Full Chapter Geometry and Analysis of Metric Spaces Via Weighted Partitions Jun Kigami PDF

Geometry and Analysis of Metric
Spaces via Weighted Partitions Jun

Kigami
Visit to download the full and correct content document:
https://textbookfull.com/product/geometry-and-analysis-of-metric-spaces-via-weighted
-partitions-jun-kigami/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Functional Analysis: An Introduction to Metric Spaces,

Hilbert Spaces, and Banach Algebras: Second Edition
Joseph Muscat
https://textbookfull.com/product/functional-analysis-an-
introduction-to-metric-spaces-hilbert-spaces-and-banach-algebras-
second-edition-joseph-muscat/
A Primer on Hilbert Space: Theory Linear Spaces,

Topological Spaces, Metric Spaces, Normed Spaces, and
Topological Groups 2nd Edition Carlo Alabiso
https://textbookfull.com/product/a-primer-on-hilbert-space-
theory-linear-spaces-topological-spaces-metric-spaces-normed-
spaces-and-topological-groups-2nd-edition-carlo-alabiso/
A Primer on Hilbert Space Theory Linear Spaces

Topological Spaces Metric Spaces Normed Spaces and
Topological Groups 2nd Edition Carlo Alabiso Ittay
Weiss
https://textbookfull.com/product/a-primer-on-hilbert-space-
theory-linear-spaces-topological-spaces-metric-spaces-normed-
spaces-and-topological-groups-2nd-edition-carlo-alabiso-ittay-
weiss/
Geometry and Martingales in Banach Spaces 1st Edition

Wojbor A Woyczynski
https://textbookfull.com/product/geometry-and-martingales-in-
banach-spaces-1st-edition-wojbor-a-woyczynski/
Fundamentals of Signal Processing in Metric Spaces with
Lattice Properties: Algebraic Approach 1st Edition
Andrey Popoff
https://textbookfull.com/product/fundamentals-of-signal-
processing-in-metric-spaces-with-lattice-properties-algebraic-
approach-1st-edition-andrey-popoff/
An Invitation to Alexandrov Geometry CAT 0 Spaces

Stephanie Alexander
https://textbookfull.com/product/an-invitation-to-alexandrov-
geometry-cat-0-spaces-stephanie-alexander/
Spaces An Introduction to Real Analysis Tom Lindstrøm
https://textbookfull.com/product/spaces-an-introduction-to-real-
analysis-tom-lindstrom/
Riemannian Geometry and Geometric Analysis 7th Edition

Jürgen Jost (Auth.)
https://textbookfull.com/product/riemannian-geometry-and-
geometric-analysis-7th-edition-jurgen-jost-auth/
Analysis on Fock spaces and mathematical theory of

quantum fields 1st Edition Asao Arai
https://textbookfull.com/product/analysis-on-fock-spaces-and-
mathematical-theory-of-quantum-fields-1st-edition-asao-arai/
Lecture Notes in Mathematics 2265
Jun Kigami
Geometry
and Analysis
of Metric Spaces
via Weighted
Partitions
Lecture Notes in Mathematics
Volume 2265
Editors-in-Chief
Jean-Michel Morel, CMLA, ENS, Cachan, France
Bernard Teissier, IMJ-PRG, Paris, France
Series Editors
Karin Baur, University of Leeds, Leeds, UK
Michel Brion, UGA, Grenoble, France
Camillo De Lellis, IAS, Princeton, NJ, USA
Alessio Figalli, ETH Zurich, Zurich, Switzerland
Annette Huber, Albert Ludwig University, Freiburg, Germany
Davar Khoshnevisan, The University of Utah, Salt Lake City, UT, USA
Ioannis Kontoyiannis, University of Cambridge, Cambridge, UK
Angela Kunoth, University of Cologne, Cologne, Germany
Ariane Mézard, IMJ-PRG, Paris, France
Mark Podolskij, University of Luxembourg, Esch-sur-Alzette, Luxembourg
Sylvia Serfaty, NYU Courant, New York, NY, USA
Gabriele Vezzosi, UniFI, Florence, Italy
Anna Wienhard, Ruprecht Karl University, Heidelberg, Germany
This series reports on new developments in all areas of mathematics and their
applications - quickly, informally and at a high level. Mathematical texts analysing
new developments in modelling and numerical simulation are welcome. The type of
material considered for publication includes:
1. Research monographs
2. Lectures on a new field or presentations of a new angle in a classical field
3. Summer schools and intensive courses on topics of current research.
Texts which are out of print but still in demand may also be considered if they fall
within these categories. The timeliness of a manuscript is sometimes more important
than its form, which may be preliminary or tentative.
More information about this series at http://www.springer.com/series/304

Jun Kigami
Geometry and Analysis

of Metric Spaces via
Weighted Partitions
Jun Kigami
Graduate School of Informatics
Kyoto University
Kyoto, Japan
ISSN 0075-8434 ISSN 1617-9692 (electronic)

Lecture Notes in Mathematics
ISBN 978-3-030-54153-8 ISBN 978-3-030-54154-5 (eBook)
https://doi.org/10.1007/978-3-030-54154-5
Mathematics Subject Classification: 30F45, 53C23
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland
AG 2020
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This monograph reflects what I have been pursuing for the last 10 years or more.
The main questions are what is the natural way to comprehend the “structure” of
a space? and how can one develop “analysis” on a space from its “structure”?
These are, of course, vague questions since the meanings of the words “structure”
and “analysis” are not quite clear. The former can be replaced by “geometry” but
this does not make things better. One of the more precise formulations is, for a given
space, what is a natural counterpart of the Euclidean metric on Rn ? Needless to say,
rich analysis on Rn has been developed under the Euclidean metric. A prototype is
a class of self-similar sets, including the Sierpinski gasket and carpet, where there is
a natural notion of partitions associated with trees and on which Brownian motion
has been constructed and studied. This monograph is just a tiny step toward this vast
frontier of geometry and analysis of metric spaces but, hopefully, it is a right one at
least.
The origin of the story is the notion of successive divisions of compact metric
spaces, which appear in many different areas of mathematics such as the construc-
tion of self-similar sets, Markov partitions associated with hyperbolic dynamical
systems, and dyadic cubes associated with a doubling metric space. The common
feature in these is to divide a space into a finite number of subsets, then divide
each subset into finitely many pieces, repeating this process again and again. In
this monograph, we generalize such successive divisions and call them partitions.
Given a partition, we consider the notion of a weight function assigning a “size”
to each piece of the partition. Intuitively, we believe that a partition and a weight
function should provide a “geometry” and an “analysis” on the space of interest. We
pursue this idea in Chaps. 2–4. In Chap. 2, the metrizability of a weight function,
i.e. the existence of a metric “adapted to” a given weight function, is shown to be
equivalent to the Gromov hyperbolicity of the graph associated with the weight
function. In Chap. 3, notions such as bi-Lipschitz equivalence, Ahlfors regularity,
the volume doubling property, and quasisymmetry are shown to be equivalent to
certain properties of weight functions. In particular, we find that quasisymmetry
and the volume doubling property are the same notion in the world of weight
functions. In Chap. 4, a characterization of the Ahlfors regular conformal dimension
v
vi Preface
of a compact metric space is given as the critical index p of p-energies associated

with the partition and the weight function corresponding to the metric.
I express my gratitude to the many colleagues who have given me insights into
these subjects. In particular, I would like to thank Professors M. Bonk and B. Kleiner
for their suggestions on the directions in which to proceed, the anonymous referees
for their valuable comments on the original manuscript, and the Springer editorial
and production teams for their sincere efforts to make this monograph much better
than its original form.
Last but not least, in this crazy moment of our history, I sincerely wish good
health and safety to all, and that I will soon be able to share thoughts face to face
with my friends around the world.
Kyoto, Japan Jun Kigami

August 23, 2020
Contents
1 Introduction and a Showcase . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1
1.2 Summary of the Main Results; the Case of 2-dim. Sphere.. . . . . . . . . . . 8
2 Partitions, Weight Functions and Their Hyperbolicity . . . . . . . . . . . . . . . . . . 17
2.1 Tree with a Reference Point . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 17
2.2 Partition .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 20
2.3 Weight Function and Associated “Visual Pre-metric”.. . . . . . . . . . . . . . . . 28
2.4 Metrics Adapted to Weight Function.. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 33
2.5 Hyperbolicity of Resolutions and the Existence
of Adapted Metrics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 41
3 Relations of Weight Functions . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
3.1 Bi-Lipschitz Equivalence . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
3.2 Thickness of Weight Functions .. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 63
3.3 Volume Doubling Property . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 66
3.4 Example: Subsets of the Square .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 75
3.5 Gentleness and Exponentiality . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 85
3.6 Quasisymmetry.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 90
4 Characterization of Ahlfors Regular Conformal Dimension. . . . . . . . . . . . 97
4.1 Construction of Adapted Metric I . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 97
4.2 Construction of Ahlfors Regular Metric I. . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 105
4.3 Basic Framework .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 108
4.4 Construction of Adapted Metric II . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 112
4.5 Construction of Ahlfors Regular Metric II .. . . . . . . .. . . . . . . . . . . . . . . . . . . . 116
4.6 Critical Index of p-Energies and the Ahlfors Regular
Conformal Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 123
4.7 Relation with p-Spectral Dimensions .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 135
4.8 Combinatorial Modulus of Curves . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 142
4.9 Positivity at the Critical Value .. . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 146
vii
viii Contents
A Fact from Measure Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 153

B List of Definitions, Notations and Conditions . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 155
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 161
Chapter 1
Introduction and a Showcase
1.1 Introduction
Successive divisions of a space have played important roles in many areas of

mathematics. One of the simplest examples is the binary division of the unit interval
[0, 1] shown in Fig. 1.1. Let Kφ = [0, 1] and divide Kφ in half as K0 = [0, 12 ]
and K1 = [ 12 , 1]. Next, K0 and K1 are divided in half again and yield Kij for each
(i, j ) ∈ {0, 1}2 . Repeating this procedure, we obtain {Ki1 ...im }i1 ,...,im ∈{0,1} satisfying
Ki1 ...im = Ki1 ...im 0 ∪ Ki1 ...im 1 (1.1.1)
for any m ≥ 0 and i1 . . . im ∈ {0, 1}m . In this example, there are two notable
properties.
The first one is the role of the (infinite) binary tree

Tb = {φ, 0, 1, 00, 01, 10, 11, 000, 001, 010, 011, . . .} = {0, 1}m ,
m≥0
where {0, 1}0 = {φ}. The vertex φ is called the root or the reference point and Tb is
called the tree with the root (or the reference point) φ. Note that the correspondence
i1 . . . im → Ki1 ...im determines a map from the binary tree to the collection of
compact subsets of [0, 1] with the property (1.1.1).
Secondly, note that Ki1 ⊇ Ki1 i2 ⊇ Ki1 i2 i3 ⊇ . . . and

Ki1 ...im is a single point (1.1.2)
m≥1
© The Editor(s) (if applicable) and The Author(s), under exclusive license 1
to Springer Nature Switzerland AG 2020
J. Kigami, Geometry and Analysis of Metric Spaces via Weighted Partitions,
Lecture Notes in Mathematics 2265, https://doi.org/10.1007/978-3-030-54154-5_1
2 1 Introduction and a Showcase
0 Kφ 1 φ
1
K0 2 K1
0 1
1 3
K00 4 K01 K10 4 K11
00 01 10 11
1 3 5 7
8 8 8 8
A partition of [0, 1] The associated tree
Fig. 1.1 A partition of the unit interval [0, 1] and the associated tree
for any infinite sequence

i1 i2 . . .. (Of course, this is the binary expansion and hence
the single point is m≥1 2imm .) In other words, there is a natural map σ : {0, 1}N →
[0, 1] given by

σ (i1 i2 . . .) = Ki1 ...im .
m≥1
Such successive divisions of a compact metric space, which may not be as

simple as the above example, appear in various areas in mathematics. One of typical
examples is a self-similar set in fractal geometry. A self-similar set is a union of
finite contracted copies of itself. Then each contracted copy is again a union of
contracted copies and so forth. Another example is the Markov partition associated
with a hyperbolic dynamical system. See [1] for details. Also, the division of a
metric measure space having the volume doubling property by dyadic cubes can
be thought of as another example of such divisions of a space. See Christ[14] for
example.
In general, let X be a compact metrizable topological space with no isolated
point. The common properties of the above examples are;
(i) There exists a tree T (i.e. a connected graph without loops) with a root φ.
(ii) For any vertex p of T , there is a corresponding nonempty compact subset of
X denoted by Xp and X = Xφ .
(iii) Every vertex p of T except φ has unique predecessor π(p) ∈ T and

Xq = Xp . (1.1.3)
p∈{p |π(p )=q}
(iv) The totality of edges of T is {(π(q), q)|q ∈ T \{φ}}.

1.1 Introduction 3
(v) For any infinite sequence (p0 , p1 , p2 , . . .) of vertices of X satisfying p0 = φ

and π(pi+1 ) = pi for any i ≥ 1,

Xpi is a single point. (1.1.4)
i≥1
See Fig. 1.2 for an illustration of the idea. Note that the properties (1.1.3)
and (1.1.4) correspond to (1.1.1) and (1.1.2) respectively. In this monograph such
{Xp }p∈T is called a partition of X parametrized by the tree T . (We will give the
precise definition in Sect. 2.2.) In addition to the “vertical” edges, which are the
edges of the tree, we provide “horizontal” edges to T to describe the combinatorial
structure reflecting the topology of X as is seen in Fig. 1.2. More precisely, a pair
(p, q) ∈ T × T is a horizontal edge if p and q have the same distance from the root
φ and Xp ∩ Xq = ∅. We call T with horizontal and vertical edges the resolution of
X associated with the partition.
Another key notion is a weight function on the tree T . Note that a metric and
a measure give weights of the subsets of X. More precisely, let d be a metric on
X inducing the original topology of X and let μ be a Radon measure on X where
μ(Xp ) > 0 for any p ∈ T . Define ρd : T → (0, 1] and ρμ : T → (0, 1] by
diam(Xp , d) μ(Xp )
ρd (p) = and ρμ (p) = ,
diam(X, d) μ(X)
.................
...........
..
...............................................................................................................................
.... ............
.........
φ
......... ... .......
.....
. ................................... ...
... .......
.....
. ........ ... .. ...........
.
.. ... ..
. ..
..... ......
..
...
.. ..
. .....
....
.. ... ....
.... ....
.....
....
....
X11 .. ......
.... ......
.....
X21 X22 ..
.
.
.
..
..
....
.....
....
....
....
....
...
... . .. ... .
. ... ...
.... ...... . .
.
.
. .
. ..
..... ... . ... ...
...
...
.... ... .
....... ...
. ...
....
.
.
.
.
.
. ....... ... . ..
.. .
.. ... ...
..
2
........................................... ..
X32
... .. . ...
.. . ........
. .
. . . . ..
..
..
...
...
X12 ....
..
.
..
.........................
...
......
......... ...
...
...
. .
..
.
..
..
. 1 3
... ........................................... .... ...
...
... ........ ......... ...
.... ....... .. ... ...
.... ....
X31
.. ...
....
.... . .....
. ..
. .. ......
X13
. ... ... ...
..... .... ... ... ..... 21 22
......
...... .... ... ..........
....... ..... ...
.. ... ......... 11
......... .... .....
......... . ........
........... .. ...........
.................
.................................. .. .................
......................................................................
12 32
X = Xφ
13 31
A partition of X The associated graphs T
Solid lines are vertical edges.
Dotted lines are horizontal edges.
Fig. 1.2 A partition and the associated graphs (up to the 2nd stage)
where diam(A, d) is the diameter of A with respect to the metric d. Then ρd (resp.
ρμ ) is though of as a natural weight of Xp associated with d (resp. μ). In both cases
where = d or = μ, the function ρ : T → (0, ∞) satisfies
ρ (π(p)) ≥ ρ (p) (1.1.5)
for any p ∈ T \{φ} and
lim ρ (pi ) = 0 (1.1.6)

i→∞
if π(pi+1 ) = pi for any i ≥ 1. (To have the second property (1.1.6) in case of
= μ, we must assume that the measure μ is non-atomic, i.e. μ({x}) = 0 for any
x ∈ X.)
As we have seen above, given a metric or a measure, we have obtained a weight
function ρ satisfying (1.1.5) and (1.1.6). In this monograph, we are interested in the
opposite direction. Namely, given a partition of a compact metrizable topological
space parametrized by a tree T , we define the notion of weight functions as the
collection of functions from T to (0, 1] satisfying the properties (1.1.5) and (1.1.6).
Then our main object of interest is the space of weight functions including those
derived from metrics and measures. Naively we believe that a partition and a weight
function essentially determine a “geometry” and/or an“analysis” of the original set
no matter where the weight function comes. It may come from a metric, a measure
or else. Keeping this intuition in mind, we are going to develop a basic theory of
weight functions in three closely related directions in this monograph.
The first direction is to study when a weight function is naturally associated
with a metric. In brief, our conclusion will be that a power of a weight function is
naturally associated with a metric if and only if the rearrangement of the resolution
T associated the weight function is Gromov hyperbolic. To be more precise, let
{Xw }w∈T be a partition of a compact metrizable topological space X with no
isolated points and let ρ : T → (0, 1] be a weight function. In Sect. 2.3, we
ρ
will define δM (·, ·), which is called the visual pre-metric associated with ρ, in the
ρ
following way: let s be the collection of w’s in T where the size ρ(w) is almost s.
ρ ρ
Define a horizontal edge of s as (w, v) ∈ s with Xw ∩ Xv = ∅. For r ∈ (0, 1),
the rearranged resolution Tρ,r associated with the weight function ρ is defined as
ρ
the vertices ∪m≥0 r m with the vertical edges from the tree structure of T and the
ρ ρ
horizontal edges of r m . Then the visual pre-metric δM (x, y), where M ≥ 1 is a
natural number, for x, y ∈ X is given by the infimum of s where x and y can be
ρ
connected by an M-consecutive horizontal edges in s . We think that a metric d
ρ
is naturally associated with the weight function ρ if and only if d and δM are bi-
Lipschitz equivalent on X × X. More precisely, we are going to use a phrase“d is
adapted to ρ” instead of “d is naturally associated with ρ”. The notion of visual
pre-metric in this monograph is a counterpart of that of visual pre-metric on the
boundary of a Gromov hyperbolic metric space, whose detailed account can be
seen in [12, 27] and [18] for example. Now the main conclusion of the first part
1.1 Introduction 5
is Theorem 2.5.12 saying that the hyperbolicity of the rearranged resolution Tρ,r is
equivalent to the existence of a metric adapted to some power of the weight function.
Moreover, if this is the case, the metric adapted to some power of the weight function
is shown to be a visual metric in Gromov’s sense.
The second direction is to establish relationships of various relations between
weight functions, metrics and measures. For example, Ahlfors regularity and the
volume property are relations between measures and metrics. For α > 0, a measure
μ is α-Ahlfors regular with respect to a metric d if and only if there exist c1 , c2 > 0
such that
c1 r α ≤ μ(Bd (x, r)) ≤ c2 r α ,
where Bd (x, r) = {y|y ∈ X, d(x, y) < r}, for any r ∈ (0, diam(X, d)] and x ∈ X.
See Definition 3.3.3 for the precise definition of the volume doubling property. On
the other hand, bi-Lipschitz and quasisymmetry are equivalence relations between
two metrics. (The precise definitions of bi-Lipschitz equivalence and quasisymmetry
are given in Definitions 3.1.9 and 3.6.1 respectively.) Regarding those relations, we
are going to claim the following relationships
bi-Lipschitz = Ahlfors regularity = being adapted (1.1.7)
and
the volume doubling property = quasisymmetry. (1.1.8)
in the framework of weight functions. To illustrate the first claim more explicitly, let
us introduce the notion of bi-Lipschitz equivalence of weight functions. Two weight
functions ρ1 and ρ2 are said to be bi-Lipschitz equivalent if and only if there exist
c1 , c2 > 0 such that
c1 ρ1 (p) ≤ ρ2 (p) ≤ c2 ρ1 (p)
for any p ∈ T . Now the first statement (1.1.7) can be resolved into three parts as
follows: let ρ1 and ρ2 be two weight functions.
Claim 1: Suppose that ρ1 = ρd1 and ρ2 = ρd2 for metrics d1 and d2 on X. Then
ρ1 and ρ2 are bi-Lipschitz equivalent if and only if d1 and d2 are bi-
Lipschitz equivalent as metrics.
Claim 2: Suppose that ρ1 = ρd and ρ2 = ρμ for a metric d on X and a Radon
measure μ on X. Then (ρ1 )α and ρ2 are bi-Lipschitz equivalent if and
only if μ is α-Ahlfors regular with respect to d.
Claim 3: Suppose that ρ1 = ρd for a metric d on X, then ρ1 and ρ2 are bi-
Lipschitz equivalent if and only if the metric d is adapted to the weight
function ρ2 .
One can find the precise statement in Theorem 1.2.11 in the case of partitions
of S 2 . The second statement (1.1.8) is rationalized in the same manner. See
Theorem 1.2.12 for the exact statement in the case of S 2 for example.
The third direction is a characterization of Ahlfors regular conformal dimension.
The Ahlfors regular conformal dimension, AR conformal dimension for short, of a
metric space (X, d) is defined as
dimAR (X, d) = inf{α|there exist a metric ρ on X and a Borel regular measure μ on

X such that ρ ∼ d and μ is α-Ahlfors regular with respect to ρ},
QS
where “ρ ∼ d” means that the two metrics ρ and d are quasisymmetric to each
QS
other. Note that α is the Hausdorff dimension of (X, d) if there exists a measure
μ which is α-Ahlfors regular with respect to d. In [13], Carrasco Piaggio has
given a characterization of Ahlfors regular conformal dimension in terms of the
critical exponent of p-combinatorial modulus of discrete path families. Given the
results from the previous part, we have obtained the ways to express the notions
of quasisymmetry and Ahlfors regularity in terms of weight functions. So we are
going to translate Carrasco Piaggio’s work into our framework. However, we are
going to use the critical exponent of p-energy instead of p-combinatorial modulus
in our work.1 Despite the difference between p-energy and p-modulus, we will
make essential use of Carrasco Piaggio’s ideas, which have been quietly embedded
in his paper [13]. Furthermore, we are going to define the notion of p-spectral
dimension and present a relation between Ahlfors regular conformal dimension
and p-spectral dimension. In particular, for p = 2, the 2-spectral dimension has
been know to appear in the asymptotic behavior of the Brownian motion and the
eigenvalue counting function of the Laplacian on certain fractals like the Sierpinski
gasket and the Sierpinski carpet. See [4, 5] and [23] for example. For the Sierpinski
carpet, we will show that the 2-spectral dimension gives an upper bound of Ahlfors
regular conformal dimension.
One of the main reasons for recent interest in Ahlfors regular conformal
dimension is its close connection with the Cannon’s conjecture, although we are
not going to pursue such a direction in this monograph.
Conjecture (Cannon) Let G be a Gromov hyperbolic group whose boundary ∂∞ G
is homeomorphic to the two dimensional sphere S 2 . Let d be a visual metric on
∂∞ G. Then (∂∞ G, d) is quasisymmetric to S 2 with the Euclidean metric.
For this conjecture, Bonk and Kleiner have given the following partial answer
in [7].
1 Thisidea of characterizing AR conformal dimension by p-energies was brought to the author by

B. Kleiner in a personal communication. In fact, Keith and Kleiner had obtained a result which is
comparable with that of Carrasco Piaggio [13]. See the discussions after [10, Proposition 1.5].
1.1 Introduction 7
Theorem 1.1.1 (Bonk-Kleiner) Let G be a Gromov hyperbolic group whose

boundary ∂∞ G is homeomorphic to S 2 and let d be a visual metric on ∂∞ G. Assume
that Ahlfors regular conformal dimension is attained, i.e. there exist a metric ρ
on ∂∞ G which is quasisymmetric to d and a Borel regular measure μ which is
dimAR (X, d)-Ahlfors regular with respect to ρ. Then (∂∞ G, d) is quasisymmetric
to S 2 with the Euclidean metric.
So, the last piece to prove Cannon’s conjecture is the attainability of Ahlfors
regular conformal dimension. Besides Canon’s conjecture, the attainability of
Ahlfors regular conformal dimension is an interesting problem by itself and not
solved except for limited classes of spaces like domains in Rn . In particular, it is
still open even in the case of the Sierpinski carpet at this moment.
There have been other works on an analytic characterization of Ahlfors regular
conformal dimension. In [11], Bourdon and Pajot have used the critical exponent
of p -cohomologies p(X, d). In general, p(X, d) ≤ dimAR (X, d) and the equality
holds if Ahlfors regular conformal dimension is attainable with a Loewner metric.
Lidquist has used the critical exponent of weak capacities Qw in [26]. He has shown
that Qw ≤ dimAR (X, d) and the equality has been shown to hold if Ahlfors regular
conformal dimension is attainable. For certain class of self-similar sets, Shimizu has
shown p∗ (G) ≤ dimAR (X, d), where p∗ (G) is the parabolic index of a blow-up G,
which is an infinite graph, of a self-similar set X in [28].
One of the ideas behind this study is to approximate a space by a series of
graphs. Such an idea has already been explored in association with hyperbolic
geometry. For example, in [15] and [11], they have constructed an infinite graph
whose hyperbolic boundary is homeomorphic to given compact metric space. They
have constructed a series of coverings of the space, which is a counterpart of our
partition, and then built a graph from the series. In [13], Carrasco Piaggio has
utilized this series of coverings to study the Ahlfors regular conformal dimension
of the space. His notion of “relative radius” essentially corresponds to our weight
functions. In our framework, the original space is homeomorphic to the analogue
of hyperbolic boundary of the resolution T of X even if it is not hyperbolic in the
sense of Gromov. See Theorem 2.5.5 for details. In other words, the resolution T of
X is a version of hyperbolic filling of the original space X. (See [9] for the notion
of hyperbolic fillings.) In this respect, our study in this monograph may be thought
of as a theory of weighted hyperbolic fillings.
The organization of this monograph is as follows. In Sect. 1.2, we give a summary
of the main results of this monograph in the case of the 2-dimensional sphere as a
showcase of the full theory. In Sect. 2.1, we give basic definitions and notations on
trees. Section 2.2 is devoted to the introduction of partitions and related notions.
In Sect. 2.3, we define the notion of weight function and the associated “visual
pre-metric”. We study our first question mentioned above, namely, when a weight
function is naturally associated with a (power of) metric in Sect. 2.4. In Sect. 2.5,
we are going to relate this question to the hyperbolicity of certain graph, called
a rearrangement above, associated with a weight function. Section 3.1 is devoted
to justifying the statement (1.1.7). In Sects. 3.2, 3.3, 3.5 and 3.6, we will study
the rationalized version of (1.1.8) as a mathematical statement. In particular, in

Sect. 3.3, we introduce the key notion of being “gentle”. In Sect. 3.4, we apply our
general theory to certain class of subsets of the square and obtain concrete (counter)
examples. From Sect. 4.1, we will start arguing a characterization of Ahlfors regular
conformal dimension. From Sect. 4.1–4.5, we discuss how to obtain a pair of a
metric d and a measure μ where μ is α-Ahlfors regular with respect to d for a
given order α. The main result of these sections is Theorem 4.5.1. In Sect. 4.6,
we will give a characterization of the Ahlfors regular conformal dimension as a
critical index p of p-energies. Then we will show the relation of the Ahlfors regular
conformal dimension and p-spectral dimension in Sect. 4.7. Additionally, we will
give another characterization of the Ahlfors regular conformal dimension by p-
modulus of curve families in Sect. 4.8. This recovers the original result by Carrasco
Piaggio [13]. Finally in Sect. B, we present the whereabouts of definitions, notations,
and conditions appearing in this monograph for reader’s sake.

Remark on the Usage of min, max, sup, inf and
Throughout this monograph, if A ⊆ [0, ∞) is empty, then we set
min A = max A = sup A = inf A = 0.
Moreover, if f : A → R and A = ∅, then we set

f (x) = 0.
x∈A
Such situations may happen if our target space (X, d) is not connected.
1.2 Summary of the Main Results; the Case of 2-dim. Sphere
In this section, we summarize our main results in this monograph in the case of a
2-dimensional sphere S 2 (or in other words, the Riemann sphere), which is denoted
by X in what follows. We use ds to denote the standard spherical geodesic metric
on X. Set
U = {A|A ⊆ X, closed, int(A) = ∅, ∂A is homeomorphic to the circle S 1 .}
First we divide X into finite number of subsets X1 , . . . , XN0 belonging to U, i.e.

N0
X= Xi .
i=1
1.2 Summary of the Main Results; the Case of 2-dim. Sphere 9
We assume that Xi ∩ Xj = ∂Xi ∩ ∂Xj if i = j . Next each Xi is divided into

finite number of its subsets Xi1 , Xi2 , . . . , XiNi ∈ U in the same manner as before.
Repeating this process, we obtain Xi1 ...ik for any i1 . . . ik satisfying

Xi1 ...ik = Xi1 ...ik j (1.2.1)
j =1,...,Ni1 ...ik
and if i1 . . .ik = j1 . . .jk , then
Xi1 ...ik ∩ Xj1 ...jk = ∂Xi1 ...ik ∩ ∂Xj1 ...jk . (1.2.2)
Note that (1.2.1) is a counterpart of (1.1.3). Next define
Tk = {i1 . . .ik |ij ∈ {1, . . . , Ni1 ...ij−1 } for any j = 1, . . . , k − 1}
for any k = 0, 1, . . ., where T0 is a one point set {φ}. Let T = ∪k≥0 Tk . Then T is
naturally though of as a (non-directed) tree whose edges are given by the totality of
(i1 . . .ik , i1 . . .ik ik+1 ). We regard the correspondence w ∈ T to Xw ∈ U as a map
from T to U, which is denoted by X . Namely, X (w) = Xw for any w ∈ T . Note
that X (φ) = X. Define
= {i1 i2 . . . |i1 . . . ik ∈ Tk for any k ≥ 0},
which is the “boundary” of the infinite tree T .

Furthermore we assume that for any i1 i2 . . . ∈

Xi1 ...ik
k=1,2,...
is a single point, which is denoted by σ (i1 i2 . . .). Note that σ is a map from to
X. This assumption corresponds to (1.1.4) and hence the map X is a partition of X
parametrized by the tree T , i.e. it satisfies the conditions (i), (ii), (iii), (iv) and (v) in
the introduction. Since X = ∪w∈Tk Xw for any k ≥ 0, this map σ is surjective.
In [8, Chapter 5], the authors have constructed “cell decomposition” associated
with an expanding Thurston map. This “cell decomposition” is, in fact, an example
of a partition formulated above.
Throughout this section, for simplicity, we assume the following conditions (SF)
and (TH), where (SF) is called strong finiteness in Definition 2.2.4 and (TH) ensures
the thickness of every exponential weight function. See Definition 3.1.19 for the
“thickness” of a weight function.
(SF)
#(σ −1 (x)) < +∞, (1.2.3)
where #(A) is the number of elements in a set A.

(TH) There exists m ≥ 1 such that for any w = i1 . . .in ∈ T , there exists v =
i1 . . . in in+1 . . .in+m ∈ T such that Xv ⊆ int(Xw ).
The main purpose of this monograph is to describe metrics and measures of X
from a given weight assigned to each piece Xw of the partition X .
Definition 1.2.1 A map g : T → (0, 1] is called a weight function if and only if it
satisfies the following conditions (G1), (G2) and (G3).
(G1) g(φ) = 1
(G2) g(i1 . . .ik ) ≥ g(i1 . . .ik ik+1 ) for any i1 . . .ik ∈ T and i1 . . .ik ik+1 ∈ T .
(G3)
lim sup g(w) = 0.

k→∞ w∈Tk
Define
G(T ) = {g|g : T → (0, 1] is a weight function.}
Moreover, we define following conditions (SpE) and (SbE), which represent “super-
exponential” and “sub-exponential” respectively:
(SpE) There exists λ ∈ (0, 1) such that
g(i1 . . .ik ik+1 ) ≥ λg(i1 . . .ik )
for any i1 . . .ik ∈ T and i1 . . .ik ik+1 ∈ T .

(SbE) There exist m ∈ N and γ ∈ (0, 1) such that
g(i1 . . .ik ik+1 . . .ik+m ) ≤ γ g(i1 . . .ik )
for any i1 . . .ik ∈ T and i1 . . .ik ik+1 . . .ik+m ∈ T .

Set
Ge (T ) = {g|g : T → (0, 1] is a weight function

satisfying (SpE) and (SbE).}.
Metrics and measures on X naturally have associated weight functions.

Definition 1.2.2 Set
D(X) = {d|d is a metric on X which produces the original topology of X,

and diam(X, d) = 1}
and
M(X) = {μ|μ is a Borel regular probability measure on X, μ({x}) = 0

for any x ∈ T and μ(O) > 0 for any non-empty open set O ⊆ X}.
For any d ∈ D(X), define gd : T → (0, 1] by gd (w) = diam(Xw , d) and for any
μ ∈ M(X), define gμ : T → (0, 1] by gμ (w) = μ(Xw ) for any w ∈ T .
From Proposition 2.3.5, we have the following fact.
Proposition 1.2.3 If d ∈ D(X) and μ ∈ M(X), then gd and gμ are weight
functions.
So a metric d ∈ D(X) has associated weight function gd . How about the converse
direction, i.e. for a given weight function g, is there a metric d such that g = gd ?
To make this question more rigorous and flexible, we define the notion of “visual
g
pre-metric” δM (·, ·) associated with a weight function g.
Definition 1.2.4 Let g ∈ G(T ). Define
g
s = {i1 . . .ik |i1 . . .ik ∈ T , g(i1 . . .ik−1 ) > s ≥ g(i1 . . .ik )}
for s ∈ (0, 1] and

g g
δM (x, y) = inf{s|there exist w(1), . . . , w(M + 1) ∈ s such that
x ∈ Xw(1) , y ∈ Xw(M+1) and
Xw(j ) ∩ Xw(j +1) = ∅ for any j = 1, . . . , M}
for M ≥ 0, x, y ∈ X. A weight function is called uniformly finite if and only if

g
sup #({v|v ∈ s , Xw ∩ Xv = ∅}) < +∞.
g
s∈(0,1],w∈s
g g g
Although δM (x, y) ≥ 0, δM (x, y) = 0 if and only if x = y and δM (x, y) =
g g
δM (y, x), the quantity δM may not satisfy the triangle inequality in general. The
g
visual pre-metric δM (x, y) is a counterpart of the visual metric defined in [8]. See
Sect. 2.3 for details.
g
If the pre-metric δM (·, ·) is bi-Lipschitz equivalent to a metric d, we consider d
as a metric which is naturally associated with the weight function g.
Definition 1.2.5 Let M ≥ 1

(1) A metric d ∈ D(X) is said to be M-adapted to a weight function g ∈ G(X) if
and only if there exist c1 , c2 > 0 such that
g
c1 d(x, y) ≤ δM (x, y) ≤ c2 d(x, y)
for any x, y ∈ X.
(2) A metric d is said to be M-adapted if and only if it is M-adapted to gd and it is
said to be adapted if it is M-adapted for some M ≥ 1.
(3) Define
DA,e (X) = {d|d ∈ D(X), gd ∈ Ge (T ) and d is adapted},

Me (X) = {μ|μ ∈ M(X), gμ ∈ Ge (T )}.
The value M really matters. See Example 3.4.9 for an example.

The following definition is used to describe an equivalent condition for the
existence of an adapted metric in Theorem 1.2.7.
g,r = m≥0 m , where the
Definition 1.2.6 Let g ∈ G(T ). For r ∈ (0, 1), define T
g
r
symbol means a disjoint union. Define the horizontal edges Eg,r
h and the vertical
v of T
edges Eg,r g,r as
g
h
Eg,r = {(w, v)|w, v ∈ r m for some m ≥ 0, w = v, Xw ∩ Xv = ∅}
and
g g
v
Eg,r = {(w, v)|w ∈ r m , v ∈ r m+1 for some m ≥ 0, Xw ⊇ Xv }
respectively.
The following theorem is a special case of Theorem 2.5.12.
Theorem 1.2.7 Let g ∈ G(X). There exist M ≥ 1, α > 0 and a metric d ∈ D(X)
such that d is M-adapted to g α if and only if the graph (Tg,r , Eg,r
h ∪E v ) is Gromov
g,r
hyperbolic for some r ∈ (0, 1). Moreover, if this is the case, then X is homeomorphic
to the hyperbolic boundary of (Tg,r , Eg,r
h ∪ E v ) and the adapted metric d is a
g,r
visual metric in the Gromov sense.
Next, we define two equivalent relations ∼ and ∼ on the collection of
BL GE
exponential weight functions. Later, we are going to identify these with known
relations according to the types of weight functions.
Definition 1.2.8 Let g, h ∈ Ge (T ).

(1) g and h are said to be bi-Lipschitz equivalent if and only if there exist c1 , c2 > 0
such that
c1 g(w) ≤ h(w) ≤ c2 g(w)
for any w ∈ T . We write g ∼ h if g and h are bi-Lipschitz equivalent.

BL
g
(2) h is said to be gentle to g if and only if there exists γ > 0 such that if w, v ∈ s
and Xw ∩ Xv = ∅, then h(w) ≤ γ h(v). We write g ∼ h if h is gentle to g.
GE
Clearly, ∼ is an equivalence relation. On the other hand, the fact that ∼ is an
BL GE
equivalence relation is not quite obvious and going to be shown in Theorem 3.5.2.
Proposition 1.2.9 The relations ∼ and ∼ are equivalent relations in Ge (T ).
BL GE
Moreover, if g ∼ h, then g ∼ h.
BL GE
Some of the properties of a weight function are invariant under the equivalence
relation ∼ as follows.
GE
Proposition 1.2.10
(1) Being uniformly finite is invariant under the equivalence relation ∼ , i.e. if g ∈
GE
Ge (T ) is uniformly finite, h ∈ Ge (T ) and g ∼ h, then h is uniformly finite.
GE
g,r is invariant under the equivalence relation ∼ .
(2) The hyperbolicity of T
GE
The statements (1) and (2) of the above theorem are the special cases of
Theorems 3.5.7 and 3.5.9 respectively.
The next theorem shows that bi-Lipschitz equivalence of weight functions can be
identified with other properties according to the types of involved weight functions.
Theorem 1.2.11
(1) For d, ρ ∈ DA,e (X), gd ∼ gρ if and only if d and g are bi-Lipschitz equivalent
BL
as metrics.
(2) For μ, ν ∈ M(X), gμ ∼ gν if and only if there exist c1 , c2 > 0 such that
BL
c1 μ(A) ≤ ν(A) ≤ c2 μ(A)
for any Borel set A ⊆ X.

(3) For g ∈ Ge (X) and d ∈ DA,e (X), g ∼ gd if and only if d is M-adapted to g for
BL
some M ≥ 1.
(4) For d ∈ DA,e (X), μ ∈ M(X) and α > 0, (gd )α ∼ gμ and gd is uniformly finite
BL
if and only if μ is α-Ahlfors regular with respect to d, i.e. there exist c1 , c2 > 0
such that
c1 r α ≤ μ(Bd (x, r)) ≤ c2 r α
for any r > 0 and x ∈ X.

The statements (1), (2), (3) and (4) of the above theorem follow from Corol-
lary 3.1.10, Theorems 3.1.4, 3.1.14 and 3.1.21 respectively.
The gentle equivalence relation is identified with “quasisymmetry” between two
metrics and with “volume doubling property” between a metric and a measure.
Theorem 1.2.12
(1) Let d ∈ DA,e (X) and let μ ∈ M(X). Then gμ ∈ Ge (T ), gd ∼ gμ and gd is
GE
uniformly finite if and only if μ has the volume doubling property with respect
to d, i.e. there exists C > 0 such that
μ(Bd (x, 2r)) ≤ Cμ(Bd (x, r))
for any r > 0 and x ∈ X.

(2) For d ∈ DA,e (X) and ρ ∈ D(X), d is quasisymmetric to ρ if and only if
ρ ∈ DA,e (X) and gd ∼ gρ .
GE
The statement (1) of the above theorem follows from Proposition 3.3.6 and
Theorem 3.3.9-(2). Note that the condition (TH) implies (TH1) appearing in
Theorem 3.2.3. Consequently every exponential weight function is thick by The-
orem 3.2.3. The statement (2) is immediate from Corollary 3.6.7.
In [8, Section 17], the authors have shown that the visual metric is quasisymmet-
ric to the chordal metric which is bi-Lipschitz equivalent to the standard geodesic
metric dS on S 2 for certain class of expanding Thurston maps. In view of their proof,
they have essentially shown a counterpart of the condition given in Theorem 1.2.12-
(2).
Next we present a characterization of the Ahlfors regular conformal dimension
using the critical index p of p-energies.
g g
Definition 1.2.13 Let g ∈ Ge (T ) and let r ∈ (0, 1). For A ⊆ r m , w ∈ r m ,
M ≥ 1 and n ≥ 0, define
g
S n (A) = {v|v ∈ r m+n , Xv ⊆ ∪u∈A Xu }
and
g g
M (w) = {v|v ∈ r m , there exists (v(0), v(1), . . . , v(M)) such that
v(0) = w and (v(i), v(i + 1)) ∈ Eg,r
h
for any i = 0, . . . , M − 1}.
g g
The set S n (A) corresponds to the refinement of A in r m+n and the set M (w) is
g
the M-neighborhood of w in r m .
g,r , M ≥ 1
Definition 1.2.14 Let g ∈ Ge (T ) and let r ∈ (0, 1). For p > 0, w ∈ T
and n ≥ 0, define

g
EM,p,w,n = inf |f (u) − f (v)|p
h ,u,v∈ g
(u,v)∈Eg,r
r m+n
g
f : r m+n → R, f |S n (w) = 1, f |g g
\S n (M (w)) =0 ,
r m+n
g
where the integer m is chosen so that w ∈ r m , and
g g
EM,p = lim inf sup EM,p,w,m .
m→∞
w∈Tg,r
By Theorem 4.6.4, we have the following characterization of the Ahlfors regular

g
conformal dimension of (X, d) in terms of EM,p .
Theorem 1.2.15 Let d ∈ DA,e (X) and set g = gd . Assume that d is uniformly
g
finite and M-adapted. If EM,p = 0, then there exist ρ ∈ DA,e (X) and μ ∈ Me (X)
such that μ is p-Ahlfors regular with respect to ρ and ρ is quasisymmetric to d.
Moreover, the Ahlfors regular conformal dimension of (X, d) is (finite and) given by
g
inf{p|EM,p = 0}.
Chapter 2
Partitions, Weight Functions and Their
Hyperbolicity
2.1 Tree with a Reference Point
In this section, we review basic notions and notations on a tree with a reference
point.
Definition 2.1.1 Let T be a countably infinite set and let A : T × T → {0, 1}
which satisfies A(w, v) = A(v, w) and A(w, w) = 0 for any w, v ∈ T . We call the
pair (T , A) a (non-directed) graph with the vertices T and the adjacent matrix A.
An element (u, v) ∈ T × T is called an edge of (T , A) if and only if A(u, v) = 1.
We will identify the adjacent matrix A with the collection of edges {(u, v)|u, v ∈
T , A(u, v) = 1}.
(1) The set {v|A(w, v) = 1} is called the neighborhood of w in (T , A). (T , A) is
said to be locally finite if the neighborhood of w is a finite set for any w ∈ T .
(2) For w0 , . . . , wn ∈ T , (w0 , w1 , . . . , wn ) is called a path between w0 and wn if
A(wi , wi+1 ) = 1 for any i = 0, 1, . . . n − 1. A path (w0 , w1 , . . . , wn ) is called
simple if and only if wi = wj for any i, j with 0 ≤ i < j ≤ n and |i − j | < n.
(3) (T , A) is called a (non-directed) tree if and only if there exists a unique simple
path between w and v for any w, v ∈ T with w = v. For a tree (T , A), the
unique simple path between two vertices w and v is called the geodesic between
w and v and denoted by wv. We write u ∈ wv if wv = (w0 , w1 , . . . , wn ) and
u = wi for some i.
In this monograph, we always fix a point in a tree as the root of the tree and call
the point the reference point.
© The Editor(s) (if applicable) and The Author(s), under exclusive license 17
to Springer Nature Switzerland AG 2020
J. Kigami, Geometry and Analysis of Metric Spaces via Weighted Partitions,
Lecture Notes in Mathematics 2265, https://doi.org/10.1007/978-3-030-54154-5_2
18 2 Partitions, Weight Functions and Their Hyperbolicity
Definition 2.1.2 Let (T , A) be a tree and let φ ∈ T . The triple (T , A, φ) is called

a tree with a reference point φ.
(1) Define π : T → T by
wn−1 if w = φ and φw = (w0 , w1 , . . . , wn−1 , wn ),

π(w) =
φ if w = φ
and set S(w) = {v|A(w, v) = 1}\{π(w)}.

(2) For w ∈ T , we define |w| = n if and only if φw = (w0 , w1 , . . . , wn ) with
w0 = φ and wn = w. Moreover, we set (T )m = {w|w ∈ T , |w| = m}.
(4) An infinite sequence of vertices (w0 , w1 , . . .) is called an infinite geodesic ray
originated from w0 if and only if (w0 , . . . , wn ) = w0 wn for any n ≥ 0. Two
infinite geodesic rays (w0 , w1 , . . .) and (v0 , v1 , . . .) are equivalent if and only
if there exists k ∈ Z such that wn+k = vn for sufficiently large n. An equivalent
class of infinite geodesic rays is called an end of T . We use to denote the
collection of ends of T .
(5) Define w as the collection of infinite geodesic rays originated from w ∈ T .
For any v ∈ T , vw is defined as the collection of elements of w passing
through v, namely
w
v = {(w, w1 , . . .)|(w, w1 , . . .) ∈ w
, wn = v for some n ≥ 1}.
Remark Strictly, the notations like π and | · | should be written as π (T ,A,φ) and
| · |(T ,A,φ) respectively. In fact, if we will need to specify the tree in question, we are
going to use such explicit notations.
One of the typical examples of a tree is the infinite binary tree. In the next
example, we present a class of trees where #(S(w)) is independent of w ∈ T .
(N)
Example 2.1.3 Let N ≥ 2 be an integer. Let Tm = {1, . . . , N}m for m ≥ 0.
(N)
(We let T0 = {φ}, where φ represents an empty sequence.) We customarily write
(N) (N)
(i1 , . . . , im ) ∈ Tm as i1 . . . im . Set T (N) = ∪m≥0 Tm . Define π : T (N) → T (N)
by π(i1 . . . im im+1 ) = i1 . . . im for m ≥ 0 and π(φ) = φ. Furthermore, define
1 if w = v, and either π(w) = v or π(v) = w,

wv =
A(N)
0 otherwise.
Then (T (N) , A(N) , φ) is a locally finite tree with a reference point φ. In particular,
(T (2) , A(2) , φ) is called the infinite binary tree.
It is easy to see that for any infinite geodesic ray (w0 , w1 , . . .), there exists a
geodesic ray originated from φ that is equivalent to (w0 , w1 , . . .). In fact, adding
the geodesic φw0 to (w0 , w1 , . . .) and removing a loop, one can obtain the infinite
geodesic ray having required property. This fact shows the following proposition.
2.1 Tree with a Reference Point 19
Proposition 2.1.4 There exists a natural bijective map from to φ.
Through this map, we always identify the collection of ends and the collection
of infinite geodesic rays originated from φ, φ .
Hereafter in this monograph, (T , A, φ) is assumed to be a locally finite tree with
a fixed reference point φ ∈ T . If no confusion can occur, we omit φ in the notations.
φ
For example, we use , and v in place of φ and v respectively.
Example 2.1.5 Let N ≥ 2 be an integer. In the case of (T (N) , A(N) , φ) defined
in Example 2.1.3, the collection of the ends is (N) = {1, . . . , N}N =
{i1 i2 i3 . . . , |ij ∈ {1, . . . , N} for any m ∈ N}. With the natural product topology,
(N) is a Cantor set, i.e. perfect and totally disconnected.
Definition 2.1.6 Let (T , A, φ) be a locally finite tree with a reference point φ.

(1) For ω = (w0 , w1 , . . .) ∈ , we define [ω]m by [ω]m = wm for any m ≥ 0.
Moreover, let w ∈ T . If φw = (w0 , w1 , . . . , w|w| ), then for any 0 ≤ m ≤ |w|,
we define [w]m = wm . For w ∈ T , we define
Tw = {v|v ∈ T , w ∈ φv}.
(2) For w, v ∈ T , we define the confluence of w and v, w ∧ v, by
w ∧ v = wmax{i|i=0,...,|w|,[w]i =[v]i }
(3) For ω, τ ∈ , if ω = τ , we define the confluence of ω and τ , ω ∧ τ , by
ω ∧ τ = [ω]max{m|[ω]m =[τ ]m } .
(4) For ω, τ ∈ , we define ρ∗ (ω, τ ) ≥ 0 by
2−|ω∧τ | if ω = τ,
ρ∗ (ω, τ ) =
0 if ω = τ.
It is easy to see that ρ∗ is a metric on and { [ω]m }m≥0 is a fundamental system

of neighborhood of ω ∈ . Moreover, { v }v∈T is a countable base of open sets.
This base of open sets has the following property.
Lemma 2.1.7 Let (T , A, φ) be a locally finite tree with a reference point φ. Let
w, v ∈ T . Then the following three conditions are equivalent to each other.
(1) w ∩ v = ∅
(2) |w ∧ v| = |w| or |w ∧ v| = |v|.
(3) v ⊆ w or w ⊆ v
Proof (2) ⇒ (3): If |w ∧ v| = |w|, then w = w ∧ v and hence w ∈ φv. Therefore
v ⊆ w . Similarly if |w ∧ v| = |v|, then w ⊆ v .
20 2 Partitions, Weight Functions and Their Hyperbolicity
(3) ⇒ (1): This is obvious.

(1) ⇒ (2): If ω ∈ w ∩ v, then there exist m, n ≥ 0 such that w = [ω]m and
v = [ω]m . It follows that
w if m ≤ n,
w∧v =
v if m ≤ n.
Hence we see that |w ∧ v| = |w| or |w ∧ v| = |v|.

With the help of the above lemma, we may easily verify the following well-
known fact. The proof is standard and left to the readers.
Proposition 2.1.8 If (T , A, φ) is a locally finite tree with a reference point φ. Then
ρ∗ (·, ·) is a metric on and ( , ρ) is compact and totally disconnected. Moreover,
if #(S(w)) ≥ 2 for any w ∈ T , then ( , ρ) is perfect.
By the above proposition, if #(S(w)) ≥ 2 for any w ∈ T , then is
(homeomorphic to) the Cantor set.
2.2 Partition
In this section, we formulate exactly the notion of a partition introduced in Sect. 1.1.
A partition is a map from a tree to the collection of nonempty compact subsets of
a compact metrizable topological space with no isolated point and it is required
to preserve the natural hierarchical structure of the tree. Consequently, a partition
induces a surjective map from the Cantor set, i.e. the collection of ends of the tree,
to the compact metrizable space.
Throughout this section, T = (T , A, φ) is a locally finite tree with a reference
point φ.
Definition 2.2.1 (Partition) Let (X, O) be a compact metrizable topological space
having no isolated point, where O is the collection of open sets, and let C(X, O)
be the collection of nonempty compact subsets of X. If no confusion can occur, we
write C(X) in place of C(X, O).
(1) A map K : T → C(X, O), where we customarily denote K(w) by Kw for
simplicity, is called a partition of X parametrized by (T , A, φ) if and only if
it satisfies the following conditions (P1) and (P2), which correspond to (1.1.3)
and (1.1.4) respectively.
(P1) Kφ = X and for any w ∈ T , #(Kw ) ≥ 2, Kw has no isolated point and

Kw = Kv .
v∈S(w)
(P2) For any ω ∈ , ∩m≥0 K[ω]m is a single point.

2.2 Partition 21
(2) Let K : T → C(X, O) be a partition of X parametrized by (T , A, φ). Define

Ow and Bw for w ∈ T by

Ow = Kw \ Kv ,
v∈(T )|w| \{w}

Bw = Kw ∩ Kv .
v∈(T )|w| \{w}
If Ow = ∅ for any w ∈ T , then the partition K is called minimal.

(3) Let K : T → C(X, O) be a partition of X. Then (w(1), . . . , w(m)) ∈ ∪k≥0 T k
is called a chain of K (or a chain for short if no confusion can occur) if and only
if Kw(i) ∩ Kw(i+1) = ∅ for any i = 1, . . . , m − 1. A chain (w(1), . . . , w(m))
of K is called a chain of K in ⊆ T if w(i) ∈ for any i = 1, . . . , m.
For subsets A, B ⊆ X, A chain (w(1), . . . , w(m)) of K is called a chain of
K between A and B if and only if A ∩ Kw(1) = ∅ and B ∩ Kw(m) = ∅.
We use CHK (A, B) to denote the collection of chains of K between A and B.
Moreover, we denote the collection of chains of K in between A and B by
CHK (A, B).
As is shown in Theorem 2.2.9, a partition can be modified so as to be minimal

by restricting it to a suitable subtree.
The next lemma is an assortment of direct consequences from the definition of
the partition.
Lemma 2.2.2 Let K : T → C(X, O) be a partition of X parametrized by
(T , A, φ).
(1) For any w ∈ T , Ow is an open set. Ov ⊆ Ow for any v ∈ S(w).
(2) Ow ∩ Kv = ∅ if w, v ∈ T and w ∩ v = ∅.
(3) If w ∩ v = ∅, then Kw ∩ Kv = Bw ∩ Bv .
Proof (1) Note that by (P1), X = ∪w∈(T )m Kw . Hence
Ow = Kw \(∪v∈(T )|w| \{w} Kv ) = X\(∪v∈(T )|w| \{w} Kv ).
So Ow is open. The rest of the statement is immediate from the property (P2).
(2) By Lemma 2.1.7, if u = w ∧ v, then |u| < |w| and |u| < |v|. Let w =
[w]|u|+1 and let v = [v]|u|+1 . Then w , v ∈ S(u) and w = v . Since Ow ⊆
Kw \Kv , it follows that Ow ∩ Kv = ∅. Using (1), we see Ow ∩ Kv = ∅.
(3) This follows immediately by the definition of Bw .

The condition (P2) provides a natural map from the ends of the tree to the
space X.
Another random document with
no related content on Scribd:
{27}
In his opinion the Ashantis had been preparing ever since the
British occupation in 1896 to reassert their independence."
The Governor was, himself, in Kumassi when the Ashantis first
attacked it, on the 25th of March, and he has given an account
of the desperate position in which the few British officials,
with their small native garrison and the refugees whom they
tried to protect, were placed. "Our force," said Sir Frederic
Hodgson, "consisted of only some 200 Hausas, while there is
reason to believe that we had not less than 15,000 Ashantis
surrounding us. In addition to our own force we had to protect
some 3,500 refugees, chiefly Mahomedan traders, Fantis, and loyal
Kumassis, none of whom we were able to take into the fort,
where every available bit of space was required for military
purposes. It was heartrending to see the efforts of these poor
people to scale the walls or break through the gate of the
fort, and we had to withdraw the Hausas from the cantonments
and draw a cordon round the refugees. It is impossible to
describe the horror of the situation with these 3,500 wretched
people huddled together without shelter under the walls of the
fort. That same night a tornado broke over Kumassi, and the scene
next morning with over 200 children was too terrible for
words. Afterwards they were able to arrange shelters for
themselves." Near the end of April, two small reinforcements
from other posts reached Kumassi; but while this strengthened
the numbers for defence, it weakened the food supply. Taking
stock of their food, the besieged decided that they could hold
out until June 23, and that if the main body then marched out,
to cut, if possible, their way through the enemy, leaving a
hundred men behind, the latter might keep the fort until July
15. This, accordingly, was done. On the 23d of June Governor
Hodgson, with all but 100 men, stole away from Kumassi, by a
road which the Ashantis had not guarded, and succeeded in
reaching the coast, undergoing great hardships and dangers in
the march. Meantime, an expedition from Cape Coast Castle was
being energetically prepared by Colonel Sir J. Willcocks, who
overcame immense difficulties and fought his way into Kumassi
on Ju]y 15, the very day on which the food-supply of the
little garrison was expected to give out. The following
account of his entry into Kumassi is from Colonel Willcocks'
official report: "Forming up in the main road, we marched
towards Kumassi, a mile distant, the troops cheering wildly
for the Queen and then followed silence. No sound came from
the direction of the fort, which you cannot see till quite
close. For a moment the hideous desolation and silence, the
headless bodies lying everywhere, the sickening smell, &c.,
almost made one shudder to think what no one dared to
utter—'Has Kumassi fallen? Are we too late?' Then a bugle
sound caught the ear—'the general salute'—the tops of the
towers appeared, and again every man in the column, white and
black, broke into cheers long sustained. The brave defenders
had at last seen us; they knew for hours' past from the firing
growing ever nearer that we were coming, yet they dared not
open their only gate; they perforce must wait, for even as we
appeared the enemy were making their last efforts to destroy
the outlying buildings, and were actually setting them on fire
until after dark, when a party of 100 men went out and treated
them to volleys and cleared them out. If I have gone too fully
into details of the final scene, the occasion was one that
every white man felt for him comes perhaps but once, and no
one would have missed it for a kingdom."
ASPHYXIATING SHELLS: Declaration against.
See (in this volume)

PEACE CONFERENCE.
ASSASSINATIONS:
Of President Barrios.

CENTRAL AMERICA (GUATEMALA): A. D. 1897-1898.
Of President Borda.

URUGUAY: A. D. 1896-1899.
Of Canovas del Castillo.

SPAIN: A. D. 1897 (AUGUST-OCTOBER).
Of Empress Elizabeth of Austria.

AUSTRIA-HUNGARY: A. D. 1898 (SEPTEMBER).
Of Governor Goebel.

KENTUCKY: A. D. 1895-1900.
Of President Heureaux.

DOMINICAN REPUBLIC: A. D. 1899.
Of King Humbert.

ITALY: A. D. 1899-1900;
and 1900 (JULY-SEPTEMBER).
Of Professor Mihaileano.

BALKAN AND DANUBIAN STATES.
Of Nâsr-ed-din, Shah of Persia.

PERSIA: A. D. 1896.
Of M. Stambouloff.

BALKAN AND DANUBIAN STATES (BULGARIA).
ASSIOUT, Nile barrage at.

EGYPT: A. D. 1898-1901.
ASSOCIATIONS BILL, The French.

FRANCE: A. D. 1901.
ASSOUAN. Nile barrage at.

EGYPT: A. D. 1898-1901.
ASSUMPTIONIST FATHERS, Dissolution of the Society of the.

FRANCE: A. D. 1899-1900 (AUGUST-JANUARY).
ATACAMA, The question concerning.

CHILE: A. D. 1894-1900.
ATBARA, Battle of the.

EGYPT: A. D. 1897-1898.
ATHENS: A. D. 1896.
The revival of Olympic games.
As the result of a movement instituted in France by the Baron

de Coubertin, an interesting attempt to give athletic sports
the spirit and semblance of the ancient Olympic games was made
at Athens in the spring of 1896. A number of wealthy Greeks in
different parts of the world joined generously in the
undertaking, one gentleman especially, M. Averoff, of
Alexandria, bearing the cost of a restoration in marble of the
stadium at Athens, for the occasion. The games were held in
April, from the 6th to the 15th, and were witnessed by a great
number of people. Besides Greek competitors, there were 42 from
Germany, 23 from England, 21 from America, 15 from France. The
great event of the occasion was the long foot-race from
Marathon to Athens, which was won by a young Greek.
The U. S. Consul at Athens, writing of the reconstruction of

the ancient stadium for the games, described the work as
follows:
{28}
"The stadium may be described as an immense open air

amphitheater constructed in a natural ravine, artificially
filled in at the end. It is in the shape of an elongated
horseshoe. The spectators, seated upon the sloping sides of
the ravine, look down into the arena below, which is a little
over 600 feet in length and about 100 feet wide at the widest
part. … The stadium, as rebuilt for the games, will consist of
(1) the arena, bounded by a marble curbing, surmounted by an
iron railing adorned with Athenian owls;
(2) a walk between this curbing and the first row of seats;
(3) a low retaining wall of marble on which rests the first
row of seats, the entire row being of marble;
(4) the seats;
(5) the underground tunnel.
In addition to these features there will be an imposing

entrance, a surrounding wall at the top of the hill, and two
supporting walls at the entrance. As far as possible, in the
reconstruction of the stadium, the old portions will be used,
where these are in a sufficient state of preservation, and an
effort will be made to reproduce, as nearly as practicable,
the ancient structure. The seats at present will not all be
made of pentelic marble, as there is neither time nor money
for such an undertaking. At the closed end of the arena,
seventeen rows will be made of pentelic marble, as well as the
first row all the way around. The remaining rows up to the
first aisle are being constructed of Pincus stone. These will
accommodate 25,000 seated spectators. From this aisle to the
top will be placed wooden benches for 30,000 seated
spectators. Add to these standing room for 5,000, and we have
the holding capacity of the stadium 60,000 without crowding."
United States Consular Reports,

March, 1896, pages 353-354.
ATLANTA: A. D. 1895.
The Cotton States and International Exposition.
An important exposition, named as above, was held with great

success at Atlanta, Georgia, from the 18th of September until
the end of the year 1895. The exhibits from Mexico and many of
the Central and South American States were extensive and
interesting; but the main interest and value of the exposition
were in its showing of the industrial resources of the
Southern States of the American Union, and of the recent
progress made in developing them.
AUSGLEICH, The.

AUSTRIA-HUNGARY: FINANCIAL RELATIONS;
and A. D. 1897 (OCTOBER-DECEMBER).
AUSTRAL ISLANDS:
Annexation to France.
The Austral or Tubuai Islands were formally annexed to France

by the Governor of Tahiti, on the 21st of August, 1900.
----------AUSTRALIA. Start--------
AUSTRALIA:
Recent extensions of Democracy in the Australian Colonies
and New Zealand.
Social experiments.
"The five colonies of the Australian continent, Tasmania, and

New Zealand constitute seven practically independent
commonwealths under the British crown. Australians and New
Zealanders have therefore been able to develop their countries
along their own lines, and have surpassed all other
Anglo-Saxon nations in the number and variety of functions
which the state is called upon to perform. … The railways
almost without exception, and all the telegraph and telephone
lines, are in the hands of the community. In the few cases in
which there is private ownership of railways, a particular
line was demanded at a certain time, and the government were
not then in a position to borrow the funds required for its
construction. Western Australia has recently purchased the
entire property of one of the two private undertakings in the
colony. A mass of sanitary and industrial legislation also has
been placed upon the statute book.
"Again, South Australia, Victoria, Western Australia, and New

Zealand lend money to settlers at low rates of interest; South
Australia sells its wines in London; Queensland facilitates
the erection of sugar mills; Victoria and South Australia have
given a bonus upon the exportation of dairy produce; South
Australia, New Zealand, and Victoria receive the produce,
grade and freeze it free of charge, or at a rate which barely
covers the expenses; Victoria contributes toward the erection
of butter factories; Victoria and New Zealand have subsidized
the mining industry; and Western Australia has adopted a
comprehensive scheme for the supply of water to the Coolgardie
gold fields. In all the colonies the national system of
primary education is compulsory and undenominational. In South
Australia, Victoria, Queensland, and New Zealand it is also free.
In the other colonies fees are charged, which may be remitted
wholly or partly if parents are unable to pay them. Assistance
is given in most cases for the promotion of secondary,
technical, and university education. New Zealand and South
Australia have appointed public trustees. New Zealand has long
possessed a department of life insurance.
"Finally, … New Zealand has adopted a system of old-age

pensions. A pension of seven shillings a week is to be given
to every person above the age of sixty-five years, provided he
or she has lived in the colony for twenty-five years, and is
able to pass a certain test in regard to sobriety and general
good conduct. … In South Australia direct taxation takes two
forms. There is an income tax at the rate of four and a half
pence in the pound up to £800, and of six-pence in the pound
above £800 of taxable amount resulting from personal
exertions, and at the rate of ninepence and one shilling in
the pound respectively on incomes from property. Incomes
between £125 and £425 enjoy exemption on £125 of the amount.
Again, there is a tax on the unimproved value of land of one
half-penny in the pound up to, and one penny above, the
capital value of £5000. …
"Similar taxation is to be found in New Zealand, and includes

both a progressive income tax and a tax on land values which
is more highly graduated than that of South Australia. … All
improvements are excluded from the assessment of the taxable
amount. … If the owner of the property is dissatisfied with
the assessment of the government, he can call upon them to buy
it of him at their own valuation. In only one case has such an
extreme step been taken; and it is pleasant to find that it
has resulted in an annual profit of nearly five per cent upon
the outlay, and that the land which formerly gave employment
to a few shepherds is now occupied by a large number of
thriving settlers.
{29}
I may add that when the government deem that an estate is not
being developed as it should be by its owners, they are
authorized by statute to purchase it—by negotiation if
possible, otherwise at a price paid by an impartial
tribunal—with a view to its subdivision into small holdings
suitable to the requirements of the community. This system of
taxation, it will be said with some truth, is based upon the
teachings of Henry George. He travelled in Australia and New
Zealand, and was listened to with attention; but, while he
looked to the ultimate absorption of the whole unearned
increment, his hearers in the antipodes dissociated themselves
from his conclusions, though they appreciated the value of his
premises. Consequently, while accepting his principles, they
did not hesitate to exempt small properties from the tax, and
to increase its rate progressively in relation to the amount
of the unimproved value. …
"One of the most hopeful signs of the day is that, with the
help of the representatives of labor in Parliament, Australian
governments have done much within recent years to mitigate the
excess of population in the large towns, and to replace the
unemployed upon the land. Of course mistakes have been made.
In some cases settlers have failed through lack of
agricultural knowledge; in others, on account of the
barrenness of the soil. In South Australia, the village
settlements, which were avowedly started as an alternative to
relief works, have been only a modified success. In New
Zealand, village settlements have produced very satisfactory
results. … In Victoria, a labor colony has been established,
with the entire support of the trades-unionists, to which the
unemployed may be sent, and at which they receive, at a very
low rate of wages, a course of instruction in agricultural
pursuits which enables them subsequently to obtain private
employment with farmers or others. In New Zealand, I found a
very strong feeling among trades-unionists that it would be to
the interest of the workingmen themselves if a penal colony
were established, on the lines of those which exist in
Germany, to which loafers might be sent, and at which they
would be compelled to work, with the alternative of
starvation."
H. De R. Walker,
Australasian Extensions of Democracy
(Atlantic Monthly, May, 1899).
See also (in this volume),

NEW ZEALAND: A. D. 1891-1900.
AUSTRALIA: Western Australia:

The Outlander problem in Australia.
"Here we have a problem in many respects similar to that which

has distracted South Africa. In several particulars the
resemblance is startlingly close. … Many of the elements of
disorder in the two continents are the same. In Western
Australia, as in the Transvaal, there is a large population of
mining residents, who complain that they are treated like
'helots'—to use Sir Alfred Milner's term—by the privileged
agricultural burghers. They urge that they are denied fair
representation, so that the burghers monopolise political
power; that the administration is in the hands of a knot of
politicians and place-hunters at Perth—I had almost written
Pretoria; that they have made the colony wealthy by their
enterprise and capital, only to see a large part of the fruits
of their industry drawn from them by excessive taxation, which
is expended mainly outside their own district; that they are
burdened by oppressive railway rates and denied access to the
port which is the natural outlet to the Goldfields, and so on.
The Kalgoorlie 'Uitlanders,' like the Johannesburgers, have
sent a petition to the Queen, signed by a larger number of
persons than those who forwarded the famous memorial which set
the ball rolling in South Africa and led to the Bloemfontein
Conference. The case is fully and temperately set forth in
this petition, and in the Manifesto of the Eastern Goldfields
Reform League of Western Australia, both of which documents
are in the last Bluebook relating to Australian Federation.
The same official compilation contains a statement to Mr.
Chamberlain from Dr. Paget Thurston, in which the parallel
between West Australia and the Transvaal is asserted with the
most uncompromising directness. 'We have here,' says the
writer, 'a Boer and Outlander question almost parallel to that
in the Transvaal. As an Outlander I appeal to you.' Dr.
Thurston adds: 'The old West Australians openly speak as if
the colony was theirs, and we were interlopers who have no
course open to us but to leave the colony if we are
dissatisfied.' This has a very familiar sound, and so has the
following: 'The great bulk of the taxation is levied through
duties on food and drink. As the Boer party includes all the
agricultural producers, and the Outlanders include the great
bulk of the consumers, this acts injuriously on us in two
ways. It puts a frightful load on the Outlander taxpayer, and
enables the Boer producer to command a very high price for his
food-stuffs. Owing to the limitation of the market by
excessive protection, many articles of common use reach famine
prices at times. In the three years I have been here, for
instance, potatoes have been £22 10s. a ton; apples, 2s. 6d. a
pound; oranges, 5s. a dozen; new-laid eggs, 4s. a dozen (at
the time of writing, 3s. 6d.). Fresh butter is practically
unobtainable for ten months in the year, and common country
wine (such as I used to buy for 3d. and 4d. a bottle in the
Canary Islands) is here 2s. a bottle. I ask you, Sir, whether
any other place in Her Majesty's Empire (not physically
inaccessible) can show prices one half as high during the past
three years?'
"Nor does the ominous kind of hint that preceded the Jameson
Raid fail to be uttered. Only three terminations, according to
Dr. Thurston, are possible if Sir John Forrest does not modify
his Krugerite policy towards the mining settlers:
'(1) Separation of the goldfields.—This would be only fair to
the goldfields; but thousands of Outlanders have settled in
the other parts of the colony, and this step would not redress
their wrongs. The practical result of this step would be
prosperity for the goldfields, but almost ruin for the rest of
the colony.
(2) Revolution.—I fear this is much more probable than is
generally thought. Unless a material change takes place
quickly there will be bloodshed in this colony.
(3) General depression, practically equivalent to bankruptcy.'
Separation, however, and the creation of a new colony, which
would include the Goldfields district and come down to the
sea, and would immediately join the Australian Federation, is
the remedy officially proposed by the representatives of the
Outlanders. …
{30}
"The Colonial Secretary has deferred his final answer to the

Goldfields Petition until the comments of the Perth Ministry
upon that document have been received and considered. But he
has sent a provisional reply to the representatives of the
petitioners in London. He sees the solution of the matter in
getting Western Australia somehow into the new Commonwealth.
In a communication to Mr. Walter Griffiths, one of the
Goldfields delegates, the Colonial Secretary says: 'The
decision of the Government of Western Australia to summon
Parliament immediately with the view to the passing of a
measure for the submission of the Commonwealth Bill to the
electors of the colony has removed the chief of the grievances
put forward in the petition and has opened up an early
prospect of obtaining the object which the petitioners had in
view. An answer will be returned to the petition after a
careful consideration of its terms and of the comments of the
Government of the colony thereon, but Mr. Chamberlain trusts
that before an answer can be returned the people of the colony
will have decided to join the Commonwealth, for the government of
which, in that event, it will be to deal with the grievances
alleged in the petition in so far as they are not exclusively
within the province of the Parliament and Government of
Western Australia.' In other words, let the Federation dispose
of the matter. But the delegates point out that this might not
remove their grievances. The Federal Parliament would have no
power to compel the dominant party in the Perth Assembly
either to redistribute seats fairly, or divide the colony, so
as to create 'Home Rule for the Rand.' True, we should have
washed our hands of the affair, and could tell the malcontent
Uitlanders that it was none of our business. But if Perth
still remained obstinate, and Coolgardie in consequence began
to carry out some of those ugly projects hinted at by Dr.
Thurston, it might become our business in an embarrassing
fashion. At any rate, it does not seem quite fair to the new
Commonwealth to start it in life with this grave question,
still unsettled, upon its hands."
S. Low,
Enigmas of Empire
(Nineteenth Century, June, 1900).
AUSTRALIA: New South Wales: A. D. 1894-1895.

Defeat of the Protectionist policy.
Adoption of a liberal tariff.
At the general elections of July, 1894, in New South Wales,

the tariff issue was sharply defined. "'Protection' was
inscribed on the banners of the ministerial party, led by the
then Premier, Sir George Dibbs, while the aggressive
opposition, led by Mr. Reid, … fought under the banner of
'free trade.' The Free Traders won the battle in that
election, as there were 63 Free Traders, 40 Protectionists,
and 22 labor members, mostly with free-trade leanings,
returned. On the reassembling of Parliament, Sir George Dibbs
was confronted with a large majority, and Mr. George H. Reid
was called to form a government on the lines suggested by the
issues of the campaign. The Council or 'upper house,'
consisting of Crown nominees for life, rejected the measures
suggested by Mr. Reid and passed by the Assembly by an
overwhelming majority, and Mr. Reid dissolved Parliament on
July 6, 1895, and appealed to the country. The election was
held on July 24, and again the issues, as set forth in the
measures, were fought out vigorously. The great leader of
protection, Sir George Dibbs, with several of his ablest
followers, was defeated, and the so-called Free Trade party
came back, much stronger than before. Thus, it was claimed
that the mandate of the people, declaring for free trade and
direct taxation, had been reaffirmed, and on the reassembling
of Parliament, on August 13, the same measure, as passed by
the Assembly and rejected by the Council, was again presented
and passed by the Assembly by a majority of 50 to 26, and
again went to the upper house. Again it was met with great
hostility, but the Government party in that chamber, having
been augmented by ten new appointments, the temper of the
house was softened and the bill was passed with some two
hundred and fifty amendments. As, the Assembly could only
accept some eighty of these without, yielding material points
… a conference was suggested, which, after several days of
discussion, agreed to a modified measure, embracing the
principle of free trade, as interpreted in this colony, and
direct taxation, and the new law goes into effect as above
stated, on January 1, 1896.
"It may be well here to remark that there are a few articles,
notably raw sugar, glucose, molasses, and treacle, upon which
the duty will be removed gradually, so as not to wantonly
disturb vested interests, but, with these exceptions, the
change is a very sweeping one."

June, 1896, page 299.
AUSTRALIA: New South Wales: A. D. 1896.

Change in the government of Norfolk Island.
Its re-annexation to New South Wales.
A change in the government of Norfolk Island was proclaimed in

November, 1896, by the Governor of New South Wales, who came
to the island, acting under directions from the British
Colonial Office, and announced that "Her Majesty's Government
has decided to appoint a resident magistrate. The object
sought is to secure the impartial administration of justice,
while leaving the local and municipal affairs of the island to
be conducted by a council representing the inhabitants. In
consideration of the fact that the Norfolk Island settlement
originally formed part of the administrative colony of New
South Wales, and that the legal business of the island and the
registration of all land titles and transfers have uniformly
been conducted by the Government departments at Sydney, Her
Majesty's Government has decided to transfer the
administration of the island to the Government of New South
Wales. The Government of New South Wales has accepted the
charge and as soon as the necessary arrangements have been
completed Norfolk Island will be administered by the governor
of New South Wales in council." "It will thus be seen that the
Pitcairn community, which, for more than one hundred years,
has governed itself by its own laws, is now abolished and that
a new era has begun. The governor's legal right to annul the
constitution given by the Queen when the community emigrated
from Pitcairn was questioned. A deputation was appointed to
wait on the governor, but he refused
to discuss the subject further."

May, 1897, page 37.
AUSTRALIA: A. D. 1897.
Conference of colonial premiers with
the British Colonial Secretary.

ENGLAND: A. D. 1897 (JUNE-JULY).
Map of Australia and Islands of the Pacific.

{31}
Federation of the Australian Colonies.
The steps by which the Union was accomplished.
Passage of the "Commonwealth of Australia Constitution Act" by
the Imperial Parliament.
"The first indication of a plan for united action among the

colonies is to be found in a proposal of Earl Grey in 1850.
The main object of the proposal was to bring about uniformity
in colonial tariffs; but, though partially adopted, it came to
nothing. From 1850 to 1860 the project of federation was
discussed from time to time in several of the colonial
legislatures, and committees on the subject were appointed.
But there seems to have been little general interest in the
question, and up to 1860 all efforts in the direction of
federation met with complete failure. Shortly after, however,
a new form of united action, less ambitious but more likely of
success, was suggested and adopted. From 1863 to 1883
conferences of colonial ministers were held at various times
to discuss certain specified topics, with a view to
introducing identical proposals in the separate colonial
legislatures. Six of these conferences were held at Melbourne
and three at Sydney; and one also was held at Hobart in 1895,
though the period of the real activity of the conference
scheme practically closed in 1883. The scheme proved a
failure, because it was found impossible to carry out the
measures concerted in the conferences. But material events
were doing more than could any public agitation to draw
attention to the advantages of closer union. The colonies were
growing in population and wealth, railroads were building and
commerce was extending. The inconveniences of border customs
duties suggested attempts at something like commercial
reciprocity between two or more colonies. New political
problems also helped to arouse public interest. Heretofore
there had been little fear of foreign aggression and, hence,
no feeling of the need of united action for common defense;
nor had there been any thought of the extension of Australian
power and interests beyond the immediate boundaries of the
different colonies. But the period from 1880 to 1890 witnessed
a change in this respect. It was during this period that much
feeling was aroused against the influx of French criminals,
escaped from the penal settlements in New Caledonia. The
difficulties in regard to New Guinea belong also to this
decade. Suspicion of the designs of Germany upon that part of
the island of New Guinea nearest the Dutch boundary led to the
annexation of its eastern portion by the Queensland
government. This action was disavowed by the British
government under Gladstone, and the fears of the colonists
were ridiculed; but almost immediately after the northern half
of New Guinea was forcibly taken possession of by Germany. The
indignation of Australians was extreme, and the opinion was
freely expressed that the colonies would have to unite to
protect their own interests. Finally, this was the time of the
French designs on the New Hebrides Islands and of German
movements with reference to Samoa. These conditions, economic
and political, affected all the colonies more or less
intimately and resulted in the first real, though loose, form
of federal union. At the instigation of the Honorable James
Service, premier of Victoria, a convention met at Sydney,
November, 1883, composed of delegates from all the colonial
governments. This convention adopted a bill providing for the
establishment of a Federal Council, with power to deal with
certain specified subjects and with such other matters as
might be referred to it by two or more colonies. … New South
Wales and New Zealand refused to agree to the bill, but it was
adopted by the other colonies; and the Imperial Parliament, in
1885, passed an act permitting such a Council to be called
into existence at the request of any three colonies, to be
joined by other colonies as they saw fit. Meetings of the
Council took place in 1886, 1888, 1889 and 1891, but very
little was accomplished. That the Federal Council was a very
weak affair is obvious. … Meanwhile, interest in a more
adequate form of federation was growing. In 1890 Sir Henry
Parkes proposed a plan for federal union of a real and
vigorous sort. At his suggestion, a conference met at
Melbourne, February 6, 1890, to decide on the best method of
getting the question into definite shape for consideration. …
Provision was made … for the calling of a convention to draw
up a constitution. … In accordance with the decision of the
conference, delegates from the several colonies convened at
Sydney, March 2, 1891; and with the work of this convention
began the third and final stage in the federation movement.
The Sydney convention formulated a bill, embodying a draft of
a federal constitution, and then resolved that provision
should be made by the several parliaments to submit it to the
people in such manner as each colony should see fit. … But
there was no sufficient external pressure to bring about an
immediate discussion and an early settlement. … The result was
that nothing was done. … Meanwhile, federation leagues had
been organized in different colonies, and in 1893 delegates
from a number of these leagues met at Bendigo, Victoria. …
After adopting the bill of 1891 as a basis of discussion, the
Bendigo conference resolved to urge the colonial governments
to pass uniform enabling acts for a new convention—its members
to be elected by popular vote—to frame a constitution which
should be submitted to the people for approval. This proposal
met with general favor and resulted in the calling of a
meeting of the premiers of all the colonies at Hobart in
January, 1895. There an enabling bill was drafted which five
premiers agreed to lay before their respective parliaments. …
It took two years to get this machinery into working order. At
length, however, the requisite authority was granted by five
colonies: New South Wales, Victoria, South Australia, Western
Australia and Tasmania, Queensland and New Zealand declining
to participate. On March 22, 1897, the second constitutional
convention assembled at Adelaide. This convention drew up a
new federal constitution, based upon the draft of 1891.
Between May 5 and September 2 the constitution was discussed
in each of the parliaments. When the convention reassembled at
Sydney on March 2, as many as 75 amendments were reported as
suggested by the different colonies. Many were of an
insignificant character and many were practically identical.
The constitution and proposed amendments were discussed in two
sessions of the convention, which finally adjourned March 16,
1898, its work then being ready to submit to the people.
{32}
In June a popular vote resulted in the acceptance of the
constitution by Victoria, Tasmania, and South Australia; but
the failure of the parent colony, New South Wales, to adopt it
blocked all hope of federal union for the moment. Recently,
however, at a conference of colonial premiers certain
amendments demanded by New South Wales were agreed to in part,
and upon a second vote the constitution, as amended, was
accepted by that colony."
W. G. Beach,
The Australian Federal Constitution
(Political Science Quarterly, December, 1899).
In August, 1899, the draft of a Constitution thus agreed upon

was transmitted to England, with addresses from the provincial
legislatures, praying that it be passed into law by the
Imperial Parliament. Early in the following year delegates
from the several colonies were sent to England to discuss with
the Colonial Office certain questions that had arisen, and to
assist in procuring the passage by Parliament of the necessary
Act. Looked at from the Imperial standpoint, a number of
objections to the draft Constitution were found, but all of
them were finally waived excepting one. That one related to a
provision touching appeals from the High Court of the
Australian Commonwealth to the Queen in Council. As framed and
adopted in Australia, the provision in question was as
follows:
"74. No appeal shall be permitted to the Queen in Council in
any matter involving the interpretation of this Constitution
or of the Constitution of a State, unless the public interests
of some part of Her Majesty's Dominions, other than the
Commonwealth or a State, are involved. Except as provided in
this section, this Constitution shall not impair any right
which the Queen may be pleased to exercise, by virtue of Her
Royal Prerogative, to grant special leave of appeal from the
High Court to Her Majesty in Council. But The Parliament may
make laws limiting the matters in which such leave may be
asked."
This was objected to on several grounds, but mainly for the

reasons thus stated by Mr. Chamberlain: "Proposals are under
consideration for securing a permanent and effective
representation of the great Colonies on the Judicial
Committee, and for amalgamating the Judicial Committee with
the House of Lords, so as to constitute a Court of Appeal from
the whole British Empire. It would be very unfortunate if
Australia should choose this moment to take from the Imperial
Tribunal the cognizance of the class of cases of greatest
importance, and often of greatest difficulty. Article 74
proposes to withdraw from the Queen in Council matters
involving the interpretation of the Constitution. It is
precisely on questions of this kind that the Queen in Council
has been able to render most valuable service to the
administration of law in the Colonies, and questions of this
kind, which may sometimes involve a good deal of local
feeling, are the last that should be withdrawn from a Tribunal
of appeal with regard to which there could not be even a
suspicion of prepossession. Questions as to the constitution
of the Commonwealth or of a State may be such as to raise a
great deal of public excitement as to the definition of the
boundaries between the powers of the Commonwealth Parliament
and the powers of the State Parliaments. It can hardly be
satisfactory to the people of Australia that in such cases,
however important and far-reaching in their consequences, the
decision of the High Court should be absolutely final. Before
long the necessity for altering the Constitution in this
respect would be felt, and it is better that the Constitution
should be enacted in such a form as to render unnecessary the
somewhat elaborate proceedings which would be required to
amend it."
Great Britain, Parliamentary Publications

(Papers by Command, April and May, 1900,
Australia—Cd. 124 and 158).
In reply, the Australian delegates maintained that they had no

authority to amend, in any particular, the instrument which
the people of the several colonies had ratified by their
votes; but the Imperial authorities were inflexible, and the
article 74 was modified in the Act which passed Parliament, on
the 7th of July, 1900, "to constitute the Commonwealth of
Australia," as may be seen by reference to the text, published
elsewhere.

CONSTITUTION OF AUSTRALIA.
The question of the Federal Capital.
By the Constitution of the Commonwealth, it is required that

the seat of government "shall be determined by the Parliament,
and shall be within territory which shall have been granted to
or acquired by the Commonwealth, and shall be vested in and
belong to the Commonwealth, and shall be in the State of New
South Wales, and be distant not less than one hundred miles
from Sydney;" and "such territory shall contain an area of not
less than one hundred square miles." "New South Wales," says a
correspondent, writing from Sydney, "is naturally anxious to
get the question decided as quickly as possible; but Victoria
will equally be inclined to procrastinate, and the new
Parliament—which cannot be more comfortable than it will be
at Melbourne—will not be in a hurry to shift. The necessity
for a new and artificial capital arises entirely out of our
provincial jealousies, and it would have been a great saving
of initial expense and a great diminution of inconvenience if
we could have used one of the old capitals for a quarter of a
century." To remove preliminary difficulties and avoid delay,
the government of New South Wales appointed a commissioner to
visit and report on the most likely places. The report of this
commissioner, made early in October, "reduces the possible
positions to three—one near Bombala in the south-east corner
of the colony at the foot of the Australian Alps, one near
Yass on the line of the railway between Sydney and Melbourne,
and one near Orange on our western line. On the whole he gives
the preference to the first named."
AUSTRALIA: New South Wales: A. D. 1900.

Old-Age Pension Act.
A letter from Sydney, November 29, 1900, announced: "The

question of the establishment of an old age pension system,
similar to that now in successful operation in New Zealand
[see (in this volume) NEW ZEALAND: A. D. 1899], has been
agitating New South Wales for several months, and to-day the
bill for that purpose became a law. There has been a desire on
the part of some members of the Legislature to hold over the
bill until the convening of the Federal Parliament, in the
hope that the measure would become universal throughout the
continent, but the majority, including the Premier, wished the
bill to be pushed through without loss of time. There is no
opposition worth mentioning. … At a mass meeting in favor of
the bill representatives of every political party, of every
Church and of every profession and trade in the community were
present.
{33}
The sentiment of the colony has never been more unanimous. …
The estimated cost of the scheme is something like £250,000 or
£300,000 a year, but this does not take into consideration the
amount which will be saved by doing away with the charitable
institutions now draining the pockets alike of the state and
of the individual. Private contributions alone amount to
£600,000 a year; all this will be saved, together with a part
of the Government's annual expenditure—about £400,000—for
public institutions. Not all pauper institutions can be
abolished, for many of the aged and friendless poor are ailing
or slightly feeble minded, and will continue to need medical
attention."
AUSTRALIA: A. D. 1900 (March).

New Zealand looking toward federation with the Australian
Commonwealth.

NEW ZEALAND: A. D. 1900 (MARCH).
AUSTRALIA: West Australia: A. D. 1900 (August).

Vote to join the Commonwealth.
The question of union with the other colonies in the

Commonwealth, from which the West Australians had previously
held aloof, was submitted to them in August (women voting for
the first time), and decided affirmatively by 44,704 against
19,691. Adding the West Australian totals to the aggregate
vote at the decisive referendum in each of the other
federating colonies, the following is the reported result:
For federation. 422,647

Against federation. 161,024
Majority. 261,623
AUSTRALIA: A. D. 1900 (September-December).

The Queen's Proclamation of the Australian Commonwealth.
Contemplated visit of the Duke and Duchess of York
to open the first session of the Federal Parliament.
Appointment of Lord Hopetoun to be Governor-General.
The first Federal Cabinet.
On the 17th of September the following proclamation of the

Australian Commonwealth was issued by the Queen:
"Whereas by an Act of Parliament passed in the sixty-third and

sixty-fourth years of Our reign, intituled 'An Act to
constitute the Commonwealth of Australia,' it is enacted that
it shall be lawful for the Queen, with the advice of the Privy
Council, to declare by Proclamation that, on and after a day
therein appointed, not being later than one year after the
passing of this Act, the people of New South Wales, Victoria,
South Australia, Queensland, and Tasmania, and also, if Her
Majesty is satisfied that the people of Western Australia have
agreed thereto, of Western Australia, shall be united in a
Federal Commonwealth, under the name of the Commonwealth of
Australia. And whereas We are satisfied that the people of
Western Australia have agreed thereto accordingly. We
therefore, by and with the advice of Our Privy Council, have
thought fit to issue this Our Royal Proclamation, and We do
hereby declare that on and after the first day of January, one
thousand nine hundred and one, the people of New South Wales,
Victoria, South Australia, Queensland, Tasmania, and Western
Australia shall be united in a Federal Commonwealth under the
name of the Commonwealth of Australia. Given at Our Court at
Balmoral, this seventeenth day of September, in the year of
our Lord one thousand nine hundred, and in the sixty-fourth
year of Our reign. God save the Queen."
At the same time, the following announcement, which caused

extreme delight in Australia, was published officially from
the Colonial Office:
"Her Majesty the Queen has been graciously pleased to assent,

on the recommendation of the Marquis of Salisbury, to the

Full Chapter Geometry and Analysis of Metric Spaces Via Weighted Partitions Jun Kigami PDF

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

Full Chapter Geometry and Analysis of Metric Spaces Via Weighted Partitions Jun Kigami PDF

Uploaded by

Copyright:

Available Formats

Geometry and Analysis of Metric

Spaces via Weighted Partitions Jun

Functional Analysis: An Introduction to Metric Spaces,

A Primer on Hilbert Space: Theory Linear Spaces,

A Primer on Hilbert Space Theory Linear Spaces

Geometry and Martingales in Banach Spaces 1st Edition

An Invitation to Alexandrov Geometry CAT 0 Spaces

Spaces An Introduction to Real Analysis Tom Lindstrøm

Riemannian Geometry and Geometric Analysis 7th Edition

Analysis on Fock spaces and mathematical theory of

More information about this series at http://www.springer.com/series/304

Geometry and Analysis

ISSN 0075-8434 ISSN 1617-9692 (electronic)

Mathematics Subject Classification: 30F45, 53C23

of a compact metric space is given as the critical index p of p-energies associated

Kyoto, Japan Jun Kigami

1 Introduction and a Showcase . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1

A Fact from Measure Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 153

Successive divisions of a space have played important roles in many areas of

Ki1 ...im = Ki1 ...im 0 ∪ Ki1 ...im 1 (1.1.1)

A partition of [0, 1] The associated tree

for any infinite sequence

Such successive divisions of a compact metric space, which may not be as

(iv) The totality of edges of T is {(π(q), q)|q ∈ T \{φ}}.

(v) For any infinite sequence (p0 , p1 , p2 , . . .) of vertices of X satisfying p0 = φ

ρ (π(p)) ≥ ρ (p) (1.1.5)

for any p ∈ T \{φ} and

lim ρ (pi ) = 0 (1.1.6)

c1 r α ≤ μ(Bd (x, r)) ≤ c2 r α ,

bi-Lipschitz = Ahlfors regularity = being adapted (1.1.7)

the volume doubling property = quasisymmetry. (1.1.8)

c1 ρ1 (p) ≤ ρ2 (p) ≤ c2 ρ1 (p)

dimAR (X, d) = inf{α|there exist a metric ρ on X and a Borel regular measure μ on

1 Thisidea of characterizing AR conformal dimension by p-energies was brought to the author by

Theorem 1.1.1 (Bonk-Kleiner) Let G be a Gromov hyperbolic group whose

the rationalized version of (1.1.8) as a mathematical statement. In particular, in

min A = max A = sup A = inf A = 0.

Moreover, if f : A → R and A = ∅, then we set

1.2 Summary of the Main Results; the Case of 2-dim. Sphere

U = {A|A ⊆ X, closed, int(A) = ∅, ∂A is homeomorphic to the circle S 1 .}

First we divide X into finite number of subsets X1 , . . . , XN0 belonging to U, i.e.

We assume that Xi ∩ Xj = ∂Xi ∩ ∂Xj if i = j . Next each Xi is divided into

and if i1 . . .ik = j1 . . .jk , then

Xi1 ...ik ∩ Xj1 ...jk = ∂Xi1 ...ik ∩ ∂Xj1 ...jk . (1.2.2)

Note that (1.2.1) is a counterpart of (1.1.3). Next define

Tk = {i1 . . .ik |ij ∈ {1, . . . , Ni1 ...ij−1 } for any j = 1, . . . , k − 1}

= {i1 i2 . . . |i1 . . . ik ∈ Tk for any k ≥ 0},

which is the “boundary” of the infinite tree T .

#(σ −1 (x)) < +∞, (1.2.3)

where #(A) is the number of elements in a set A.

lim sup g(w) = 0.

G(T ) = {g|g : T → (0, 1] is a weight function.}

g(i1 . . .ik ik+1 ) ≥ λg(i1 . . .ik )

for any i1 . . .ik ∈ T and i1 . . .ik ik+1 ∈ T .

g(i1 . . .ik ik+1 . . .ik+m ) ≤ γ g(i1 . . .ik )

for any i1 . . .ik ∈ T and i1 . . .ik ik+1 . . .ik+m ∈ T .

Ge (T ) = {g|g : T → (0, 1] is a weight function

Metrics and measures on X naturally have associated weight functions.

D(X) = {d|d is a metric on X which produces the original topology of X,

M(X) = {μ|μ is a Borel regular probability measure on X, μ({x}) = 0

for s ∈ (0, 1] and

for M ≥ 0, x, y ∈ X. A weight function is called uniformly finite if and only if

Definition 1.2.5 Let M ≥ 1

ρ (π(p)) ≥ ρ (p) (1.1.5)

lim ρ (pi ) = 0 (1.1.6)

Hence we see that |w ∧ v| = |w| or |w ∧ v| = |v|.