Professional Documents
Culture Documents
PDF Differential Geometry of Manifolds Second Edition Lovett Ebook Full Chapter
PDF Differential Geometry of Manifolds Second Edition Lovett Ebook Full Chapter
https://textbookfull.com/product/differential-geometry-of-
manifolds-second-edition-lovett-stephen-t/
https://textbookfull.com/product/lectures-on-the-geometry-of-
manifolds-3rd-edition-liviu-i-nicolaescu/
https://textbookfull.com/product/from-differential-geometry-to-
non-commutative-geometry-and-topology-neculai-s-teleman/
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/
https://textbookfull.com/product/differential-geometry-and-lie-
groups-a-computational-perspective-gallier-j/
https://textbookfull.com/product/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/
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
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
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.
Preface vii
Acknowledgements xiii
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
v
vi Contents
Bibliography 424
Index 429
Preface
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.
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
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.
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
~x = (x1 , x2 , . . . , xn ),
1
2 1. Analysis of Multivariable Functions
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
y x
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:
y
z
x
y
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
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
1.1. Functions from Rn to Rm 7
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
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 ,
8 1. Analysis of Multivariable Functions
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
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 .
10 1. Analysis of Multivariable Functions
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
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.
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
1.2. Continuity, Limits, and Differentiability 11
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 ,
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
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
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
t
(1, 2) Du~ F (−1, 2)
~u
F
s
t=
±1
t=
±2
1
=±
s=
±2
s
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.
(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),
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.’
451. Cp. Ars Amat. iii. 419.
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.,
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.
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:
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.’
361. Cp. Ars Amat. iii. 291.
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.
405. Cp. Ars Amat. iii. 179.
407. Cp. Ars Amat. iii. 185.
413-416. De Vita Monachorum, p. 186.
421-428. De Vita Monachorum, p. 189.
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,
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.
967 f. Cp. Rem. Amoris, 97 f.
969 f. Cp. Rem. Amoris, 101 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.’
979. Cp. Ars Amat. iii. 597.
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.,
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,
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.’