High Dimensional Probability VIII: The Oaxaca Volume Nathael Gozlan All Chapter Instant Download

Download the full version of the textbook now at textbookfull.
com
High Dimensional Probability VIII: The Oaxaca

Volume Nathael Gozlan
https://textbookfull.com/product/high-dimensional-
probability-viii-the-oaxaca-volume-nathael-gozlan/
Explore and download more textbook at https://textbookfull.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
Coulson and Richardson’s Chemical Engineering, Fourth

Edition: Volume 3A: Chemical and Biochemical Reactors and
Reaction Engineering R. Ravi
https://textbookfull.com/product/coulson-and-richardsons-chemical-
engineering-fourth-edition-volume-3a-chemical-and-biochemical-
reactors-and-reaction-engineering-r-ravi/
textbookfull.com
VIII International Scientific Siberian Transport Forum

TransSiberia 2019 Volume 2 Zdenka Popovic
https://textbookfull.com/product/viii-international-scientific-
siberian-transport-forum-transsiberia-2019-volume-2-zdenka-popovic/
textbookfull.com
High-dimensional Microarray Data Analysis: Cancer Gene

Diagnosis and Malignancy Indexes by Microarray Shuichi
Shinmura
https://textbookfull.com/product/high-dimensional-microarray-data-
analysis-cancer-gene-diagnosis-and-malignancy-indexes-by-microarray-
shuichi-shinmura/
textbookfull.com
Teach Yourself VISUALLY iPhone 6s Covers iOS 9 and all

models of iPhone 6s 6 and iPhone 5 Guy Hart-Davis
https://textbookfull.com/product/teach-yourself-visually-
iphone-6s-covers-ios-9-and-all-models-of-iphone-6s-6-and-iphone-5-guy-
hart-davis/
textbookfull.com
Animalism: new essays on persons, animals, and identity
1st Edition Stephan Blatti
https://textbookfull.com/product/animalism-new-essays-on-persons-
animals-and-identity-1st-edition-stephan-blatti/
textbookfull.com
Training for Change Transforming Systems to be Trauma

Informed Culturally Responsive and Neuroscientifically
Focused Alisha Moreland-Capuia
https://textbookfull.com/product/training-for-change-transforming-
systems-to-be-trauma-informed-culturally-responsive-and-
neuroscientifically-focused-alisha-moreland-capuia/
textbookfull.com
Who Is Vera Kelly Rosalie Knecht
https://textbookfull.com/product/who-is-vera-kelly-rosalie-knecht/
textbookfull.com
New Essays on Aristotle s Organon Routledge Monographs in

Classical Studies 1st Edition António Pedro Mesquita
https://textbookfull.com/product/new-essays-on-aristotle-s-organon-
routledge-monographs-in-classical-studies-1st-edition-antonio-pedro-
mesquita/
textbookfull.com
The Rise And Development Of FinTech: Accounts Of

Disruption From Sweden And Beyond Robin Teigland
https://textbookfull.com/product/the-rise-and-development-of-fintech-
accounts-of-disruption-from-sweden-and-beyond-robin-teigland/
textbookfull.com
Unlocking the Secrets of Prompt Engineering Gilbert
Mizrahi
https://textbookfull.com/product/unlocking-the-secrets-of-prompt-
engineering-gilbert-mizrahi/
textbookfull.com
Progress in Probability
74
Nathael Gozlan
Rafał Latała
Karim Lounici
Mokshay Madiman
Editors
High Dimensional
Probability VIII
The Oaxaca Volume
Volume 74
Series Editors
Steffen Dereich, Universität Münster, Münster, Germany
Davar Khoshnevisan, The University of Utah, Salt Lake City, UT, USA
Andreas E. Kyprianou, University of Bath, Bath, UK
Sidney I. Resnick, Cornell University, Ithaca, NY, USA
Progress in Probability is designed for the publication of workshops, seminars

and conference proceedings on all aspects of probability theory and stochastic
processes, as well as their connections with and applications to other areas such
as mathematical statistics and statistical physics.
More information about this series at http://www.springer.com/series/4839

Nathael Gozlan • Rafał Latała • Karim Lounici •
Mokshay Madiman
Editors
High Dimensional
Probability VIII
The Oaxaca Volume
Editors
Nathael Gozlan Rafał Latała
MAP 5 Institute of Mathematics
Université Paris Descartes University of Warsaw
Paris, France Warsaw, Poland
Karim Lounici Mokshay Madiman

Centre de Mathématiques Appliquées Department of Mathematical Sciences
Ecole Polytechnique University of Delaware
Palaiseau, France Newark, DE, USA
ISSN 1050-6977 ISSN 2297-0428 (electronic)

ISBN 978-3-030-26390-4 ISBN 978-3-030-26391-1 (eBook)
https://doi.org/10.1007/978-3-030-26391-1
© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved 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 book is published under the imprint Birkhäuser, www.birkhauser-science.com, by the registered
company Springer Nature Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The history of the High-Dimensional Probability (HDP) conferences dates back to

the 1975 International Conference on Probability in Banach Spaces in Oberwolfach,
Germany. After eight Probability in Banach Spaces meetings, in 1994 it was
decided to give the series its current name: the International Conference on High-
Dimensional Probability.
The present volume is an outgrowth of the Eighth High-Dimensional Probability
Conference (HDP VIII), which was held at the Casa Matemática Oaxaca (Mexico)
from May 28th to June 2nd, 2017. The scope and quality of the talks and contributed
papers amply demonstrate that, now more than ever, high-dimensional probability
is a very active area of mathematical research.
High-Dimensional Probability has its roots in the investigation of limit theorems
for random vectors and regularity of stochastic processes. It was initially motivated
by the study of necessary and sufficient conditions for the boundedness and
continuity of trajectories of Gaussian processes and the extension of classical limit
theorems, such as laws of large numbers, laws of the iterated logarithm and central
limit theorems, to Hilbert and Banach space-valued random variables and empirical
processes.
This resulted in the creation of powerful new tools: the methods of high-
dimensional probability and especially its offshoots, the concentration of measure
phenomenon and generic chaining techniques, were found to have a number of
applications in various areas of mathematics, as well as statistics and computer
science. These include random matrix theory, convex geometry, asymptotic geomet-
ric analysis, nonparametric statistics, empirical process theory, statistical learning
theory, compressed sensing, strong and weak approximations, distribution function
estimation in high dimensions, combinatorial optimization, random graph theory,
stochastic analysis in infinite dimensions, and information and coding theory.
In recent years there has been substantial progress in the area. In particu-
lar, numerous important results have been obtained concerning the connections
between various functional inequalities related to the concentration of measure
phenomenon, application of generic chaining methods to study the suprema of
stochastic processes and norms of random matrices, Malliavin–Stein theory of
v
vi Preface
Gaussian approximation, various stochastic inequalities and their applications in

high-dimensional statistics and computer science. This breadth is duly reflected by
the diverse contributions in the present volume.
The majority of the papers gathered here were presented at HDP VIII. The
conference participants wish to express their gratitude for the support provided
by the BIRS-affiliated mathematics research center Casa Matemática Oaxaca. In
addition, the editors wish to thank Springer-Verlag for publishing the proceedings.
The book begins with a dedication to our departed and esteemed colleague,
Jørgen Hoffmann-Jørgensen, whom we lost in 2017. This is followed by a collection
of contributed papers that are divided into four general areas: inequalities and
convexity, limit theorems, stochastic processes, and high-dimensional statistics. To
give readers an idea of their scope, in the following we briefly describe them by
subject area and in the order they appear in this volume.
Dedication to Jørgen Hoffmann-Jørgensen (1942–2017)
• Jørgen Hoffmann-Jørgensen, by M. B. Marcus, G. Peskir and J. Rosiński.
This paper honors the memory, scientific career and achievements of Jørgen
Hoffmann-Jørgensen.
Inequalities and Convexity
• Moment estimation implied by the Bobkov-Ledoux inequality, by W. Bednorz and
G. Głowienko. The authors derive general bounds for exponential Orlicz norms
of locally Lipschitz functions using the Bobkov-Ledoux entropic form of the
Poincaré inequality.
• Polar isoperimetry I—the case of the plane, by S. G. Bobkov, N. Gozlan,
C. Roberto and P.-M. Samson. This is the first part of a lecture notes series
and offers preliminary remarks on the plane isoperimetric inequality and its
applications to the Poincaré and Sobolev type inequalities in dimension one.
• Iterated Jackknives and two-sided variance inequalities, by O. Bousquet and
C. Houdré. The authors revisit selected classical variance inequalities, such as
the Efron–Stein inequality, and present refined versions.
• A probabilistic characterization of negative definite functions, by F. Gao. The
author proves using Fourier transform tools that a continuous function f on Rn
is negative definite if and only if it is polynomially bounded and satisfies the
inequality
Ef (X − Y ) ≤ Ef (X + Y )
for all i.i.d. random vectors X and Y in Rn .

• Higher order concentration in presence of Poincaré type inequalities, by F. Götze
and H. Sambale. The authors obtain sharpened forms of the concentration
of measure phenomenon that typically apply to differentiable functions with
centered derivatives up to the order d − 1 and bounded derivatives of order d.
Preface vii
• Rearrangement and Prékopa–Leindler type inequalities, by J. Melbourne. The

author obtains rearrangement sharpenings of several classical Prékopa–Leindler
type functional inequalities.
• Generalized semimodularity: order statistics, by I. Pinelis. The author
introduces a notion of generalized n-semimodularity, which extends that of
(sub/super)modularity, and derives applications to correlation inequalities for
order statistics.
• Geometry of np -balls: Classical results and recent developments, by J. Prochno,
C. Thäle and N. Turchi. The paper presents a survey of asymptotic theorems for
uniform measures on np -balls and cone measures on np -spheres.
• Remarks on superconcentration and Gamma calculus. Applications to spin
glasses, by K. Tanguy. This paper explores applications of Bakry-Emery Γ2
calculus to refined variant inequalities for several spin systems models.
Limit Theorems
• Asymptotic behavior of Renyi entropy in the central limit theorem, by
S. G. Bobkov and A. Marsiglietti. The authors explore the asymptotic behavior
and monotonicity of Renyi entropy along convolutions in the central limit
theorem.
• Uniform-in-bandwidth functional limit laws for multivariate empirical processes,
by P. Deheuvels. The author provides uniform-in-bandwidth functional limit laws
for multivariate local empirical processes, with statistical applications to kernel
density estimation.
• Universality of limiting spectral distribution under projective criteria, by
F. Merlevède and M. Peligrad. The authors study the limiting empirical spectral
distribution of an n × n symmetric matrix with dependent entries. For a class of
generalized martingales, they show that the asymptotic behavior of the empirical
spectral distribution depends only on the covariance structure.
• Exchangeable pairs on Wiener chaos, by I. Nourdin and G. Zheng. In this paper,
the authors propose a new proof of a quantitative form of the fourth moment
theorem in Gaussian approximation based on the construction of exchangeable
pairs of Brownian motions.
Stochastic Processes
• Permanental processes with kernels that are equivalent to a symmetric matrix,
by M. B. Marcus and J. Rosen. The authors consider α-permanental processes
whose kernel is of the form

u(x, y) = u(x, y) + f (y), x, y ∈ S,
where u is symmetric and f has some good properties. In turn, they define con-
ditions that determine whether the kernel
u is symmetrizable or asymptotically
symmetrizable.
• Pointwise properties of martingales with values in Banach function spaces,
by M. Veraar and I. Yaroslavtsev. In this paper, the authors consider local
viii Preface
martingales with values in a UMD Banach function space and prove that
such martingales have a version which is a martingale field. Moreover, a new
Burkholder–Davis–Gundy type inequality is obtained.
High-Dimensional Statistics
• Concentration inequalities for randomly permuted sums, by M. Albert. The
author proves a deviation inequality for random permutations and uses it to
analyze the second kind error rate in a test of independence.
• Uncertainty quantification for matrix compressed sensing and quantum tomog-
raphy problems, by A. Carpentier, J. Eisert, D. Gross and R. Nickl. The authors
construct minimax optimal non-asymptotic confidence sets for low-rank matrix
recovery algorithms such as the Matrix Lasso and Dantzig selector.
• Uniform-in-bandwidth estimation of the gradient lines of a density, by D. Mason
and B. Pelletier. This paper exploits non parametric statistical techniques to
estimate the gradient flow of a stochastic differential equation. The results can
be of interest in clustering applications or the analysis of stochastic gradient
schemes.
Paris, France Nathael Gozlan

Warsaw, Poland Rafał Latała
Palaiseau, France Karim Lounici
Newark, DE, USA Mokshay Madiman
Contents
1 Jørgen Hoffmann-Jørgensen (1942–2017) . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1

Michael B. Marcus, Goran Peskir, and Jan Rosiński
2 Moment Estimation Implied by the Bobkov-Ledoux Inequality . . . . . . 9
Witold Bednorz and Grzegorz Głowienko
3 Polar Isoperimetry. I: The Case of the Plane . . . . . . . .. . . . . . . . . . . . . . . . . . . . 21
Sergey G. Bobkov, Nathael Gozlan, Cyril Roberto, and Paul-Marie
Samson
4 Iterated Jackknives and Two-Sided Variance Inequalities .. . . . . . . . . . . . 33
Olivier Bousquet and Christian Houdré
5 A Probabilistic Characterization of Negative Deﬁnite Functions . . . . . 41
Fuchang Gao
6 Higher Order Concentration in Presence of Poincaré-Type
Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 55
Friedrich Götze and Holger Sambale
7 Rearrangement and Prékopa–Leindler Type Inequalities . . . . . . . . . . . . . 71
James Melbourne
8 Generalized Semimodularity: Order Statistics. . . . . .. . . . . . . . . . . . . . . . . . . . 99
Iosif Pinelis
9 Geometry of np -Balls: Classical Results and Recent
Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 121
Joscha Prochno, Christoph Thäle, and Nicola Turchi
10 Remarks on Superconcentration and Gamma Calculus:
Applications to Spin Glasses. . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 151
Kevin Tanguy
ix
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
x Contents
11 Asymptotic Behavior of Rényi Entropy in the Central Limit

Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 169
Sergey G. Bobkov and Arnaud Marsiglietti
12 Uniform-in-Bandwidth Functional Limit Laws for Multivariate
Empirical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 201
Paul Deheuvels
13 Universality of Limiting Spectral Distribution Under
Projective Criteria .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 241
Florence Merlevède and Magda Peligrad
14 Exchangeable Pairs on Wiener Chaos. . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 277
Ivan Nourdin and Guangqu Zheng
15 Permanental Processes with Kernels That Are Not Equivalent
to a Symmetric Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 305
Michael B. Marcus and Jay Rosen
16 Pointwise Properties of Martingales with Values
in Banach Function Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 321
Mark Veraar and Ivan Yaroslavtsev
17 Concentration Inequalities for Randomly Permuted Sums . . . . . . . . . . . . 341
Mélisande Albert
18 Uncertainty Quantification for Matrix Compressed Sensing
and Quantum Tomography Problems .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 385
Alexandra Carpentier, Jens Eisert, David Gross, and Richard Nickl
19 Uniform in Bandwidth Estimation of the Gradient Lines
of a Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 431
David Mason and Bruno Pelletier
Jørgen Hoffmann-Jørgensen (1942–2017)
Michael B. Marcus, Goran Peskir, and Jan Rosiński
Jørgen Hoffmann-Jørgensen, docent emeritus in the Department of Mathematics at

Aarhus University, Denmark, died on the 8th of December 2017. He was 75 years
old. He is survived by Karen, his wife of fifty years, his mother Ingeborg, his brother
Bent and his niece Dorthe.
He was a devoted teacher and advisor, a wonderful, friendly person, and a very
fine and prolific mathematician. His ties to Aarhus are legendary. Jørgen received his
magister scientiarum degree from the Institute of Mathematics at Aarhus University
in 1966. He began his research and teaching there in the previous year and continued
through the academic ranks, becoming docent in 1988.
With a stroke of good luck he began his career as a probabilist under the most
auspicious circumstances. Kiyoshi Itô was a professor at Aarhus from 1966 to 1969.
Ron Getoor, who had been with Itô at Princeton, came to Aarhus as a visiting
professor in the spring semester of 1969. Jørgen began his research career in the
presence of these outstanding probabilists. He often commented that, more than any
other mathematician, Itô had the greatest influence on his work.
There was widespread interest in sums of independent Banach space valued
random variables at that time. The famous paper of Itô and Nisio, ‘On the
convergence of sums of independent Banach space valued random variables’,
appeared in 1968. Jean-Pierre Kahane’s book, ‘Some random series of functions’
M. B. Marcus
CUNY Graduate Center, New York, NY, USA
e-mail: [email protected]
G. Peskir
Department of Mathematics, The University of Manchester, Manchester, UK
J. Rosiński ()
Department of Mathematics, University of Tennessee, Knoxville, TN, USA
© Springer Nature Switzerland AG 2019 1

N. Gozlan et al. (eds.), High Dimensional Probability VIII,
Progress in Probability 74, https://doi.org/10.1007/978-3-030-26391-1_1
2 M. B. Marcus et al.
(first edition), mostly dealing with random Fourier series, also came out in 1968.
Functional analysts in the circle of Laurent Schwartz were using properties of sums
of independent Banach space valued random variables to classify Banach spaces.
Engaged in this work, Jørgen published his most cited papers, ‘Sums of
independent Banach space valued random variables’, as a publication of the Institute
of Mathematics in Aarhus in 1972, and a paper with the same title, in Studia
Mathematica in 1974 (cf. [9]). The two papers overlap but each has material that
is not in the other. They contain the important and very useful relationship, between
the norm of the maximal term in a series and the norm of the series, that is now
commonly referred to as ‘Hoffmann-Jørgensen’s inequality’.
Continuing in this study, Jørgen collaborated on two important papers; with
Gilles Pisier on the law of large numbers and the central limit theorem in Banach
spaces [12], and with Richard Dudley and Larry Shepp on the lower tails of
Gaussian seminorms [13]. He returned repeatedly to the topics of these and his
other early papers, examining them in more general and abstract spaces. In this vein
Jørgen reexamined the concept of weak convergence from a new perspective that
completely changed the paradigm of its applications in statistics. He formulated his
new definition of weak convergence in the 1980s1. This is now referred to as ‘weak
convergence in Hoffmann-Jørgensen’s sense’.
Jørgen remained an active researcher throughout his life. He was completing a
paper with Andreas Basse-O’Connor and Jan Rosiński on the extension of the Itô-
Nisio theorem to non-separable Banach spaces, when he died.
Jørgen was also a very fine teacher and advisor with great concern for his
students. He wrote 10 sets of lecture notes for his courses, 2,620 pages in total, and
a monumental 1,184 page, two volume, ‘Probability with a view toward Statistics’,
published by Chapman and Hall in 1994. He was the principal advisor of seven
Ph.D. students.
Reflecting the interest in sums of independent Banach space valued random
variables, and the related field of Gaussian processes in Europe, Laurent Schwarz
and Jacques Neveu organized an auspicious conference on Gaussian Processes in
Strasbourg in 1973. This stimulated research and collaborations that continue to
this day. The Strasbourg conference was followed, every two or three years, by
nine conferences on Probability in Banach Spaces and eight conferences on High
Dimensional Probability. The last one was in Oaxaca, Mexico in 2017. The change
in the conference name reflected a broadening of the interests of the participants.
Jørgen was one of a core group, many of whom attended the 1973 conference,
who took part in all or most of the eighteen conferences throughout their careers,
and often were the conference organizers and editors of the conference proceedings.
Most significantly, Jørgen was the principal organizer of three of these conferences
in the beautiful, serene, conference center in Sandbjerg, Denmark in 1986, 1993
and 2002, and was an editor of the proceedings of these conferences. Moreover, his
1 Some authors have claimed, as we did in [14], that this definition was introduced in Jørgen’s paper
Probability in Banach space [10] in 1977. However, after a careful reading of this paper, we do not
think that this is correct.
1 Jørgen Hoffmann-Jørgensen (1942–2017) 3
influence on the study of probability in Europe extended beyond these activities. In

total, Jørgen served on the conference committees of eighteen meetings in Croatia,
Denmark, Italy, France and Germany. Jørgen also served as an editor of the Journal
of Theoretical Probability.
Jørgen was one of the mathematicians at Aarhus University who made Aarhus
a focal point for generations of probabilists. But it was not only the research that
brought them to Aarhus. Just as important was Jørgen’s warmth and wit and not
least of all the wonderful hospitality he and his wife Karen extended to all of them.
Who can forget the fabulous Danish meals at their house, and then, sitting around
after dinner, exchanging mathematical gossip and arguing politics, with the mating
calls of hump backed whales playing in the background2.
We now present some of Jørgen’s better known results. This is not an attempt to
place him in the history of probability but merely to mention some of his work that
has been important to us and to give the reader a glimpse of his achievements.
Hoffmann-Jørgensen’s Inequality Let (Xn ) be a sequence of independent sym-
metric random variables with values in a Banach space E with norm · . We define

n
Sn = Xj , N = sup Xn , M = sup Sn .
n n
j =1
Hoffmann-Jørgensen’s inequality states that
P(M ≥ 2t + s) ≤ 2P(N ≥ s) + 8P2 (M ≥ t) (1.1)
for all t, s > 0.

Note that since probabilities are less than 1 and the last term in this inequality
is a square it suggests that if M has sufficient regularity the distribution of M is
controlled by the distribution of N. This is a remarkable result.
Jørgen gives this inequality in his famous paper [9]. He does not highlight it. It
simply appears in the proof of his Theorem 3.1 which is:
Theorem 1 Let (Xn ) be a sequence of independent E-valued random variables
such that
P(M < ∞) = 1 and E(N p ) < ∞
for some 0 < p < ∞. Then E(M p ) < ∞.
2 The material up to this point has appeared in [14].

This is how he uses the inequality to prove this theorem. Assume that the
elements of (Xn ) are symmetric and let R(t) = P(M ≥ t) and Q(t) = P(N ≥ t)
for t ≥ 0. Using the relationship
∞
E(M p ) = px p−1 R(x)dx,
0
and similarly for N and Q, it follows from (1.1) that for A > 0
A A/3
px p−1 R(x)dx = p 3p px p−1 R(3x)dx (1.2)
0 0
A/3 A/3
≤ 2p 3p px p−1 Q(x)dx + 8p 3p px p−1 R 2 (x)dx
0 0
A/3
≤ 2p 3p E(N p ) + 8p 3p px p−1 R 2 (x)dx.
0
Choose t0 > 0 such that R(t0 ) < (16p3p )−1 . The condition that P(M < ∞) = 1
implies that t0 < ∞. Then choose A > 3t0 . Note that
A/3 t0 A/3
px p−1 R 2 (x)dx = px p−1 R 2 (x)dx + px p−1 R 2 (x)dx
0 0 t0
A/3
p
≤ t0 + R(t0 ) px p−1 R(x)dx. (1.3)
t0
Combining (1.2) and (1.3) we get

A A/3
p 1
px p−1
R(x)dx ≤ 2p 3 E(N p p
) + t0 + px p−1 R(x)dx. (1.4)
0 2 0
It follows from (1.4) that when the elements of (Xn ) are symmetric and E(N p ) <
∞, then E(M p ) < ∞. Eliminating the condition that (Xn ) is symmetric is routine.
Inequalities for sums of independent random variables that relate the sum to
the supremum of the individual terms are often referred to as Hoffmann-Jørgensen
type inequalities. Jørgen’s original inequality has been generalized and extended.
Many of these results are surveyed in [5] which obtains Hoffmann-Jørgensen type
inequalities for U statistics. See [4] for a more recent treatment of Hoffmann-
Jørgensen type inequalities in statistics.
Weak Convergence in Hoffmann-Jørgensen’s Sense The classic concept of
convergence in distribution, dating back to de Moivre’s central limit theorem in
1737, admits the following well-known characterisation, traditionally referred to as
weak convergence (cf. [3]).
Let (, F, P) be a probability space, let S be a metric (topological) space, and let
B(S) be the Borel σ -algebra on S. Let X1 , X2 , . . . and X be measurable functions
from to S with respect to F and B(S). If
lim Ef (Xn ) = Ef (X) (1.5)

n→∞
for every bounded continuous function f : S → R, then we say that Xn converges

weakly to X, and following Jørgen’s notation, write
∼
Xn → X (1.6)
as n → ∞. The expectation E in (1.5) is defined as the (Lebesgue-Stieltjes) integral

with respect to the (σ -additive) probability measure P.
The state space S in classical examples is finite dimensional, e.g. R or Rn for
n ≥ 2. The main motivation for Jørgen’s reconsideration of (1.5) and (1.6) comes
from the empirical processes theory. Recall that the empirical distribution function
is given by
1
n
Fn (t, ω) := I (ξi (ω) ≤ t) (1.7)
n
i=1
for n ≥ 1, t ∈ [0, 1] and ω ∈ , where ξ1 , ξ2 , . . . are independent and identically

distributed random variables on taking values in [0, 1] and having the common
distribution function F . In this setting, motivated by the classical central limit
theorem, one forms the empirical process
√
Xn (t, ω) := n Fn (t, ω)−F (t) (1.8)
and aims to establish that Xn converges ‘weakly’ to a limiting process X (of a

Brownian bridge type) as n → ∞. A substantial difficulty arises immediately
because the mapping Xn : → S is not measurable when S is taken to be the
set of all right-continuous functions x : [0, 1] → R with left-limits, equipped with
the supremum norm x∞ = sup t ∈[0,1] |x(t)| as a natural choice.
Skorokhod solved this measurability problem in 1956 by creating a different
metric on S, for which the Borel σ -algebra coincides with the cylinder σ -algebra,
so that each Xn is measurable. For more general empirical processes
√ 1
n
Xn (f, ω) := n f (Xi (ω))−Ef (X1 ) (1.9)
n
i=1
indexed by f belonging to a family of functions, there is no obvious way to

extend the Skorokhod approach. Jørgen solved this measurability problem in the
most elegant way by simply replacing the first expectation E in (1.5) by the outer
expectation E∗ , which is defined by
E∗ Y = inf { EZ | Z ≥ Y is measurable } (1.10)
where Y is any (not necessarily measurable) function from to R, and leaving

the second expectation E in (1.5) unchanged (upon assuming that the limit X is
measurable).
This definition of weak convergence in Hoffmann-Jørgensen’s sense is given for
the first time in his monograph [11, page 149]. Although [11] was published in
1991, a draft of the monograph was available in Aarhus and elsewhere since 1984.
Furthermore, the first paper [1] which uses Jørgen’s new definition was published in
1985. Jørgen’s definition of weak convergence became standard soon afterwards. It
continues to be widely used.
It is now known that replacing the first E in (1.5) by E∗ is equivalent to replacing
it by EQ where Q is any finitely additive extension of P from F to 2 (see
Theorem 4 in [2] for details). This revealing equivalence just adds to both simplicity
and depth of Jørgen’s thought when opting for E∗ in his celebrated definition.
Hoffmann-Jørgensen’s Work on Measure Theory As measure theory matured,
difficult measurability problems arose in various areas of mathematics that could not
be solved in general measure spaces. Consequently, new classes of measure spaces
were introduced, such as analytic spaces, also called Souslin spaces, defined by
Lusin and Souslin and further developed by Sierpiński, Kuratowski and others. For
many years analytic spaces received little attention until important applications were
found in potential theory by Choquet and group representation theory by Mackey.
Analytic spaces were also found to be important in the theory of convex sets, and
other branches of mathematics.
Stimulated by these developments, Jørgen undertook a deep study of analytic
spaces early in his academic career, resulting in his monograph ‘The Theory
of Analytic Spaces’ [7]. This monograph contains many original, and carefully
presented results, that are hard to find elsewhere. For example, from Jørgen’s Section
Theorem, [7, Theorem 1, page 84], one can derive all of the most commonly used
section and selection theorems in the literature.
The final chapter of the monograph is devoted to locally convex vector spaces,
where it is shown that all of the locally convex spaces that are of interest to
researchers are analytic spaces. As Jørgen wrote “The importance of analytic spaces
lies in the fact that even though the category is sufficiently small to exclude all
pathological examples . . . , it is sufficiently large to include all (or almost all)
interesting and important examples of topological measure spaces.”
In one of his first papers [6] listed in Mathematical Reviews and Zentralblatt,
Jørgen investigates extensions of regenerative events to continuous state spaces, a
problem proposed to him by P.-A. Meyer. In his subsequent paper [8], he makes the
surprising observation that the existence of a measurable modification of a stochastic
process depends only on its 2-dimensional marginal distributions. He then gives
necessary and sufficient conditions for the existence of such a modification for the
process (Xt )t ∈T with values in a complete separable metric space K, expressed in

terms of the kernel
Q(s, t, A) = P((Xs , Xt ) ∈ A)
where T is a separable metric space, s, t ∈ T , and A ∈ B(K 2 ). Jørgen’s interest in

measure theory aspects of probability continued throughout his career.
References
1. N.T. Andersen, The central limit theorem for nonseparable valued functions. Z. Wahrsch. Verw.
Gebiete 70, 445–455 (1985)
2. P. Berti, P. Rigo, Convergence in distribution of nonmeasurable random elements. Ann. Probab.
32, 365–379 (2004)
3. P. Billingsley, Convergence of Probability Measures (Willey, New York, 1968)
4. E. Giné, R. Nickl, Mathematical Foundations of Infinite-Dimensional Statistical Models
(Cambridge University Press, New York, 2016)
5. R. Giné, E. Latała, J. Zinn, Exponential and moment inequalities for U-statistics, in High
Dimensional Probability II (Seattle 1999). Programs and Probability, vol. 47 (Birkhäuser,
Boston, 2000), pp. 13–38
6. J. Hoffmann-Jørgensen, Markov sets. Math. Scand. 24, 145–166 (1969)
7. J. Hoffmann-Jørgensen, The Theory of Analytic Spaces, vol. 10 (Matematisk Institut, Aarhus
Universitet, Aarhus, 1970)
8. J. Hoffmann-Jørgensen, Existence of measurable modifications of stochastic processes. Z.
Wahrsch. Verw. Gebiete 25, 205–207 (1973)
9. J. Hoffmann-Jørgensen, Sums of independent Banach space valued random variables. Stud.
Math. 52, 159–186 (1974)
10. J. Hoffmann-Jørgensen, Probability in Banach Space. Lecture Notes in Mathematics, vol. 598
(Springer, Berlin, 1977), pp. 1–186
11. J. Hoffmann-Jørgensen, Stochastic Processes on Polish Spaces, vol. 39 (Matematisk Institut,
Aarhus Universitet, Aarhus, 1991)
12. J. Hoffmann-Jørgensen, G. Pisier, The law of large numbers and the central limit theorem in
Banach spaces. Ann. Probab. 4, 587–599 (1976)
13. J. Hoffmann-Jørgensen, L.A. Shepp, R.M. Dudley, On the lower tail of Gaussian seminorms.
Ann. Probab. 7, 319–342 (1979)
14. M.B. Marcus, G. Peskir, J. Rosiński, Jørgen Hoffmann-Jørgensen (1942–2017), vol. 54
(Danish Mathematical Society, Matilde, 2018), pp. 14–15
Chapter 2
Moment Estimation Implied
by the Bobkov-Ledoux Inequality
Witold Bednorz and Grzegorz Głowienko
Abstract In this paper we consider a probability measure on the high dimensional

Euclidean space satisfying Bobkov-Ledoux inequality. Bobkov and Ledoux have
shown in (Probab Theory Related Fields 107(3):383–400, 1997) that such entropy
inequality captures concentration phenomenon of product exponential measure and
implies Poincaré inequality. For this reason any measure satisfying one of those
inequalities shares the same concentration result as the exponential measure. In
this paper using B-L inequality we derive some bounds for exponential Orlicz
norms for any locally Lipschitz function. The result is close to the question posted
by Adamczak and Wolff in (Probab Theory Related Fields 162:531–586, 2015)
regarding moments estimate for locally Lipschitz functions, which is expected to
result from B-L inequality.
Keywords Concentration of measure · Poincaré inequality · Sobolev inequality
Subject Classification 60E15, 46N30
2.1 The Bobkov-Ledoux Inequality
Let μ be a probability measure on Rd . We assume that μ satisfies Bobkov-Ledoux

inequality i.e. with fixed D > 0, for any positive, locally Lipschitz function f such
that |∇f |∞ f/2 we have
Entμ f 2 DEμ |∇f |22 . (2.1)
As noticed by Bobkov and Ledoux in [3] this modification of log-Sobolev inequality

is satisfied by product exponential measure, but more importantly, it implies
W. Bednorz () · G. Głowienko

Institute of Mathematics, University of Warsaw, Warszawa, Poland
e-mail: [email protected]; [email protected]
© Springer Nature Switzerland AG 2019 9

N. Gozlan et al. (eds.), High Dimensional Probability VIII,
Progress in Probability 74, https://doi.org/10.1007/978-3-030-26391-1_2
10 W. Bednorz and G. Głowienko
subexponential concentration. It is also quite easy to show that it implies Poincaré

inequality. For any smooth function g we may take f = 1 + g and > 0 such that
|∇f |∞ f/2, which allows us to apply (2.1). In the next step divide both sides of
inequality by 2 , consider standard Taylor expansion and take limit with tending
to 0. As a result
D
Varμ g Eμ |∇g|22 , (2.2)
2
which is exactly the Poincaré inequality. Finally just notice that any locally Lipschitz
function f such that both f and |∇f |2 are square integrable w.r.t. μ may be
approximated in (2.2) by smooth functions. The result means that B-L inequality
(2.1) is stronger than Poincaré inequality (2.2), nevertheless both inequalities imply
concentration phenomenon of product exponential measure, therefore any measure
satisfying one of those inequalities shares the same concentration result. See [3] for
more details regarding this subtle connection.
As we are dealing with big number of constants in the following section, it would
be wise to adopt some useful convention. Therefore, let us denote by D numeric
constant which may vary from line to line, but importantly, it is comparable to D
from log-Sobolev inequality (2.1). Similarly let C be constant comparable to 1 and
by C(α) denote one that depends on α only.
In [4] it was noticed by E. Milman that, Poincaré inequality (2.2) implies the
following estimate for p 1
√
f − Eμ f p D p|∇f |2 p , (2.3)
with f locally Lipschitz. It is easy to see that above results with the following bound
√ √
f − Eμ f p D p d|∇f |∞ p .
Adamczak and Wolff has conjectured in [1] that Bobkov-Ledoux inequality (2.1)
imply
√ √
f − Eμ f p D p|∇f |2 p + Cp|∇f |∞ p .
They also proved following weaker form of the conjecture

√ √
f − Eμ f p D p|∇f |2 p + Cp|∇f |∞ ∞ . (2.4)
Their result is based on tricky modification of given function so that (2.1) could be
used. In our paper we are trying to understand this phenomenon and apply its more
advanced form.
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
2 Moment Estimation Implied by the Bobkov-Ledoux Inequality 11
2.2 Bounds for Moments
In this section we investigate possible estimates for gpα , with a given α > 0,
when we know that g α is globally Lipschitz. This bounds will be useful when we
start dealing with the exponential Orlicz norms.
Theorem 2.1 If measure μ satisfies (2.1), function g is non-negative, locally
Lipschitz and p 1, then
for 0 < α 1
1 1 1 1 √
gpα 2 α max p α |∇g α |∞ α
∞
, g2α , αp 2 D |∇g|2 pα
and in case of α > 1
1 1 1 1 1 √
gpα max 2 α p α |∇g α |∞ α
∞
, 2 α g2α , αp 2 D |∇g|2 pα .
Proof Consider g α to be a non-negative Lipschitz function, otherwise estimate is

trivial. Note that in case of p 2 there is also nothing to prove, therefore we may
take p > 2. For simplicity let us assume that |∇g α |∞ ∞ = 1. If it happens to be
gαpα 2p |∇g α |∞ ∞
(2.5)
then proof is once again trivial, therefore assume that
gαpα > 2p |∇g α |∞ ∞

, (2.6)
then following the idea of the proof of (2.4) from [1] we define function h =
1
max{g, c}, where c = gpα /2 α . Obviously, for 2 t p
|∇hαt /2|∞ t |∇hα |∞

= .
hαt /2 2 hα
Due to our definition h c and |∇hα |∞ |∇g α |∞ , which gives us
|∇hα |∞ 2|∇g α |∞
.
hα gαpα
Combining above with (2.6) we get
|∇hαt /2|∞ t 1
.
hαt /2 ∞ 2p 2
Therefore, we may apply (2.1) to the function hαt /2 and thus by the Aida Stroock
[2] argument i.e.
d α 2 2 2/t −1 D αt 2/t −1
h t = 2 Ehαt Ent(hαt /2 )2 Eh E|hαt /2−α ∇hα |22 ,
dt t 2
combined with Hölder inequality with exponents t/(t − 2) and t/2 applied to the
last term, gives us
d α 2 D αt 2/t −1 αt 1−2/t 2/t D 2

h t Eh Eh E|∇hα |t2 = |∇hα |2 t
.
dt 2 2
The moment function (as function of t) is non-decreasing, therefore for 2 t p
we get
D
hα 2p − hα 22 (p − 2)|∇hα |2 2p . (2.7)
2
Now we have to consider two cases. First suppose that α 1 and then
|∇hα |2 p α|∇g|2 hα−1 p αcα−1 |∇g|2 p
and combining this with (2.7), we infer
α2 D
hα 2p hα 22 + (p − 2)c2α−2|∇g|2 2p .
2
Now observe that hα 2p g α 2p and furthermore
1 α 2
hα 22 c2α + g α 22 g p + g α 22 ,
4
which combined together gives us
3 α 2 α2 D
g p g α 22 + (p − 2)c2α−2|∇g|2 2p . (2.8)
4 2
Noting that the case of
gαpα 2gα2α , (2.9)
is another trivial part, we assume conversely getting
1 1 α 2
g α 22 = g2α
2α g2α
pα = g p
4 4
which together with (2.8) implies that
g2α 2
pα α D(p − 2)c
2α−2
|∇g|2 2p . (2.10)
Reminding that cα = 2−1 gαpα we infer
g2pα 2 α −2 α 2 D(p − 2)|∇g|2 2p

2
and rewriting it in simplified form

1 √ 1
gpα 2 α α D p 2 |∇g|2 p . (2.11)
Combining together (2.5), (2.9), and (2.11) implies the result in the case of 0 < α
1.
Consider now case of α > 1, following the same reasoning as in previous case,
up to the (2.7) after that Hölder inequality is used, we get
|∇hα |2 p α|∇g|2 hα−1 p α|∇g|2 pα hα−1

pα .
Therefore, by (2.7)
h2α D
h2pα (1 − 2α
) α 2 (p − 2)|∇g|2 2pα . (2.12)
hpα
2α 2
Again, either (2.9) holds or we have
1 α 2 1 1
h2α α 2
2α = h 2 c
2α
+ g α 22 = g p + g α 2p = g2α
pα .
4 4 2
pα gpα , we get
Since obviously h2α 2α
h2α
h2pα (1 − 2α
) 2−1 g2pα
h2α
pα
and combining above with (2.12) gives us

√ 1
gpα α D p 2 |∇g|2 p . (2.13)
Clearly (2.5), (2.9), and (2.13) cover the case of α > 1, which ends whole proof.
Next step of the reasoning is to apply previous result to g = |f −Eμ f | and combine
it with Poincaré inequality. Let us gather everything together in form of
Corollary 2.1 If measure μ satisfies (2.1), function f is locally Lipschitz and p

1, then for 0 < α 1
1
1 1 α
f − Eμ f pα 2 α max p α∇|f − Eμ f |α ∞ ∞
,
√ 1√

D |∇f |2 2 , αp 2 D |∇f |2 pα .
and in case of α > 1

1
1 1 α
f − Eμ f pα max 2 α p α∇|f − Eμ f |α ∞ ∞
,
1 √ 1√

2 α α D |∇f |2 2α , αp 2 D |∇f |2 pα .
Proof If we fix g = |f − Eμ f | then by the Poincaré inequality

√
f − Eμ f 2α (α ∨ 1) D |∇f |2 2(α∨1)
.
Note also that
|∇g|2 pα
= |∇f |2 pα
,
then applying Theorem 2.1 statement easily follows.
2.3 Bounds for Exponential Orlicz Norms
First, let us recall the notion of exponential Orlicz norms. For any α > 0
f ϕ(α) = inf{s > 0 : Eμ exp(|f |α /s α ) 2}.
Obviously, f ϕ(α) is a norm in case of α 1 only, otherwise there is a

1
problem with the triangle inequality. Moreover, we have f ϕ(α) = |f |α ϕ(1)
α
.
Nevertheless, in case of 0 < α < 1 one can use
1
f + gϕ(α) = |f + g|α ϕ(1)
α
1 1
|f |α + |g|α ϕ(1)
α
(|f |α ϕ(1) + |g|α ϕ(1) ) α
1 1
= (f αϕ(α) + gαϕ(α) ) α 2 α −1 (f ϕ(α) + gϕ(α) ).
f kα
It is worth to know that f ϕ(α) is always comparable with supk1 k 1/α
. More
precisely, observe that for all k 1 and a positive g
gkα
kα
gkα
ϕ(α) .
k!
Note that, just by the definition of gϕ(α) , there exists k 1 for which
gkα
kα
2−k gkα
ϕ(α) .
k!
Let us denote the set of such k 1 by J (g, α) and note that for any k ∈ J (g, α)
1 1 1
(k!)− kα gkα gϕ(α) 2 α (k!)− kα gkα . (2.14)
1
Next let M e be such a constant that (k!) k k/M for all k 1. We have
following crucial observation namely for all k ∈ J (g, α)
1 gkα
gϕ(α) (2M) α 1
. (2.15)
kα
Therefore, we may use Theorem 2.1 in order to obtain
Corollary 2.2 If μ satisfies (2.1) and g is non-negative locally Lipschitz function,
then for any k ∈ J (g, α) in case of 0 < α 1
1 √
, k − α g2α , αk 2 − α
1 1 1 1
gϕ(α) (4M) α max |∇g α |∞ α
∞
D |∇g|2 kα
.
and for 1 < α 2

1 √
, k − α g2α , 2− α αk 2 − α
1 1 1 1 1
gϕ(α) (4M) α max |∇g α |∞ α
∞
D |∇g|2 kα
.
Note that set J (g, α) is stable with respect to g → h, where h = max{g, c} i.e. if c
is comparable to gϕ(α) there exists C 1 such that for k ∈ J (g, α)
hkα 1
kα
k hkα
ϕ(α) ,
k! C
which means that we cannot easily improve the result using the trick.
In the same way as we have established Corollary 2.1 we can deduce the
following result.
Random documents with unrelated
content Scribd suggests to you:
and the May sunshine lit up the budding trees and shrubs. The touch
of spring and new-born life was everywhere. The Ambassador
snapped off a spray of honeysuckle which grew along the fence
protecting his parking from his neighbor’s, and tucked the spray in
his buttonhole as he entered the waiting motor. “Drive to the club,”
he directed briefly, as the car moved off.
CHAPTER IV
THE BROKEN APPOINTMENT
ELEANOR THORNTON turned in bed and stretched herself
luxuriously. It was good to be young and to be sleepy. For a few
seconds she dozed off again; then gradually awoke, and, too
comfortable to move, let her thoughts wander where they would. In
her mind’s eye she reviewed the events of the past months, and,
despite herself, her lips parted in a happy smile. She had come to
Washington in November to visit her friend, Cynthia Carew, and,
delighted with the reception accorded her, had invited her cousin,
Mrs. Gilbert Truxton, to chaperon her, and, on her acceptance, had
rented a small furnished residence near Dupont Circle for the winter.
Mrs. Winthrop and Cynthia Carew, whom she had known at
boarding school, took her everywhere with them, and her cousin,
Mrs. Truxton, belonging as she did to an old aristocratic family of the
District, procured her entrée to the exclusive homes of the “cave-
dwellers,” as the residential circle was sometimes called.
Born also with the gifts of charm and tact, Eleanor’s wild rose
beauty had made an instant impression, and she was invited
everywhere. The butler’s tray was filled with visiting cards, which
many newcomers, anxious for social honors, longed to have left at
their doors.
Eleanor was one of the older girls at Dobbs Ferry during Cynthia’s
first year at that boarding school. They had taken an immense liking
to each other, which later blossomed into an intimate friendship.
After her graduation she and Cynthia had kept up their
correspondence without a break, and, true to her promise, given
years before, she had left Berlin and journeyed to Washington to be
present at Cynthia’s début.
After the death of her mother, Eleanor had been adopted by an
indulgent uncle, Mr. William Fitzgerald, of New York, and on his
death had inherited a comfortable fortune.
In many ways the winter had brought numerous triumphs in its
train, enough to spoil most natures. But Eleanor was too well poised
to lose her head over adulation. She had sounded the depths of
social pleasantries, and found them shallow. In every country she
had visited all men had been only too ready to be at her beck and
call—except one. The dreamy eyes hardened at the thought, and the
soft lips closed firmly. She had made the advances, and he had not
responded. A situation so unique in her experience had made an
indelible impression. Angry with herself for even recalling so
unpleasant an episode, she touched the bell beside the bed; then,
placing her pillow in a more comfortable position, she leaned back
and contemplated her surroundings with speculative eyes.
Her individuality had stamped itself upon the whole room. A
picture or two, far above the average, a few choice books, whose
dainty binding indicated a taste and refinement quite unusual; one
or two Chinese vases, old when the Revolutionary War began; an
ivory carving of the Renaissance; a mirror in whose lustrous depths
Venetian beauties had seen their own reflections hundreds of years
ago. All these things gave sure indication of study and travel, and a
maturity of thought and taste which, oddly enough, seemed rather
to enhance Eleanor’s natural charm.
A discreet knock sounded on her door. “Bon jour, Mademoiselle,”
exclaimed the maid, entering with the breakfast tray.
“Bon jour, Annette,” responded Eleanor, rousing herself, then
lapsing into English, which her maid spoke with but a slight accent.
“Put the tray here beside me. Must I eat that egg?” she made a
slight grimace.
“But yes, Mademoiselle.” The Frenchwoman stepped to the
window and raised the shade. “Madame Truxton gave orders to Fugi
to tell the cook that he must send you a more substantial breakfast.
She does not approve of rolls and coffee. I think she wishes you to
eat as she does.”
Eleanor shuddered slightly. “Did—did she have beefsteak and
fried onions this morning?” she inquired.
“But yes, Mademoiselle,” Annette’s pretty features dimpled into a
smile, “and she ate most heartily.”
“Not another word, Annette, you take away my appetite. Is Mrs.
Truxton waiting to see me?”
“No, Mademoiselle; she was up at six o’clock and had her
breakfast at half-past seven.” Annette paused in the act of laying out
a supply of fresh lingerie. “What have the Americans on their
conscience that they cannot sleep in the morning?”
“You cannot complain of my early rising,” laughed Eleanor,
glancing at the clock, whose hands pointed to a quarter to twelve.
“Ah, Mademoiselle, you have lived so long away from America
that you have acquired our habits.”
“You may take the tray, Annette; I have even less appetite than
usual to-day.” Eleanor waited until it had been removed, then sprang
out of bed. “Come back in fifteen minutes,” she called.
It did not take her long to complete her toilette, and when the
maid returned she was seated before her dressing table.
“What news to-day, Annette?” she asked, as the Frenchwoman,
with skilful fingers, arranged her wavy hair, which fell far below her
waist.
“Madame and Fugi——” began the maid.
“I don’t want household details,” broke in Eleanor impatiently.
“Tell me of some outside news, if there is any.”
“Oh, indeed, yes; news the most startling. Senator Carew——”
she paused to contemplate her handiwork.
“Well, what about him?” inquired Eleanor listlessly.
“He is dead.”
“Dead!” The handglass slipped from Eleanor’s grasp and fell
crashing to the hearth. Annette pounced upon it.
“Oh, Mademoiselle, the glass is broken. Quelle horreur!”
“Bother the glass.” Eleanor’s foot came down with an
unmistakable stamp. “Tell me at once of Senator Carew’s death. I
cannot believe it!”
“It is only too true,” Annette was a privileged character and
deeply resented being hurried, also her volatile French nature
enjoyed creating a sensation. She had eagerly read the morning
paper, and had refrained from telling Eleanor the news until she
could get her undivided attention. “Senator Carew was found dead
in his carriage early this morning on his return from the dance at
Mrs. Owen’s”—Annette had no reason to complain, Eleanor was
giving her full attention to the story—“he had been stabbed.”
The maid’s hand accidentally touched Eleanor’s bare neck, and
she felt the taut muscles quiver. Covertly she glanced into the mirror
and studied the lovely face. But Eleanor’s expression told her
nothing. Her cheeks were colorless and her eyes downcast.
After a barely perceptible pause Annette continued her story.
“The coachman has been arrested——” a knock interrupted her and
she hastened to open the door, returning in an instant with a note.
“Fugi says the messenger boy is waiting for an answer,
Mademoiselle.”
Eleanor tore it open and read the hastily scrawled lines.
Dear Eleanor:
I suppose you have been told of last night’s terrible tragedy.
Cynthia is prostrated. She begs pitifully to see you. Can you come to
us for a few days? Your presence will help us both.
Affectionately,
Charlotte Winthrop.
Eleanor read the note several times, then walked thoughtfully
over to her desk.
Dearest Mrs. Winthrop: [she wrote] It is awful. I will come as soon as
possible.
Devotedly,
Eleanor.
“Give this to Fugi, Annette, then come back and pack my small
steamer trunk,” as the maid hastened out of the room; she picked
up a silk waist preparatory to putting it on, but her toilette was
doomed to another interruption.
“Well, my dear, may I come in?” asked a pleasant voice from the
doorway.
“Indeed you may, Cousin Kate,” Eleanor stepped across the room
and kissed the older woman affectionately. Mrs. Truxton’s ruddy face
lighted with an affectionate smile as she returned her greeting. She
did not altogether approve of her young cousin, many of her “foreign
ways” as she termed it, offended her, but Eleanor’s lovable
disposition had won a warm place in her regard.
Mrs. Truxton seated herself in one of the comfortable lounging
chairs and contemplated the disheveled room and Eleanor’s oriental
silk dressing gown with disapproval.
“Do you know the time?” she inquired pointedly.
“Nearly one,” answered Eleanor, as she discarded her dressing
gown for a silk waist. “Lunch will soon be ready. I hope you have a
good appetite.”
“Yes, thank you; I’ve been out all the morning,” reproachfully.
“Mrs. Douglas has asked me to dine with her this evening, and, I
think, Eleanor, if it will not interfere with your arrangements, that I
will accept the invitation.”
“Do so by all means,” exclaimed Eleanor heartily. “I hope she
won’t talk you deaf, dumb, and blind.”
“She is rather long-winded,” admitted Mrs. Truxton, tranquilly.
“On the telephone this morning she took up twenty minutes telling
me of the arrival here of her nephew, Douglas Hunter—good
gracious, child——” as Eleanor’s silver powder box rolled on the floor
with a loud bang—“how you startle one.”
“I beg your pardon,” Eleanor was some seconds picking it up, for
her fingers fumbled clumsily. “What were you saying, Cousin Kate?”
replacing the silver on the dressing table.
“Mercy, child, how inattentive you are! I was only remarking that
Douglas Hunter is no stranger to Washington. He was raised here, as
he belongs to one of the first families of Georgetown.”
“I never heard of a ‘second’ family in Georgetown,” smiled
Eleanor; then, seeing her cousin’s offended expression, she hastily
changed the subject. “Have you heard the shocking news of Senator
Carew’s—” she hesitated for a moment—“tragic death?”
“Indeed I have. Washington is talking of nothing else. Why are
you packing, Annette?” as the servant entered.
“Mrs. Winthrop has just written and asked me to spend a few
days with them,” explained Eleanor hurriedly, “so suppose you invite
Miss Crane to stay with you in my absence.”
“Of course you cannot very well decline to go,” said Mrs. Truxton
thoughtfully. “Still, I hate to have you mixed up in such an affair,
Eleanor.”
“Nonsense, Cousin Kate, you must not look at it in that light,”
Eleanor patted the fat shoulder nearest her affectionately. “Cynthia
told me yesterday that Senator Carew had said he was going to
discharge the coachman, Hamilton (a surly brute, I always thought
him), for drunkenness. I have no doubt he committed the murder
from revenge, and while under the influence of liquor.”
“I sincerely trust that is the correct solution of the mystery,” Mrs.
Truxton looked dubious, “but there has been one fearful scandal in
that family already, Eleanor, and I very much doubt if Senator Carew
was killed by a servant.”
“Why, what do you mean?” Eleanor wheeled around in her chair
and faced her abruptly.
“Time will show.” Mrs. Truxton shook her head mysteriously.
“Oh, nonsense!” exclaimed Eleanor impatiently.
As Mrs. Truxton opened her lips to reply, Annette reëntered the
room.
“Pardon, madame, you are wanted at the telephone,” and as Mrs.
Truxton lifted herself carefully out of her chair and walked out of the
room, she handed a package to Eleanor. “This has just come for you,
Mademoiselle; the boy who left it said there was no answer.”
“Annette! Annette!” came Mrs. Truxton’s shrill voice from the
lower hall.
“Coming, Madame, coming,” and the maid hastened out of the
room shutting the door behind her.
Left alone, Eleanor turned the sealed package over curiously. The
address was written in an unknown hand. Quickly breaking the red
sealing wax and tearing off the paper, she removed the pasteboard
cover and a layer of cotton. A startled exclamation escaped her as
she drew out the contents of the box—a necklace of large rubies and
smaller diamonds in an antique setting.
Eleanor, who knew the value of jewels, realized from their color
and size that the rubies were almost priceless, and in the pure joy of
beholding their beauty laid the necklace in the palm of her left hand
and along her bare arm. After contemplating the effect for a
moment, a thought occurred to her, and she pulled out the
remaining cotton in the box and found at the bottom a small card.
She picked it out and read the message written on the card.
“The appointment was not kept. Well done.”
The card fluttered to the floor unheeded. The pigeon blood
rubies made a crimson stain on Eleanor’s white arm, strong wrist,
and supple fingers.
CHAPTER V
MUTE TESTIMONY
DOUGLAS HUNTER sighed involuntarily as he left busy Fourteenth
Street, and walked down Massachusetts Avenue. Twelve years’
absence makes a great difference in the ever-shifting population of
Washington. He felt like another Rip Van Winkle as he gazed at each
passer-by in his search for a familiar face. Even the streets had
changed, and he was almost appalled by the grandeur of some of
the huge white palaces erected by multimillionaires on
Massachusetts and New Hampshire Avenues, and the Avenue of the
Presidents. He had spent part of the morning motoring about the
city with one of his cousins, and the outward and visible signs of
wealth had staggered him. What had become of the unpretentious,
generous-hearted hospitality, and the old world manners and courtly
greeting of the former host and hostess who had ruled so long at
the National Capital? Had Mammon spoiled the old simplicity, and
had Washington become but a suburb of New York and Chicago? It
truly seemed as if plutocracy had displaced aristocracy.
As Douglas approached the Carew residence he glanced keenly
at the handsome old mansion and at the numerous idlers loafing in
the vicinity drawn there by idle curiosity. A policeman stood on guard
in the driveway, and a number of photographers loitered near by,
cameras in hand, waiting patiently to snapshot any member of the
Carew family who might incautiously venture out of doors.
The house itself, a handsome square red brick and stone
trimmed four-storied building, stood some distance back from the
sidewalk with beautifully kept lawns divided by the carriage drive.
The blinds were drawn and the ominous black streamer over the bell
presented a mournful spectacle. It was the finest residence in that
once fashionable locality, and Douglas decided that he preferred its
solid, home-like architecture to the more ornate and pretentious
dwellings in other parts of the city. As the years went by Senator
Carew had added improvements until the residence was one of the
most delightful in Washington.
As Douglas turned into the walk, a large touring car wheeled into
the driveway, and as it purred softly by him, he stepped back
respectfully and raised his hat to the tired-faced man sitting alone in
the tonneau. He did not need to glance at the small coat-of-arms of
the United States emblazoned on the polished door, or at the two
Secret Service men following on their motor cycles, to recognize the
distinguished occupant of the car.
As the motor stopped under the porte-cochère, the colored butler
ran down the steps, and the President leaned forward and placed a
note in the bowing and scraping negro’s hand; then the big car
continued on down the driveway and out into the street.
Douglas waited where he was for a few minutes before mounting
the short flight of steps. The hall door was opened several inches on
his approach, and Joshua solemnly extended his card tray, which
Douglas waved aside.
“I called to see Mr. Brett; is he here?” he asked.
“Yessir,” Joshua opened the door still further, and inspected him
carefully.
“Take my card to him and ask if he can spare me a few minutes,”
and he dropped his visiting card on the tray.
“Walk in, suh,” exclaimed Joshua, impressed by Douglas’ well-
groomed appearance; then he hesitated, embarrassed by a sudden
idea.
“I’ll wait here,” volunteered Douglas, stepping inside the square
hall.
“All right, suh,” Joshua closed the front door, “just a moment,
suh,” and he stepped softly across the hall and into a room. Douglas
glanced about him curiously and caught a glimpse of spacious rooms
and lofty ceilings. It was a double house, and to the right of the
entrance was the drawing-room, and back of that another large
room, which Douglas took to be the dining room, judging from the
glittering silver pieces on a high sideboard of which he had a glimpse
through the door leading into the square hall. Across from the
drawing-room was the room into which Joshua had disappeared, and
back of that a broad circular staircase which ran up to the top floor.
Douglas was idly gazing out of the glass panel of the front door
when Joshua returned, followed by a middle-aged man with a keen,
clever face.
“Is it really you, Mr. Hunter?” he asked, as they shook hands
warmly. “I could hardly believe my eyes when I saw your card. Come
this way,” and he conducted Douglas into the room he had just left,
and closed the door softly behind them.
“When did you arrive in Washington?” he inquired, motioning
Douglas to take a chair near the window and dropping into one
opposite him.
“Yesterday.” Douglas leaned back and studied his surroundings.
His eyes traveled over the handsome carved rosewood bookcases
which lined the walls, at the large desk table, and the comfortable
leather-covered revolving desk chair. The desk silver, drop lights, and
large upholstered davenport pushed invitingly before the huge
fireplace with its shining brass fire-dogs and fender, each told a tale
of wealth and artistic taste—two assets not often found together. His
eyes returned to Brett, and he smiled involuntarily as he caught the
other intently regarding him.
Brett smiled in return. “I was wondering why you looked me up
so soon,” he admitted candidly. “Don’t think I’m not glad to see
you”—hastily—“but I remember of old that you seldom do things
without a motive.”
“On the contrary, I am here this afternoon to find a motive—for
Senator Carew’s tragic death.” The smile vanished from Douglas’
clear-cut features. “One moment,” as Brett opened his mouth to
speak. “After reading the account of the Senator’s death in the
morning papers, I went down to headquarters to get what additional
facts I could, and they told me that you had been put on the case.
So I decided to look you up in person, and here I am.”
“May I ask why you take such an interest in this case?”
“Certainly, Brett; I was coming to that. Senator Carew used his
influence to get me in the Diplomatic Service, and during the past
twelve years he has shown me many kindnesses, such as seeing
that I was detailed to desirable posts, and helped me to secure
promotion.”
“He wouldn’t have done that, Mr. Hunter, if you hadn’t made
good,” broke in Brett quickly.
“I saw him last at Delmonico’s in New York on my way to Japan a
little over a year ago,” continued Douglas. “He asked me to lunch
with him, and evinced great interest in the mystery of the Jewel
Custom Fraud which he, in some way, knew I had had a hand in
exposing.”
“Sure he did. I told the department about your assistance when I
was in Paris. If it hadn’t been for you, I’d never have landed those
swindlers. They led me a pretty dance over the Atlantic.”
“We worked together then,” said Douglas thoughtfully, “and, on
the strength of our past success, I’m going to ask you to take me on
as a sort of advisory partner in this Carew case.”
“Suppose you first tell me the reason for making such a request.”
“In the first place I owe a debt of gratitude to Senator Carew. For
the sake of his friendship with my father years ago, he has taken a
great interest in me. Secondly, I am in Washington at his request.”
Brett looked his interest, and Douglas went on rapidly: “Some
time ago I received a note from him asking me to apply for leave of
absence from Tokio and to come direct to Washington, saying that
he wished to see me on important business.”
“Did he state the nature of that business?” inquired Brett eagerly.
“No. I at once followed his suggestion and applied to the State
Department for leave. It was granted, and I hastened home as fast
as steamer and train could bring me.”
“Did you see Senator Carew?”
“Unfortunately, no. I only reached Washington late last night. I
expected to see the Senator this morning, instead of which I read of
his mysterious death in the morning papers.”
Brett mused for a few minutes, then roused himself. “I am only
too glad to have your assistance, Mr. Hunter.”
“Good!” ejaculated Douglas, well pleased. “Suppose you tell me
all the facts in the case so far discovered.”
Brett leaned back in his chair and crossed his legs. “On the face
of things it looks as if the negro driver, Hamilton, was guilty.”
“Tell me what leads you to think that?” inquired Douglas quickly.
“He is the worst type of negro, a vicious brute with a taste for
liquor. I have inquired about him and examined him thoroughly and
am really puzzled, Hunter, to find out why Senator Carew ever
employed him.”
“Is he an old family servant?”
“No. He has only been in Carew’s employ about a year I am told.
He knows how to handle horses, and took excellent care of the
Senator’s valuable stable.”
“That probably explains why he was kept on,” said Douglas. “I’ve
been told that Carew was hipped about his horses.”
“Yes. I gathered from Mrs. Winthrop that Hamilton has been
drinking steadily, and his conduct to the other servants grew
intolerable. Senator Carew had to discharge him.”
“When did that happen?”
“Yesterday afternoon.”
“Then, how was it that he was driving the carriage last night?”
“Oh, Carew gave him a week’s notice, said he couldn’t fill his
place at once, and told him to stay on. Joshua tells me that Hamilton
uttered some ugly threats in the kitchen that evening, but that the
servants paid no attention to his black humor, as they saw he had
been drinking.”
“I see in the papers that Hamilton vehemently declares his
innocence.”
“He does,” agreed Brett, checking his remarks off on his fingers;
“he declares he did not see Senator Carew after being discharged by
him; that no one was in the carriage when he drove away from the
stable at midnight; that he went directly to Mrs. Owen’s residence;
and that he does not know when or how Senator Carew’s body was
secreted in the carriage.”
“The plot thickens,” muttered Douglas. “Do you believe his
statements?”
“I do, and I don’t. The servants all declare that he was half
drunk; therefore, I doubt if he was in a condition to pay much
attention to anything, or that his statements can be relied on. He
was sobered by the shock of finding Carew’s body in his carriage,
and, when I arrested him, collapsed from fright.”
“Well, judging from the facts you have just told me, I don’t much
believe he killed Carew.”
“Why not?” argued Brett. “Hamilton was apparently half out of
his mind from rage and drink, and his brute nature made him seek
revenge. It’s quite possible Carew entered the carriage thinking it
would not be safe for his niece to drive home alone from the dance,
and Hamilton took that opportunity to kill him.”
“I read in the evening paper that Hamilton was told to stop at
the house for one of the maids, but, instead, drove directly from the
stable to the dance,” said Douglas. “Therefore Carew did not enter
the carriage at this door.”
“Hamilton may have been too befogged with drink to have
remembered the order,” suggested the detective.
“I grant you, Brett,” said Douglas thoughtfully, “that the negro
may have the nature, the desire, and the opportunity to commit
murder—but why select such a weapon?”
“Probably picked up the first thing at hand,” grunted Brett.
“But a desk file is not the ‘first thing at hand’ in a stable,”
remarked Douglas calmly. “In fact, it’s the last thing you would
expect to find there.”
“I don’t know about that; perhaps it was thrown away in a
wastepaper basket, and Hamilton may have picked it out of the ash
pile,” suggested Brett.
“What did the file look like?”
“It is of medium size, the slender steel being very sharp, the
round solid base being silver. I’ve shown it to several jewelers, and
they all say it’s like hundreds of others, rather expensive, but
popular with their well-to-do customers, and that they have no
means of tracing it back to any particular owner. It was something
like that one,” pointing to an upright file on Senator Carew’s desk.
Douglas leaned over and took it up. “An ideal weapon,” he said
softly, balancing it in his hand as his fingers closed over the round
heavy base. He removed the cork which was used to guard the
sharp point and felt it with his thumb. “It must have taken a shrewd
blow to drive the file through overcoat and clothing so that it would
cause instant death.”
“The Senator wore no overcoat.” Douglas looked his surprise.
After a moment’s silence Brett edged his chair closer to his
companion and lowered his voice. “You recollect how it rained last
night?”
“In torrents. I have seldom seen such a cloudburst,” admitted
Douglas.
“It commenced to rain about ten-thirty,” continued Brett, “and it
did not stop until after three o’clock. Hamilton drove twice in that
drenching rain to Mrs. Owen’s and back again, first taking Miss
Carew to the dance and returning with her. Senator Carew’s body
was discovered on the last trip home. Miss Carew told her aunt that
no one was in the carriage with her when she made the first trip to
the dance. Senator Carew’s body was not removed until after my
arrival here this morning, and I then made a thorough examination
of the carriage and, with the coroner’s assistance, of the body as
well”—he paused and cleared his throat—“I found Senator Carew’s
clothes were absolutely dry—as I said before, he wore no overcoat—
now, how did Carew get into that carriage in that soaking downpour
without getting wet?” asked Brett, settling back in his chair.
“Perhaps he was first murdered and then carried out and put into
the carriage.”
“Perhaps so, but I doubt it.”
“He may have entered the carriage at the stable when Hamilton
was not around.”
“I thought of that,” returned Brett, “and as soon as it was
daylight examined the yard and the alley. The concrete walk from
the house to the stable is being laid now and cannot be used, so
that one has to tread on the ground, which is extremely soft and
muddy. The alley is a long one, and Carew’s stable is about in the
center of it, and the rain, settling in the holes of the uneven cobbles,
made walking very unpleasant. I am telling you all these details
because of another discovery I made,” went on Brett slowly; “the
Senator’s shoes had been recently polished and the blacking was not
even stained.”
Douglas leaned back and bit his thumb nail, a childish habit of
which he had never been able to break himself.
“Where did Carew spend the evening?” he asked finally.
“That is what I have not been able to find out,” growled Brett.
“Mrs. Winthrop told me she had not seen her brother since
breakfast. That he went to the Capitol as usual in the morning. She
was told on entering the house just before dinner that he would not
return for that meal, but they did not state where he was going.”
“Upon my word it’s a very pretty problem,” commented Douglas
softly.
“It is,” agreed Brett, rising and slowly pacing the room. He
glanced piercingly at Douglas, who was thoughtfully contemplating a
life-size portrait of one of Carew’s ancestors which hung above the
mantel over the fireplace. Douglas Hunter’s clear-cut features, broad
forehead, and square jaw indicated cleverness and determination.
When Douglas smiled the severe lines relaxed and his smooth-
shaven face was almost boyish. He had a keen sense of the
ridiculous, which prevented him from taking himself too seriously. In
the past Brett had conceived a high regard for the other’s quick wit
and indomitable courage.
“This is Senator Carew’s study or library,” he said, stopping
before the desk, “and I was giving the room my special attention
when you came in.”
“Have you met with any success?” inquired Douglas quickly.
“So far only one thing—it may be a clew or it may not; under this
writing pad I found this blotter,” holding up a square white sheet; “it
has been used only once, first on one side then on the other, so that
by holding it in front of this mirror you can read quite clearly, see
——”
Douglas rose, stepped behind Brett, and peeped over his
shoulder into the silver-mounted mirror, which the latter had
removed from its place on the mantel.
The large, bold writing was fairly legible. “What do you make out
of it?” asked Brett impatiently.
Obediently Douglas read the words aloud:
“‘Am writing in case I don’t see you before you’—” the writing
ceased.
“He must have been interrupted,” explained Brett, “and clapped
down the blotter on top of the sheet so that whoever entered
couldn’t see the written words. Now look at the other side,” and he
turned over the blotter on which were traced only a few words:
“‘I have discovered——’” read Douglas.
“What do you think of it?” asked Brett, putting the blotter in an
inner pocket of his coat.
“It depends on when it was written”—Douglas’ eyes strayed to
the door. Surely Brett had closed it when they entered, now it stood
partly open into the hall. He pointed silently to it, and by common
impulse both men stepped out into the hall.
Listening intently they heard a faint rap on one of the doors in
the upper hall; then a high-pitched, quivering voice reached them:
“Eleanor, Eleanor, I’m so glad you’ve come. I’m nearly sick with
misery. They quarreled, Eleanor, they quarreled——” her voice
caught in a sob—the door slammed shut.
The two men glanced at each other, their eyes asked the same
question. Who quarreled?
CHAPTER VI
CIRCUMSTANTIAL EVIDENCE
ASLIGHT sound behind him caused Douglas to wheel swiftly
around. A pretty woman, with astonishment written largely in her
round eyes, stood regarding the two men. She was carrying a
handbag.
“Whom do you wish to see?” asked Brett sharply.
“No one, Monsieur,” replied Annette, her accent denoting her
nationality. “I am Miss Thornton’s maid.”
Douglas started. “Eleanor—Miss Thornton!” Was it possible that
she could mean the Eleanor Thornton he used to know?
“I am taking her bag to her room as she is spending the night
here,” added the servant.
“Indeed.” Brett inspected her keenly. “When did Miss Thornton
enter the house?”
“A few minutes ago, Monsieur,” vaguely. “Joshua showed
Mademoiselle in while I stopped a moment to speak with the
chauffeur, and he left the front door open so that I could enter.”
At that moment the butler appeared from the dining room
carrying a tray on which were glasses and a pitcher of ice water.
“Joshua, is this Miss Thornton’s maid?” asked Brett.
“Yessir,” Joshua ducked his head respectfully as he answered the
detective. “Annette, Miss Eleanor done hab her same room next do’
ter Miss Cynthia’s. Yo’ kin take her things right upstairs, and tell Miss
Eleanor I done got der ice water fo’ her.”
With a half curtsey Annette stepped past the two men, and ran
quickly up the staircase.
“Stop a moment, Joshua,” ordered Brett, as the butler started to
follow the maid. “Who opened the door into the library a few
moments ago?”
“’Deed I dunno, suh; I been so busy takin’ in cyards I ain’t
noticed particular.”
“Who has been in the hall besides yourself?” persisted Brett.
“Ain’t no one,” began Joshua, then paused. “Now I do recollect
dat Marse Philip cum in right smart time ergo, suh. He axed fo’ yo’,
and I tole him yo’ was in de lib’ary. I ’specks he mighter been
alookin’ fo’ yo’.”
“Ah, indeed; where is Mr. Winthrop now?”
“Ah dunno, suh.”
“Well, find him, Joshua, and tell him I wish to see him—at once.”
Brett’s pleasant voice had deepened, and Joshua blinked nervously.
“Yessir, I’ll tell him, suh, ’deed ah will,” he mumbled, as he
started upstairs.
As Douglas and Brett walked across the hall to enter the library a
man stepped out of the drawing-room.
“Are you looking for me, Mr. Brett?”
The question was asked courteously enough, and Douglas was
the more astonished to encounter a hostile stare as the newcomer
glanced at him.
“I hope you can give me a few minutes of your time,” said Brett;
“will you be so good as to step into the library?” and he stood aside
to allow Philip Winthrop to enter first. Douglas followed them into
the room and locked the door. As the key clicked slightly Winthrop
frowned, and his pale face flushed.
“That is only a precaution against eavesdroppers,” explained
Douglas quickly.
“Mr. Winthrop, this is Mr. Douglas Hunter, who is assisting me in
my efforts to unravel the mystery surrounding Senator Carew’s
death, and with your permission will be present at this interview.“
“Why, certainly,” exclaimed Winthrop, with well simulated
heartiness; “won’t you both sit down?” and he dropped into the
revolving desk chair. Douglas picked out his old seat in the window
and turned his back to the light the better to face Winthrop and
Brett, who also sat near the desk.
“When will they hold the inquest, Mr. Brett?” questioned
Winthrop.
“The coroner, Dr. Penfield, told me to-morrow.”
“Has Hamilton a lawyer to look out for his interests?”
“That’s not absolutely necessary at the inquest, Mr. Winthrop. At
present the negro is simply held on suspicion. If the inquest so
decides, he will be charged with the murder and held for the grand
jury.”
Douglas had been busy scanning Winthrop’s face intently. He
noted the heavy lines in the handsome face, and the unnatural
brilliancy of his eyes. It was apparent to both men, by Winthrop’s
thick speech and unsteady hands, which kept fingering the desk
ornaments nervously, that he had been drinking heavily.
“Where did you last see Senator Carew?”
“In this room yesterday afternoon.”
“Did you see him alone, or were others present?”
“He was alone.”
“Did he show you a letter which he was then writing?” inquired
Douglas at a venture, and was startled at the effect of his question
on Winthrop. The latter whitened perceptibly, and pulled his short
black mustache to hide his twitching lips.
“I know nothing about any letter,” he stammered.
Brett did not press the point, but asked instead: “Where did you
spend last night?”
“I dined here with my mother and cousin.”
“And afterwards?” put in Douglas.
“I went to the Alibi Club soon after dinner.”
“How late did you stay there?”
“Most of the night,” was the evasive reply.
“Please mention the exact hour you left the club,” persisted Brett.
“I really cannot recollect the exact time; I did not reach this
house until after two this morning. We had a pretty gay time at the
club, and I was in no condition to remember the hour,” and he
smiled deprecatingly.
Again Brett did not press the question. He turned over the pages
of his small memorandum book in which he had been making
entries.
“Have you any idea where Senator Carew dined and spent the
evening?”
“No,” came the emphatic answer. “He asked me to tell my mother
not to expect him at dinner, that was all.”
“Ah, indeed. Have you any idea when the Senator left the
house?”
“No, I left him here, and went up to my room, where I stayed
until dinner was announced.”
“Where is your room?”
“Third floor, back.”
“Who has rooms on the next floor?”
“Senator Carew’s bedroom, bath, and sitting-room are over this
part of the house; Miss Cynthia Carew occupies the suite of rooms
across the hall from his rooms. My mother and I have the third floor
to ourselves.” Winthrop plucked nervously at the desk pad. “Talking
is dry work; won’t you and Mr. Hunter join me, I’ll ring for Joshua.”
“One moment,” Brett’s tone was peremptory and, with an
unmistakable scowl, Winthrop sank down in his chair and leaned
heavily on the desk. “What members of the family were in the house
yesterday afternoon?”
Winthrop thought for a moment before replying. “No one but my
uncle and myself,” he said reluctantly. “My mother and Miss Carew
went out early to some bridge party, and did not return until just
before dinner.”
“I see.” Brett leaned back in his chair and contemplated Winthrop
thoughtfully.
“Mr. Winthrop,” asked Douglas, breaking the short silence, “were
you and your uncle always on good terms?”
“Why, yes.” Winthrop’s twitching fingers closed unconsciously on
the slender desk file, and as he spoke his shifting eyes dropped from
Douglas’ clear gaze, and fell on the sharp steel desk ornament in his
hand. With a convulsive shudder he dropped it and sprang to his
feet. “What’s all this questioning about?” he demanded loudly. “I’ve
had enough of this, you——” his hands clinched, and the blood
flamed his pale face, a gurgle choked his utterance, and before Brett
could reach him he fell prone across the desk.
CHAPTER VII
A PIECE OF ORIENTAL SILK
“I’Mopened
glad you could come back, Mr. Hunter,” said Brett, as Joshua
the library door of the Carew residence and admitted
Douglas. “Can you stay here all night?”
“If necessary,” replied Douglas, glancing at him in surprise.
“I think it would be best. Mrs. Winthrop is completely unstrung;
her niece, Miss Carew, prostrated from shock, and Mr. Philip
Winthrop in bed with a bad attack of delirium tremens. In such a
household your presence to-night might be invaluable if anything
else were to happen—not that I am anticipating any further trouble
or tragedies.”
“Very well, I will stay,” agreed Douglas.
“’Deed I’se mighty glad ter hyar dat,” volunteered Joshua, who
hovered just inside the door on the pretext of arranging some
furniture. “But I dunno whar I’ll put yo’, suh. Miss Eleanor, she’s in
de gues’ chambah, an’ Annette’s in de room back ob her’s, and de
nusses fo’ Marse Philip has der spar rooms in der third flo’.”
“Never mind, Joshua, I can camp out in this room. That sofa
looks very comfortable,” and Douglas pointed to the large
upholstered davenport which faced the empty fireplace.
“Just a moment, Joshua,” exclaimed Brett, as the old butler
moved toward the door. “Did you see Senator Carew leave the house
yesterday afternoon?”
“No, suh.”
“Did he take luncheon here?”
“No, suh; he cum in ’bout three o’clock; leastways dat was when
he rung fo’ me, an’ I reckon he’d only jes’ arrived, ’cause he had his
hat an’ coat on his arm.”
“What did he want with you?”
“He axed me why Hamilton hadn’t called fo’ him at de Capitol as
ordered, an’ when I tole him dat Hamilton was a-sittin’ in de stable
doin’ nuffin, he said I was ter go right out an’ send him to de library
—which I done.”
“Did you see Senator Carew after that?”
“Yessir. After ’bout fifteen minutes Hamilton cum out lookin’
mighty black an’ mutterin’ under his breff. Den Marse James rung fo’
me ag’in, an’ sent me to tell Marse Philip dat he wanted ter see him
to onst.”
“Was there anything unusual in Senator Carew’s manner?”
inquired Douglas, who had been listening attentively to the old
darky’s statements.
“He seemed considerable put out, dat was all,” responded
Joshua, after due reflection.
“Was Senator Carew irritable and quick-tempered?”
“Mostly he was real easy-going, but sometimes he had flare-ups,
an’ den it was bes’ ter keep outer his way.”
“Did you find Mr. Winthrop?”
“Yessir. I gib him de message, an’ he went right down to de
lib’ary.”
“Do you know how long Senator Carew and young Winthrop
remained in this room?”
“No, suh. I went ter de fron’ doo’, an’ while in de hall I heard a
regular ruction goin’ on inside dis room.”
“Could you hear what was said?” demanded Brett eagerly.

High Dimensional Probability VIII: The Oaxaca Volume Nathael Gozlan All Chapter Instant Download

Uploaded by

Copyright:

Available Formats

High Dimensional Probability VIII: The Oaxaca Volume Nathael Gozlan All Chapter Instant Download

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

High Dimensional Probability VIII: The Oaxaca Volume Nathael Gozlan All Chapter Instant Download

Uploaded by

Copyright:

Available Formats

Download the full version of the textbook now at textbookfull.

High Dimensional Probability VIII: The Oaxaca

Explore and download more textbook at https://textbookfull.com

Coulson and Richardson’s Chemical Engineering, Fourth

VIII International Scientific Siberian Transport Forum

High-dimensional Microarray Data Analysis: Cancer Gene

Teach Yourself VISUALLY iPhone 6s Covers iOS 9 and all

Training for Change Transforming Systems to be Trauma

Who Is Vera Kelly Rosalie Knecht

New Essays on Aristotle s Organon Routledge Monographs in

The Rise And Development Of FinTech: Accounts Of

Progress in Probability is designed for the publication of workshops, seminars

More information about this series at http://www.springer.com/series/4839

Karim Lounici Mokshay Madiman

ISSN 1050-6977 ISSN 2297-0428 (electronic)

© Springer Nature Switzerland AG 2019

The history of the High-Dimensional Probability (HDP) conferences dates back to

Gaussian approximation, various stochastic inequalities and their applications in

for all i.i.d. random vectors X and Y in Rn .

• Rearrangement and Prékopa–Leindler type inequalities, by J. Melbourne. The

Paris, France Nathael Gozlan

1 Jørgen Hoffmann-Jørgensen (1942–2017) . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1

11 Asymptotic Behavior of Rényi Entropy in the Central Limit

Michael B. Marcus, Goran Peskir, and Jan Rosiński

Jørgen Hoffmann-Jørgensen, docent emeritus in the Department of Mathematics at

© Springer Nature Switzerland AG 2019 1

influence on the study of probability in Europe extended beyond these activities. In

Hoffmann-Jørgensen’s inequality states that

P(M ≥ 2t + s) ≤ 2P(N ≥ s) + 8P2 (M ≥ t) (1.1)

for all t, s > 0.

P(M < ∞) = 1 and E(N p ) < ∞

for some 0 < p < ∞. Then E(M p ) < ∞.

2 The material up to this point has appeared in [14].

Combining (1.2) and (1.3) we get

lim Ef (Xn ) = Ef (X) (1.5)

for every bounded continuous function f : S → R, then we say that Xn converges

as n → ∞. The expectation E in (1.5) is defined as the (Lebesgue-Stieltjes) integral

for n ≥ 1, t ∈ [0, 1] and ω ∈ , where ξ1 , ξ2 , . . . are independent and identically

and aims to establish that Xn converges ‘weakly’ to a limiting process X (of a

indexed by f belonging to a family of functions, there is no obvious way to

expectation E∗ , which is defined by

E∗ Y = inf { EZ | Z ≥ Y is measurable } (1.10)

where Y is any (not necessarily measurable) function from  to R, and leaving

process (Xt )t ∈T with values in a complete separable metric space K, expressed in

where T is a separable metric space, s, t ∈ T , and A ∈ B(K 2 ). Jørgen’s interest in

Witold Bednorz and Grzegorz Głowienko

Abstract In this paper we consider a probability measure on the high dimensional

Keywords Concentration of measure · Poincaré inequality · Sobolev inequality

Subject Classification 60E15, 46N30

2.1 The Bobkov-Ledoux Inequality

Let μ be a probability measure on Rd . We assume that μ satisfies Bobkov-Ledoux

Entμ f 2 DEμ |∇f |22 . (2.1)

As noticed by Bobkov and Ledoux in [3] this modification of log-Sobolev inequality

W. Bednorz () · G. Głowienko

© Springer Nature Switzerland AG 2019 9

subexponential concentration. It is also quite easy to show that it implies Poincaré

They also proved following weaker form of the conjecture

2.2 Bounds for Moments

and in case of α > 1

for n ≥ 1, t ∈ [0, 1] and ω ∈ , where ξ1 , ξ2 , . . . are independent and identically

where Y is any (not necessarily measurable) function from to R, and leaving

d α 2 D αt 2/t −1 αt 1−2/t 2/t D 2