . . . . . . . . . . . . . . . . . . . . . . . . 66
3.2 Differentiable Maps between Manifolds . . . . . . . . . . . . . . . . . 79
3.3 Tangent Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4 The Differential of a Differentiable Map . . . . . . . . . . . . . . . . 91
3.5 Manifolds with Boundaries . . . . . . . . . . . . . . . . . . . . . . . 97
3.6 Immersions, Submersions, and Submanifolds . . . . . . . . . . . . . . 102
3.7 Orientability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4 Multilinear Algebra 119

4.1 Hom Space and Dual . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.2 Bilinear Forms and Inner Products . . . . . . . . . . . . . . . . . . . 128
4.3 Adjoint, Self-Adjoint, and Automorphisms . . . . . . . . . . . . . . . 136
4.4 Tensor Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.5 Components of Tensors over V . . . . . . . . . . . . . . . . . . . . . 151
4.6 Symmetric and Alternating Products . . . . . . . . . . . . . . . . . . 160
4.7 Algebra over a Field . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
v
vi Contents
5 Analysis on Manifolds 179

5.1 Vector Bundles on Manifolds . . . . . . . . . . . . . . . . . . . . . . 180
5.2 Vector and Tensor Fields on Manifolds . . . . . . . . . . . . . . . . . 189
5.3 Lie Bracket and Lie Derivative . . . . . . . . . . . . . . . . . . . . . 196
5.4 Differential Forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
5.5 Pull-Backs of Covariant Tensor Fields . . . . . . . . . . . . . . . . . 216
5.6 Lie Derivative of Tensor Fields . . . . . . . . . . . . . . . . . . . . . 221
5.7 Integration on Manifolds - Definition . . . . . . . . . . . . . . . . . . 225
5.8 Integration on Manifolds - Applications . . . . . . . . . . . . . . . . 235
5.9 Stokes’ Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
6 Introduction to Riemannian Geometry 251

6.1 Riemannian Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
6.2 Connections and Covariant Differentiation . . . . . . . . . . . . . . . 267
6.3 Vector Fields along Curves; Geodesics . . . . . . . . . . . . . . . . . 279
6.4 Curvature Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
6.5 Ricci Curvature and Einstein Tensor . . . . . . . . . . . . . . . . . . 301
7 Applications of Manifolds to Physics 307

7.1 Hamiltonian Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . 308
7.2 Special Relativity; Pseudo-Riemannian Manifolds . . . . . . . . . . . 321
7.3 Electromagnetism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
7.4 Geometric Concepts in String Theory . . . . . . . . . . . . . . . . . 340
7.5 Brief Introduction to General Relativity . . . . . . . . . . . . . . . . 348
A Point Set Topology 363

A.1 Metric Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
A.2 Topological Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
B Calculus of Variations 403

B.1 Formulation of Several Problems . . . . . . . . . . . . . . . . . . . . 403
B.2 Euler-Lagrange Equation . . . . . . . . . . . . . . . . . . . . . . . . 404
B.3 Several Dependent Variables . . . . . . . . . . . . . . . . . . . . . . . 409
B.4 Isoperimetric Problems and Lagrange Multipliers . . . . . . . . . . . 410
C Further Topics in Multilinear Algebra 415

C.1 Binet-Cauchy and k-Volume of Parallelepipeds . . . . . . . . . . . . 415
C.2 Volume Form Revisited . . . . . . . . . . . . . . . . . . . . . . . . . 418
C.3 Hodge Star Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
Bibliography 424
Index 429
Preface
Purpose of this Book

This book is the second in a pair of books which together are intended to bring
the reader through classical differential geometry into the modern formulation of
the differential geometry of manifolds. The first book in the pair, by Banchoff and
Lovett, entitled Differential Geometry of Curves and Surfaces [6], introduces the
classical theory of curves and surfaces, only assuming the calculus sequence and
linear algebra. This book continues the development of differential geometry by
studying manifolds – the natural generalization of regular curves and surfaces to
higher dimensions. Though a background course in analysis is useful for this book,
we have provided all the necessary analysis results in the text. Though [6] provides
many examples of one- and two-dimensional manifolds that lend themselves well to
visualization, this book does not rely on [6] and can be read independently.
Taken on its own, this book provides an introduction to differentiable manifolds,
geared toward advanced undergraduate or beginning graduate readers in mathemat-
ics, retaining a view toward applications in physics. For readers primarily interested
in physics, this book may fill a gap between the geometry typically offered in under-
graduate programs and that expected in physics graduate programs. For example,
some graduate programs in physics first introduce electromagnetism in the context
of a manifold. The student who is unaccustomed to the formalism of manifolds
may be lost in the notation at worst or, at best, be unaware of how to do explicit
calculations on manifolds.
What is Differential Geometry?

Differential geometry studies properties of and analysis on curves, surfaces, and
higher dimensional spaces using tools from calculus and linear algebra. Just as the
introduction of calculus expands the descriptive and predictive abilities of nearly
every scientific field, so the use of calculus in geometry brings about avenues of
inquiry that extend far beyond classical geometry.
Though differential geometry does not possess the same restrictions as Euclidean
geometry on what types of objects it studies, not every conceivable set of points
falls within the purview of differential geometry. One of the underlying themes
vii
viii Preface
of this book is the development and description of the types of geometric sets on
which it is possible to “do calculus.” This leads to the definition of differentiable
manifolds. A second, and somewhat obvious, theme is how to actually do calculus
(measure rates of change of functions or interdependent variables) on manifolds. A
third general theme is how to “do geometry” (measure distances, areas and angles)
on such geometric objects. This theme leads us to the notion of a Riemannian
manifold.
Applications of differential geometry outside of mathematics first arise in me-
chanics in the study of the dynamics of a moving particle or system of particles.
The study of inertial frames is in common to both physics and differential geome-
try. Most importantly, however, differential geometry is necessary to study physical
systems that involve functions on curved spaces. For example, just to make sense of
directional derivatives of the surface temperature at a point on the earth (a sphere)
requires analysis on manifolds. The study of mechanics and electromagnetism on
a curved surface also requires analysis on a manifold. Finally, arguably the most
revolutionary application of differential geometry to physics came from Einstein’s
theory of general relativity, in which spacetime becomes curved in the presence of
mass/energy.
Organization of Topics
A typical calculus sequence analyzes one variable real functions (R → R), paramet-
ric curves (R → Rn ), multivariable functions (Rn → R) and vector fields (R2 → R2
or R3 → R3 ). This does not quite reach the full generality necessary for the defi-
nition of manifolds. Chapter 1 presents the analysis of functions f : Rn → Rm for
any positive integers n and m.
Chapter 2 discusses the concept and calculus of variable frames. Variable frames
arise naturally when using curvilinear coordinates, in the differential geometry of
curves (see Chapters 1, 3, and 8 of [5]), and, in physics, in the mechanics of a mov-
ing particle. In special relativity, of critical importance are momentarily comoving
reference frames (MCRFs), which are yet other examples of variable frames. Im-
plicit in our treatment of variable frames is a view toward Lie algebras. However,
to retain the chosen level of this book, we do not develop that theory here.
Chapter 3 defines the category of differentiable manifolds. Manifolds serve as the
appropriate and most complete generalization to higher dimensions of regular curves
and regular surfaces. The chapter also introduces the definition for the tangent space
on a manifold and attempts to provide the underlying intuition behind the abstract
definitions.
Before jumping into the analysis on manifolds, Chapter 4 introduces some neces-
sary background in multilinear algebra. We focus on bilinear forms, dual spaces, au-
tomorphisms of nondegenerate bilinear forms, and tensor products of vector spaces.
Chapter 5 then develops the analysis on differentiable manifolds, including the
differentials of functions between manifolds, vector fields, differential forms, and
Preface ix
integration.
Chapter 6 introduces Riemannian geometry without any pretention of being
comprehensive. One can easily take an entire course on Riemannian geometry, the
proper context in which one can do both calculus and geometry on a curved space.
The chapter introduces the notions of metrics, connections, geodesics, parallel trans-
port and the curvature tensor.
Having developed the technical machinery of manifolds, in Chapter 7 we apply
our the theory to a few areas in physics. We consider the Hamiltonian formulation
of dynamics, with a view toward symplectic manifolds; the tensorial formulation of
electromagnetism; a few geometric concepts involved in string theory, namely the
properties of the world sheet which describes a string moving in a Minkowski space;
and some fundamental concepts in general relativity.
In order to be rigorous and still only require the standard core in most under-
graduate math programs, three appendices provide any necessary background from
topology, calculus of variations, and a few additional results from multilinear alge-
bra. The reader without any background in analysis would be served by consulting
Appendix A on point set topology before Chapter 3.
A Comment on Using the Book

Because of the intended purpose of the book, it can serve well either as a textbook
or for self-study. The conversational style attempts to introduce new concepts in
an intuitive way, explaining why we formulate certain definitions as we do. As a
mathematics text, this book provides proofs or references for all theorems. On the
other hand, this book does not supply all the physical theory and discussion behind
the all the application topics we broach.
Each section concludes with an ample collection of exercises. Problems marked
with (*) indicate difficulty which may be related to technical ability, insight, or
length.
As mentioned above, this book only assumes prior knowledge of multivariable
calculus and linear algebra. A few key results presented in this textbook rely on
theorems from the theory of differential equations but either the calculations are all
spelled out or a reference to the appropriate theorem has been provided. Therefore,
except in the case of exercises about geodesics, experience with differential equations
is helpful though not necessary.
From the perspective of a faculty person using this as a course textbook, the
author intends every section to correspond to one 60-minute lecture period. With
the assumption of a 16-week semester, a course using this book should find the time
to cover all main sections and the appendices on topology. If a faculty knows that
his or her students have enough analysis or topology, Chapter 1 or Appendix A can
be skipped.
x Preface
Notation
It has been said jokingly that “differential geometry is the study of things that are
invariant under a change of notation.” A quick perusal of the literature on differ-
ential geometry shows that mathematicians and physicists usually present topics in
this field in a variety of different ways. One could argue that notational differences
have contributed to a communication gap between mathematicians and physicists.
In addition, the classical and modern formulations of many differential geometric
concepts vary significantly. Whenever different notations or modes of presentation
exist for a topic (e.g. differentials, metric tensor, tensor fields), this book attempts
to provide an explicit coordination between the notation variances.
As a comment on vector and tensor notation, this book consistently uses the
following conventions. A vector or vector function in a Euclidean vector space is
~
denoted by ~v , X(t) ~
or X(u, v). Vectors in an arbitrary vector space, curves on
manifolds, tangent vectors to a manifold, vector fields or tensor fields have no over-
right-arrow designation and are written, for example, as v, γ, X or T . A fair
number of physics texts use a bold font like g or A to indicate tensors or tensor
fields. Therefore, when discussing tensors taken from a physics context, we also use
that notation.
Different texts also employ a variety of notations to express the coordinates of a
vector with respect to a given basis. In this textbook, we regularly use the following
notation. If V is a vector space with an ordered basis B = (e1 , e2 , . . . , en ), then the
coordinates of a vector v ∈ V with respect to B are denoted by [v]B . More precisely,
 
v1
 v2 
[v]B =  .  if and only if v = v1 e1 + v2 e2 + · · · + vn en .
 
 .. 
vn
As a point of precision, when discussing coordinates we must use an ordered basis

since the order of vectors in the n-tuple matters for associating the correct coordi-
nate.
Beginning in Chapter 2, we switch from this typical notation to writing the in-
dices of coordinates in a superscript. So we will refer to the coordinates of v ∈ V
with respect to B as (v i ). This switch in notation from that developed in intro-
ductory linear algebra courses is standard in differential geometry and multilinear
algebra. The reason for this switch is explained fully in Section 4.1. In this context,
the superscript is not a power but an index. This modified notation is particularly
useful to recognize the difference between a (contravariant) vector and a dual vector
(also called covector) and then to use Einstein’s summation convention. This new
notation is standard in differential geometry, including applications in physics.
For linear transformations and their associated matrices, this book uses the
following convention. Suppose also that W is a vector space with a basis B 0 and
0
that T is a linear transformation T : V → W . Then we denote by [T ]B B the matrix
Preface xi
representing T with respect to the basis B on V and the basis B 0 on W . We recall

that this matrix is defined as the matrix such that
0
[T (v)]B0 = [T ]B
B [v]B
for all v ∈ V .
The authors of [6] chose the following notations for certain specific objects of
interest in differential geometry of curves and surfaces. Often γ indicates a curve
parametrized by X(t) ~ ~
while writing X(t) ~
= X(u(t), v(t)) indicates a curve on a
surface. The unit tangent and the binormal vectors of a curve in space are written
in the standard notation T~ (t) and B(t)~ but the principal normal is written P~ (t),
reserving N ~ (t) to refer to the unit normal vector to a curve on a surface. For a
plane curve, U ~ (t) is the vector obtained by rotating T~ (t) by a positive angle of π/2.
Furthermore, we denote by κg (t) the curvature of a plane curve to identify it as the
geodesic curvature of a curve on a surface. When these concepts occur in this text,
we use the same conventions as [6].
Occasionally, there arise irreconcilable discrepancies in habits of notation, e.g.,
how to place the signs on a Minkowski metric, how one defines θ and φ in spherical
coordinates, what units to use in electromagnetism, etc. In these instances the text
makes a choice that best suits its purpose and philosophical leanings, and indicates
commonly used alternatives.
Changes in the Second Edition

The second edition of this text arose from feedback from students and faculty using
this book and the author seeing room for improvement of his personal experience
teaching from it.
As a first major change to benefit faculty using this book, the second edition
commits that each section should correspond to one 60-minute lecture period. Con-
sequently, some of the sections in the first edition were split in two. Part of the
reorganization required the creation of a few new sections to cover topics, which
the author felt had been too compressed in the first edition, e.g., orientability of
manifolds, the Lie derivative of vector fields, applications of integration.
The centrality of multilinear algebra in this text’s approach encouraged us to
take that content out of the appendices in the first edition to become Chapter 4
in the current edition. This may feel like an interlude between Chapter 3, which
defines manifolds and differentiable maps between them, and Chapter 5, which stud-
ies the analysis on manifolds. Nonetheless, hopefully the location on this content
makes sense since it first becomes necessary in Chapter 5. Having a regular chapter
on multilinear algebra allows for a more natural introduction to tensors and the
notation for tensor component notation.
Woven throughout, the second edition attempts to improve the presentation
style and better foreshadow certain topics. For example, Equation (2.11) about
xii Preface
how to decompose the partial derivatives in a frame of vector fields augurs the
definition of a connection on a manifold.
Most of the exercises remained the same, though we improved the statements of
some and modified the challenge level of the computations for others. In addition,
we added a few new interesting problems.
Acknowledgments
I would first like to thank Thomas Banchoff my teacher, mentor and friend. After
one class, he invited me to join his team of students on developing electronic books
for differential geometry and multivariable calculus. Despite ultimately specializing
in algebra, the exciting projects he led and his inspiring course in differential ge-
ometry instilled in me a passion for differential geometry. His ability to introduce
differential geometry as a visually stimulating and mathematically interesting topic
served as one of my personal motivations for writing this book.
I am grateful to the students and former colleagues at Eastern Nazarene College.
In particular I would like to acknowledge the undergraduate students who served
as a sounding board for the first draft of this manuscript: Luke Cochran, David
Constantine, Joseph Cox, Stephen Mapes, and Chris Young. Special thanks are due
to my colleagues Karl Giberson, Lee Hammerstrom and John Free. In addition, I
am indebted to Ellie Waal who helped with editing and index creation.
The continued support from my colleagues at Wheaton College made writing
the first edition of this text a gratifying project. In particular, I must thank Terry
Perciante, Chair of the Department of Mathematics and Computer Science, for
his enthusiasm and his interest. I am indebted to Dorothy Chapell, Dean of the
Natural and Social Sciences, and to Stanton Jones, Provost of the College, for their
encouragement and for a grant which freed up my time to finish writing. I am also
grateful to Thomas VanDrunen and Darren Craig.
In preparation for the second edition, I need to thank Cole Adams, Dave Broad-
dus, Matthew McMillan, and Edwin Townsend for insightful feedback on the first
edition. Their comments on exercises and content significantly improved on the
first edition.
Finally, I am always grateful to my wife Carla Favreau Lovett and my daughter
Anne. While I was absorbed in writing the first edition, they braved the significant
time commitment and encouraged me at every step. They also continue to kindly
put up with my occasional geometry musings such as how to see the Gaussian
curvature in the reflection of the Cloud Gate in Chicago.
xiii
CHAPTER 1
Analysis of Multivariable Functions
Manifolds provide a generalization to the concept of a curve or a surface, objects

introduced in the usual calculus sequence. Parametrized curves into Rn are con-
tinuous functions from an interval of R to Rn ; parametrized surfaces in R3 involve
continuous functions from R2 to R3 . In order to generalize the study of curves and
surfaces to the theory of manifolds, we need a solid foundation in the analysis of
multivariable functions f : Rn → Rm .
1.1 Functions from Rn to Rm

Let U be a subset of Rn and let f : U → Rm be a function from U to Rm . Writing
the input variable as
~x = (x1 , x2 , . . . , xn ),
we denote the output assigned to ~x by f (~x) or f (x1 , . . . , xn ). Since the codomain

of f is Rm , the images of f are m-tuples so we can write
f (~x) = (f1 (~x), f2 (~x), . . . , fm (~x))

= (f1 (x1 , x2 , . . . , xn ), f2 (x1 , x2 , . . . , xn ), . . . , fm (x1 , x2 , . . . , xn )) .
The functions fi : U → R, for i = 1, 2, . . . , m, are called the component functions

of f .
We sometimes use the notation f~(~x) to emphasize the fact that the codomain
Rm is a vector space and that any operation on m-dimensional vectors is permitted
on functions f~ : Rn → Rm . Therefore, some authors call such functions vector
functions of a vector variable.
In any Euclidean space Rn , the standard basis is the set of vectors written as
1
2 1. Analysis of Multivariable Functions
{~e1 , ~e2 , . . . , ~en }, where  

0
 .. 
.
 
1
~ei =  
.
 .. 
0
with the only nonzero entry 1 occurring in the ith coordinate. If no basis is explicitly
specified for Rn , then it is assumed that one uses the standard basis.
At this point, a remark is in order concerning the differences in notations between
calculus and linear algebra. In calculus, one usually denotes an element of Rn as an
n-tuple and writes this element on one line as (x1 , x2 , . . . , xn ). On the other hand,
in order to reconcile vector notation with the usual manner we multiply a matrix
by a vector, in linear algebra we denote an element of Rn as a column vector
 
x1
 x2 
~x =  .  .
 
 .. 
xn
At first pass, we might consider these differences of notation as an unfortunate result
of history. However, the difference between column vectors and row vectors is not a
mere variance of notation: one represents the coordinates of an element in a vector
space V with respect to some basis, while the other represents the coordinates of an
element in the dual vector space V ∗ , a concept which we develop later. In the rest
of this book, we will write the components of a vector function on one line as per
the n-tuple notation, but whenever a vector or vector function appears in a linear
algebraic context, we write it as a column vector.
In the typical calculus sequence, we encounter vector functions or vector-valued
functions in the following contexts.
Example 1.1.1 (Curves in Rn ). A parametrized curve into n-dimensional space is a
continuous function ~x : I → Rn , where I is some interval of R. Parametrized curves
are vector functions of a single variable. We can view the independent variable as
coming from a one-dimensional real vector space.
Example 1.1.2 (Nonlinear Coordinate Changes). A general change of coordinates
in R2 is a function F : U → R2 , where U is the subset of R2 in which the coor-
dinates are defined. For example, the change from polar coordinates to Cartesian
coordinates is given by the function F : R2 → R2 defined by
F (r, θ) = (r cos θ, r sin θ).
Example 1.1.3. In a multivariable calculus, we encounter functions F : Rn → R,
written as F (x1 , x2 , . . . , xn ). All such functions are just examples of vector functions
of a vector variable with a codomain of R.
1.1. Functions from Rn to Rm 3
Example 1.1.4. As an example of a function from R2 to R3 , consider the function

2v(1 − u2 ) 4uv 1 − v2
F (u, v) = , , .
(1 + u2 )(1 + v 2 ) (1 + u2 )(1 + v 2 ) 1 + v 2
Notice that the component functions satisfy
4v 2 (1 − u2 )2 + 16u2 v 2 + (1 + u2 )2 (1 − v 2 )2
F12 + F22 + F32 =
(1 + u2 )2 (1 + v 2 )2
4v 2 (1 + u2 )2 + (1 + u2 )2 (1 − v 2 )2
=
(1 + u2 )2 (1 + v 2 )2
(1 + u2 )2 (1 + v 2 )2
= = 1.
(1 + u2 )2 (1 + v 2 )2
Thus, the image of F lies on the unit sphere S2 = {(x, y, z) ∈ R3 | x2 + y 2 + z 2 = 1}.

Note that F does not surject onto S2 . Assuming x2 + y 2 + z 2 = 1, if F (u, v) =
(x, y, z), then in particular
r
1 − v2 1−z
z= ⇐⇒ v =
1 + v2 1+z
which implies that −1 < z ≤ 1, and hence, the point (0, 0, −1) is not in the range
of F . Furthermore, since
2v 2 2v p
z2 + = 1, and thus = 1 − z2,
1 + v2 1 + v2
for any fixed z, we have
1 − u2 p 2u p
x= 1 − z2 and 1 − z2.
y=
1 + u2 1 + u2
√
But then, if y = 0, it is impossible to obtain x = − 1 − z 2 . Consequently, the
image of F is
p
F (R2 ) = S2 − {(x, y, z) ∈ S2 | x = − 1 − z 2 with z < 1}.
Figure 1.1 shows the image of F over the rectangle (x1 , x2 ) ∈ [−2, 5] × [0.5, 5].
There are a few different ways to visualize functions, particularly when n and
m are less than or equal to 3. Recall that the graph of a function f : Rn → Rm is
the subset of Rn × Rm = Rn+m defined by
{(x1 , . . . , xn , y1 , . . . , ym ) ∈ Rn+m | (y1 , . . . , ym ) = f (x1 , . . . , xn )}.
We can visualize this explicitly when m + n ≤ 3 with a three dimensional graphic.

When m = 1, we recover the usual method to depict functions f : R → R and
f : R2 → R.
y x
Figure 1.1: Portion of the image for Example 1.1.4.
For functions F : R2 → R (respectively F : R3 → R), another way to attempt to

visualize F is by plotting together (or in succession if one has dynamical graphing
capabilities) a collection of level curves (respectively surfaces) defined by F (x, y) =
ci (respectively F (x, y, z) = ci ) for a discrete set of values ci . This is typically called
a contour diagram of F . Figure 1.2 depicts a contour diagram of 2y/(x2 + y 2 + 1)
with c = 0, ±0.2, ±0.4, ±0.6, ±0.8.
In multivariable calculus or in a basic differential geometry course ([5]), one
typically uses yet another technique to visualize functions of the form f~ : R → Rm ,
for m = 2 or 3. By plotting the points that consist of the image of f~ we see a plane
or space curve. In doing so, we lose visual information about how fast one travels
along the curve. Figure 1.3 shows the image of the so-called space cardioid , given
by the function
f~(t) = (1 − cos t) cos t, (1 − cos t) sin t, sin t .

Similarly, in the study of surfaces, it is common to depict a function F~ : R2 → R3

by plotting its image in R3 . (The graph of a function of the form R2 → R3 is a
subset of R5 , which is quite difficult to visualize no matter what computer tools one
has at one’s disposal!)
We define the usual operations on functions as expected.
Definition 1.1.5. Let f~ and ~g be two functions defined over a subset U of Rn with
codomain Rm . Then we define the following functions:
1. (f~ + ~g ) : U → Rm , where (f~ + ~g )(~x) = f~(~x) + ~g (~x).

y
z
x
y
Figure 1.2: A contour diagram. Figure 1.3: A space curve.
2. (f~ · ~g ) : U → R, where (f~ · ~g )(~x) = f~(~x) · ~g (~x).
3. If m = 3, (f~ × ~g ) : U → R3 , where (f~ × ~g )(~x) = f~(~x) × ~g (~x).
Definition 1.1.6. Let f~ be a function from a subset U ⊂ Rn to Rm , and let ~g

be a function from V ⊂ Rm to Rs . If the image of f~ is a subset of V , then the
composition function ~g ◦ f~ is the function U → Rs defined by
(~g ◦ f~)(~x) = ~g f~(~x) .

Out of the vast variety of possible functions one could study, the class of linear
functions serves a fundamental role in the analysis of multivariable functions. We
remind the reader of various properties of linear functions.
Definition 1.1.7. A function F : Rn → Rm is called a linear function if
F (~x + ~y ) = F (~x) + F (~y ) for all ~x, ~y ∈ Rn ,

F (k~x) = kF (~x) for all k ∈ R and all ~x ∈ Rn .
If a function F : Rn → Rm is linear, then
F (~0) = F (~0 − ~0) = F (~0) − F (~0) = ~0,
and hence F maps the origin of Rn to the origin of Rm .

If B = {f~1 , f~2 , . . . , f~n } is a basis of Rn , then any vector ~v ∈ Rn can be written
uniquely as a linear combination of vectors in B as
~v = c1 f~1 + c2 f~2 + · · · + cn f~n .

One often writes the coefficients in linear algebra as the column vector
 
c1
 c2 
[~v ]B =  .  .
 
 .. 
cn
If the basis B is not specified, one assumes that the coefficients are given in terms
of the standard basis. If F is a linear function, then
F (~v ) = c1 F (f~1 ) + · · · + cn F (f~n ),
hence, to know all outputs of F one needs to know the coefficients of [~v ]B and the
output of the basis vectors of B. Suppose also that B 0 = {w ~ 1, w ~ m } is a basis
~ 2, . . . , w
of Rm . If the B 0 -coordinates of the outputs of the vectors in B are
     
a11 a12 a1n
 a21   a22   a2n 
[F (f~1 )]B0 =  .  , [F (f~2 )]B0 =  .  , · · · , [F (f~n )]B0 =  .  ,
     
 ..   ..   .. 
am1 am2 amn
then the image of the vector ~v ∈ Rn is given by
        
a11 a12 a1n a11 a12 ··· a1n c1
 a21   a22   a2n   a21 a22 ··· a2n   c2 
[F (~v )]B0 = c1  .  + c2  .  + · · · + cn  .  =  . ..   ..  .
        
.. ..
 ..   ..   ..   .. . . .  . 
am1 am2 amn am1 am2 ··· amn cn
The matrix  
a11 a12 ··· a1n
 a21 a22 ··· a2n 
A= .
 
.. .. .. 
 .. . . . 
am1 am2 ··· amn
is called the B, B -matrix representing the linear function F and is denoted by [F ]B
0
B0 .
Therefore,
[F (~v )]B0 = [F ]B
B0 [~
v ]B
for all ~v ∈ Rn .
Given a linear function F : Rn → Rm , one calls the image of F the set Im F =
F (Rn ), also called the range. The kernel of F is the zero set
ker F = {~u ∈ Rn | F (~u) = ~0}.
The image Im F is a vector subspace of the codomain Rm and the kernel is a
subspace of the domain Rn . The rank of F is the dimension dim(Im F ) and can
be shown to be equal to the size of the largest nonvanishing minor of any matrix
representing F , which is independent of the bases. The image of F cannot have a
greater dimension than either the domain or the codomain, so
rank F ≤ min{m, n},
and one says that F has maximal rank if rank F = min{m, n}. It is not hard to
show that a linear function F : Rn → Rm is surjective if and only if rank F = m
and F is injective if and only if rank F = n.
The rank is also useful in determining the linear dependence between a set of
vectors. If {~u1 , ~u2 , . . . , ~un } is a set of vectors in Rm , then the matrix

A = ~u1 ~u2 · · · ~un ,
where the ~ui are viewed as column vectors, represents a linear function F : Rn →
Rm , with
Im F = Span{~u1 , ~u2 , . . . , ~un }.
Thus, the set of vectors {~u1 , ~u2 , . . . , ~un } is linearly independent if and only if
rank F = n.
In the case of n = m, the determinant provides an alternative characterization
to linear independence. If F is a linear function from Rn to itself with associated
matrix A, then | det A| is the n-volume of the image under F of the unit n-cube.
Consequently, if the columns of A are not linearly independent, the n-volume of
this parallelopiped will be 0. This leads one to a fundamental summary theorem in
linear algebra.
Theorem 1.1.8. For a linear function F : Rn → Rn with associated square matrix
A, the following statements are equivalent:
1. rank F = n.
2. det A 6= 0.
3. Im F = Rn .
4. ker F = {~0}.
5. The column vectors of A are linearly independent.
6. The column vectors of A form a basis of Rn .
7. The column vectors of A span Rn .
8. F has an inverse function.
We remind the reader that matrix multiplication is defined in such a way so
that if A is the matrix for a linear function F : Rn → Rm and B is the matrix for
a linear function G : Rp → Rn , then the product AB is the matrix representing the
composition F ◦ G : Rp → Rm . In other words,
[F ◦ G]A B A
C = [F ]C [G]B ,
where A, B and C are bases on Rp , Rn , and Rm respectively. Furthermore, if

m = n and rank F = n, then the matrix A−1 is the matrix that represents the
inverse function of F .
A particularly important case of matrices representing linear transformations is
the change of basis matrix . Let B and B 0 be two bases on Rn . The change of basis
matrix from B to B 0 coordinates is M = [id]B n n
B0 , where id : R → R is the identity
n
transformation. In other words, for all ~v ∈ R ,
[~v ]B0 = M [~v ]B .

If B = {f~1 , f~2 , . . . , f~n }, then M = [f~1 ]B0 [f~2 ]B0 · · · [f~n ]B0 .
Problems
1.1.1. Consider the function F in Example 1.1.4. Prove algebraically that if the domain
is restricted to R × (0, +∞), it is injective. What is the image of F in this case?
1.1.2. Let F : R2 → R2 be the function defined by F (s, t) = (s2 − t2 , 2st), and let
G : R2 → R2 be the function defined by G(u, v) = (2u2 − 3v, uv + v 3 ). Calculate
the component functions of F ◦ G and of G ◦ F .
~ : [0, 2π] × [0, π] → R3 , with
1.1.3. Show that the function X
~ 1 , x2 ) = (cos x1 sin x2 , sin x1 sin x2 , cos x2 ),
X(x
defines a mapping onto the unit sphere in R3 . Which points on the unit sphere
have more than one preimage?
1.1.4. Consider the function F from R3 to itself defined by
F (x1 , x2 , x3 ) = (x1 + 2x2 + 3x3 , 4x1 + 5x2 + 6x3 , 7x1 + 8x2 + 9x3 ).
Prove that this is a linear function. Find the matrix associated to F (with respect
to the standard basis). Find the rank of F , and if the rank is less than 3, find
equations for the image of F .
1.1.5. Consider a line L in Rn traced out by the parametric equation ~ x(t) = t~a + ~b.
Prove that for any linear function F : Rn → Rm , the image F (L) is either a line
or a point.
1.1.6. Let F : Rn → Rm be a linear function, and let L1 and L2 be parallel lines in Rn .
Prove that F (L1 ) and F (L2 ) are either both points or both lines in Rm . If F (L1 )
and F (L2 ) are both lines, prove that they are parallel.
1.1.7. Let F : Rn → Rm be a linear function represented by a matrix A with respect
to a basis B on Rn and a basis B0 on Rm . Prove that F maps every pair of
perpendicular lines in Rn to another pair of perpendicular lines in Rm if and only
if AT A = λIn for some nonzero real number λ.
~ be a nonzero vector in Rn . Define the function F : Rn → R as
1.1.8. Let ω
F (~ ~ ·~
x) = ω x.
Prove that F is a linear function. Find the matrix associated to F (with respect
to the standard basis).
1.2. Continuity, Limits, and Differentiability 9
~ be a nonzero vector in R3 . Define the function F : R3 → R3 as

1.1.9. Let ω
F (~ ~ ×~
x) = ω x.
Prove that F is a linear function. Find the matrix associated to F (with respect
to the standard basis). Prove that rank F = 2.
1.2 Continuity, Limits, and Differentiability

Intuitively, a function is called continuous if it preserves “nearness.” A rigorous
mathematical definition for continuity for functions from Rn to Rm is hardly any
different for functions from R → R.
In calculus of a real variable, one does not study functions defined over a discrete
set of real values because the notions behind continuity and differentiability do not
make sense over such sets. Instead, one often assumes the function is defined over
some interval. Similarly, for the analysis of functions Rn to Rm , one does not study
functions defined from any subset of Rn into Rm . One typically considers functions
defined over what is called an open set in Rn , a notion we define now.
Definition 1.2.1. The open ball around ~x0 of radius r is the set
Br (~x0 ) = ~x ∈ Rn : k~x − ~x0 k < r .

A subset U ⊂ Rn is called open if for all ~x ∈ U there exists an r > 0 such that
Br (~x) ⊂ U .
Intuitively speaking, the definition of an open set U in Rn implies that at every
point p ∈ U it is possible to “move” in any direction by at least a little amount and
still remain in U . This means that in some sense U captures the full dimensionality
of the ambient space Rn . This is why, when studying the analysis of functions from
Rn to Rm , we narrow our attention to functions F : U → Rm , where U is an open
subset of Rn .
The reader is encouraged to consult Subsection A.1.2 in Appendix A for more
background on open and closed sets. The situation in which we need to consider
an open set U and a point ~x0 in U is so common that another terminology exists
for U in this case.
Definition 1.2.2. Let ~x0 ∈ Rn . Any open set U in Rn such that ~x0 ∈ U is called
an open neighborhood , or more simply, a neighborhood, of ~x0 .
We are now in a position to formally define continuity.
Definition 1.2.3. Let U be an open subset of Rn , and let F be a function from U
into Rm . The function F is called continuous at the point ~x0 ∈ U if F (~x0 ) exists
and if, for all ε > 0, there exists a δ > 0 such that for all ~x ∈ R,
k~x − ~x0 k < δ =⇒ kF (~x) − F (~x0 )k < .
The function F is called continuous on U if it is continuous at every point of U .
With the language of open balls, one can rephrase the definition of continuity
as follows. Let U be an open subset of Rn . A function F : U → Rm is continuous
at a point ~x0 if for all ε > 0 there exists a δ > 0 such that
F (Bδ (~x0 )) ⊂ Bε (F (~x0 )). (1.1)
Sections A.1.2 and A.1.4 in Appendix A provide a comprehensive discussion

about open sets in a metric space, a generalization of Rn , and continuity of functions
between metric spaces. We point out that using the language of open sets, Definition
1.2.3 can be rephrased once more in a manner that lines up with the definition of
continuity of functions between topological spaces.
Proposition 1.2.4. Let F : U → Rm be a function, where U ⊂ Rn is open. The

function F is continuous if and only if F −1 (V ) is open for all open sets V ∈ Rm .
Proof. Suppose the function F is continuous. Let V be open in Rm and let ~x0 ∈
F −1 (V ), which means that F (~x0 ) ∈ V . Since V is open, there exists ε > 0 such that
Bε (F (~x0 )) ⊂ V . By (1.1), there exists δ > 0 such that F (Bδ (~x0 )) ⊂ Bε (F (~x0 )).
This means that Bδ (~x0 ) ⊆ F −1 (V ), which, since ~x0 was arbitrary in V , implies that
F −1 (V ) is open.
Conversely, suppose that F −1 (V ) is open for all open sets V ⊂ Rm . Let ~x0 ∈ U
be any point and let ε > 0 be a real number. Consider the open ball Bε (F (~x0 )). By
hypothesis, F −1 (Bε (F (~x0 ))) is open in Rn . Since ~x0 ∈ F −1 (Bε (F (~x0 ))), we deduce
that there exists δ > 0 such that Bδ (~x0 ) ⊂ F −1 (Bε (F (~x0 ))). This is equivalent to
(1.1), so the proposition follows.
Example 1.2.5. Consider the function F : Rn → Rn defined by

(
~x/k~xk, if ~x 6= ~0,
F (~x) =
~0, if ~x = ~0.
This function leaves ~0 fixed and projects the rest of Rn onto the unit sphere. If
~x 6= ~0, then
~x ~x0 ~x ~x ~x ~x0
kF (~x) − F (~x0 )k = − ≤ − + − .
k~xk k~x0 k k~xk k~x0 k k~x0 k k~x0 k
However,
~x ~x 1 1 | k~x0 k − k~xk | 1
− = − k~xk = k~xk ≤ k~x − ~x0 k,
k~xk k~x0 k k~xk k~x0 k k~xk k~x0 k k~x0 k
and thus,
2
kF (~x) − F (~x0 )k ≤ k~x − ~x0 k.
k~x0 k
Consequently, given any ε > 0 and setting

1
δ = min k~x0 k, εk~x0 k ,
2
we know that ~x 6= ~0 and also that kF (~x) − F (~x0 )k < ε. Hence, F is continuous at
all ~x0 6= ~0.
On the other hand, if ~x0 = ~0, for all ~x 6= ~x0 ,
kF (~x) − F (~0)k = kF (~x) − ~0k = kF (~x)k = 1,
which can never be less than ε if ε ≤ 1.
Example 1.2.6. As a contrast to Example 1.2.5, consider the function

(
~x, if all components of ~x are rational,
F (~x) =
~0, otherwise.
The function F is obviously continuous at ~0, with δ = ε satisfying the requirements

of Definition 1.2.3. On the other hand, if ~x0 6= ~0, then in Bδ (~x0 ), for any δ > 0,
one can always find an ~x that has either all rational components or has at least one
irrational component. Thus, if ε < k~x0 k, for all δ > 0, we have
F (Bδ (~x0 )) 6⊂ Bε (F (~x0 )).
Thus, F is discontinuous everywhere except at ~0.
The following theorem implies many other corollaries concerning continuity of

multivariable functions.
Theorem 1.2.7. Let U be an open subset of Rn , let F : U → Rm be a function, and

let Fi , with i = 1, . . . , m, be the component functions. The function F is continuous
at the point ~a ∈ U if and only if, for all i = 1, . . . , m, the component function
Fi : U → R is continuous at ~a.
Proof. Suppose that F is continuous at ~a. Thus, for all ε > 0, there exists a δ > 0
such that k~x − ~ak < δ implies kF (~x) − F (~a)k < ε. Since
p
kF (~x) − F (~a)k = (F1 (~x) − F1 (~a)1 )2 + · · · + (Fm (~x) − Fm (~a))2
≥ |F (~x)i − F (~a)i |,
then for all ε > 0, having k~x − ~ak < δ implies that
|Fi (~x) − Fi (~a)| ≤ kF (~x) − F (~a)k < ε
for any i. Hence, each function Fi : U → R is continuous at ~a.

Conversely, suppose that all the functions Fi are continuous at ~a. Thus, for any √ ε
and for all i, there exist δi > 0 such that k~x−~ak < δi implies |Fi (~x)−Fi (~a)| < ε/ m.
Then taking δ = min(δ1 , . . . , δm ), if k~x − ~ak < δ, then
p
kF (~x) − F (~a)k = |F1 (~x) − F1 (~a)|2 + · · · + |Fm (~x) − Fm (~a)|2
r
ε2 ε2
≤ + ··· + = ε.
m m
Thus F is continuous.
If U is an open set containing a point ~a, then the set U − {~a} is called a
deleted neighborhood of ~a. If a function F is a function into Rm defined on a deleted
neighborhood of a point ~a ∈ Rn , it is possible to define the limit of F at ~a. The limit
~ such that if F (~a) were L,
of F at ~a is the value L ~ then F (~a) would be continuous
at ~a. We make this more precise as follows.
Definition 1.2.8. Let ~a ∈ Rn . Let F be a function from an open subset U − {~a} ⊂
~ and we write
Rn into Rm . The limit of F at ~a is defined as the point L,
~
lim F (~x) = L,
x→~
~ a
if for all ε there exists a δ such that

~
F (Bδ (~a) − {~a}) ⊂ Bε (L).
We point out right away that a function F : U → Rm , where U is open in Rn is

continuous at ~a ∈ U if and only if
lim F (~x) = F (~a).

x→~
~ a
Key results in calculus and analysis are the limit laws along with their implications
for continuity.
Theorem 1.2.9. Let U be an open set in Rn , let ~a ∈ U , and let F and G be
functions from U − {~a} to Rm and w : U − {~a} → R. Suppose that the limits of F ,
G, and w at ~a exist. Then

lim (F (~x) + G(~x)) = lim F (~x) + lim G(~x)
x→~
~ a ~x→~
a x→~
~ a

lim (w(~x)F (~x)) = lim w(~x) lim F (~x)
x→~
~ a ~x→~
a x→~
~ a

lim (F (~x) · G(~x)) = lim F (~x) · lim G(~x) (dot product)
x→~
~ a x→~
~ a x→~
~ a
lim kF (~x)k = lim F (~x) .

x→~
~ a x→~
~ a
t
(1, 2) Du~ F (−1, 2)
~u
F
s
t=
±1
t=
±2
1
=±
s=
±2
s
Figure 1.4: Example 1.2.12.
Proof. (Left as an exercise for the reader.)
Theorem 1.2.10. Let U be an open set in Rn , let F and G be functions from U

to Rm , let w : U → R, and suppose that F, G, and w are all continuous at ~a ∈ U .
Then the functions kF k, F + G, wF , and F · G are also continuous at ~a. If m = 3,
then the vector function F × G is also continuous at ~a.
Proof. (Left as an exercise for the reader.)
Similar to most multivariable calculus courses, before addressing partial deriva-

tives, we introduce the notion of a directional derivative, which measures the rate
of change of a function in a given direction.
Definition 1.2.11. Let F be a function from an open subset U ⊂ Rn into Rm , let

~x0 ∈ U be a point, and let ~u be a unit vector. The directional derivative of F in
the direction ~u at the point ~x0 is
F (~x0 + h~u) − F (~x0 )

D~u F (~x0 ) = lim
h→0 h
whenever the limit exists.
Another way to understand this definition is to consider the curve ~γ : (−ε, ε) →

Rm , for some ε > 0, defined by ~γ (t) = F (~x0 + t~u). Then D~u F (~x0 ) is equal to the
derivative ~γ 0 (0).
We note that though F is a multivariable function, the definition of D~u F (~x0 )
reduces to a single variable, vector-valued function before taking a derivative.
Example 1.2.12. Consider the function F (s, t) = (s2 − t2 , 2st) from R2 to itself.
We will calculate
√ the directional derivative of F at ~x0 = (1, 2) in the direction of
~u = (1/2, − 3/2).
We can picture this kind of function by plotting a discrete set of coordinate
lines mapped under F (see Figure 1.4). However, for functions like F that are not
injective, even this method of picturing F can be misleading since every point in
the codomain can have multiple preimages.
Now,
√ √
1 2 3 2 1 3
F (~x0 + t~u) = (1 + t) − (2 − t) , 2(1 + t)(2 − t)
2 2 2 2
√ √ √
= − 3 + (2 + 4 3)t − 2t2 , 4 + (4 − 2 3)t − 2 3t2 ,
so
√ √ √ √ √
D~u F (~x0 ) = (2 + 4 3) − 4t, (4 − 2 3) − 4 3t = (2 + 4 3, 4 − 2 3).
t=0
Figure 1.4 shows the curve F (~x0 + t~u) and illustrates the directional derivative
as being the derivative of F (~x0 + t~u) at t = 0. The figure shows that though ~u must
be a unit vector, the directional derivative is usually not.
Let F be a function from an open set U ⊂ Rn to Rm . For any point ~x0 ∈ U ,
the directional derivative of F in the direction ~uk at ~x0 is called the kth partial
derivative of F at ~x0 . The kth partial derivative of F is itself a vector function
possibly defined on a smaller set than U . Writing

F (~x) = F1 (x1 , . . . , xn ), . . . , Fm (x1 , . . . , xn ) ,
some common notations for the kth partial derivative D~uk F are
∂F
Fxk , , Dk F, F,k .
∂xk
In the last notation, the comma distinguishes the derivative operation from an
index. It is not hard to show that
∂F ∂F
1 ∂Fm
(~x) = (x1 , . . . , xn ), . . . , (x1 , . . . , xn ) .
∂xk ∂xk ∂xk
Example 1.2.13. Consider the real-valued function f (x1 , x2 ) defined by
 2
 x1 x2 , if (x , x ) 6= (0, 0),
1 2
f (x1 , x2 ) = x21 + x42
0, otherwise.

See Figure 1.5. We study the behavior of f near ~x = ~0.

Another random document with
no related content on Scribd:
117. Cp. Metam. i. 106, ‘Et quae deciderant patula Iovis arbore
glandes’: ‘patule glandes’ is nonsense.
119. Cp. Metam. i. 103.
128. A play on the word ‘regula’: ‘re’ has been taken away and
there remains only ‘gula.’
145. Cp. Metam. viii. 830.
147. Metam. viii. 835.
151 ff. Cp. Metam. viii. 837 ff.
163. Cp. Ars Amat. iii. 647.
165 f. Cp. Conf. Amantis, Prol. 473 ff.
175. Ars Amat. iii. 503 f., but Ovid has ‘Gorgoneo saevius,’ for
‘commota lenius.’
177. Cp. Metam. viii. 465, ‘Saepe suum fervens oculis dabat ira
ruborem.’ The reading ‘oculis’ is necessary to the sense and appears
in one manuscript.
179. Cp. Ovid, Ars Amat. iii. 509.
215. ‘corrodium’ (or ‘corredium’) is the allowance made from the
funds of a religious house for the sustentation of a member of it or of
someone else outside the house: see Ducange under ‘conredium’
and New Engl. Dict. ‘corrody.’ Gower himself perhaps had in his later
life a corrody in the Priory of Saint Mary Overey, of which he was a
benefactor.
302. The reference is to Ecclus. xix. 27, ‘Amictus corporis et risus
dentium et ingressus hominis enunciant de eo.’ Cp. Confessio
Amantis, i. 2705, margin.
305-310. Aurora, (MS. Bodley 822) f. 65,
‘Est nigra coruus auis et predo cadaueris, illum

Quem male denigrat ceca cupido notans.
. . . . .
Sub uolucrum specie descripsit legifer illos,
Quos mundanus honos ad scelus omne trahit.
Hunc aliquem tangit qui religionis amictum
Se tegit, ut cicius possit honore frui.’
(MS. Univ. Coll. 143: ‘libido’ for ‘cupido,’ ‘amictu’ for ‘amictum,’
‘maius’ for ‘cicius’).
311. Cp. Ovid, Ars Amat. iii. 249, ‘Turpe pecus mutilum,’ &c. The
word ‘monstrum’ in Gower came probably from a corruption in his
copy of Ovid.
327 ff. With this chapter compare Mirour de l’Omme, 21133 ff.
The capital letters of ‘Paciens,’ ‘Castus,’ ‘Luxus,’ &c. are supplied by
the editor, being clearly required by the sense.
354. Apocapata, ‘cut short’: cp. ‘per apocapen,’ v. 820.
363 f. The habit described is that of the Canons of the order of St.
Augustine.
395. Cp. Neckam, De Vita Monachorum, p. 175 (Rolls Series, 59,
vol. ii),
‘Vovistis, fratres, vovistis; vestra, rogamus,

Vivite solliciti reddere vota deo.
397. De Vita Monachorum, p. 176.

403 f. De Vita Monachorum, p. 177.
405-430. Most of this is taken from Neckam, De Vita
Monachorum, p. 176.
425. Ovid, Ars Amat. ii. 465.
427. foret, ‘should be,’ i.e. ‘ought to be.’
431-446. Taken with slight alterations from De Vita Monachorum,
pp. 187, 188.
449 Cp. Ovid, Fasti, ii. 85,
‘Saepe sequens agnam lupus est a voce retentus.’
Our author has interchanged the sexes for the purpose of his
argument, the man being represented as a helpless victim.
450. The subject to be supplied must be ‘agnus.’
453 f. Tristia, i. 6. 9 f.
461-466. De Vita Monachorum, p. 188.
469-490. Nearly the whole of this is taken from Neckam, p. 178.
537 f. Cp. Ovid, Rem. Amoris, 235 f.,
‘Adspicis ut prensos urant iuga prima iuvencos,

Et nova velocem cingula laedat equum?’
575. Cp. Amores, iii. 4. 17.

587. ‘Genius’ is here introduced as the priest of Venus and in l.
597 in the character of a confessor, as afterwards in the Confessio
Amantis. The reference to the ‘poets’ in the marginal note can hardly
be merely to the Roman de la Rose, where Genius is the priest and
confessor of Nature, but the variation ‘secundum Ouidium’ of the
Glasgow MS. does not seem to be justified by any passage of Ovid.
The connexion with Venus obviously has to do with the classical idea
of Genius as a god who presides over the begetting of children: cp.
Isid. Etym. viii. 88. The marginal note in S is written in a hand
probably different from that of the text, but contemporary.
617 f. Cp. Ars Amat. ii. 649 f.,
‘Dum novus in viridi coalescit cortice ramus,

Concutiat tenerum quaelibet aura, cadet.’
623. Spiritus est promptus, &c. Gower apparently took this text to
mean, ‘the spirit is ready to do evil, and the flesh is weak’: cp. Mirour,
14165.
624. Cp. Mirour, 16768.
637. For this use of ‘quid’ cp. that of ‘numquid,’ ii. Prol. 59, and v.
279.
648. Rev. xiv. 4, ‘Hi sequuntur agnum ... quocunque ierit.’
657 f. Apparently referring to Rev. xii. 14.
659. Cp. the Latin Verses after Confessio Amantis, v. 6358.
681 f. Cp. Ovid, Pont. iv. 4. 3 f.
689 ff. Cp. Mirour de l’Omme, 21266, margin.
699. fore: used here and elsewhere by our author for ‘esse’; see
below, l. 717, and v. 763.
715. Acephalum. This name was applied in early times to
ecclesiastics who were exempt from the authority of the bishop: see
Ducange. The word is differently used in iii. 956, and by comparison
with that passage we might be led to suppose that there was some
reference here to the ‘inopes’ and ‘opem’ of the next line.
723 ff. Compare with this the contemporary accounts of the
controversy between FitzRalph, archbishop of Armagh, and the
Mendicant Friars, who are said to have bribed the Pope to confirm
their privileges (Walsingham, i. 285), and the somewhat prejudiced
account of their faults in Walsingham, ii. 13. The influence of the
Dominican Rushook, as the king’s confessor was the subject of
much jealousy in the reign of Richard II.
735 ff. Cp. Mirour de l’Omme, 21469 ff.
736. sepulta: used elsewhere by Gower for ‘funeral rites,’ e.g. i.
1170. The meaning is that the friar claims to perform the funeral
services for the dead bodies of those whose confessor he has been
before death. Perhaps however we should take ‘sepulta’ here as
equivalent to ‘sepelienda.’
769. Hos. iv. 8: cp. Mirour, 21397, where the saying is attributed
to Zephaniah.
777 f. Cp. Ovid, Tristia, i. 9. 7 f.
781. Tristia, i. 9. 9.
784. Cp. Fasti, v. 354.
788. See Mirour, 21625 ff. and note.
795. ‘Prioris’ in S, but it is evidently an adjective here.
813 ff. Cp. Mirour, 21499 ff.
847. The wording is suggested by 1 Cor. ix. 24, ‘ii qui in stadio
currunt, omnes quidem currunt, sed unus accipit bravium.’
864. Titiuillus: see note in Dyce’s edition of Skelton, vol. ii. pp.
284 f.
869. Cp. Job ii. 4, ‘Pellem pro pelle, et cuncta quae habet homo,
dabit pro anima sua.’
872. vltima verba ligant. As in a bargain the last words are those
that are binding, so here the last word mentioned, namely ‘demon,’ is
the true answer to the question.
874. ‘Men sein, Old Senne newe schame,’ Conf. Amantis, iii.
2033.
903. Cp. Ovid, Metam. ii. 632, ‘Inter aves albas vetuit consistere
corvum.’ Gower’s line seems to have neither accidence nor syntax.
953 f. Fasti, ii. 219 f.
959. A reference to Ps. lxxii. 5, ‘In labore hominum non sunt, et
cum hominibus non flagellabuntur.’ The same passage is alluded to
in Walsingham’s chronicle (i. 324), where reference is made to the
fact that the friars were exempted from the poll-tax. The first half of
this psalm seems to have been accepted in some quarters as a
prophetic description of the Mendicants.
963. There is no variation of reading here in the MSS., but the
metre cannot be regarded as satisfactory. A fifteenth (or sixteenth)
century reader has raised a slight protest against it in the margin of
S, ‘at metrum quomodo fiet.’
969. Cp. Ps. lxxii. 7, ‘Prodiit quasi ex adipe iniquitas eorum:
transierunt in affectum cordis.’
971 ff. Cp. Mirour, 21517 ff.,
‘Mal fils ne tret son pris avant,
Par ce qant il fait son avant
Q’il ad bon piere,’ &c.

1059-1064. These six lines are taken without change from
Aurora, (MS. Bodley 822) f. 65.
1072. ‘lingua’ was here the original reading, but was altered to
‘verba’ in most of the copies. H and G have ‘verba’ over an erasure.
1081. In G we have ‘adepcio’ by correction from ‘adopcio.’
1090. adheret: meant apparently for pres. subj. as if from a verb
‘adherare.’
1099 f. Cp. Aurora, f. 19 vo,
‘Sarra parit, discedit Agar; pariente fideles

Ecclesia populos, dat synagoga locum.’
1103. Odium: written thus with a capital letter in H, but not in the
other MSS.
1143 ff. Cp. Mirour de l’Omme, 21403 ff. and note.
1145 ff. These lines are partly from Neckam’s Vita Monachorum,
p. 192:
‘Porticibus vallas operosis atria, quales

Quotque putas thalamos haec labyrinthus
habet.
. . . . .
Ostia multa quidem, variae sunt mille fenestrae,
Mille columnarum est marmore fulta domus.’
Gower alters the first sentence by substituting ‘valuas’ for the verb
‘vallas.’ ‘It has folding-doors, halls, and bed-chambers as various
and as many as the labyrinth.’
1161. ‘historia parisiensis’ in the MSS. I cannot supply a
reference.
1175 f. From De Vita Monachorum, p. 193.
1189 ff. The reference is to the Speculum Stultorum, where
Burnel the Ass, after examining the rules of all the existing orders
and finding them in various ways unsatisfactory to him, comes to the
conclusion that he must found an order of his own, the rules of which
shall combine the advantages of all the other orders. Members of it
shall be allowed to ride easily like the Templars, to tell lies like the
Hospitallers, to eat meat on Saturday like the Benedictines of Cluny,
to talk freely like the brothers of Grandmont, to go to one mass a
month, or at most two, like the Carthusians, to dress comfortably like
the Praemonstratensians, and so on. What is said here by our author
expresses the spirit of these rules rather than the letter.
1197 f. The text here gives the original reading, found in TH₂ and
remaining unaltered in S. CHG have ‘et si’ written over an erasure,
and in the next line ‘Mechari cupias’ is written over erasure in G,
‘Mechari cupias ordine’ in C, and ‘ordine’ alone in H. The other MSS.
have no erasures.
1212. CHG have this line written over an erasure.
1214. Written over erasure in CHG, the word ‘magis’ being still
visible in G as the last word of the line in the earlier text. The
expression ‘Linquo coax ranis’ is said to have been used by Serlo on
his renunciation of the schools: see Leyser, Hist. Poet. p. 443.
1215. The word ‘mundi’ is over erasure in CHG.
1221*-1232*. These lines are written over erasure in CHG.
1225. A planta capiti, ‘from foot to head’: more correctly, v. 116,
‘Ad caput a planta.’
LIB. V.
45. Architesis. It must be assumed that this word means ‘discord,’
the passage being a series of oppositions.
53. Est amor egra salus, &c. Compare the lines which follow our
author’s Traitié, ‘Est amor in glosa pax bellica, lis pietosa,’ &c., and
Alanus de Insulis, De Planctu Naturae, p. 472 (Rolls Series, 59, vol
ii).
79 ff. There is not much construction here; but we must suppose
that after this loose and rambling description the general sense is
resumed at l. 129.
98. Nec patet os in eis: cp. Chaucer, Book of the Duchess, 942.
104. Nec ... vix: cp. l. 153 and vii. 12.
121 f. Cp. Ovid, Her. iv. 71 f.
123 f. Cp. Fasti, ii. 763.
165. From Metam. vii. 826, but quoted without much regard to the
sense. In the original there is a stop after ‘est,’ and ‘subito collapsa
dolore’ is the beginning of a new sentence of the narrative.
169 f. Cp. Rem. Amoris, 691 f.
171. Cp. Her. iv. 161.
193. Cp. Her. v. 149. For ‘O, quia’ cp. i. 59.
209. Cp. Metam. x. 189.
213. Cp. Her. vii. 179. We have here a curious example of the
manner in which our author adapts lines to his use without regard to
the original sense.
221. Cp. Her. ii. 63.
257 ff. Cp. Mirour de l’Omme, 23920, Conf. Amantis, iv. 1634.
280. Numquid. This seems to be used here and in some other
passages to introduce a statement: cp. ii. Prol. 59, iv. 637. Rather
perhaps it should be regarded as equivalent to ‘Nonne’ and the
clause printed as a question: so vii. 484, 892, &c. For ‘num’ used
instead of ‘nonne’ cp. ii. 306.
299. S has in the margin in a later hand, ‘Nota de muliere bona.’
The description is taken of course from Prov. xxxi.
333. In the margin of S, as before, ‘Nota de muliere mala et eius
condicionibus.’
341 ff. Cp. Neckam, De Vita Monachorum, p. 186.
359 f. Cp. Ovid, Ars Amat. iii. 289, 294. Presumably ‘bleso’ in l.
360 is a mistake for ‘iusso.’
367 f. Ars Amat. iii. 311 f.
376. Cp. Ars Amat. i. 598.
383 f. This reference to Ovid seems to be with regard to what
follows about the art of preserving and improving beauty. Some of it
is from the Ars Amatoria, and some from Neckam, De Vita
Monachorum. For ‘tenent,’ meaning ‘belong,’ cp. iii. 584.
399-402. Taken with slight changes from Ars Amat. iii. 163-166.
403. Cp. Metam. ii. 635.
450. The line (in the form ‘Illa quidem fatuos,’ &c.) is written over
an erasure in the Glasgow MS.
454. ‘interius’ is written over an erasure in HG.
461. Vt quid, ‘Why.’
501. The reading ‘nos,’ which is evidently right, appears in CG as
a correction of ‘non.’
510. ‘While one that is stained with its own filth flies from the
field.’
520. Cp. Mirour de l’Omme, 23701 ff.
556. The neglect of the burden of a charge, while the honour of it
is retained, is a constant theme of denunciation by our author: cp. iii.
116, and below, ll. 655 ff.
557 ff. With this account of the labourers cp. Mirour de l’Omme,
26425 ff. It is noticeable that there is nothing here about the
insurrection.
593. Cp. Metam. vi. 318.
597. H punctuates here ‘salua. que.’
613. A quotation from Pamphilus: cp. Mirour, 14449.
659. maioris, ‘of mayor.’
693 f. Cp. Aurora, f. 36,
‘Dupla die sexta colleccio facta labore

Ostendit quia lux septima nescit opus.’
703. The capitals which mark the personification of ‘Fraus’ and

‘Vsura’ are due to the editor. ‘Fraus’ corresponds to ‘Triche’ in the
Mirour de l’Omme: see ll. 25237 ff.
731. Nonne, used for ‘Num,’ as also in other passages, e.g. vi.
351, 523, vii. 619.
745 ff. Cp. Mirour de l’Omme, 25741 ff.
In l. 745 SG have the reading ‘foris’ as a correction from ‘foras.’
760 ff. Cp. Chaucer, Cant. Tales, C 472 ff.
775. See note on l. 280.
785 f. The readings ‘fraus’ for ‘sibi’ and ‘surripit’ for ‘fraus capit’
are over erasure in CG.
812. ‘Thethis,’ (‘Thetis,’ or ‘Tethis’) stands several times for
‘water’ (properly ‘Tethys’): cp. vii. 1067. The line means that the
water is so abundant in the jar that it hardly admits the presence of
any malt (‘Cerem’ for ‘Cererem’).
835 ff. It is difficult to say who is the bad mayor of London to
whom allusion is here made. The rival leaders in City politics were
Nicholas Brembre and John of Northampton. The former was lord
mayor in the years 1377, 1378, and again in 1383 and 1384, when
he was elected against his rival (who had held the office in 1381,
1382) in a forcible and unconstitutional manner which evoked many
protests. Brembre, who belonged to the Grocers’ company,
represented the interests of the greater companies and was of the
Court party, a special favourite with the king, while John of
Northampton, a draper, engaged himself in bitter controversy with
the Fishmongers, who were supported by the Grocers, and was
popular with the poorer classes. In the Cronica Tripertita Gower
bitterly attacks Brembre (who was executed by sentence of the so-
called ‘Merciless Parliament’ in 1388), and we might naturally
suppose that he was the person referred to here; but that passage
was written before the political events which led to that invective and
in all probability not later than 1382, and the references to the low
origin of the mayor in question, ll. 845-860, do not agree with the
circumstances of Nicholas Brembre. Political passion in the City ran
high from the year 1376 onwards, and the person referred to may
have been either John of Northampton or one of the other mayors,
who had in some way incurred Gower’s dislike: cp. Mirour, 26365 ff.
877. Cp. Conf. Amantis, v. 7626,
‘It floureth, bot it schal not greine

Unto the fruit of rihtwisnesse.’
915 f. Ovid, Tristia, i. 5. 47 f.

922. Cp. Prov. xxv. 15, ‘lingua mollis confringet duritiam,’ and the
verses at the beginning of the Confessio Amantis,
‘Ossibus ergo carens que conterit ossa loquelis

Absit.’
953 f. Ars Amat. ii. 183 f., but Ovid has ‘Numidasque leones.’
957 f. Rem. Amoris, 447 f. (but ‘ceratas’ for ‘agitatas’).
965 f. Pont. iii. 7. 25 f.
971 f. Cp. Rem. Amoris, 729 f., ‘Admonitus refricatur amor,’ &c.
973. Cp. Rem. Amoris, 623.
975 f. Cp. Rem. Amoris, 731 f., ‘Ut pene extinctum cinerem si
sulfure tangas, Vivet,’ &c. The reading ‘sub’ must be a mistake on
the part of our author for ‘si.’
981. Ars Amat. iii. 373.
983 f. Ars Amat. iii. 375 f., but Ovid has ‘iratos et sibi quisque
deos.’
985 f. Cp. Ars Amat. iii. 501 f.
990. Fasti, iii. 380, absurdly introduced here.
991 f. Cp. Conf. Amantis, Latin Verses before Prol. 499.
1003 f. Cp. Tristia, ii. 141 f.
LIB. VI.
1-468. With this section of the work compare Mirour, 24181 ff.
11. Ps. xiv. 3.
89-94. From Aurora, (MS. Bodley 822) f. 66, where however the
reading is ‘sapit’ in l. 94 (for ‘rapit’).
95-98. Aurora, f. 65, where we find ‘in nocte’ for ‘in noctem’ and
‘reprobi’ for ‘legis’ (l. 97).
101 f. Cp. Aurora, 64 f.,
‘Inprouisus adest cum pullos tollere miluus

Esurit, in predam non sine fraude ruit.’
This is adapted by our author to his own purpose, but as his

meaning is altogether different, some obscurity results, and he does
not make it clear to us how the biter is bit.
113. Metam. v. 606.
115-118. Cp. Metam. vi. 527 ff.
133. In the Glasgow MS. ‘locuplex’ has been altered to the more
familiar ‘locuples.’
141 f. Is. v. 8, ‘Vae qui coniungitis domum ad domum et agrum
agro copulatis usque ad terminum loci: numquid habitabitis vos soli
in medio terrae?’ The same text is quoted in the Mirour, 24541 ff.
144. By comparison with Mirour, 24580 ff. we may see that the
dissipation of the property by the son is here alleged as a proof that
it has been ill acquired:
‘Qu’ils font pourchas a la senestre

Le fin demoustre la verrour.’
176. forum, i.e. the market price.

188. que foret equa, ‘(the balance) which should be fair’: so also
‘foret’ below, l. 190.
203. Basiliscus: cp. Mirour, 3748 ff.
209 f. Ovid, Pont. ii. 3. 39 f. (but ‘lasso’ for ‘lapso’).
217. nam nemo dolose Mentis, &c. ‘for no man of a crafty mind
can have sure speech.’
225. tenebrescunt, ‘darken.’ So other inceptives are used
transitively, e.g. ‘ditescere,’ ii. 607, Cron. Trip. iii. 119.
233 f. ‘And this lex, legis, from ledo, ledis, as ius from iurgo,
administers justice at this present time.’ It is meant that the
administration of law, as we see it, suggests the above etymologies.
The use of ‘isto’ for ‘hoc’ is quite regular.
249 ff. Cp. Mirour, 24349 ff., and see Pulling, Order of the Coif,
ch. iv.
269. The reference is to Ecclus. xx. 31, ‘Xenia et dona excaecant
oculos iudicum.’
274. ‘Fear puts to flight the discernment of justice.’
313-326. These fourteen lines are taken with some alterations
(not much for the better) from Neckam, De Vita Monachorum, pp.
180 f.
327 f. Cp. De Vita Monachorum, p. 182,
‘Sic mihi, divitibus si quando defuit hostis;

Hos terit et quassat saepe ruina gravis.’
Where, it would seem, we ought to read ‘Dic mihi.’

329 ff. De Vita Monachorum, p. 181. Most of the lines 329-348
are borrowed.
351. ‘Nonne’ for ‘Num,’ as often: cp. v. 731.
355 f. Cp. De Vita Monachorum, p. 182,
‘Iustitiae montes virtutumque ardua nullus

Scandet, dum mundi rebus onustus erit.’

359-372. Most of these lines are borrowed with slight alterations
from De Vita Monachorum, p. 191.
389. Cp. De Vita Monachorum, p. 192, ‘Cur ampla aedificas
busto claudendus in arcto?’
397. De Vita Monachorum, p. 193,
‘Et cecidit Babylon, cecidit quoque maxima Troia

Olim mundipotens, aspice, Roma iacet.’
419 ff. Cp. Mirour, 24817-25176.

421 f. For the idea contained in ‘vnccio’ and ‘vncta’ cp. iii. 1376.
433. ‘The word comes receives its beginning not from vice but
from vicium.’ That is, apparently, the prefix which makes ‘comes’ into
‘vicecomes’ is to be derived from ‘vicium.’
439 f. Cp. Mirour, 25166 ff.
445 ff. With this compare the corresponding lines in the Carmen
super multiplici viciorum Pestilencia, under the head of ‘Avarice’ (246
ff.),
‘Vendere iusticiam nichil est nisi vendere Cristum,’

&c.
463 f. Cp. Mirour, 24973 ff.

467 f. Vt Crati bufo, &c.: cp. Mirour, 24962 f.
498. Cp. Mirour, 22835 f.
522. The insertion which is found after this line in the Digby MS.
(and in no other) consists of eight lines taken from the original text of
the passage 545-580, which was rewritten by the author: see ll.
561*-566* and 579* f.
523 ff. ‘Can a house be built without timber? But of what use is
timber to the builder if it be not hewn?’ ‘Nonne’ for ‘Num,’ as
frequently: see note on v. 731. It seems that ‘sibi’ refers to the builder
rather than to the house; in any case, it has no reflexive sense.
Finally ‘ligna’ is here used as a singular feminine: all the MSS. have
‘foret’ in l. 524 and ‘valet’ in 525.
The idea of the passage seems to be that good laws are as the
material, and the ruler as the builder of the house.
529 ff. Cp. Conf. Amantis, vii. 2695 ff.
545-580. It is certain that the passage preserved to us in the
Dublin and Hatfield MSS. is that which was originally written in those
books which now exhibit an erasure; for in several places words are
legible underneath the present text of these latter MSS. For example
in S ‘maior’ is visible as the last word of the original l. 547, and
‘locuta,’ ‘aula,’ similarly in ll. 549, 551. The chief difference
introduced is in the direction of throwing more responsibility on the
king, who however is still spoken of as a boy. Thus instead of ‘Stat
puer immunis culpe,’ we have ‘Rex puer indoctus morales negligit
actus’ (or more strongly still ‘respuit’).
The text of 545*-580* follows the Dublin MS. (T) with corrections
from H₂. Neither text is very correct: both omit a word in l. 549*,
which I supply by conjecture, and both read ‘omnes’ in l. 561*. There
are some obvious errors in T, as ‘sinis’ for ‘sinit’ in l. 554*, ‘Tempe’ for
‘Tempora’ in l. 559*, which have been passed over without notice.
Cap. viii. Heading. The ensuing Epistle to the young king, which
extends as far as l. 1200, assumes a more severely moral form
owing to the alteration of the preceding passage, the exclusion of all
compliment (‘regnaturo’ in this heading for ‘excellentissimo’) and the
substitution of ‘doctrine causa’ for ‘in eius honore.’ (The readings
‘excellentissimo,’ ‘in eius honore’ no doubt are to be found in the
Hatfield MS., but I have accidentally omitted to take note of them.)
629 f. Neckam, De Vita Monachorum, p. 185,
‘Quid tibi nobilitas et clarum nomen avorum,

Si vitiis servus factus es ipse tuis?’
640. ‘vix’ is sometimes used by our author (apparently) in the

sense of ‘paene.’
696. Ovid, Rem. Amoris, 526.
710. iudiciale, ‘judgement,’ used as a substantive: cp. iii. 1692.
718. culpe ... sue, ‘for their fault,’ i.e. the fault of his ministers.
719-722. Cp. Aurora, (MS. Bodley 822) f. 65,
‘Euolat ancipiter ad prede lucra, suisque

Deseruit dominis in rapiendo cybum.
Sic multi dominis famulando suis, ad eorum
Nutum pauperibus dampna ferendo nocent.’
725. presul, ‘the bishop.’

740. The expression ‘Cuius enim’ for ‘Eius enim’ occurs more
than once, e.g. l. 1238: cp. vii. 372. It is found also in the Confessio
Amantis, Latin Verses after vii. 1984, but was there corrected in the
third recension.
765. stabiles: apparently used in a bad sense.
793 f. Cp. Aurora, f. 96 vo,
‘Exiguus magnum vicit puer ille Golyam,

Nam virtus humilis corda superba domat.’
816. Ovid, Amores, i. 8. 62, ‘Crede mihi, res est ingeniosa dare.’
839 f. Cp. Aurora, f. 95 vo.
846. Fasti, ii. 226.
875-902. This passage of twenty-six lines is taken with few
alterations from the Aurora, f. 76.
876. bella: in the original ‘corda’ (or ‘colla’ MS. Univ. Coll. 143).
883. noctibus: in the original ‘nutibus.’
884. Spirant: so in the original according to MS. Bodley 822, but
‘Spirent’ in MS. Univ. Coll. 143.
886. acuum ferrum: in the original ‘minitans ferrum.’ Apparently
our author took ‘acus’ to mean a spear or javelin. The choice of the
word in this passage is unfortunate.
887 ff. ‘vincit,’ ‘tenet’ (or ‘teret,’ MS. Univ. Coll. 143), ‘consurgit’ in
the original.
891. In the original, ‘Rex hoc consilium grata bibit aure, puellas
Preparat,’ &c.
892. ‘genis’ in the original.
894. ‘furit’ for ‘fugat’ is the reading of the original, and we find this
in several MSS. of our text, but in the Glasgow MS. this has been
corrected to ‘fugat,’ which is the reading of S.
898. In the original, ‘Vultus que geminus ridet in ore decor,’ (or
‘Vultus et geminus,’ &c., MS. Univ. Coll. 143).
907. Aurora, f. 100.
947-950. Taken from the description of Saul at the battle of
Gilboa, Aurora, f. 100 vo.
971 ff. Cp. Praise of Peace, 78 ff.
985-992. From Aurora, f. 64 vo,
‘Alta petens aquila uolat alite celsius omni,

Quisque potens, tumidus corde, notatur ea:
Vt sacra testantur cythariste scripta prophete,
In celum tales os posuere suum.
Pennatum griphes animal, pedibusque quaternis
Inuitos homines carpit, abhorret equos:
Designatur in his facinus crudele potentum,
Qui mortes hominum cum feritate bibunt.’
986. Our author no doubt read ‘mundus corde’ here in the

Aurora.
987. citharistea: properly no doubt ‘cithariste,’ to be taken with
‘prophete,’ as in the Aurora.
990. ‘horret equos’ seems to represent the ‘equis vehementer
infesti’ of Isidore, Etym. xii. 2.
1019-1024. From Neckam, De Vita Monachorum, p. 185, with
slight variations.
1037. esse: as substantive, ‘existence.’
1041-1050. Taken with slight changes from Aurora, f. 108.
1066. fugat: used as subjunctive; so also iii. 1498, 2078.
1085 f. From De Vita Monachorum, p. 184.
1159* ff. That this was the text which stood originally in S is
proved partly by the fact that the original heading of the chapter
stands still as given here in the Table of Chapters, f. 5, and also by
the traces of original coloured initials at ll. 1175 and 1199. A
considerable part of the erased chapter reappears in the poem ‘Rex
celi deus,’ &c., addressed to Henry IV: see p. 343.
1189 f. Si tibi ... cupias conuertere ... Te. These words appear in
S as a correction of the rewritten text by a second erasure and in
another hand.
Cap. xix. Heading. The original form, as given by DLTH₂, is still to
be found in the Table of Chapters in S.
1201. Cp. Ovid, Metam. vii. 585 f.,
‘veluti cum putria motis

Poma cadunt ramis agitataque ilice glandes.’
1204 ff. Note the repeated use of ‘modo’ in the sense of ‘now’:
cp. 1210, 1218, 1222, 1232, 1235, 1243, 1263, 1280, &c. The usual
word for ‘formerly’ is ‘nuper’; see 1241, 1245, 1279, &c.
1205. Metam. ii. 541.
1223. Oza, that is Uzzah (2 Sam. vi.), who is selected as a type
of carnal lust, apparently on the strength of the quite gratuitous
assumption adopted in Lib. III. 1885 ff. Apparently ‘luxus’ in the next
line is genitive, in spite of the metre: cp. ‘excercitus,’ i. 609, ‘ducatus,’
Cron. Trip. iii. 117.
1236. Giesi, i.e. Gehazi.
1238. Cuius enim: cp. note on l. 740.
1243. Liberius: pope from 352-366 a. d. He is mentioned here as
a type of unfaithfulness to his charge, because he was induced to
condemn Athanasius.
1251. defunctis, ‘for the dead,’ that is, to bury them charitably, as
Tobit did.
1261. Cp. John xii. 24.
1267. Perhaps an allusion to Wycliffe, who seems to be referred
to as a new Jovinianus in a later poem, p. 347.
1268. dant dubitare, ‘cause men to doubt.’
1273. Troianus: i.e. Trajan, whose name is so spelt regularly by
our author.
1277. Valentinianus: cp. Conf. Amantis, v. 6398 ff.
1284. Leo: cp. Conf. Amantis, Prol. 739.
1286. Tiberii: i.e. Tiberius Constantinus; cp. Conf. Amantis, ii. 587
ff.
1306. quis, for ‘quisquam’: so also ‘quem’ in l. 1308; cp. i. 184.
1321 f. Cp. Conf. Amantis, vii. 2217 ff.: ‘relinquendo’ is used for
‘relinquens,’ as i. 304, 516, &c.
1323. Cp. Conf. Amantis, v. 6372 ff., Mirour, 18301 ff.
1330. Vix si: cp. iv. 218, Cron. Trip. iii. 444.
1345. Cp. Ovid, Amores, i. 9. 1.
1357 f. ‘She is silent as a jackdaw, chaste as a pigeon, and
gentle as a thorn.’
1361 f. Perhaps an allusion to the case of Edward III and Alice
Perrers.
LIB. VII.
5. Cp. Conf. Amantis, Prol. 595 ff.
9. modo, ‘now’: cp. note on vi. 1204.
12. nec ... vix. For this combination of ‘vix’ with a negative cp. v.
104, 153.
42. dicunt ... volunt, ‘say that they wish’: cp. ii. 200 f.
47 f. Cp. Conf. Amantis, v. 49 ff.; so below, ll. 61 ff.
123. Rev. ii. 25, ‘id quod habetis tenete, donec veniam.’

PDF Differential Geometry of Manifolds Second Edition Lovett Ebook Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PDF Differential Geometry of Manifolds Second Edition Lovett Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

Differential geometry of manifolds

Second Edition Lovett

Differential geometry of manifolds Second Edition

Lectures on the Geometry of Manifolds 3rd Edition Liviu

From Differential Geometry to Non commutative Geometry

Introduction to Riemannian manifolds Corrected version

Differential geometry and Lie groups A computational

Differential Geometry Connections Curvature and

Differential Geometry Connections Curvature and

Elementary differential equations Second Edition

CRYPTOGRAPHY: THEORY AND PRACTICE, FOURTH EDITION

© 2020 by Taylor & Francis Group, LLC

No claim to original U.S. Government works

Printed on acid-free paper

International Standard Book Number-13: 978-0-367-18046-1 (Hardback)

Visit the Taylor & Francis Web site at

and the CRC Press Web site at

1 Analysis of Multivariable Functions 1

4 Multilinear Algebra 119

5 Analysis on Manifolds 179

6 Introduction to Riemannian Geometry 251

7 Applications of Manifolds to Physics 307

A Point Set Topology 363

B Calculus of Variations 403

C Further Topics in Multilinear Algebra 415

Purpose of this Book

What is Differential Geometry?

A Comment on Using the Book

As a point of precision, when discussing coordinates we must use an ordered basis

representing T with respect to the basis B on V and the basis B 0 on W . We recall

Changes in the Second Edition

Analysis of Multivariable Functions

Manifolds provide a generalization to the concept of a curve or a surface, objects

1.1 Functions from Rn to Rm

we denote the output assigned to ~x by f (~x) or f (x1 , . . . , xn ). Since the codomain

f (~x) = (f1 (~x), f2 (~x), . . . , fm (~x))

The functions fi : U → R, for i = 1, 2, . . . , m, are called the component functions

{~e1 , ~e2 , . . . , ~en }, where  

Example 1.1.4. As an example of a function from R2 to R3 , consider the function

Thus, the image of F lies on the unit sphere S2 = {(x, y, z) ∈ R3 | x2 + y 2 + z 2 = 1}.

{(x1 , . . . , xn , y1 , . . . , ym ) ∈ Rn+m | (y1 , . . . , ym ) = f (x1 , . . . , xn )}.

We can visualize this explicitly when m + n ≤ 3 with a three dimensional graphic.

Figure 1.1: Portion of the image for Example 1.1.4.

For functions F : R2 → R (respectively F : R3 → R), another way to attempt to

f~(t) = (1 − cos t) cos t, (1 − cos t) sin t, sin t .

Similarly, in the study of surfaces, it is common to depict a function F~ : R2 → R3

1. (f~ + ~g ) : U → Rm , where (f~ + ~g )(~x) = f~(~x) + ~g (~x).

Figure 1.2: A contour diagram. Figure 1.3: A space curve.

2. (f~ · ~g ) : U → R, where (f~ · ~g )(~x) = f~(~x) · ~g (~x).

3. If m = 3, (f~ × ~g ) : U → R3 , where (f~ × ~g )(~x) = f~(~x) × ~g (~x).

Definition 1.1.6. Let f~ be a function from a subset U ⊂ Rn to Rm , and let ~g

(~g ◦ f~)(~x) = ~g f~(~x) .

F (~x + ~y ) = F (~x) + F (~y ) for all ~x, ~y ∈ Rn ,

If a function F : Rn → Rm is linear, then

F (~0) = F (~0 − ~0) = F (~0) − F (~0) = ~0,

and hence F maps the origin of Rn to the origin of Rm .

~v = c1 f~1 + c2 f~2 + · · · + cn f~n .

rank F ≤ min{m, n},

where A, B and C are bases on Rp , Rn , and Rm respectively. Furthermore, if

~ be a nonzero vector in R3 . Define the function F : R3 → R3 as

1.2 Continuity, Limits, and Differentiability

F (Bδ (~x0 )) ⊂ Bε (F (~x0 )). (1.1)

Sections A.1.2 and A.1.4 in Appendix A provide a comprehensive discussion