Professional Documents
Culture Documents
Full Chapter Distribution Theory Applied To Differential Equations 1St Edition Adina Chirila PDF
Full Chapter Distribution Theory Applied To Differential Equations 1St Edition Adina Chirila PDF
https://textbookfull.com/product/introduction-to-complex-theory-
of-differential-equations-savin/
https://textbookfull.com/product/an-invitation-to-applied-
mathematics-differential-equations-modeling-and-computation-1st-
edition-carmen-chicone/
https://textbookfull.com/product/applied-stochastic-differential-
equations-1st-edition-simo-sarkka-and-arno-solin/
https://textbookfull.com/product/applied-differential-equations-
with-boundary-value-problems-1st-edition-vladimir-a-dobrushkin/
Theory of Differential Equations in Engineering and
Mechanics 1st Edition Kam Tim Chau
https://textbookfull.com/product/theory-of-differential-
equations-in-engineering-and-mechanics-1st-edition-kam-tim-chau/
https://textbookfull.com/product/introduction-to-partial-
differential-equations-peter-j-olver/
https://textbookfull.com/product/from-ordinary-to-partial-
differential-equations-1st-edition-giampiero-esposito-auth/
https://textbookfull.com/product/ordinary-differential-equations-
and-boundary-value-problems-volume-i-advanced-ordinary-
differential-equations-1st-edition-john-r-graef/
https://textbookfull.com/product/elementary-differential-
equations-second-edition-roberts/
Adina Chirilă
Marin Marin
Andreas Öchsner
Distribution
Theory Applied
to Differential
Equations
Distribution Theory Applied to Differential
Equations
Adina Chirilă Marin Marin Andreas Öchsner
• •
123
Adina Chirilă Marin Marin
Department of Mathematics Department of Mathematics
and Computer Sciences and Computer Sciences
Transilvania University of Braşov Transilvania University of Braşov
Braşov, Romania Braşov, Romania
Andreas Öchsner
Faculty of Mechanical Engineering
Esslingen University of Applied Sciences
Esslingen am Neckar, Baden-Württemberg
Germany
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Initial Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Test Functions and Regularization . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Seminorms and Locally Convex Spaces . . . . . . . . . . . . . . . . . . 11
2.3.1 Locally Convex Spaces . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 Convex and Balanced Sets . . . . . . . . . . . . . . . . . . . . . 13
2.3.3 Absorbing Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Duals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 Reflexive Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5 The Inductive Limit Topology . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3 Convex and Lower-Semicontinuous Functions . . . . . . . . . . . . . . . . 25
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.2 Convex Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.3 Lower Semicontinuous Functions . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Convexity and Lower Semicontinuity . . . . . . . . . . . . . . . . . . . . 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4 The Subdifferential of a Convex Function . . . . . . . . . . . . . . . . . . . . 37
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.2 The Conjugate Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 The Additivity of the Subdifferential . . . . . . . . . . . . . . . . . . . . 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
v
vi Contents
5 Evolution Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 The Resolvent and the Yosida Approximation . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6 Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ 73
6.1 Fundamental Spaces in the Theory of Distributions . . ........ 73
6.1.1 On Some Properties of the Spaces C m ðXÞ
and C1 ðXÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
6.2 The Space of Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
6.3 The Dual of C 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
6.4 The Derivative of a Distribution . . . . . . . . . . . . . . . . . . . . . . . 87
6.5 Distributions as Generalized Functions . . . . . . . . . . . . . . . . . . . 90
6.6 On Some Spaces of Distributions . . . . . . . . . . . . . . . . . . . . . . . 96
6.7 The Primitive of a Distribution . . . . . . . . . . . . . . . . . . . . . . . . 96
6.7.1 Structure Theorems . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.8 Extras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
6.8.1 Higher-Order Primitives . . . . . . . . . . . . . . . . . . . . . . . 109
6.8.2 The Local Structure of Distributions . . . . . . . . . . . . . . 112
6.9 Convolutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6.9.1 The Direct Product of Distributions . . . . . . . . . . . . . . . 117
6.9.2 Convolution of Distributions . . . . . . . . . . . . . . . . . . . . 121
6.9.3 Convolution of Functions and Distributions:
Regularization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.9.4 Convolution Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
7 Tempered Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.1 The Schwartz Space of Infinitely Differentiable Functions
Rapidly Decreasing at Infinity . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.2 Tempered Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.3 The Fourier Transform in SðRn Þ . . . . . . . . . . . . . . . . . . . . . . . 140
7.3.1 The Inverse Fourier Transform . . . . . . . . . . . . . . . . . . 142
7.3.2 Properties of the Fourier Transform . . . . . . . . . . . . . . . 144
7.4 Fourier Transform of Tempered Distributions . . . . . . . . . . . . . . 145
7.5 The Fourier Transform of a Distribution with Compact
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.6 The Product of a Distribution by a C 1 Function . . . . . . . . . . . 151
7.7 The Space of Multipliers of S 0 ðRn Þ . . . . . . . . . . . . . . . . . . . . . 152
7.8 Some Results on Convolutions with Tempered Distributions . . . 156
7.9 The Paley-Wiener-Schwartz Theorem . . . . . . . . . . . . . . . . . . . . 160
7.10 A Result on the Fourier Transform of a Convolution
of Two Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Contents vii
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Chapter 1
Introduction
The concept of distributions is a relatively recent one since it was used for the first
time in 1951 in a study of the French mathematician L. Schwarz [43]. Distributions
are a generalization of the concept of functions, hence some mathematicians call
them today “generalized functions”. This generalization was motivated by practical
situations. For instance, a boundary value problem from the theory of differential
equations (or even a Cauchy problem) has a solution if the right-hand side of the
problem is at least of class C 1 . But in the phenomena modelled by these problems
these conditions are not satisfied. In most of these phenomena, the functions are not
even continuous, very often having discontinuities of the second kind.
We will see that the same problem, formulated in the context of the theory of
distributions will not impose restrictions on the right-hand side, these being auto-
matically satisfied. We anticipate that every distribution is differentiable (in a sense
that we will discuss and which does not differ too much from the differentiation of
a function) of every order. In the context of the theory of distributions the regularity
conditions are weaker. For example, the functions do not need to be continuous [1,
2, 4].
The theory of distributions, as it is structured today, has the disadvantage that it
employs very sophisticated mathematical concepts, see for example [13, 18, 20, 21,
31, 35, 42]. Therefore, this theory is less accessible to the researchers in mathematics.
Most of the concepts used in the theory of distributions are at the intersection of other
subjects of mathematics.
The fundamentals of the theory of distributions are based on advanced topics in
functional analysis [3, 5, 6, 9, 24, 30, 36, 37, 47, 49], topology, mathematical physics
[7, 10, 11, 14, 16, 17, 19, 22, 23, 25, 32, 34, 38–40, 45, 46, 48] differential geometry
and so on [8, 15, 26–29, 33, 44]. We may consider that the origins of this theory
coincide with Sobolev’s trial to define generalized solutions for hyperbolic partial
differential equations in the 1940s. Sobolev’s idea was employed by L. Schwarz in
the 1960s, when the concept of distributions was used for the first time. In 1971
Schwarz [43] published a book in which he laid the foundations of the unified theory
of distributions by using the method of linear functionals. Afterwards, Sikorski [41]
introduced a theory of distributions by using fundamental sequences of continuous
functions. This method is similar to that used by Dedekind [12] to introduce the real
numbers as cuts in the set of natural numbers.
The way in which Schwarz introduced the concept of distributions seems more
natural and for this reason we use it in the sequel. Since the contributions of Schwarz
intersect with the ones of Sobolev, many concepts and results from the theory of
distributions use the names of these two mathematicians, even though the theory
as it is today used, is almost completely different from that initiated by these two
scientists.
The main qualitative improvement expected from this theory is the simplification
of the mathematical theory in order to make it more accesible. Starting with the
initiators of this theory, the concepts have become more complex.
References
16. L.C. Evans, Partial Differential Equations, 4th edn. (American Mathematical Society, 2008)
17. I. Fonseca, G. Leoni, Modern Methods in the Calculus of Variations: L p Spaces (Springer,
2007)
18. F.G. Friedlander, M. Joshi, Introduction to the Theory of Distributions, 2nd edn. (Cambridge
University Press, 1999)
19. D. Gilbarg, N.S. Trudinger, Elliptic Partial Differential Equations of Second Order, 2nd edn.
(Springer, 2001)
20. D.D. Haroske, H. Triebel, Distributions, Sobolev Spaces, Elliptic Equations (European Math-
ematical Society, 2008)
21. L. Hörmander, The Analysis of Linear Partial Differential Operators I: Distribution Theory
and Fourier Analysis, 2nd edn. (Springer, 1990)
22. T. Kato, Nonstationary flows of viscous and ideal fluids in R3 . J. Funct. Anal. 9, 296–305
(1972)
23. T. Kato, On the eigenfunction of many-particle systems in quantum mechanics. Comm. Pure
Appl. Math. 10, 151–177 (1957)
24. Y. Komura, On linear topological spaces. Kumamoto J. Sci. A 5, 148–157 (1962)
25. L. Lichtenstein, Grundlagen der Hydromechanik (Springer, 1929)
26. M. Marin, H. Arabnia, Equations of Evolution (Elliot & Fitzpatrick, Athens, USA, 2010)
27. M. Marin, A. Öchsner, Complements of Higher Mathematics (Springer, 2018)
28. M. Marin, A. Öchsner, Essentials of Partial Differential Equations (Springer, 2019)
29. S.D. Micu, Introduction to Finite Element Method (Romanian), Publications of the Centre for
Nonlinear Analysis and its Applications
30. G. Minty, Monotone operators in Hilbert space. Duke Math. J. 29, 341–346 (1962)
31. D. Mitrea, Distributions, Partial Differential Equations, and Harmonic Analysis (Springer,
2013)
32. L. Onsager, Statistical hydrodynamics. Nuovo Cimento (9), 6(Supplemento, 2(Convegno Inter-
nazionale di Meccanica Statistica)), 279–287 (1949)
33. A. Öchsner, Computational Statics and Dynamics: An Introduction Based on the Finite Element
Method (Springer, Singapore, 2020)
34. N.S. Papageorgiou, Weak solutions of differential equations in Banach spaces. Bull. Austral.
Math. Soc. 33, 407–418 (1986)
35. I. Richards, H. Youn, Theory of Distributions: A Non-technical Introduction (Cambridge Uni-
versity Press, 1995)
36. R.T. Rockafellar, Convex Analysis, No. 28 (Princeton University Press, Princeton, N.J., 1970)
37. W. Rudin, Functional analysis (Tata-McGraw Hill Publishing, 1974)
38. V. Scheffer, An inviscid flow with compact support in space-time. J. Geom. Anal. 3(4), 343–401
(1993)
39. A.I. Shnirelman, On the nonuniqueness of weak solution of the Euler equation. Comm. Pure
Appl. Math. 50(12), 1261–1286 (1997)
40. R.E. Showalter, Hilbert space methods for partial differential equations. Electron. J. Diff. Eqns.,
Monograph 01 (1994)
41. R. Sikorski, The elementary theory of distributions. Bull. Acad. Pol., Warszawa (1957)
42. R. Strichartz, A Guide to Distribution Theory and Fourier Transforms (CRC Press, 1994)
43. L. Schwartz, Theorie des Distributions (Herman & Cie, Paris, 1971)
44. E. Stein, Singular Integral and Differentiability Properties of Functions (Princeton University
Press, 1970)
45. L. Szekelyhidi, Weak solutions of the Euler equations: non-uniqueness and dissipation.
Journees Equations aux derivees partielles, Roscoff, 1-5 juin 2015
46. L. Szekelyhidi Jr., E. Wiedemann, Young measures generated by ideal incompressible fluid
flows. Arch. Rational Mech. Anal. (2012)
47. H. Triebel, Theory of Function Spaces (Birkhäuser, 1983)
48. E. Wiedemann, Existence of weak solutions for the incompressible Euler equations. Ann. Inst.
H. Poincare Anal. Non Lineaire 28(5), 727–730 (2011)
49. L.C. Young, Lecture on the Calculus of Variations and Optimal Control Theory (American
Mathematical Society, 1980)
Chapter 2
Preliminaries
Abstract This chapter presents some concepts that are used in the next chapters. One
section is dedicated to seminorms and locally convex spaces, as well as to convex,
balanced and absorbing sets. Another section presents duals and reflexive spaces.
Finally, the inductive limit topology is presented.
2.1 Introduction
Let X be a set. We denote by 2 X the set of all subsets of X . This includes the empty
set.
Definition 2.1.1 Let X be a set and T be a subset of 2 X . Then the pair (X, T ) is a
topological space if it satisfies the following properties:
(i) ∅ ∈ T , X ∈ T ;
(ii) if U ∈ T , V ∈ T , then U ∩ V ∈ T ;
(iii) if is an arbitrary set and Uλ ∈ T for all λ ∈ , then ∪λ∈ Uλ ∈ T .
Proposition 2.1.1 Let (X, T ) be a Hausdorff space. Then every sequence has at
most one limit point.
Definition 2.1.3 We call a pair (X,d) a metric space if d : X × X → [0, ∞) has the
three properties below:
(i) d(x, y) = 0 ⇔ x = y;
(ii) d(y, x) = d(x, y), ∀x, y ∈ X ;
(iii) d(x, y) ≤ d(x, z) + d(z, y), ∀x, y, z ∈ X .
Definition 2.1.5 Let (X, d) be a metric space. Then it is called complete if every
Cauchy sequence converges.
Definition 2.1.7 Let (X, · ) be a normed space. Then it is called a Banach space
if it is complete under the metric d : X × X → [0, ∞) defined by
Definition 2.1.8 Let X be a K vector space. We call a map (x, y) → (x, y) X from
X × X to K a sesquilinear form if for all x, y, x1 , y1 , x2 , y2 ∈ X and all α ∈ K, the
following holds true:
(i) (αx, y) X = α(x, y) X , (x, αy) X = ᾱ(x, y) X ;
(ii) (x1 + x2 , y) X = (x1 , y) X + (x2 , y) X , (x, y1 + y2 ) X = (x, y1 ) X + (x, y2 ) X .
Definition 2.1.10 Let X be a K vector space. The pair (X, (·, ·) X ), where (·, ·) X is
a scalar product, is called a pre-Hilbert space.
Definition 2.1.11 A pre-Hilbert space (X, (·, ·) X ) which is complete under the norm
induced by the scalar product is called a Hilbert space.
F(x) = f (x), ∀x ∈ Y,
(2.1.5)
F(x) ≤ p(x), ∀x ∈ X.
x = y on Y,
(2.1.6)
x X = y Y .
Observation 2.2.1 The support of f is the smallest relative closed set outside of
which f is identically zero.
This is a C ∞ function for which the support is the closed unit ball. We can divide
by a constant, namely Rn β(x)d x. Then we obtain another C ∞ function, which we
denote by α. Its support is the closed unit ball and we have
8 2 Preliminaries
α(x)d x = 1. (2.2.3)
Rn
Note that αε (x) ∈ Cc∞ (Rn ). Moreover, its support is Bε (0), the closed ball with
center at the origin and radius ε. We have
αε (x)d x = 1. (2.2.5)
Rn
Note that the family of test functions (αε )ε>0 can be used to regularize discon-
tinuous functions, like integrable ones or L p functions [1, 4, 5]. In other words, we
can show that such functions can be approximated by test functions.
Definition 2.2.2 A function f defined in is locally integrable if f is integrable
(in the sense of Lebesgue [3]) on every compact subset K ⊂ .
We denote by χ K the characteristic function of K , which is equal to 1 on K and
0 outside of K . Hence, f is locally integrable in if, for every compact set K ⊂ ,
the product f · χ K is integrable on .
Definition 2.2.3 We consider that u is a locally integrable function on Rn . Then the
convolution of u and αε is defined by
u ε (x) = u(x − y)αε (y)dy = u(y)αε (x − y)dy. (2.2.6)
Rn Rn
of all closed balls with center x ∈ K and radius ε. If x ∈ Rn and d(x, K ) > ε, then
x∈/ K ε so that αε (x − y) = 0 for all y ∈ K . Hence, the second integral in (2.2.6) is
equal to zero. We deduce that the support of u ε is contained in K ε .
Step 3 Let u be a continuous function. We assume that L is an arbitrary fixed
compact subset of Rn . We evaluate
u ε (x) − u(x) = [u(x − y) − u(x)] αε (y)dy. (2.2.8)
Rn
for all x ∈ L and for all |y| < δ. By considering that ε ≤ δ, we obtain
|u ε (x) − u(x)| ≤ |u(x − y) − u(x)| αε (y)dy < σ, (2.2.10)
Rn
Given σ > 0, we consider a continuous function v with compact support, such that
σ
||u − v|| p < . (2.2.12)
3
It follows from inequalities (2.2.11) and (2.2.12) that
σ
||u ε − vε || p ≤ ||u − v|| p < . (2.2.13)
3
Then we evaluate
Therefore, we have
y
δ −1
|∂ j φ(x)| ≤ ∂jα dy = δ −1 ∂ j α(t)dt ≤ C( j, n)a −1 . (2.2.20)
δn Rn δ Rn
Corollary 2.2.4 Let K ⊂ Rn be compact. Let ( j )1≤ j≤k be a finite open covering
of K . Then there are functions φ j ∈ Cc∞ (Rn ) such that 0 ≤ φ j ≤ 1, 1 ≤ j ≤ k and
k
j=1 φ j = 1 on a neighbourhood of K .
Proof We denote by K̊ j the interior of K j . Then we can find compact sets (K j )1≤ j≤k
such that K j ⊂ j and
k
K ⊂ K̊ j . (2.2.21)
j=1
φ1 = ψ1 , φ j = ψ j (1 − ψ1 ) · · · (1 − ψ j−1 ), j = 2, . . . , k. (2.2.22)
Note that the functions (φ j )1≤ j≤k are called a C ∞ partition of unity subordinated
to the covering ( j )1≤ j≤k of K .
Let K be a field, for example, the field of real numbers R or the field of complex
numbers C. Let E be a vector space over K . We consider that R+ is the set of all
nonnegative real numbers.
Moreover,
x∞ = sup |xi | (2.3.2)
1≤i≤n
defines a norm on Rn .
Example 2 Let 1 < p < ∞. We consider that L p (I ) is the space of functions f
defined on I and such that | f | p is integrable on I in the Lebesgue sense [3]. We
define the following seminorm on L p (I )
b
1
p
f p = | f (t)| p dt . (2.3.3)
a
(x, y) ∈ E × E → x + y ∈ E
(2.3.4)
(λ, x) ∈ K × E → λx ∈ E
Moreover, a closed ball with center x0 ∈ E and radius r > 0 is the set
x → a + x,
(2.3.7)
x → λx
are called translation and homothety and are homeomorphisms on E. Hence, it suf-
fices to have a fundamental system of neighbourhoods of the origin in order to have
a fundamental system of neighbourhoods of a point a ∈ E.
Let ( pi )i∈I be a family of seminorms defined on E. We define for every x0 ∈ E,
ε ∈ R, ε > 0 and F a finite part of I
The set V (x0 , ε, F) is the intersection of the balls with center x0 and radius ε,
corresponding to the seminorms pi with i ∈ F.
The family of all sets V (x0 , ε, F) when ε runs through the set of all positive real
numbers and F runs through all finite subsets of I defines a fundamental system of
neighbourhoods of x0 of a topology on E compatible with the vector space structure
of E. The space E endowed with this topology is called a locally convex topological
vector space.
If E is a locally convex topological vector space and ( pi )i∈I is a family of semi-
norms defining its topology, then E is a Hausdorff space [3] if and only if to any pair
x, y ∈ E with x = y we associate a seminorm pk such that pk (x) = pk (y).
Note that, given a set A, its convex hull is always well defined. Clearly, it suffices
to consider the intersection of the family of all convex sets which contain A. Note
that this family is nonempty because it contains the whole space E.
The convex hull of A will be denoted by (A). This can be characterized as the set
of all x ∈ E which can be represented as a finite sum x = i∈F αi xi with xi ∈ A,
αi ∈ R+ , i∈F αi = 1, where F is a finite set of indices which depend on x.
Definition 2.3.5 Let A be a subset of E. The smallest balanced convex subset b (A)
of E containing A is called the balanced convex hull of A.
Note that the balanced convex hull of A is the intersection of all balanced convex
sets containing A. Moreover, we have the following characterization
x ∈ b (A) ⇔ x = αi xi , (2.3.9)
i∈F
Hence, we obtain
q(x + y) ≤ q(x) + q(y). (2.3.13)
Theorem 2.3.1 Let E be a topological vector space. The following conditions are
equivalent:
(i) E is locally convex;
(ii) there is a fundamental system of convex neighbourhoods of the origin;
(iii) there is a fundamental system of absorbing balanced convex neighbourhoods
of the origin.
Proof Step 1. We show that (i) implies (ii). If the topology of E is defined by a
family of seminorms ( pi )i∈I , then the sets
with F a finite part of I and 0 < ε < 1, form a fundamental system of convex
neighbourhoods of the origin.
Step 2. We show that (iii) implies (i). Let V be an absorbing balanced convex
subset of E which belongs to a fundamental system of neighbourhoods of the origin.
For every such V we have a corresponding seminorm pV . Hence, the family of
seminorms obtained in this way defines the topology of E.
Step 3. We show that (ii) implies (iii). Let V be a convex neighbourhood of the
origin. It suffices to prove that the set
U = ∩|λ|=1 λV (2.3.17)
is an absorbing balanced convex neighbourhood of the origin. Note that the map
(μ, x) → μx is continuous at the point (0, 0). Hence, there are ε > 0 and V a
neighbourhood of zero in E such that
16 2 Preliminaries
μx ∈ V, ∀ |μ| ≤ ε, ∀x ∈ V . (2.3.18)
μW ⊂ V, ∀ |μ| ≤ 1. (2.3.19)
Hence, we have
μW ⊂ V, ∀ |μ| = 1. (2.3.20)
p
The elements of L loc () are called pth power locally integrable functions. We con-
sider the seminorm 1
p
pk ( f ) = | f (x)| p d x . (2.3.24)
K
The family of seminorms ( pk ), where k runs over all compact subsets of , defines a
p
locally convex topology on L loc (). Such a topology can be defined by a countable
p
family of seminorms. Then we can prove that L loc is a Frechet space.
2.4 Duals
Let E be a vector space over a field K . Its algebraic dual E ∗ is the vector space of all
linear maps x ∗ : E ∗ → K . We denote by x ∗ (x) or (x, x ∗ ) the value of x ∗ on x ∈ E.
Let E be a topological vector space over a field K . Then the dual E of E is the
subspace of E ∗ consisting of all continuous linear maps (or functionals) on E.
For every x ∗ ∈ E ∗ we set
px ∗ (x) = x, x ∗ . (2.4.1)
Therefore, A is bounded.
Definition 2.4.2 Let B be a subset of a topological vector space E. The polar set
B ◦ of B is the subset of E defined as
B ◦ = {x ∈ E | x, x ≤ 1, ∀x ∈ B}. (2.4.3)
Therefore, A◦ is convex.
Let x ∈ A◦ and λ ∈ K , |λ| ≤ 1. Then
x, λx = |λ| x, x ≤ 1. (2.4.5)
Let B denote the family of all bounded subsets of E. Then the family of semi-
norms ( p A◦ ) A∈B defines a Hausdorff locally convex topology on E , called the strong
topology of E . Note that a sequence (x j ) converges strongly to zero in E if and
only if the sequence x j (x) converges uniformly to zero on every bounded set of
E. Hence, the strong topology on E is also called the topology of uniform conver-
gence on bounded sets of E. We denote by E b the dual E endowed with the strong
topology.
Example Let E be a normed space. Let E be its dual endowed with the norm
x E = sup x, x . (2.4.9)
x E ≤1
Then E is a Banach space [3]. Therefore, the strong topology on E coincides with
the one defined by the above norm.
Let E be the dual of the Banach space E . We consider the following norm
x E = sup x , x (2.4.10)
x E ≤1
Then we consider on E the finest locally convex topology for which the identity map
E i → E is continuous for every i = 1, 2, . . .. This topology represents the inductive
limit topology of E defined by the subspaces E i . Moreover, the space E endowed
with this topology is called the inductive limit of the spaces (E i )i∈N .
Let V be a convex set. We want V to be a neighbourhood of zero in the inductive
limit topology. Then it is necessary and sufficient to consider that every intersec-
tion V ∩ E i is a neighbourhood of zero in E i for all i = 1, 2, . . .. Moreover, by
considering all convex hulls ∞
V = ∪i=1 Vi (2.5.2)
Proposition 2.5.1 We consider that E is the inductive limit of (E i )i∈N and that F is
any locally convex space. Then a linear map u : E → F is continuous if and only if
the restriction u i of u is continuous from E i into F for all i.
Lemma 2.5.1 We assume that F is a locally convex space and G is a closed subspace
of F. Let V be a balanced convex open neighbourhood of zero in G. We consider
x ∈ F such that x ∈
/ G. Therefore, there is a balanced convex open neighbourhood
of zero W in F such that x ∈
/ W and W ∩ G = V .
Proof Note that G is closed in F. Then there is a balanced convex open neighbour-
hood of zero V0 in F such that
(x + V0 ) ∩ G = ∅ and V0 ∩ G ⊂ V. (2.5.4)
2.5 The Inductive Limit Topology 21
which is impossible.
Proof Step 1. We will prove that the topology induced by E on every E i coincides
with the given topology of E i . Let Vi be a balanced convex neighbourhood of zero
in E i . It suffices to prove that there is a neighbourhood of zero V in E such that
Vi = V ∩ E i , ∀i. (2.5.7)
We consider
V = ∪∞
k=0 Vi+k . (2.5.9)
xn ∈ A ∩ E in and xn ∈
/ E in−1 . (2.5.10)
22 2 Preliminaries
Note that there is a sequence (Vn ) of balanced convex open neighbourhoods of zero
in E in such that
xn ∈
/ nVn and Vn ∩ E in−1 = Vn−1 . (2.5.11)
We consider V = ∪∞
n=1 Vn . Hence, V is a neighbourhood of zero in E such that
V ∩ E in = Vn and xn ∈
/ nV. (2.5.12)
We will denote by M() the space of all Radon measures on . This is the
topological dual of Cc (). Note that a linear map
μ : Cc () → C (2.5.16)
ust within the Arctic region, but nearly on the limits of what
geographers call the Atlantic Ocean, lies an island which,
since its colonization in the ninth century, has not ceased to
excite the interest of the explorer and the man of science.
Iceland—which measures about 300 miles at its greatest length,
from east to west, and about 200 miles at its greatest breadth, from
north to south—is situated in lat. 63° 23’-66° 33’ N., and long. 13°
22’-24° 35’ W.; at a distance of 600 miles from the nearest point of
Norway, 250 from the Faröe Isles, 250 from Greenland, and above
500 miles from the northern extremity of Scotland. As early as the
eighth Christian century it was discovered by some European
emigrants; though, indeed, the Landnana Book, one of the earliest of
the island-records, asserts that they found the memorials of a yet
earlier settlement in various Christian relics, such as wooden
crosses, which appeared to be of Irish origin. At all events, the first
really successful attempt at colonization was made by Ingolf, a
Norwegian, who planted himself and his followers at Reikiavik in
874. In the following century a somewhat extensive immigration took
place of Norwegians who resented the changes of polity introduced
by Harold Haarfager, and all the habitable points on the coast were
occupied by about 950 a.d. Fifty years afterwards, though not
without much opposition, Christianity was legally established, and
the bishoprics of Holar and Skalholt were founded. The government
assumed the character of an aristocratic republic, with a popular
assembly, called the Althing, meeting every summer in the valley of
Thingvalla. Commerce was encouraged, and the Icelanders early
distinguished themselves by the boldness of their maritime
enterprise, and the extent of their ocean fisheries.
About the year 932 they discovered Greenland, and about 986 a
portion of the North American coast, which they called “Vineland.”
They did not confine their voyages to the north, but sent their ships
even as far south as the Mediterranean. From 1150 to 1250 is rightly
considered the most flourishing period of Icelandic literature and
commerce. After the conquest of the island by Haco VI. of Norway,
much of the old spirit seemed to die out. When Norway was united to
Denmark in 1380, Iceland was included in the bond, and it is still
regarded as a dependency of the latter kingdom. In 1540 it
embraced the principles of Lutheran Protestantism. Its population at
one time numbered 100,000, but it gradually diminished until, in
1840, it was reduced to 57,094; but a slow increase has taken place
of late years, and it now amounts to about 70,000. The language
spoken is the old Norse.
Iceland is a fifth part larger than Ireland, and its superficial area is
estimated at 39,207 square miles. Not more than 4000 miles,
however, are habitable, all the rest being ice and lava; for the island
seems to be little more than a mass of trachyte, snow-shrouded and
frost-bound, resting on a sea of fire. It consists of two vast parallel
table-lands, the foundations of ranges of lofty mountains, most of
which are active volcanoes; and these table-lands strike across the
centre of the island, from north-east to south-west, at a distance from
one another of ninety to one hundred miles. Their mountainous
summits are not pyramidal, as is generally the case in Europe, but
rounded like domes, as in the Andes of South America. Their sides,
however, are broken up by precipitous masses of tufa and
conglomerate, intersected by deep ravines of the gloomiest
character. They are covered with a thick shroud of ice and snow, but
in their wombs seethe the fiery elements which ever and anon break
forth into terrible activity. The eastern table-land and its mountain
range is the most extensive, and contains Oërafa, the culminating
point of Iceland. It is visible from a great distance at sea, like a white
cloud suspended above the island. Its height is 6426 feet, and it
springs from a vast mountain-mass; no fewer than 3000 square
miles being perpetually burdened with ice and snow, at an altitude
varying from 3000 to 6000 feet.
AN ICELANDIC LANDSCAPE.
A very considerable portion of the island is occupied by the large
glaciers which descend from the mountains, like frozen torrents,
pushing forward into the lowlands, and even to the margin of the
sea. These act as almost impassable barriers to communication
between the various inhabited districts.
We have spoken of the two ranges of table-lands as about ninety
to one hundred miles apart. The interspace forms a low broad valley,
which opens at either extremity on the sea—an awful waste, a region
of desolation, where man is utterly powerless; where the elements of
fire and frost maintain a perpetual antagonism; where blade of grass
is never seen, nor drop of water; where bird never wings its way, and
no sign of life can be detected. It seems a realization of Dante’s
“circle of ice” in the “Inferno.” The surface consists of lava streams,
fissured by innumerable crevices; of rocks piled on rocks; of dreary
glaciers, relieved by low volcanic cones. It is supposed that some
remote portions of the inaccessible interior are less barren, because
herds of reindeer have been seen feeding on the Iceland moss that
fringes the borders of this dreary region. But there is no reason to
believe that it can ever be inhabited by man.
HARBOUR OF REIKIAVIK.
The Icelander is honest, temperate, hospitable, possessed with a
fervent spirit of patriotism, and strongly wedded to the ancient
usages. He is also industrious; and though his industry is but scantily
remunerated, he earns enough to satisfy his simple tastes. In the
interior his chief dependence is on his cattle; and as grass is the
main produce of his farm, his anxiety during the haymaking season
is extreme. A bad crop would be almost ruin. He is, however, wofully
ignorant of agriculture as a science, does but little for the
improvement of the soil, and employs implements of the most
primitive character. The process of haymaking in Iceland is thus
described:—
The best crops are gathered from the “tún,” a kind of home-park
or paddock, comprising the lands contiguous to the farmstead—the
only portion of his demesne to which the owner gives any special
attention, and on the improvement of which he bestows any labour.
This “tún” is enclosed within a wall of stone or turf, and averages an
extent of two or three acres, though sometimes it reaches to ten. Its
surface is usually a series of closely-packed mounds, like an
overcrowded graveyard, with channels or water-runs between, about
two feet deep. Hither every person employed on the farm, or whom
the farmer can engage, resorts, with short-bladed scythe and rake,
and proceeds to cut down the coarse thick grass, and rake it up into
little heaps.
Afterwards the mowers hasten to clear the neighbouring hill sides
and undrained marshes.
This primitive haymaking, so unlike the systematic operation
which bears that name in England, is carried on throughout the
twenty-four hours of the long summer day. The hay, when sufficiently
dried, is made up into bundles, and tied with cords and thongs, and
packed on the back of ponies, which carry it to the clay-built stalls or
sheds prepared for it. And a curious sight it is to see a long string of
hay-laden ponies returning home. Each pony’s halter is made fast to
the tail of his predecessor; and the little animals are so
overshadowed and overwhelmed by their burdens, that their hoofs
and the connecting ropes alone are visible, and they seem like so
many animated haycocks, feeling themselves sufficiently made up,
and leisurely betaking themselves to their resting-places.
During the protracted winter the Icelander, of course, can attend
to no out-of-door labour, and passes his time within his hut, which, in
many parts of the island, is not much superior to an Irish “cabin.”
The lower part is built of rough stones up to a height of four feet,
and between each course a layer of turf is placed, which serves
instead of mortar, and helps to keep out the cold. The roof, made of
any available wood, is covered with turf and sods. On the southern
side the building is ornamented with doors and gable-ends, each of
which is crowned by a weathercock. These doors are the entrances
to the dwelling-rooms and various offices, such as the cow-shed,
store-house, and smithy. The dwelling-rooms are connected by a
long, dark, narrow passage, and are separated from each other by
strong walls of turf. As each apartment has its own roof, the building
is, in effect, an aggregate of several low huts, which receive their
light through small windows in the front, or holes in the roof, covered
with a piece of glass or skin. The floors are of stamped earth; the
fireplace is made of a few stones, rudely packed together, while the
smoke escapes through a hole in the roof, or through a cask or
barrel, with the ends knocked out, which acts as chimney.
In some parts of the island lava is used instead of stones, and
instead of wood the rafters are made of the ribs of whales. A horse’s
skull is the best seat provided for a visitor. Too often the same room
serves as the dining, sitting, and sleeping place for the whole family,
and the beds are merely boxes filled with feathers or sea-weed.
There are, however, a few houses of a superior character, in which
the arrangements are not much unlike those of a good old-fashioned
English farm-house; the walls being wainscotted with deal, and the
doors and staircase of the same material. A few prints and
photographs, some book-shelves, one or two little pictures, decorate
the sitting-room, and a neat iron stove, and massive chests of
drawers, furnish it sufficiently.
From the houses we turn to the churches. In Reikiavik the church
is a stone building, the only stone building in the town; but this is
exceptional: most of the churches are not much better than the
houses. We will be content, therefore, with a visit to the Reikiavik
sanctuary, which is a neat and unpretending erection, capable of
accommodating three or four hundred persons. The Icelanders are
not opposed to a “decent ritual,” and the Lutheran minister wears a
black gown with a ruff round his neck. The majority of the
congregation, here as everywhere else, consists of women; some
few dressed in bonnets, and the rest wearing the national black silk
skull-cap, set jauntily on one side of the head, with a long black
tassel drooping to the shoulder, or else a quaint mitre-like structure
of white linen, almost as imposing as the head-dress of a Normandy
peasant. The remainder of an Icelandic lady’s costume, we may add,
consists of a black bodice, fastened in front with silver clasps, over
which is drawn a cloth jacket, gay with innumerable silver buttons;
round the neck goes a stiff ruff of velvet, embroidered with silver
lace; and a silver belt, often beautifully chased, binds the long dark
wadmal petticoat round the waist. Sometimes the ornaments are of
gold, instead of silver, and very costly.
Towards the end of the Lutheran service, the preacher descends
from the pulpit, and attiring himself in a splendid crimson velvet
cope, turns his back to the congregation, and chants some Latin
sentences.
Though still retaining in their ceremonies, says a recent traveller,
a few vestiges of the old religion, though altars, candles, pictures,
and crucifixes yet remain in many of their churches, the Icelanders
are stanch Protestants, and singularly devout, innocent, and pure-
hearted. Crime, theft, debauchery, cruelty are unknown amongst
them; they have neither prison, gallows, soldiers, nor police; and in
the manner of their lives mingles something of a patriarchal
simplicity, that reminds one of the Old World princes, of whom it has
been said, that they were “upright and perfect, eschewing evil, and in
their hearts no guile.”
In the rural districts, if such a phrase can properly be applied to
any part of Iceland, the church is scarcely distinguishable from any
other building, except by the cross planted on its roof. It measures,
generally, from eight to ten feet in width, and from eighteen to
twenty-four feet in length; but of this space about eight feet are
devoted to the altar, which is divided off by a wooden partition
stretching across the church, just behind the pulpit. The communion-
table is nothing more than a small wooden chest or cupboard, placed
at the end of the building, between two small square windows, each
formed by a single common-sized pane of glass. Over the table is
suspended a sorry daub, on wood, intended to represent the Last
Supper. The walls, which are wainscotted, are about six feet high;
and stout beams of wood stretch from side to side. On these are
carelessly scattered a number of old Bibles, psalters, and loose
leaves of soiled and antiquated manuscripts. The interior of the roof,
the rafters of which rest on the walls, is also lined with wood.
Accommodation, in the shape of a few rough benches, is provided
for a congregation of thirty or forty souls.
Poor as are the churches, the pastors are still poorer. The best
benefice in the island is worth not much more than £40 per annum;
the average value is £10. The bishop himself does not receive more
than £200. The principal support of the clergy, therefore, is derived
from their glebe-land, eked out by the small fees charged for
baptisms, marriages, and funerals.
Such being the case, the reader will not be surprised to learn that
the Icelandic clergy live miserably and work hard. They assist in the
haymaking; they hire themselves out as herdsmen; they act as the
leaders of the caravans of pack-horses which carry the produce of
the island to the ports, and return loaded with domestic necessaries;
and they distinguish themselves as blacksmiths, as veterinarians,
and shoers of horses.
Dr. Henderson gives an interesting and graphic account of a visit
he paid to one of these “poor parsons,” John Thorlukson, who, while
supporting himself by drudgery of this painful kind, translated
Milton’s “Paradise Lost” and Pope’s “Essay on Man” into Icelandic.
“Like most of his brethren, at this season of the year,” says Dr.
Henderson, “we found him in the meadow assisting his people at
haymaking. On hearing of our arrival he made all the haste home
which his age and infirmity would allow, and bidding us welcome to
his lowly abode, ushered us into the humble apartment where he
translated my countrymen into Icelandic. The door was not quite four
feet in height, and the room might be about eight feet in length by six
in breadth. At the inner end was the poet’s bed; and close to the
door, over against a small window, not exceeding two feet square,
was a table where he committed to paper the effusions of his muse.
On my telling him that my countrymen would not have forgiven me,
nor could I have forgiven myself, had I passed through this part of
the island without paying him a visit, he replied that the translation of
Milton had yielded him many a pleasant hour, and often given him
occasion to think of England.”
It is true that this passage was written some fifty-five years ago,
but the condition of the clergy of Iceland has not much improved in
the interval.