Get ebook downloads in full at ebookname.

com

Exploratory Multivariate Analysis by Example Using

R 1st Edition Francois Husson

https://ebookname.com/product/exploratory-multivariate-
analysis-by-example-using-r-1st-edition-francois-husson/

OR CLICK BUTTON

DOWNLOAD EBOOK

Explore and download more ebook at https://ebookname.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Multivariate Analysis of Ecological Data using CANOCO 5

2nd Edition Petr Šmilauer

https://ebookname.com/product/multivariate-analysis-of-
ecological-data-using-canoco-5-2nd-edition-petr-smilauer/

Swift by Example 1st Edition Scalzo

https://ebookname.com/product/swift-by-example-1st-edition-
scalzo/

Node js By Example 1st Edition Tsonev

https://ebookname.com/product/node-js-by-example-1st-edition-
tsonev/

Mind from Body Experience from Neural Structure 1st

Edition Don M. Tucker

https://ebookname.com/product/mind-from-body-experience-from-
neural-structure-1st-edition-don-m-tucker/
Phase transformations of elements under high pressure
CRC 2005 1st Edition E. Yu Tonkov

https://ebookname.com/product/phase-transformations-of-elements-
under-high-pressure-crc-2005-1st-edition-e-yu-tonkov/

The Iron Dragon s Daughter 1st Edition Michael Swanwick

https://ebookname.com/product/the-iron-dragon-s-daughter-1st-
edition-michael-swanwick/

The Aesthetics of Loss German Women s Art of the First

World War 1st Edition Claudia Siebrecht

https://ebookname.com/product/the-aesthetics-of-loss-german-
women-s-art-of-the-first-world-war-1st-edition-claudia-siebrecht/

Visual Guide to Hedge Funds 1st Edition Richard C.

Wilson

https://ebookname.com/product/visual-guide-to-hedge-funds-1st-
edition-richard-c-wilson/

A Dictionary of the Avant Gardes 3rd Edition Richard

Kostelanetz

https://ebookname.com/product/a-dictionary-of-the-avant-
gardes-3rd-edition-richard-kostelanetz/
Heroes From Hercules to Superman Meyer

https://ebookname.com/product/heroes-from-hercules-to-superman-
meyer/
 Exploratory Multivariate Analysis
by Example Using R

K11614_FM.indd 1 10/18/10 3:04 PM

 Chapman & Hall/CRC
Computer Science and Data Analysis Series

The interface between the computer and statistical sciences is increasing, as each discipline
seeks to harness the power and resources of the other. This series aims to foster the integration
between the computer sciences and statistical, numerical, and probabilistic methods by
publishing a broad range of reference works, textbooks, and handbooks.

SERIES EDITORS
David Blei, Princeton University
David Madigan, Rutgers University
Marina Meila, University of Washington
Fionn Murtagh, Royal Holloway, University of London

Proposals for the series should be sent directly to one of the series editors above, or submitted to:

Chapman & Hall/CRC

4th Floor, Albert House
1-4 Singer Street
London EC2A 4BQ
UK

Published Titles

Bayesian Artificial Intelligence, Second Edition Introduction to Machine Learning

Kevin B. Korb and Ann E. Nicholson and Bioinformatics
Clustering for Data Mining: Sushmita Mitra, Sujay Datta,
A Data Recovery Approach Theodore Perkins, and George Michailidis
Boris Mirkin Microarray Image Analysis:
Computational Statistics Handbook with An Algorithmic Approach
®
MATLAB , Second Edition Karl Fraser, Zidong Wang, and Xiaohui Liu
Wendy L. Martinez and Angel R. Martinez Pattern Recognition Algorithms for
Correspondence Analysis and Data Data Mining
Coding with Java and R Sankar K. Pal and Pabitra Mitra
Fionn Murtagh R Graphics
Design and Modeling for Computer Paul Murrell
Experiments R Programming for Bioinformatics
Kai-Tai Fang, Runze Li, and Agus Sudjianto Robert Gentleman
®
Exploratory Data Analysis with MATLAB Semisupervised Learning for
Wendy L. Martinez and Angel R. Martinez Computational Linguistics
Exploratory Multivariate Analysis by Steven Abney
Example Using R Statistical Computing with R
François Husson, Sébastien Lê, and Maria L. Rizzo
Jérôme Pagès
Introduction to Data Technologies
Paul Murrell

K11614_FM.indd 2 10/18/10 3:04 PM

 Exploratory Multivariate Analysis
by Example Using R

François Husson
Sébastien Lê
Jérôme Pagès

K11614_FM.indd 3 10/18/10 3:04 PM

 CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2011 by Taylor and Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper

10 9 8 7 6 5 4 3 2 1

International Standard Book Number: 978-1-4398-3580-7 (Hardback)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made
to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all
materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all
material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not
been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in
any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, micro-
filming, and recording, or in any information storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.
copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-
8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that
have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identi-
fication and explanation without intent to infringe.

Library of Congress Cataloging‑in‑Publication Data

Husson, François.
Exploratory multivariate analysis by example using R / François Husson, Sébastien Lê, Jérôme
Pagès.
p. cm. -- (Chapman & Hall/CRC computer science & data analysis)
Summary: “An introduction to exploratory techniques for multivariate data analysis, this book
covers the key methodology, including principal components analysis, correspondence analysis,
mixed models, and multiple factor analysis. The authors take a practical approach, with examples
leading the discussion of the methods and many graphics to emphasize visualization. They present
the concepts in the most intuitive way possible, keeping mathematical content to a minimum
or relegating it to the appendices. The book includes examples that use real data from a range of
scientific disciplines and implemented using an R package developed by the authors.”-- Provided
by publisher.
Includes bibliographical references and index.
ISBN 978-1-4398-3580-7 (hardback)
1. Multivariate analysis. 2. R (Computer program language) I. Lê, Sébastien. II. Pagès, Jérôme. III.
Title. IV. Series.

QA278.H87 2010
519.5’3502855133--dc22 2010040339

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com

K11614_FM.indd 4 10/18/10 3:04 PM

Contents

Preface xi

1 Principal Component Analysis (PCA) 1

1.1 Data — Notation — Examples . . . . . . . . . . . . . . . . . 1
1.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.1 Studying Individuals . . . . . . . . . . . . . . . . . . . 2
1.2.2 Studying Variables . . . . . . . . . . . . . . . . . . . . 3
1.2.3 Relationships between the Two Studies . . . . . . . . 5
1.3 Studying Individuals . . . . . . . . . . . . . . . . . . . . . . 5
1.3.1 The Cloud of Individuals . . . . . . . . . . . . . . . . 5
1.3.2 Fitting the Cloud of Individuals . . . . . . . . . . . . 7
1.3.2.1 Best Plane Representation of NI . . . . . . . 7
1.3.2.2 Sequence of Axes for Representing NI . . . . 9
1.3.2.3 How Are the Components Obtained? . . . . 10
1.3.2.4 Example . . . . . . . . . . . . . . . . . . . . 10
1.3.3 Representation of the Variables as an Aid for
Interpreting the Cloud of Individuals . . . . . . . . . . 11
1.4 Studying Variables . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4.1 The Cloud of Variables . . . . . . . . . . . . . . . . . 13
1.4.2 Fitting the Cloud of Variables . . . . . . . . . . . . . . 14
1.5 Relationships between the Two Representations NI and NK 16
1.6 Interpreting the Data . . . . . . . . . . . . . . . . . . . . . . 17
1.6.1 Numerical Indicators . . . . . . . . . . . . . . . . . . . 17
1.6.1.1 Percentage of Inertia Associated with a
Component . . . . . . . . . . . . . . . . . . . 17
1.6.1.2 Quality of Representation of an Individual or
Variable . . . . . . . . . . . . . . . . . . . . . 18
1.6.1.3 Detecting Outliers . . . . . . . . . . . . . . . 19
1.6.1.4 Contribution of an Individual or Variable to
the Construction of a Component . . . . . . 19
1.6.2 Supplementary Elements . . . . . . . . . . . . . . . . . 20
1.6.2.1 Representing Supplementary Quantitative
Variables . . . . . . . . . . . . . . . . . . . . 21
1.6.2.2 Representing Supplementary Categorical
Variables . . . . . . . . . . . . . . . . . . . . 22
1.6.2.3 Representing Supplementary Individuals . . 23

v
vi Exploratory Multivariate Analysis by Example Using R

1.6.3 Automatic Description of the Components . . . . . . . 24

1.7 Implementation with FactoMineR . . . . . . . . . . . . . . . 25
1.8 Additional Results . . . . . . . . . . . . . . . . . . . . . . . . 26
1.8.1 Testing the Significance of the Components . . . . . . 26
1.8.2 Variables: Loadings versus Correlations . . . . . . . . 27
1.8.3 Simultaneous Representation: Biplots . . . . . . . . . 27
1.8.4 Missing Values . . . . . . . . . . . . . . . . . . . . . . 28
1.8.5 Large Datasets . . . . . . . . . . . . . . . . . . . . . . 28
1.8.6 Varimax Rotation . . . . . . . . . . . . . . . . . . . . 28
1.9 Example: The Decathlon Dataset . . . . . . . . . . . . . . . 29
1.9.1 Data Description — Issues . . . . . . . . . . . . . . . 29
1.9.2 Analysis Parameters . . . . . . . . . . . . . . . . . . . 31
1.9.2.1 Choice of Active Elements . . . . . . . . . . 31
1.9.2.2 Should the Variables Be Standardised? . . . 31
1.9.3 Implementation of the Analysis . . . . . . . . . . . . . 31
1.9.3.1 Choosing the Number of Dimensions to
Examine . . . . . . . . . . . . . . . . . . . . 32
1.9.3.2 Studying the Cloud of Individuals . . . . . . 33
1.9.3.3 Studying the Cloud of Variables . . . . . . . 36
1.9.3.4 Joint Analysis of the Cloud of Individuals and
the Cloud of Variables . . . . . . . . . . . . . 39
1.9.3.5 Comments on the Data . . . . . . . . . . . . 43
1.10 Example: The Temperature Dataset . . . . . . . . . . . . . . 44
1.10.1 Data Description — Issues . . . . . . . . . . . . . . . 44
1.10.2 Analysis Parameters . . . . . . . . . . . . . . . . . . . 44
1.10.2.1 Choice of Active Elements . . . . . . . . . . 44
1.10.2.2 Should the Variables Be Standardised? . . . 45
1.10.3 Implementation of the Analysis . . . . . . . . . . . . . 46
1.11 Example of Genomic Data: The Chicken Dataset . . . . . . 51
1.11.1 Data Description — Issues . . . . . . . . . . . . . . . 51
1.11.2 Analysis Parameters . . . . . . . . . . . . . . . . . . . 52
1.11.3 Implementation of the Analysis . . . . . . . . . . . . . 52

2 Correspondence Analysis (CA) 59

2.1 Data — Notation — Examples . . . . . . . . . . . . . . . . . 59
2.2 Objectives and the Independence Model . . . . . . . . . . . . 61
2.2.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.2.2 Independence Model and χ2 Test . . . . . . . . . . . . 62
2.2.3 The Independence Model and CA . . . . . . . . . . . 64
2.3 Fitting the Clouds . . . . . . . . . . . . . . . . . . . . . . . . 65
2.3.1 Clouds of Row Profiles . . . . . . . . . . . . . . . . . . 65
2.3.2 Clouds of Column Profiles . . . . . . . . . . . . . . . . 66
2.3.3 Fitting Clouds NI and NJ . . . . . . . . . . . . . . . . 68
2.3.4 Example: Women’s Attitudes to Women’s Work in France
in 1970 . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Contents vii

2.3.4.1 Column Representation (Mother’s Activity) . 70

2.3.4.2 Row Representation (Partner’s Work) . . . . 72
2.3.5 Superimposed Representation of Both Rows and
Columns . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.4 Interpreting the Data . . . . . . . . . . . . . . . . . . . . . . 77
2.4.1 Inertias Associated with the Dimensions (Eigenvalues) 77
2.4.2 Contribution of Points to a Dimension’s Inertia . . . . 80
2.4.3 Representation Quality of Points on a Dimension or
Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.4.4 Distance and Inertia in the Initial Space . . . . . . . . 82
2.5 Supplementary Elements (= Illustrative) . . . . . . . . . . . 83
2.6 Implementation with FactoMineR . . . . . . . . . . . . . . . 86
2.7 CA and Textual Data Processing . . . . . . . . . . . . . . . . 88
2.8 Example: The Olympic Games Dataset . . . . . . . . . . . . 92
2.8.1 Data Description — Issues . . . . . . . . . . . . . . . 92
2.8.2 Implementation of the Analysis . . . . . . . . . . . . . 94
2.8.2.1 Choosing the Number of Dimensions to
Examine . . . . . . . . . . . . . . . . . . . . 95
2.8.2.2 Studying the Superimposed Representation . 96
2.8.2.3 Interpreting the Results . . . . . . . . . . . . 96
2.8.2.4 Comments on the Data . . . . . . . . . . . . 100
2.9 Example: The White Wines Dataset . . . . . . . . . . . . . . 101
2.9.1 Data Description — Issues . . . . . . . . . . . . . . . 101
2.9.2 Margins . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.9.3 Inertia . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.9.4 Representation on the First Plane . . . . . . . . . . . 106
2.10 Example: The Causes of Mortality Dataset . . . . . . . . . . 109
2.10.1 Data Description — Issues . . . . . . . . . . . . . . . 109
2.10.2 Margins . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.10.3 Inertia . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
2.10.4 First Dimension . . . . . . . . . . . . . . . . . . . . . 115
2.10.5 Plane 2-3 . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.10.6 Projecting the Supplementary Elements . . . . . . . . 121
2.10.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 125

3 Multiple Correspondence Analysis (MCA) 127

3.1 Data — Notation — Examples . . . . . . . . . . . . . . . . . 127
3.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3.2.1 Studying Individuals . . . . . . . . . . . . . . . . . . . 128
3.2.2 Studying the Variables and Categories . . . . . . . . . 129
3.3 Defining Distances between Individuals and Distances between
Categories . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
3.3.1 Distances between the Individuals . . . . . . . . . . . 130
3.3.2 Distances between the Categories . . . . . . . . . . . . 130
3.4 CA on the Indicator Matrix . . . . . . . . . . . . . . . . . . 132
viii Exploratory Multivariate Analysis by Example Using R

3.4.1 Relationship between MCA and CA . . . . . . . . . . 132

3.4.2 The Cloud of Individuals . . . . . . . . . . . . . . . . 133
3.4.3 The Cloud of Variables . . . . . . . . . . . . . . . . . 134
3.4.4 The Cloud of Categories . . . . . . . . . . . . . . . . . 135
3.4.5 Transition Relations . . . . . . . . . . . . . . . . . . . 138
3.5 Interpreting the Data . . . . . . . . . . . . . . . . . . . . . . 140
3.5.1 Numerical Indicators . . . . . . . . . . . . . . . . . . . 140
3.5.1.1 Percentage of Inertia Associated with a
Component . . . . . . . . . . . . . . . . . . . 140
3.5.1.2 Contribution and Representation Quality of
an Individual or Category . . . . . . . . . . . 141
3.5.2 Supplementary Elements . . . . . . . . . . . . . . . . . 142
3.5.3 Automatic Description of the Components . . . . . . . 143
3.6 Implementation with FactoMineR . . . . . . . . . . . . . . . 145
3.7 Addendum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
3.7.1 Analysing a Survey . . . . . . . . . . . . . . . . . . . . 148
3.7.1.1 Designing a Questionnaire: Choice of Format 148
3.7.1.2 Accounting for Rare Categories . . . . . . . . 150
3.7.2 Description of a Categorical Variable or a
Subpopulation . . . . . . . . . . . . . . . . . . . . . . 150
3.7.2.1 Description of a Categorical Variable by a
Categorical Variable . . . . . . . . . . . . . . 150
3.7.2.2 Description of a Subpopulation (or a
Category) by a Quantitative Variable . . . . 151
3.7.2.3 Description of a Subpopulation (or a
Category) by the Categories of a Categorical
Variable . . . . . . . . . . . . . . . . . . . . . 152
3.7.3 The Burt Table . . . . . . . . . . . . . . . . . . . . . . 154
3.8 Example: The Survey on the Perception of Genetically
Modified Organisms . . . . . . . . . . . . . . . . . . . . . . . 155
3.8.1 Data Description — Issues . . . . . . . . . . . . . . . 155
3.8.2 Analysis Parameters and Implementation with
FactoMineR . . . . . . . . . . . . . . . . . . . . . . . . 158
3.8.3 Analysing the First Plane . . . . . . . . . . . . . . . . 159
3.8.4 Projection of Supplementary Variables . . . . . . . . . 160
3.8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 162
3.9 Example: The Sorting Task Dataset . . . . . . . . . . . . . . 162
3.9.1 Data Description — Issues . . . . . . . . . . . . . . . 162
3.9.2 Analysis Parameters . . . . . . . . . . . . . . . . . . . 164
3.9.3 Representation of Individuals on the First Plane . . . 164
3.9.4 Representation of Categories . . . . . . . . . . . . . . 165
3.9.5 Representation of the Variables . . . . . . . . . . . . . 166
Contents ix

4 Clustering 169
4.1 Data — Issues . . . . . . . . . . . . . . . . . . . . . . . . . . 169
4.2 Formalising the Notion of Similarity . . . . . . . . . . . . . . 173
4.2.1 Similarity between Individuals . . . . . . . . . . . . . 173
4.2.1.1 Distances and Euclidean Distances . . . . . . 173
4.2.1.2 Example of Non-Euclidean Distance . . . . . 174
4.2.1.3 Other Euclidean Distances . . . . . . . . . . 175
4.2.1.4 Similarities and Dissimilarities . . . . . . . . 175
4.2.2 Similarity between Groups of Individuals . . . . . . . 176
4.3 Constructing an Indexed Hierarchy . . . . . . . . . . . . . . 177
4.3.1 Classic Agglomerative Algorithm . . . . . . . . . . . . 177
4.3.2 Hierarchy and Partitions . . . . . . . . . . . . . . . . . 179
4.4 Ward’s Method . . . . . . . . . . . . . . . . . . . . . . . . . 179
4.4.1 Partition Quality . . . . . . . . . . . . . . . . . . . . . 180
4.4.2 Agglomeration According to Inertia . . . . . . . . . . 181
4.4.3 Two Properties of the Agglomeration Criterion . . . . 183
4.4.4 Analysing Hierarchies, Choosing Partitions . . . . . . 184
4.5 Direct Search for Partitions: K-means Algorithm . . . . . . . 185
4.5.1 Data — Issues . . . . . . . . . . . . . . . . . . . . . . 185
4.5.2 Principle . . . . . . . . . . . . . . . . . . . . . . . . . 186
4.5.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . 187
4.6 Partitioning and Hierarchical Clustering . . . . . . . . . . . . 187
4.6.1 Consolidating Partitions . . . . . . . . . . . . . . . . . 188
4.6.2 Mixed Algorithm . . . . . . . . . . . . . . . . . . . . . 188
4.7 Clustering and Principal Component Methods . . . . . . . . 188
4.7.1 Principal Component Methods Prior to AHC . . . . . 189
4.7.2 Simultaneous Analysis of a Principal Component Map
and Hierarchy . . . . . . . . . . . . . . . . . . . . . . . 189
4.8 Example: The Temperature Dataset . . . . . . . . . . . . . . 190
4.8.1 Data Description — Issues . . . . . . . . . . . . . . . 190
4.8.2 Analysis Parameters . . . . . . . . . . . . . . . . . . . 190
4.8.3 Implementation of the Analysis . . . . . . . . . . . . . 191
4.9 Example: The Tea Dataset . . . . . . . . . . . . . . . . . . . 197
4.9.1 Data Description — Issues . . . . . . . . . . . . . . . 197
4.9.2 Constructing the AHC . . . . . . . . . . . . . . . . . . 197
4.9.3 Defining the Clusters . . . . . . . . . . . . . . . . . . . 199
4.10 Dividing Quantitative Variables into Classes . . . . . . . . . 202

Appendix 205
A.1 Percentage of Inertia Explained by the First Component or by
the First Plane . . . . . . . . . . . . . . . . . . . . . . . . . . 205
A.2 R Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
A.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 210
A.2.2 The Rcmdr Package . . . . . . . . . . . . . . . . . . . 214
A.2.3 The FactoMineR Package . . . . . . . . . . . . . . . . 216
x Exploratory Multivariate Analysis by Example Using R

Bibliography of Software Packages 221

Bibliography 223

Index 225
Preface

Qu’est-ce que l’analyse des données ? (English: What is data analysis?)

As it is usually understood in France, and within the context of this book,
the expression analyse des données reflects a set of statistical methods whose
main features are to be multidimensional and descriptive.
The term multidimensional itself covers two aspects. First, it implies
that observations (or, in other words, individuals) are described by several
variables. In this introduction we restrict ourselves to the most common
data, those in which a group of individuals is described by one set of variables.
But, beyond the fact that we have many values from many variables for each
observation, it is the desire to study them simultaneously that is characteristic
of a multidimensional approach. Thus, we will use those methods each time
the notion of profile is relevant when considering an individual, for example,
the response profile of consumers, the biometric profile of plants, the financial
profile of businesses, and so forth.
From a dual point of view, the interest of considering values of individuals
for a set of variables in a global manner lies in the fact that these variables are
linked. Let us note that studying links between all the variables taken two-by-
two does not constitute a multidimensional approach in the strict sense. This
approach involves the simultaneous consideration of all the links between vari-
ables taken two-by-two. That is what is done, for example, when highlighting
a synthetic variable: such a variable represents several others, which implies
that it is linked to each of them, which is only possible if they are themselves
linked two-by-two. The concept of synthetic variable is intrinsically multi-
dimensional and is a powerful tool for the description of an individuals ×
variables table. In both respects, it is a key concept within the context of this
book.
One last comment about the term analyse des données since it can have
at least two meanings — the one defined previously and another broader one
that could be translated by “statistical investigation”. This second meaning
is from a user’s standpoint; it is defined by an objective (to analyse data)
and says nothing about the statistical methods to be used. This is what the
English term data analysis covers. The term data analysis, in the sense of a set
of descriptive multidimensional methods, is more of a French statistical point
of view. It was introduced in France in the 1960s by Jean-Paul Benzécri and
the adoption of this term is probably related to the fact that these multivariate
methods are at the heart of many “data analyses”.

xi
xii Exploratory Multivariate Analysis by Example Using R

To Whom Is This Book Addressed?

This book has been designed for scientists whose aim is not to become statis-
ticians but who feel the need to analyse data themselves. It is therefore
addressed to practitioners who are confronted with the analysis of data. From
this perspective it is application-oriented; formalism and mathematics writing
have been reduced as much as possible while examples and intuition have been
emphasised. Specifically, an undergraduate level is quite sufficient to capture
all the concepts introduced.
On the software side, an introduction to the R language is sufficient, at
least at first. This software is free and available on the Internet at the following
address: http://www.r-project.org/.
Content and Spirit of the Book
This book focuses on four essential and basic methods of multivariate ex-
ploratory data analysis, those with the largest potential in terms of applica-
tions: principal component analysis (PCA) when variables are quantitative,
correspondence analysis (CA) and multiple correspondence analysis (MCA)
when variables are categorical and hierarchical cluster analysis. The geo-
metric point of view used to present all these methods constitutes a unique
framework in the sense that it provides a unified vision when exploring mul-
tivariate data tables. Within this framework, we will present the principles,
the indicators, and the ways of representing and visualising objects (rows and
columns of a data table) that are common to all those exploratory methods.
From this standpoint, adding supplementary information by simply projecting
vectors is commonplace. Thus, we will show how it is possible to use categor-
ical variables within a PCA context where variables that are to be analysed
are quantitative, to handle more than two categorical variables within a CA
context where originally there are two variables, and to add quantitative vari-
ables within an MCA context where variables are categorical. More than
the theoretical aspects and the specific indicators induced by our geometrical
viewpoint, we will illustrate the methods and the way they can be exploited
using examples from various fields, hence the name of the book.
Throughout the text, each result correlates with its R command. All these
commands are accessible from FactoMineR, an R package developed by the
authors. The reader will be able to conduct all the analyses of the book as
all the datasets (as well as all the lines of code) are available at the following
Web site address: http://factominer.free.fr/book. We hope that with
this book, the reader will be fully equipped (theory, examples, software) to
confront multivariate real-life data.
The authors would like to thank Rebecca Clayton for her help in the transla-
tion.
1
Principal Component Analysis (PCA)

1.1 Data — Notation — Examples

Principal component analysis (PCA) applies to data tables where rows are
considered as individuals and columns as quantitative variables. Let xik be
the value taken by individual i for variable k, where i varies from 1 to I and
k from 1 to K.
Let x̄k denote the mean of variable k calculated over all individual instances
of I:
I
1X
x̄k = xik ,
I i=1

and sk the standard deviation of the sample of variable k (uncorrected):

v
u I
u1 X
sk = t (xik − x̄k )2 .
I i=1

Data subjected to a PCA can be very diverse in nature; some examples

are listed in Table 1.1.
This first chapter will be illustrated using the “orange juice” dataset chosen
for its simplicity since it comprises only six statistical individuals or observa-
tions. The six orange juices were evaluated by a panel of experts according
to seven sensory variables (odour intensity, odour typicality, pulp content, in-
tensity of taste, acidity, bitterness, sweetness). The panel’s evaluations are
summarised in Table 1.2.

1.2 Objectives
The data table can be considered either as a set of rows (individuals) or as a
set of columns (variables), thus raising a number of questions relating to these
different types of objects.
2 Exploratory Multivariate Analysis by Example Using R
TABLE 1.1
Some Examples of Datasets
Field Individuals Variables xik
Ecology Rivers Concentration of pollutants Concentration of pollu-
tant k in river i
Economics Years Economic indicators Indicator value k for year
i
Genetics Patients Genes Expression of gene k for
patient i
Marketing Brands Measures of satisfaction Value of measure k for
brand i
Pedology Soils Granulometric composition Content of component k
in soil i
Biology Animals Measurements Measure k for animal i
Sociology Social classes Time by activity Time spent on activity k
by individuals from so-
cial class i

TABLE 1.2
The Orange Juice Data
Odour Odour Pulp Intensity Acidity Bitter- Sweet-
intensity typicality of taste ness ness
Pampryl amb. 2.82 2.53 1.66 3.46 3.15 2.97 2.60
Tropicana amb. 2.76 2.82 1.91 3.23 2.55 2.08 3.32
Fruvita fr. 2.83 2.88 4.00 3.45 2.42 1.76 3.38
Joker amb. 2.76 2.59 1.66 3.37 3.05 2.56 2.80
Tropicana fr. 3.20 3.02 3.69 3.12 2.33 1.97 3.34
Pampryl fr. 3.07 2.73 3.34 3.54 3.31 2.63 2.90

1.2.1 Studying Individuals

Figure 1.1 illustrates the types of questions posed during the study of individ-
uals. This diagram represents three different situations where 40 individuals
are described in terms of two variables: j and k. In graph A, we can clearly
identify two distinct classes of individuals. Graph B illustrates a dimension of
variability which opposes extreme individuals, much like graph A, but which
also contains less extreme individuals. The cloud of individuals is therefore
long and thin. Graph C depicts a more uniform cloud (i.e., with no specific
structure).
Interpreting the data depicted in these examples is relatively straightfor-
ward as they are two-dimensional. However, when individuals are described
by a large number of variables, we require a tool to explore the space in which
these individuals evolve. Studying individuals means identifying the similari-
ties between individuals from the point of view of all the variables. In other
words, to provide a typology of the individuals: which are the most similar
individuals (and the most dissimilar)? Are there groups of individuals which
are homogeneous in terms of their similarities? In addition, we should look
Principal Component Analysis 3
A B C

2
1.0
1.0

1
0.5
0.5
Variable k

Variable k

Variable k
0
−1.0 −0.5 0.0

−0.5 0.0

−1 −2
−1.0
−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 −3 −2 −1 0 1 2
Variable j Variable j Variable j

FIGURE 1.1
Representation of 40 individuals described by two variables: j and k.

for common dimensions of variability which oppose extreme and intermediate

individuals.
In the example, two orange juices are considered similar if they were eval-
uated in the same way according to all the sensory descriptors. In such cases,
the two orange juices have the same main dimensions of variability and are
thus said to have the same sensory “profile”. More generally, we want to know
whether or not there are groups of orange juices with similar profiles, that is,
sensory dimensions which might oppose extreme juices with more intermediate
juices.

1.2.2 Studying Variables

Following the approach taken to study the individuals, might it also be possi-
ble to interpret the data from the variables? PCA focuses on the linear rela-
tionships between variables. More complex links also exist, such as quadratic
relationships, logarithmics, exponential functions, and so forth, but they are
not studied in PCA. This may seem restrictive, but in practice many relation-
ships can be considered linear, at least for an initial approximation.
Let us consider the example of the four variables (j, k, l, and m) in Fig-
ure 1.2. The clouds of points constructed by working from pairs of variables
show that variables j and k (graph A) as well as variables l and m (graph F)
are strongly correlated (positively for j and k and negatively for l and m).
However, the other graphs do not show any signs of relationships between
variables. The study of these variables also suggests that the four variables
are split into two groups of two variables, (j, k) and (l, m) and that, within
one group, the variables are strongly correlated, whereas between groups, the
variables are uncorrelated. In exactly the same way as for constructing groups
of individuals, creating groups of variables may be useful with a view to syn-
thesis. As for the individuals, we identify a continuum with groups of both
very unusual variables and intermediate variables, which are to some extent
4 Exploratory Multivariate Analysis by Example Using R

linked to both groups. In the example, each group can be represented by one
single variable as the variables within each group are very strongly correlated.
We refer to these variables as synthetic variables.

A B C

0.0

0.0
1.0 0.5

−0.4

−0.4
Variable k

Variable l

Variable l
0.0

−0.8

−0.8
−0.5
−1.0

−1.2

−1.2
−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0
Variable j Variable j Variable k

D E F
1.0

1.0

1.0
0.4 0.6 0.8

0.4 0.6 0.8

0.4 0.6 0.8

Variable m

Variable m

Variable m
0.2

0.2

0.2
0.0

0.0

0.0

−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 −1.2 −0.8 −0.4 0.0
Variable j Variable k Variable l

FIGURE 1.2
Representation of the relationships between four variables: j, k, l, and m,
taken two-by-two.

When confronted with a very small number of variables, it is possible to

draw conclusions from the clouds of points, or from the correlation matrix
which groups together all of the linear correlation coefficients r(j, k) between
the pairs of variables. However, when working with a great number of vari-
ables, the correlation matrix groups together a large quantity of correlation
coefficients (190 coefficients for K = 20 variables). It is therefore essential to
have a tool capable of summarising the main relationships between the vari-
ables in a visual manner. The aim of PCA is to draw conclusions from the
linear relationships between variables by detecting the principal dimensions
of variability. As you will see, these conclusions will be supplemented by the
definition of the synthetic variables offered by PCA. It will therefore be eas-
ier to describe the data using a few synthetic variables rather than all of the
original variables.
In the example of the orange juice data, the correlation matrix (see Ta-
ble 1.3) brings together the 21 correlation coefficients. It is possible to group
the strongly correlated variables into sets, but even for this reduced number
of variables, grouping them this way is tedious.
Principal Component Analysis 5

TABLE 1.3
Orange Juice Data: Correlation Matrix
Odour Odour Pulp Intensity Acidity Bitter- Sweet-
intensity typicality of taste ness ness
Odour intensity 1.00 0.58 0.66 −0.27 −0.15 −0.15 0.23
Odour typicality 0.58 1.00 0.77 −0.62 −0.84 −0.88 0.92
Pulp content 0.66 0.77 1.00 −0.02 −0.47 −0.64 0.63
Intensity of taste −0.27 −0.62 −0.02 1.00 0.73 0.51 −0.57
Acidity −0.15 −0.84 −0.47 0.73 1.00 0.91 −0.90
Bitterness −0.15 −0.88 −0.64 0.51 0.91 1.00 −0.98
Sweetness 0.23 0.92 0.63 −0.57 −0.90 −0.98 1.00

1.2.3 Relationships between the Two Studies

The study of individuals and the study of variables are interdependent as
they are carried out on the same data table: studying them jointly can only
reinforce their respective interpretations.
If the study of individuals led to a distinction between groups of individ-
uals, it is then possible to list the individuals belonging to only one group.
However, for high numbers of individuals, it seems more pertinent to char-
acterise them directly by the variables at hand: for example, by specifying
that some orange juices are acidic and bitter whereas others have a high-pulp
content.
Similarly, when there are groups of variables, it may not be easy to inter-
pret the relationships between many variables and we can make use of specific
individuals, that is, individuals who are extreme from the point of view of
these relationships. In this case, it must be possible to identify the individu-
als. For example, the link between acidity-bitterness can be illustrated by the
opposition between two extreme orange juices: Fresh Pampryl (orange juice
from Spain) versus Fresh Tropicana (orange juice from Florida).

1.3 Studying Individuals

1.3.1 The Cloud of Individuals
An individual is a row of the data table, that is, a set of K numerical values.
The individuals thus evolve within a space RK called “the individual’s space”.
If we endow this space with the usual Euclidean distance, the distance between
two individuals i and l is expressed as:
v
uK
uX
d(i, l) = t (xik − xlk )2 .
k=1
6 Exploratory Multivariate Analysis by Example Using R

If two individuals have similar values within the table of all K variables, they
are also close in the space RK . Thus, the study of the data table can be
conducted geometrically by studying the distances between individuals. We
are therefore interested in all of the individuals in RK , that is, the cloud
of individuals (denoted NI ). Analysing the distances between individuals is
therefore tantamount to studying the shape of the cloud of points. Figure 1.3
illustrates a cloud of point is within a space RK for K = 3.

FIGURE 1.3
Flight of a flock of starlings illustrating a scatterplot in RK .

The shape of cloud NI remains the same even when translated. The data
are also centred, which corresponds to considering xik − x̄k rather than xik .
Geometrically, this is tantamount to coinciding the centre of mass of the cloud
GI (with coordinates x̄k for k = 1, ..., K) with the origin of reference (see
Figure 1.4). Centring presents technical advantages and is always conducted
in PCA.
The operation of reduction (also referred to as standardising), which con-
sists of considering (xik − x̄k )/sk rather than xik , modifies the shape of the
cloud by harmonising its variability in all the directions of the original vectors
(i.e., the K variables). Geometrically, it means choosing standard deviation
sk as a unit of measurement in direction k. This operation is essential if the
variables are not expressed in the same units. Even when the units of mea-
surement do not differ, this operation is generally preferable as it attaches
the same importance to each variable. Therefore, we will assume this to be
the case from here on in. Standardised PCA occurs when the variables are
Principal Component Analysis 7

FIGURE 1.4
Scatterplot of the individuals in RK .

centred and reduced, and unstandardised PCA when the variables are only
centred. When not otherwise specified, it may be assumed that we are using
standardised PCA.
Comment: Weighting Individuals
So far we have assumed that all individuals have the same weight. This applies
to almost all applications and is always assumed to be the case. Neverthe-
less, generalisation with unspecified weights poses no conceptual or practical
problems (double weight is equivalent to two identical individuals) and most
software packages, including FactoMineR envisage this possibility (FactoMineR
is a package dedicated to Factor Analysis and Data Mining with R, see Sec-
tion A.2.3 in the Appendix). For example, it may be useful to assign a different
weight to each individual after having rectified a sample. In all cases, it is
convenient to consider that the sum of the weights is equal to 1. If supposed
to be of the same weight, each individual will be assigned a weight of 1/I.

1.3.2 Fitting the Cloud of Individuals

1.3.2.1 Best Plane Representation of NI
The aim of PCA is to represent the cloud of points in a space with reduced
dimensions in an “optimal” manner, that is to say, by distorting the distances
between individuals as little as possible. Figure 1.5 gives two representations
of three different fruits. The viewpoints chosen for the images of the fruits on
the top line make them difficult to identify. On the second row, the fruits can
be more easily recognised. What is it which differentiates the views of each
fruit between the first and the second lines? In the pictures on the second line,
8 Exploratory Multivariate Analysis by Example Using R

the distances are less distorted and the representations take up more space
on the image. The image is a projection of a three-dimensional object in a
two-dimensional space.

FIGURE 1.5
Two-dimensional representations of fruits: from left to right an avocado, a
melon and a banana, each row corresponds to a different representation.

For a representation to be successful, it must select an appropriate view-

point. More generally, PCA means searching for the best representational
space (of reduced dimension) thus enabling optimal visualisation of the shape
of a cloud with K dimensions. We often use a plane representation alone,
which can prove inadequate when dealing with particularly complex data.
To obtain this representation, the cloud NI is projected on a plane of RK
denoted P , chosen in such a manner as to minimise distortion of the cloud
of points. Plane P is selected so that the distances between the projected
points might be as close as possible to the distances between the initial points.
Since, in projection, distances can only decrease, we try to make the projected
distances as high as possible. By denoting Hi the projection of the individual
i on plane P , the problem consists of finding P , with:
I
X
OHi2 maximum.
i=1

The convention for notation uses mechanical terms: O is the centre of gravity,
OHi is a vector and the criterion is the inertia of the projection of NI . The
criterion which consists of increasing the variance of the projected points to a
maximum is perfectly appropriate.
Remark
If the individuals are weighted with different weights pi , the maximised crite-
PI
rion is i=1 pi OHi2 .

In some rare cases, it might be interesting to search for the best axial
representation of cloud NI alone. This best axis is obtained in the same way:
Principal Component Analysis 9
PI
find the component u1 when i=1 OHi2 are maximum (where Hi is the pro-
jection of i on u1 ). It can be shown that plane P contains component u1 (the
“best” plane contains the “best”component): in this case, these representa-
tions are said to be nested. An illustration of this property is presented in
Figure 1.6. Planets, which are in a three-dimensional space, are traditionally
represented on a component. This component determines their positions as
well as possible in terms of their distances from one other (in terms of inertia
of the projected cloud). We can also represent planets on a plane according
to the same principle: to maximise the inertia of the projected scatterplot
(on the plane). This best plane representation also contains the best axial
representation.

ne
Su ury

s
r
rn

nu
te

tu
o
c
tu

ut
pi
n

ra

ep
er
Sa

Pl
Ju

U
M

N
M h
Ve s
s
rt
ar
nu
Ea

Uranus
Mars
Saturn Earth Sun
Mercury Venus Neptune

Jupiter

Pluto

FIGURE 1.6
The best axial representation is nested in the best plane representation of the
solar system (18 February 2008).

We define plane P by two nonlinear vectors chosen as follows: vector u1

which defines the best axis (and which is included in P ), and vector u2 of
the plane P orthogonal to u1 . Vector u2 corresponds to the vector which
expresses the greatest variability of NI once that which is expressed by u1 is
removed. In other words, the variability expressed by u2 is the best coupling
and is independent of that expressed by u1 .

1.3.2.2 Sequence of Axes for Representing NI

More generally, let us look for nested subspaces of dimensions s = 1 to S
so that each subspace is of maximum inertia for the given dimension s. The
10 Exploratory Multivariate Analysis by Example Using R
PI 2
subspace of dimension s is obtained by maximising i=1 (OHi ) (where Hi
is the projection of i in the subspace of dimension s). As the subspaces
are nested, it is possible to choose vector us as the vector of the orthogonal
subspace for all of the vectors ut (with 1 ≤ t < s) which define the smaller
subspaces.
The first plane (defined by u1 , u2 ), i.e., the plane of best representation, is
often sufficient for visualising cloud NI . When S is greater than or equal to 3,
we may need to visualise cloud NI in the subspace of dimension S by using a
number of plane representations: the representation on (u1 , u2 ) but also that
on (u3 , u4 ) which is the most complementary to that on (u1 , u2 ). However, in
certain situations, we might choose to associate (u2 , u3 ) for example, in order
to highlight a particular phenomenon which appears on these two components.

1.3.2.3 How Are the Components Obtained?

Components in PCA are obtained through diagonalisation of the correlation
matrix which extracts the associated eigenvectors and eigenvalues. The eigen-
vectors correspond to vectors us which are each associated with the eigenvalues
of rank s (denoted λs ), as the eigenvalues are ranked in descending order. The
eigenvalue λs is interpreted as the inertia of cloud NI projected on the compo-
nent of rank s or, in other words, the “explained variance” for the component
of rank s.
If all of the eigenvectors are calculated (S = K), the PCA recreates a basis
for the space RK . In this sense, PCA can be seen as a change of basis in which
the first vectors of the new basis play an important role.

Remark
When variables are centred but not standardised, the matrix to be diago-
nalised is the variance–covariance matrix.

1.3.2.4 Example
The distance between two orange juices is calculated using their seven sensory
descriptors. We decided to standardise the data to attribute each descriptor
equal influence. Figure 1.7 is obtained from the first two components of the
PCA and corresponds to the best plane for representing the cloud of individu-
als in terms of projected inertia. The inertia projected on the plane is the sum
of two eigenvalues, that is, 86.82% (= 67.77% + 19.05%) of the total inertia
of the cloud of points.
The first principal component, that is, the principal axis of variability
between the orange juices, separates the two orange juices Tropicana fr. and
Pampryl amb. According to data Table 1.2, we can see that these orange
juices are the most extreme in terms of the descriptors odour typicality and
bitterness: Tropicana fr. is the most typical and the least bitter while Pampryl
amb. is the least typical and the most bitter. The second component, that
is, the property that separates the orange juices most significantly once the
Principal Component Analysis 11

Pampryl fr.

2
Dim 2 (19.05%)
1
Tropicana fr.
Fruvita fr.
0
Pampryl amb.
-1

Joker amb.
Tropicana amb.
-2

-4 -2 0 2 4
Dim 1 (67.77%)

FIGURE 1.7
Orange juice data: plane representation of the scatterplot of individuals.

main principal component of variability has been removed, identifies Tropicana

amb., which is the least intense in terms of odour, and Pampryl fr., which is
among the most intense (see Table 1.2).
Reading this data is tedious when there are a high number of individuals
and variables. For practical purposes, we will facilitate the characterisation
of the principal components by using the variables more directly.

1.3.3 Representation of the Variables as an Aid for

Interpreting the Cloud of Individuals
Let Fs denote the coordinate of the I individuals on component s and Fs (i)
its value for individual i. Vector Fs is also called the principal component of
rank s. Fs is of dimension I and thus can be considered as a variable. To
interpret the relative positions of the individuals on the component of rank s,
it may be interesting to calculate the correlation coefficient between vector Fs
and the initial variables. Thus, when the correlation coefficient between Fs
and a variable k is positive (or indeed negative), an individual with a positive
coordinate on component Fs will generally have a high (or low, respectively)
value (relative to the average) for variable k.
In the example, F1 is strongly positively correlated with the variables
odour typicality and sweetness and strongly negatively correlated with the
variables bitter and acidic (see Table 1.4). Thus Tropicana fr., which has the
highest coordinate on component 1, has high values for odour typicality and
sweetness and low values for the variables acidic and bitter. Similarly, we
can examine the correlations between F2 and the variables. It may be noted
that the correlations are generally lower (in absolute value) than those with
the first principal component. We will see that this is directly linked to the
percentage of inertia associated with F2 which is, by construction, lower than
12 Exploratory Multivariate Analysis by Example Using R

that associated with F1 . The second component can be characterised by the

variables odour intensity and pulp content (see Table 1.4).

TABLE 1.4
Orange Juice Data: Correlation between
Variables and First Two Components
F1 F2
Odour intensity 0.46 0.75
Odour typicality 0.99 0.13
Pulp content 0.72 0.62
Intensity of taste −0.65 0.43
Acidity −0.91 0.35
Bitterness −0.93 0.19
Sweetness 0.95 −0.16

To make these results easier to interpret, particularly in cases with a high

number of variables, it is possible to represent each variable on a graph, using
its correlation coefficients with F1 and F2 as coordinates (see Figure 1.8).
Variables factor map (PCA)
1.0

Odour intensity
0.62 Pulpiness
Dimension 2 (19.05%)

Intensity of taste
0.5

Acidity
Bitterness
Odour typicality
0.0

0.72
Sweetness
-0.5
-1.0

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

Dimension 1 (67.77%)

FIGURE 1.8
Orange juice data: visualisation of the correlation coefficients between vari-
ables and the principal components F1 and F2 .

We can now interpret the joint representation of the cloud of individuals

with this representation of the variables.
Remark
A variable is always represented within a circle of radius 1 (circle represented
in Figure 1.8): indeed, it must be noted that F1 and F2 are orthogonal (in
the sense that their correlation coefficient is equal to 0) and that a variable
cannot be strongly related to two orthogonal components simultaneously. In
the following section we shall examine why the variable will always be found
within the circle of radius 1.
Principal Component Analysis 13

1.4 Studying Variables

1.4.1 The Cloud of Variables
Let us now consider the data table as a set of columns. A variable is one of the
columns in the table, that is, a set of I numerical values, which is represented
by a point of the vector space with I dimensions, denoted RI (and known as
the “variables’ space”). The vector connects the origin of RI to the point. All
of these vectors constitute the cloud of variables and this ensemble is denoted
NK (see Figure 1.9).

O
1

FIGURE 1.9
The scatterplot of the variables NK in RI . In the case of a standardised PCA,
the variables k are located within a hypersphere of radius 1.

The scalar product between two variables k and l is expressed as:

I
X
xik × xil = kkk × klk × cos(θkl ).
i=1

with kkk and klk the norm for variable k and l, and θkl the angle produced by
the vectors representing variables k and l. Since the variables used here are
centred, the norm for one variable is equal to its standard deviation multiplied
by the square root of I, and the scalar product is expressed as follows:
I
X
(xik − x̄k ) × (xil − x̄l ) = I × sk × sl × cos(θkl ).
i=1
14 Exploratory Multivariate Analysis by Example Using R

On the right-hand side of the equation, we can identify covariance between

variables k and l.
Similarly, by dividing each term in the equation by the standard deviations
sk and sl of variables k and l, we obtain the following relationship:

r(k, l) = cos(θkl ).

This property is essential in PCA as it provides a geometric interpretation of

the correlation. Therefore, in the same way as the representation of cloud NI
can be used to visualise the variability between individuals, a representation of
the cloud NK can be used to visualise all of the correlations (through the angles
between variables) or, in other words, the correlation matrix. To facilitate
visualisation of the angles between variables, the variables are represented by
vectors rather than points.
Generally speaking, the variables being centred and reduced (scaled to
unit variance) have a length with a value of 1 (hence the term “standardised
variable”). The vector extremities are therefore on the sphere of radius 1 (also
called “hypersphere” to highlight the fact that, in general, I > 3), as shown
in Figure 1.9.
Comment about the Centring
In RK , when the variables are centred, the origin of the axes is translated
onto the mean point. This property is not true for NK .

1.4.2 Fitting the Cloud of Variables

As is the case for the individuals, the cloud of variables NK is situated in a
space RI with a great number of dimensions and it is impossible to visualise the
cloud in the overall space. The cloud of variables must therefore be adjusted
using the same strategy
PK as for the cloud of individuals. We maximise an
2
equivalent criterion k=1 (OHk ) with Hk , the projection of variable k on
the subspace with reduced dimensions. Here too, the subspaces are nested
and we can identify a series of orthogonal axes S which define the subspaces
for dimensions s = 1 to S. Vector vs therefore belongs to a given subspace
and is orthogonal to the vectors vt which make up the PKsmaller subspaces. It
can therefore be shown that the vector vs maximises k=1 (OHks )2 where Hks
is the projection of variable k on vs .
Remark
In the individual space RK , centring the variables causes the origin of the
axes to shift to a mean point: the maximised criterion is therefore interpreted
as a variance; the projected points must be as dispersed as possible. In RI ,
centring has a different effect, as the origin is not the same as the mean point.
The projected points should be as far as possible from the origin (although not
necessarily dispersed), even if that means being grouped together or merged.
This indicates that the position of the cloud NK with respect to the origin is
important.
Principal Component Analysis 15

Vectors vs (s = 1, ..., S) belong to the space RI and consequently can be

considered new variables. The correlation coefficient r(k, vs ) between variables
k and vs is equal to the cosine of the angle θks between Ok and vs when variable
k is centred and scaled, and thus standardised. The plane representation
constructed by (v1 , v2 ) is therefore pleasing as the coordinates of a variable k
correspond to the cosine of the angle θk1 and that of angle θk2 , and thus the
correlation coefficients between variables k and v1 , and between variables k
and v2 . In a plane representation such as this, we can therefore immediately
visualise whether or not a variable k is related to a dimension of variability
(see Figure 1.10).
PK 2
By their very construction, variables vs maximise criterion k=1 (OHks ) .
s
Since the projection of a variable k is equal to the cosine of angle θk , the
criterion maximises:
XK K
X
cos2 θks = r2 (k, vs ).
k=1 k=1

The above expression illustrates that vs is the new variable which is the most
strongly correlated with all of the initial variables K (with the orthogonality
constraint of vt already found). As a result, vs can be said to be a synthetic
variable. Here, we are experiencing the second aspect of the study of variables
(see Section 1.2.2).

A
HA
HB
HA
D HB HD

HD HC
HC

FIGURE 1.10
Projection of the scatterplot of the variables on the main plane of variabil-
ity. On the left: visualisation in space RI ; on the right: visualisation of the
projections in the principal plane.

Remark
When a variable is not standardised, its length is equal to its standard deviation.
16 Exploratory Multivariate Analysis by Example Using R

In an unstandardised PCA, the criterion can be expressed as follows:

K
X K
X
2
(OHks ) = s2k r2 (k, vs ) .
k=1 k=1

This highlights the fact that, in the case of an unstandardised PCA, each
variable k is assigned a weight equal to its variance s2k .

It can be shown that the axes of representation NK are in fact eigenvec-

tors of the scalar products matrix between individuals. This property is, in
practice, only used when the number of variables exceeds the number of in-
dividuals. We will see in the following that these eigenvectors are deducted
from those of the correlation matrix.
The best plane representation of the cloud of variables corresponds exactly
to the graph representing the variables obtained as an aid to interpreting the
representation of individuals (see Figure 1.8). This remarkable property is not
specific to the example but applies when carrying out any standardised PCA.
This point will be developed further in the following section.

1.5 Relationships between the Two Representations NI

and NK
So far we have looked for representations of NI and NK according to the same
principle and from one single data table. It therefore seems natural for these
two analyses (NI in RK and NK in RI ) to be related.
The relationships between the two clouds NI and NK are brought to-
gether under the general term of “relations of duality”. This term refers to
the dual approach of one single data table, by considering either the lines or
the columns. This approach is also defined by “transition relations” (calcu-
lating the coordinates in one space from those in the other). Where Fs (i) is
the coordinate of individual i and Gs (k) the coordinate of variable k of the
component of rank s, we obtain the following equations:

K
1 X
Fs (i) = √ xik Gs (k),
λs k=1
I
1 X
Gs (k) = √ (1/I) xik Fs (i).
λs i=1

This result is essential for interpreting the data, and makes PCA a rich
and reliable experimental tool. This may be expressed as follows: individuals
Principal Component Analysis 17

are on the same side as their corresponding variables with high values, and
opposite their corresponding variables with low values. It must be noted that
xik are centred and carry both positive and negative values. This is one of the
reasons why individuals can be so far from the variable for which they carry
low values. Fs is referred to as the principal component of rank s; λs is the
variance of Fs and its square root the length of Fs in RI ; vs is known as the
standardised principal component.
The total inertias of both clouds are equal (and equal to K for standardised
PCA) and furthermore, when decomposed component by component, they
are identical. This property is remarkable: if S dimensions are enough to
perfectly represent NI , the same is true for NK . In this case, two dimensions
are sufficient. If not, why generate a third synthetic variable which would not
differentiate the individuals at all?

1.6 Interpreting the Data

1.6.1 Numerical Indicators
1.6.1.1 Percentage of Inertia Associated with a Component
The first indicators that we shall examine are the ratios between the projected
inertias and the total inertia. For component s:
PI 1 s 2 PK 2
i=1 I (OHi ) k=1 (OHks ) λs
PI 1 2
= PK = PK .
i=1 I (Oi) k=1 Ok 2 s=1 λs
PK
In the most usual case, when the PCA is standardised, s=1 λs = K.
When multiplied by 100, this indicator gives the percentage of inertia (of NI
in RK or of NK in RI ) expressed by the component of rank s. This can be
interpreted in two ways:
1. As a measure of the quality of data representation; in the example,
we say that the first component expresses 67.77% of data variability
(see Table 1.5). In a standardised PCA (where I > K), we often
compare λs with 1, the value below which the component of rank
s, representing less data than a stand alone variable, is not worthy
of interest.
2. As a measure of the relative importance of the components; in the
example, we say that the first component expresses three times more
variability than the second; it affects three times more variables but
this formulation is truly precise only when each variable is perfectly
correlated with a component.
18 Exploratory Multivariate Analysis by Example Using R

Because of the orthogonality of the axes (both in RK and in RI ), these iner-

tia percentages can be added together for several components; in the example,
86.82% of the data are represented by the first two components (67.77% +
19.05% = 86.82%).

TABLE 1.5
Orange Juice Data: Decomposition of Variability
per Component
Eigenvalue Percentage of Cumulative
variance of variance percentage
Comp 1 4.74 67.77 67.77
Comp 2 1.33 19.05 86.81
Comp 3 0.82 11.71 98.53
Comp 4 0.08 1.20 99.73
Comp 5 0.02 0.27 100.00

Let us return to Figure 1.5: the pictures of the fruits on the first line cor-
respond approximately to a projection of the fruits on the plane constructed
by components 2 and 3 of PCA, whereas the images on the second line cor-
respond to a projection on plane 1-2. This is why the fruits are easier to
recognise on the second line: the more variability (i.e., the more information)
collected on plane 1-2 when compared to plane 2-3, the easier it is to grasp the
overall shape of the cloud. Furthermore, the banana is more easy to recognise
in plane 1-2 (the second line), as it retrieves greater inertia on plane 1-2. In
concrete terms, as the banana is a longer fruit than a melon, this leads to
more marked differences in inertia between the components. As the melon is
almost spherical, the percentages of inertia associated with each of the three
components are around 33% and therefore the inertia retrieved by plane 1-2
is nearly 66% (as is that retrieved by plane 2-3).

1.6.1.2 Quality of Representation of an Individual or Variable

The quality of representation of an individual i on the component s can be
measured by the distance between the point within the space and the projec-
tion on the component. In reality, it is preferable to calculate the percentage
of inertia of the individual i projected on the component s. Therefore, when
θis is the angle between Oi and us , we obtain:

Projected inertia of i on us
qlts (i) = = cos2 θis .
Total inertia of i
Using Pythagoras’ theorem, this indicator is combined for multiple compo-
nents and is most often calculated for a plane.
The quality of representation of a variable k on the component of rank s
is expressed as:
Projected inertia of k on vs
qlts (k) = = cos2 θks .
Total inertia of k
Principal Component Analysis 19

This last quantity is equal to r2 (k, vs ), which is why the quality of represen-
tation of a variable is only very rarely provided by software. The representa-
tional quality of a variable in a given plane is obtained directly on the graph
by visually evaluating its distance from the circle of radius 1.

1.6.1.3 Detecting Outliers

Analysing the shape of the cloud NI also means detecting unusual or remark-
able individuals. An individual is considered remarkable if it has extreme
values for multiple variables. In the cloud NI , an individual such as this is far
from the cloud’s centre of gravity, and its remarkable nature can be evaluated
from its distance from the centre of the cloud in the overall space RK .
In the example, none of the orange juices are particularly extreme (see
Table 1.6). The two most extreme individuals are Tropicana ambient and
Pampryl fresh.

TABLE 1.6
Orange Juice Data: Distances from the Individuals to the Centre of the
Cloud
Pampryl amb. Tropicana amb. Fruvita fr. Joker amb. Tropicana fr. Pampryl fr.
3.03 1.98 2.59 2.09 3.51 2.34

1.6.1.4 Contribution of an Individual or Variable to the

Construction of a Component
Outliers have an influence on analysis, and it is interesting to know to what
extent their influence affects the construction of the components. Further-
more, some individuals can influence the construction of certain components
without being remarkable themselves. Detecting those individuals that con-
tribute to the construction of a principal component helps to evaluate the
component’s stability. It is also interesting to evaluate the contribution of
variables in constructing a component (especially in nonstandardised PCA).
To do so, we decompose the inertia of a component individual by individual
(or variable by variable). The inertia explained by the individual i on the
component s is:
2
(1/I) (OHis )
× 100.
λs
Distances intervene in the components by their squares, which augments the
roles of those individuals at a greater distance from the origin. Outlying
individuals are the most extreme on the component, and their contributions
are especially useful when the individuals’ weights are different.
Remark
These contributions are combined for a multiple individuals.
When an individual contributes significantly (i.e., much more than the
20 Exploratory Multivariate Analysis by Example Using R

others) to the construction of a principal component (for example Tropicana

ambient and Pampryl fresh; for the second component, see Table 1.7), it is
not uncommon for the results of a new PCA constructed without this indi-
vidual to change substantially: the principal components can change and new
oppositions between individuals may appear.

TABLE 1.7
Orange Juice Data: Contribution of
Individuals to the Construction of the
Components
Dim.1 Dim.2
Pampryl amb. 31.29 0.08
Tropicana amb. 2.76 36.77
Fruvita fr. 13.18 0.02
Joker amb. 12.63 8.69
Tropicana fr. 35.66 4.33
Pampryl fr. 4.48 50.10

Similarly, the contribution of variable k to the construction of component

s is calculated. An example of this is presented in Table 1.8.

TABLE 1.8
Orange Juice Data: Contribution of Variables to the
Construction of the Components
Dim.1 Dim.2
Odour intensity 4.45 42.69
Odour typicality 20.47 1.35
Pulp content 10.98 28.52
Taste intensity 8.90 13.80
Acidity 17.56 9.10
Biterness 18.42 2.65
Sweetness 19.22 1.89

1.6.2 Supplementary Elements

We here define the concept of active and supplementary (or illustrative) el-
ements. By definition, active elements contribute to the construction of the
principal components, contrary to supplementary elements. Thus, the inertia
of the cloud of individuals is calculated on the basis of active individuals, and
similarly, the inertia of the cloud of variables is calculated on the basis of
active variables. The supplementary elements make it possible to illustrate
the principal components, which is why they are referred to as “illustrative
elements”. Contrary to the active elements, which must be homogeneous, we
can make use of as many illustrative elements as possible.
Principal Component Analysis 21

1.6.2.1 Representing Supplementary Quantitative Variables

By definition, a supplementary quantitative variable plays no role in calcu-
lating the distances between individuals. They are represented in the same
way as active variables; to assist in interpreting the cloud of individuals (Sec-
tion 1.3.3). The coordinate of the supplementary variable k 0 on the component
s corresponds to the correlation coefficient between k 0 and the principal com-
ponent s (i.e., the variable whose values are the coordinates of the individuals
projected on the component of rank s). k 0 can therefore be represented on
the same graph as the active variables.
More formally, the transition formulae can be used to calculate the coor-
dinate of the supplementary variable k 0 on the component of rank s:

1 X
Gs (k 0 ) = √ xik0 Fs (i) = r(k, Fs ),
λs
i∈{active}

where {active} refers to the set of active individuals. This coordinate is cal-
culated from the active individuals alone.
In the example, in addition to the sensory descriptors, there are also physic-
ochemical variables at our disposal (see Table 1.9). However, our stance re-
mains unchanged; namely, to describe the orange juices based on their sensory
profiles. This problem can be enriched using the supplementary variables since
we can now link sensory dimensions to the physicochemical variables.

TABLE 1.9
Orange Juice Data: Supplementary Variables
Glucose Fructose Saccharose Sweetening pH Citric Vitamin C
power acid
Pampryl amb. 25.32 27.36 36.45 89.95 3.59 0.84 43.44
Tropicana amb. 17.33 20.00 44.15 82.55 3.89 0.67 32.70
Fruvita fr. 23.65 25.65 52.12 102.22 3.85 0.69 37.00
Joker amb. 32.42 34.54 22.92 90.71 3.60 0.95 36.60
Tropicana fr. 22.70 25.32 45.80 94.87 3.82 0.71 39.50
Pampryl fr. 27.16 29.48 38.94 96.51 3.68 0.74 27.00

The correlations circle (Figure 1.11) represents both the active and sup-
plementary variables. The main component of variability opposes the orange
juices perceived as acidic/bitter, slightly sweet and somewhat typical with the
orange juices perceived as sweet, typical, not very acidic and slightly bitter.
The analysis of this sensory perception is reinforced by the variables pH and
saccharose. Indeed, these two variables are positively correlated with the first
component and lie on the side of the orange juices perceived as sweet and
slightly acidic (a high pH index indicates low acidity). One also finds the re-
action known as “saccharose inversion” (or hydrolysis): the saccharose breaks
down into glucose and fructose in an acidic environment (the acidic orange
juices thus contain more fructose and glucose, and less saccharose than the
average).
22 Exploratory Multivariate
Variables Analysis by Example Using R
factor map (PCA)

1.0
Odour.intensity
Sweetening.power
Pulpiness

Intensity.of.taste

0.5
Acidity Fructose
Glucose Saccharose
Dim 2 (19.05%)
Bitterness
Odour.typicality

0.0 Citric.acid Sweetness

pH
Vitamin.C
-0.5
-1.0

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

Dim 1 (67.77%)

FIGURE 1.11
Orange juice data: representation of the active and supplementary variables.

Remark
When using PCA to explore data prior to a multiple regression, it is advisable
to choose the explanatory variables for the regression model as active variables
for PCA, and to project the variable to be explained (the dependent variable)
as a supplementary variable. This gives some idea of the relationships between
explanatory variables and thus of the need to select explanatory variables.
This also gives us an idea of the quality of the regression: if the dependent
variable is appropriately projected, it will be a well-fitted model.

1.6.2.2 Representing Supplementary Categorical Variables

In PCA, the active variables are quantitative by nature but it is possible to
use information resulting from categorical variables on a purely illustrative
basis (= supplementary), that is, not used to calculate the distances between
individuals.
The categorical variables cannot be represented in the same way as the
supplementary quantitative variables since it is impossible to calculate the
correlation between a categorical variable and Fs . Information about cate-
gorical variables lies within their categories. It is quite natural to represent
a categorical variable at the barycentre of all the individuals possessing that
variable. Thus, following projection on the plane defined by the principal
components, these categories remain at the barycentre of the individuals in
their plane representation. A categorical variable can thus be regarded as the
mean individual obtained from the set of individuals who have it. This is
therefore the way in which it will be represented on the graph of individuals.
The information resulting from a supplementary categorical variable can
Discovering Diverse Content Through
Random Scribd Documents
 ‘The suitor of truth? Thou?’ Thus they mocked.
‘Nay! Merely a poet!
An animal, a cunning, preying, stealing one,
Which must lie,
Which must lie, consciously, voluntarily,
Longing for prey,
Disguised in many colours,
A mask unto itself,
A prey unto itself.
That—the suitor of truth?
Only a fool! a poet!
Only a speaker in many colours
Speaking in many colours out of fools’ masks,
Stalking about on deceitful word bridges,
On deceitful rain-bows,
Between false heavens
Wandering, stealing about—
Only a fool! a poet!

(Italics mine.)

Thus, it is the Deep, the Unique, the Abyss within, that is the great
Isolator. Nietzsche was indeed “the eagle that long, long gazeth
benumbed into abysses, into its own abysses!”
And he spoke in parables. Give heed—so Zarathustra counsels his
disciples—to the times when your spirits speak in parables, for in
these times is the origin of your virtue.
I said I would not vindicate Nietzsche. But what if his deification of
force-humanity, of master-humanity, were Oberfläche, “surface,”
mask, which he “feigned” or wore, in order to protect his pearls from
sows, his holy of holies from hounds? What if this—scandalizing the
scandalous!—were but picture and parable which Nietzsche flaunted
to the people that they might wreak their vengeance thereupon?
And the parable is so pertinently chosen that it says everything to
men of sense and seriousness, hides everything from fools; that the
pearls can be recognized if right eyes behold, but protectingly
concealed from rude eyes and awkward hands.
 Of course, Nietzsche was a homicide! So must we be! If thy right
eye offend thee, pluck it out, and cast it from thee; if thy right hand
offend thee, hew it off, and cast it from thee. And there are things
more offensive than an eye or a hand! These are the weaknesses
which we pamper and grow in ourselves: thought-lessness which we
wink at; old pet habits which have come to be just too dear for
anything, especially for us to knife; above all, sickly sentiments, self-
pity, from which even all our joys cannot rescue us—so that we do
not have the courage to join those warriors who turn their weapons
against their own selves, and to swear an “I will,” that is hard as
steel, against all these softnesses and humors and self-
commiserations. Surely, it were well to be force-men, master-men,
so that we would not coddle our impotency or carry on a pleasure-
pain play with our weakness.
Yes, in these “stillest hours” there is also a “still” homicide and
interment, a plucking out and a hacking off, and the warrior-hero
does not betray the least pathos as he does this—there is no
plaintive note in his voice. The greatest thing about the dying
Socrates, sipping away at his cup of hemlock, was the total absence
of pathos and self-pity. Ah, if we but took half the pains to marshal
forces of will in ourselves, that we now devote to conserving our
weak wills, and to adducing all sorts of plausible reasons for their
impuissance! If we but actually learned Herrenmoral, master-
morality, that were indeed masterful and understood mastership! We
are called to be masters by our creator, not only masters of the
earth, but also masters of the spirit. And mastership is a great
sacred thing, which we ought to learn from world-masters. We ought
to be hammers in life and not anvils. The great calamity among men
is that they shrink from being hammers, and call the virtue of the
anvil that lets itself be struck by the name of “patience.”
It is just not true that Christianity abhors master-morality and
preaches a Sclavenmoral, a slave-morality. Yes it is true of the
cowardly and inert thing that men call Christianity, this religion of the
study-chair and the barracks which can make use of no master,
because it summons just those powers to rule whose whole strength
consists only in the weakness of others. But there is a Christianity
which has been outright mighty force, outright master-instinct, this
kingly Christianity, in whose presence a Pilate, and a Herod, with the
entire host of their war-slaves, were feeble folk indeed; a Christianity
of love and gentleness and meekness,—aye, aye, sir! But one can
have gentleness in the heart,—and yet lay on with a club! That was
indeed master-morality when the Son of Man made himself master
of the Sabbath; when he with a whip of cords scourged the money-
changers and mammonists out of the Temple! That was a force-man
and a master-man who hurled his, “Get thee behind me, Satan!”
against the weak heart of Peter.
How would it do for our churches to have a new festival, a festival
of “the stillest hour,” memorializing the “invention of new values,
around which the world revolves, noiselessly revolves”? Noises
enough, often enough Höllenlärm, have there been in our churches,
are yet, God knows! But it is not noise that rules the world. It is
stillness which ultimately is the spiritual and moral might of the men
who will possess the kingdom of earth. What if even the history of
peoples “feigns a surface,” wears a mask, for those who having eyes
see not, having ears hear not? What if men mistake Höllenlärm for
messages of great occurrences in history, and on this account hold
themselves aloof from those phenomena and experiences in which
something new, a life of the heart, presses on to its birth-hour? Yet
the human race will not always need or require noise and masks as
its history rolls on. The more men kill what is really worthy of death,
the less will they set out to kill each other. The more powerfully the
will becomes conscious of its calling to master, the more strenuously
men strive after greatness, human greatness, the more ridiculous
will it come to seem to them in the course of time that the force of
man should be sought in the force of his muscles, the mastership of
man in the hoarded prerogative of powder and lead. The day will yet
come—as come it shall—when we will estimate our life, not
according to its noisiest, but according to its stillest hours. And then
a great and pure life will be created by what is done in the heart of
man.
 The Birth of a Poem
(Translated from the Russian of Maximilian Voloshin by A. S. K.)

In my soul is a fragrant dusk of coming thunder...

Heat-lightnings coil there like blue-birds...
Lighted windows burn...
And fibres, long,
Slow-singing,
Grow in the gloom...
O the odor of flowers that reaches a scream!
Lo! lightning in a white zig-zag...
And at once all became bright and great...
How radiant is the night!
Words dance, then flash in couples
In an enamored harmony.
Out of the womb of consciousness, from the bottom of the labyrinth
—
Visions crowd in a quailing host...
And the verse blossoms into a hyacinth-flower,
Cold, fragrant, white.
 Editorials

Why Socialists Went to War

W e have listened with much interest to the excuses for the

German Socialists who went to war, as well as to the attacks
on them for doing so. Now, though hesitating to obtrude
our ignorance into the muddle of a complicated discussion, we
can’t refrain from offering a suggestion.
The bottom reason for sudden activity under the stress of
unusual circumstances is to be found, not in a conscious mental
decision, but in the previously-formed habits of the individual
mind. We are referring partly to the mob-emotion which has
swept away so many even of the greatest souls of Europe. We are
thinking more of the essence of Socialism, and the sort of
emotional method which has been produced among its adherents
—the material upon which mob-psychology had to work.
There is no essential difference between the method of German
Imperialism and the method of German Socialism; the only
difference lies in the objectives. Both insist on the supreme
importance of the state, both work through cohesive organization
and the almost unquestioning following of leaders. The habit of
obedience, the instinct for organization, the gregarious mode of
action—these are the very qualities of the individual German
which have made it possible for the German Social Democratic
Party to grow to such size and strength. What more inevitable,
when the mobilization order went up, when flags flew and drums
beat, than that the individual German Socialist should in his
excitement shoulder his gun and march to war?
Of course, we don’t really know anything about it, and we
haven’t the resources to make anything like a scientific
investigation. But we strongly suspect that the morals of
organized humanity will remain inferior to the morals of the
individual until the individual habit of mind becomes one which
denies to organized humanity supreme authority over the will.
G. H. S.

Even Galsworthy!

I n Scribner’s Magazine for November, Mr. Galsworthy has a

stunning article on the War. And then at its close:—“Your
Prussian supermen of Nietzsche’s cult...!”

Another New Poet

M r. Scharmel Iris is a young Italian poet, born in Florence, who

at the tender age of ten, and later, was praised by Ruskin,
Swinburne, Francis Thompson, Edmund Gosse, and other
men who may be assumed to know what good poetry is. Ruskin
wrote: “He is a youth of genius and his poems are marvelously
beautiful. His heart has felt the pathos of life and he has set this
pathos to music.” Swinburne said: “He writes with imaginative
ardor, and impassioned is the word which best illustrates his
utterance. He is genuine and sincere, and his lovely poems display
energy of emotion and a true sense of poetic restraint.” Thompson
was more superlative: “I believe Scharmel Iris to be a poet of the
first rank,” he stated. “His poems are sublime in conception, rich
in splendid imagery, full of remarkable metaphors and new
figures, and musical in expression.” Of course it has been difficult
for a young man of such talent to find a publisher or a public; but
at last a volume of his work is to be brought out by the Ralph
Fletcher Seymour Company. The book will be called Lyrics of a
Lad, and will be ready about Christmas time. Beside a preface by
Maurice Francis Egan and an interesting title-page decoration by
Michele Greco, it will have a frontispiece portrait by Eugene R.
Hutchinson, the photographer who should never be referred to by
any noun except “artist.” Personally, we love Mr. Iris’s work; we
use the verb thoughtfully, because his poetry is not merely the
sort which interests or attracts; it remains in your mind as part of
that art treasure-house which is your religion and your life.

Prizes for Poetry

A n interesting announcement comes from Poetry in regard to

two prize offers. One—the Helen Haire Levinson prize of two
hundred dollars for the best poetry by a citizen of the United
States published in the magazine during its second year—has
been awarded to Mr. Carl Sandburg for his Chicago Poems. This is
a particularly gratifying decision, for Mr. Sandburg’s is a new voice
which must be reckoned with in American poetic production. The
second is a one hundred dollar offer for the best war or peace
poem on the present European situation, and has been given to
Miss Louise Driscoll of Catskill, New York, for a poem called Metal
Checks, which appears in the November issue.
 My Friend, the Incurable

A t dusk I pass an ugly red building with shrieking fat black letters
on its façade—Home for Incurables. Shrill grass, narcotic
carnations, hazy figures in rocking chairs and on the balconies,
melting in the liquid gold of autumn twilight—a harmony of discord
that screams for the spiritual brush of Kandinsky. There are no signs
of pain or grief on the faces of the doomed: a profound calmness
they bear, a resolute quiescence, reminding us of Dante after he had
seen hell or of Andreyev’s resurrected Lazarus. “To be sure, they are
quite happy,” explained the obliging Doctor. “These men and women
have come to be free of struggles, of doubts, and of the anguish of
hopes. The knowledge of their fate, the ultimate, irrevocable truth, is
a relieving balm for the tired spirits—nay, even for the hopeless
bodies, for as soon as they cease fighting their disease they learn to
adapt themselves to that disease, to consider it an inseparable part
of their existence. I can show you a number of patients who are
actually in love with their affliction, who would resent the idea of
being turned normal. Look at the hilarious face of that fellow yonder
at the fountain; he is intoxicated with sunset, and appears to be the
happiest of mortals, despite his terrible disease. A queer case, an
un-American case.”
The doctor uttered a fearful Latin term and told me the history of
that patient. A European, he has been for many years afflicted with
something like “sentimentalomania,” a peculiarly Continental ailment.
Skilful physicians had tried in vain to cure him; change of climate
and environment had been of no avail: even in Siberian tundras and
in foggy London his disposition remained unaltered. In despair he
went to Berlin, where, he was advised, the gravest case of
sentimentality would be annihilated; the reaction proved almost
fatal, for the Spree and the Sieges Allee made such a nauseating
impression upon the poor fellow that his illness was complicated by
a severe outbreak of Germanophobia. As a last resort, the famous
specialist, Herr Dr. Von Bierueberalles, bade him taste the influence
of the sanest atmosphere on earth, that of the States. When even
the harshest and most practical American treatment had failed to
knock out the unfortunate’s folly, he was pronounced hopeless and
offered a place among the incurables, which offer he willingly
accepted, and acquiesced. He has since become accustomed to his
disease and bears it rather with defiant joy.
At times, when I seek relief from practical values and sane
standards, I come to have a chat with my friend, the Incurable.
Henceforth he will have the floor.

With whom do I side in the War? Why, of course, with Germany!

Perhaps my attitude shows that I have not been completely cured
from the Prussophobia that I had contracted in Berlin; as it is, I
sincerely wish to see the German boot victorious on the whole
continent and over the mouldy Britons, a rude, dreamless, wingless
Napoleon brooding over old napping Europe. Picture the ruined
cathedrals of Belgium and France “restored” into comfortable
barracks for the braves of the Fatherland; picture the boulevards of
Paris and Brusselles, the quays of the Neva and the Thames, ornated
with the statues of the most Christian Wilhelm and of his illustrious
ancestors down to the Great Elector of Brandenburg; picture the
excellent Schutzman reigning supreme, physically and spiritually,
from Vladivostock to Glasgow,—think what an abyss of hatred, of
stirring electrifying hatred will arise among the rotting nations, and
out of hatred self consciousness, endeavors, cravings, to be
crystallized in torrents of new art creations! As for Germany, I have
no fear for the duration of her hegemony; she will undoubtedly
choke from indigestion. But oh, how I dread the reverse outcome!
The victory of the Allies will push Progress a century backward; it
will strengthen the tottering absolutism in Russia; it will swell the
piggish arrogance of the French bourgeois; it will augment the
insular hypocricity of the English Philistine; it will still more, if it is
possible, vulgarize international diplomacy and greed, arousing the
appetites of the so-called Democracies.
Democracy—who was it that recently stated with charming aplomb
that “Individualism and democracy are synonymous terms?” Yes, I
recall: it came from the pen of the author of Incense and Splendor
and To the Innermost. I confess this statement, especially when
considering its authorship, came to me as a revelation. To me the
word “democracy,” as many another beautiful word, has lost its
original lofty meaning and has come to rhyme with mediocrity, with
the strangling of the Few of the Mountain by the Many of the Valley.
Could you name many great things that the most democratized
countries, like America and Switzerland, have produced outside of
Schweitzer-cheese and Victrolas? Has there ever been a great
individualist who appeared as a child of his age, as an outgrowth
and a reflection of a democracy? I do not know of such instances. Of
course, I grant that the writer of that statement put into the word
“Democracy” a higher, a more idealistic meaning. Words, like music,
like practically every medium of art, express the author’s personality,
and, provided he is an artist, he binds us to share his interpretation.
Take, for example, that popular song, “Oh, You Beautiful Doll”;
apparently there is nothing tragic in it, yet my emotions were stirred
when I heard its French interpretation by Olga Petrova (it was before
the kind American entrepreneurs had forced her to perform stunts in
Panthea). She had managed to put so much sorrow and tenderness
into “O Ma Grande Belle Poupée!” that one forgot the triteness of
the words and felt gripping sadness. Or take a less vulgar illustration
—Gertrude Stein’s Tender Buttons.[1] It is an exquisite little thing in
cream covers, with a green moon in the center, implying the yolk of
an egg with which “something is the matter,” and it gave me rare
pleasure to witness the first attempt to revolutionize the most
obsolete and inflexible medium of Art—words. The author has
endeavored to use language in the same way as Kandinsky uses his
colors: to discard conventional structure, to eliminate
understandable figures and forms, and to create a “spiritual
harmony,” leaving to the layman the task of discovering the “innerer
Klang.” Both iconoclasts have admirably succeeded; both the
“Improvisations” and the little “essays” on roast-beef and seltzer-
bottles have given me the great joy of cocreating, allowing me to
interpret them in my own autonomous way. Says the Painter:[2]
The apt use of a word, repetition of this word, twice, three times or even more
frequently, will not only tend to intensify the inner harmony but also to bring to
light unsuspected spiritual properties of the word itself. Further than that, frequent
repetition of a word deprives the word of its original external meaning.

Gertrude Stein has beautifully followed this recipe. Words, plain

everyday words, have lost their “external meaning” under her skilful
manipulation, and in their grotesque arrangement, frequent
repetition, and intentional incoherence they have come to serve as
quaint ephemeral sounds of a suggestive symphony, or, if you
please, cacophony. The Tender Buttons arouse in the sympathetic
reader a limitless amount of moods, from scherzo to maestoso. I
shall recall for you a few lines of one peculiar motive:
(From A Substance in a Cushion.)

What is the use of a violent kind of delightfulness if there is no pleasure in not

getting tired of it.
(From Red Roses.)

A cool red rose and a pink cut pink, a collapse and a sole hole, a little less hot.
Aider, why aider why whow, whow stop touch, aider whow, aider stop the
muncher, muncher munchers.
(From Breakfast.)

What is a loving tongue and pepper and more fish than there is when tears many
tears are necessary.
Why is there more craving than there is in a mountain.... Why is there so much
useless suffering. Why is there.

Do you not feel the deep melancholy underlying these incongruities?

I could quote places that would bring you into a totally different
mood, most hilarious at times. These “exaggerated cranberries,” to
paraphrase an expression of one of my incurable colleagues, should
be chanted to the music of another great iconoclast, Schoenberg.
But I observe an indulgent sneer on your face. Of course, I am an
Incurable—Adieu!
Ibn Gabirol.

[1] Tender Buttons, by Gertrude Stein [Claire Marie, New York].

[2] The Art of Spiritual Harmony, by W. Kandinsky [Houghton Mifflin, Boston].

 London Letter
E. Buxton Shanks

London, Sept. 11, 1914.

W e are all soldiers now and literature, for the time, has
disappeared. The publishing business is at a standstill, reviews
are cutting down their size, and all the best poets are
sedulously learning to form fours in the squares of London. It is, by
itself, a remarkable thing, which will have an effect on all of us when
the war stops and we begin to write again. To leave your pens and
paper, to know that you have before you in the day, not an endless
struggle with rhythm, rhyme, and editors, but a few hours’ drilling
that is laborious and terminable—it is a rousing experience for a
poet, mentally as well as physically.
Meanwhile the literary result of the war is nothing but disastrous.
All our more or less “official” poets—Mr. Bridges, Mr. Newbolt, Mr.
Binyon, Mr. Watson, Mr. Phillips, and so on—have come forward with
amazing arrays of abstract nouns. Mr. Bridges, who is almost the
worst as well as almost the best of living poets, printed a copy of
verses in The Times which rhymed far less often than is proper in a
ceremonial piece and ended thus:

Up, careless, awake!

Ye peacemakers, fight!
ENGLAND STANDS FOR HONOUR:
GOD DEFEND THE RIGHT.

Mr. William Watson has been prodigal of poetry and has reached his
highest level in a poem which contains the following singular lines:—
 We bit them in the Bight,
The Bight of Heligoland.

It is a very sad business. These gentlemen have retired to their

studies, determined to feel what is proper, and they come out having
done their best; but they will be heartily ashamed of it—I hope—in a
few months. Unfortunately, Mr. John Lane has collected their verses
in a volume and is selling their shame for charity. Three good poems
have come out of the welter, one by Mr. G. K. Chesterton—The Wife
of Flanders, a very fine composition—and two by Mr. De La Mare.
The trouble is that a poet does not feel war fever very acutely in a
general sense. Patriotic poetry is nearly always bad. If there is a
worthy reference to the Armada in Elizabethan poetry, it has
escaped me; and the English resistance to Napoleon has never been
a very happy subject for English writers. The good poetry that is
provoked by war is of a different character: it is personal, visual, and
concrete. It never expresses any general aspect of war, but only
such subjects as have been personally observed and felt by the poet.
I would give as instances Rudyard Kipling and the German poet
Liliencron, both of whom have written well about soldiers and
fighting, but foolishly about War and Patriotism.
Yet any poet going about the streets today must see and feel a
quantity of poetical things. A week or so ago, I saw an endless
baggage-train belonging to the artillery, as it passed through Barnet.
It had come from Worcester, commandeering horses and wagons on
the way; it was going to Brentwood and thence—God knows! It was
very long and uneven—the carts had bakers’ and butchers’ names
on them—the horses were ridden with halters and sacks for saddles
—the men were tired and dishevelled. I spoke to one of them who
was watering his horse at a trough, offered to bring him beer from a
public-house close by; but someone had given him tea farther back
on the road and he would rot. He thanked me and rode away,
drooping very much over his horse’s neck. It was all a poem in itself
or it gave me the emotions of a poem, because it had none of the
conventional glitter of war. It was poetical because it was business-
like, just as our khaki service uniforms are more beautiful than the
bright clothes the troops wear in peace.
If the war-poets would confine themselves to real and tangible
things like this, they might well express the experience through
which we are now passing. But they seem unhappily obsessed with
the idea of expressing an obstreperous valour and self-confidence
and bluster which the nation is very far from feeling. The nation, so
far as I can gauge it, is showing an obstinate, workmanlike silence
and does not either make light of, or grumble at, the hardships it has
to suffer: the baggage-train of which I have spoken was a very
adequate symbol of this. But no one is ever so greatly out of touch
with the people as a popular poet.
At the beginning of the war, the musical in London were shocked
by an announcement that no German or Austrian music would be
played at the famous Queen’s Hall Promenade Concerts. We were
naturally a little upset, as we depend on these performances for
solid and regular entertainment: and it seemed hard and
unnecessary to renounce Bach, Mozart, Beethoven, Wagner, and
even Schönberg. Luckily good sense and humour killed the absurd
idea, but not before a French and Russian programme had been
substituted for the first Wagner night. Now, much as I shrink from
the thought of having to hear Tschaikowsky instead of Wagner, I do
believe that we have a cause for national resentment against the
second of these composers. His ridiculous and windy prose-works
have been among the writings which have provoked the war. With
Nietzsche, and with the renegade Englishman, Houston Stewart
Chamberlain, he has encouraged the notion that there is a special
Teutonic culture which is superior to any other and which deserves
to be spread at any cost. Such an idea has never appealed to the
true Germans (e. g., Goethe, who knew what he owed to France and
England), but it has been useful to the Prussian soldiers, who have
debased and vulgarized true German culture. Perhaps I am
exceeding the duties of a London letter-writer and becoming an
advocate; but I think I am giving you an accurate account of the
feelings of those here who admire German poetry and music. I am
not a Chauvinist in art—few people are. I read Goethe impenitently
in the public trains and trams, to the disgust of my neighbours, and I
continue to sing German songs, a little out of tune: unless my
Territorial uniform is served out to me very soon, I shall probably be
arrested as a spy.
 New York Letter
George Soule

S ome years ago a good woman, who would like to be foster-

mother to all struggling heroes, was sitting at midnight in her
down-town flat. Suddenly there was a noise at the front door,
someone leapt up the staircases two steps at a time, and rushed
into her room shouting “I’ve got it! I’ve got it!” She turned around
and saw the dark face of a young actor, shining with excitement. He
immediately burst into a superb interpretation of a passage from
Hamlet. He had been working over it for two weeks without being
able to satisfy himself, but it had come to him that evening. He could
not wait to let his good friend know, had jumped on an elevated
train, and after being carried two stations too far in his elation, was
there with his prize.
No, this is not the beginning of a magazine story, nor is it a
passage from the biography of a deceased European celebrity. It is
the simple truth about a young American dramatist who is known
only to a few;—and he is of New England stock!
Later the young Hamlet, having completed his acting
apprenticeship, began to write, and went into the real estate
business to support himself. Nobody wanted his plays; they were too
“highbrow.” So he began to build a theatre of his own. The
managers’ trust put every difficulty in his way, and finally, when the
building was nearly done and the company was engaged, succeeded
in crushing him. The next attempt was a repertory company on the
East side, but this wiped out what little was left of his resources
before it got fairly started. One play was produced on Broadway;—it
ran two weeks. Last year another was rehearsed for nine weeks, but
it was withdrawn on the day of the dress rehearsal, because the
author refused to make a change insisted on by the manager. Now
the writer has retired to his farm in the Connecticut hills, where he
and a companion have with their own hands built a little theatre. In
this, on Sunday afternoons during the summer, he reads from his
fifteen manuscript plays to such few people as can get there to hear
him. And as he reads, there is on his face much the same
enthusiasm as on the night years ago when he got his passage from
Hamlet.
I visited Butler Davenport for the third time last Sunday. The main
house is a rambling mid-Victorian affair, with queer crannies and
cupola rooms from which one can look far across the hills to the
Sound. On its left is an old farmhouse of the eighteenth century,
furnished as Mr. Butler’s grandfather left it, and with a musty smell
which no old-furniture shop could counterfeit. Between the two is an
old-fashioned garden, in midsummer filled with larkspur, cosmos,
and a hundred other flowers which few but our grandmothers could
name. At the intersection of walks at its center is a crab-apple tree,
surrounded by a bench. A formal garden with high, thick cedar
hedges, bird-houses, unsuspected grass walks and an avenue of
woodbine arches lies on the other side of the main house. In the
rear, stretching out towards the wide valley, is a long, hedged walk
ending in an arch, between fields of wild flowers. Down it one could
go to any kind of distant mystery.
The theatre is a simple, strong little building behind the old
farmhouse. Its most expert bit of carpentry is the balcony, but that
is, of course, unpretentious. The seats are ordinary kitchen chairs,
and there is nothing on the stage but a reading desk. But the luxury
of sitting between wide-open doors in the hill-breeze, full of grass
odors and wing sounds, is better than the comfort of plush seats and
much gilded fresco.
This time, however, as there were only four of us, we sat out
under an apple tree. Except for a moment when a tragic passage
was interrupted to shoo away a loud-voiced and ill-mannered hen, it
was the most nearly perfect theatre I have known.
 And the play? It is impossible to do more than hint at the nature
of unpublished plays. This one dealt with the “white slave” question,
but in a way infinitely superior to the melodrama of The Lure or The
Fight. There was another, of subtler treatment, called Deferred
Payment, showing the natural retribution seeking out a man who
looked for everything in a woman except companionship. Keeping
Up Appearances—the one actually produced—pictures a middle-class
family engaged in a tragic struggle with the pocket book on account
of the false ideals of the community. Justice, written before
Galsworthy’s play of the same name, draws a parallel between
society’s persecution of a woman who is consecrated to a fine love
without marriage, and society’s punishment of the unfortunate
victims of prostitution. Mr. Davenport’s best work is in The
Importance of Coming and Going, a satirical tragi-comedy which
contrasts the exaggerated emphasis we lay on death with the casual
way we regard birth. When a person who never should have come
into the world leaves it, perhaps gladly, we weep copiously and buy
showy funerals; but mothers let their daughters marry any kind of
man of wealth or position, without giving them any insight into the
mysteries of birth.
Mr. Davenport’s plays do not rank with Ibsen’s or even with
Galsworthy’s. But thousands of worse plays have been produced and
have succeeded—simply because they contained no ideas. Mr.
Davenport is master of a technique which would make it easy for
him to write a popular success if he did not insist on saying
something. One manager has told him that he is ten years ahead of
his time, but that if he were only European his work could be
produced. A publisher wrote him that his plays could be issued in
book form if he were only well-known. Mr. Davenport’s question, “My
dear Mr. ——, how am I to become well known?” has not elicited a
reply.
This man’s spirit will remain just as eager and strong as when he
began; he may get before the public eventually. Even this year
hopeful new plans are under way. But whether he ever succeeds or
not, he will have found in life a thousand times more than the
obtuse millions who are deaf to him. It would be an insult to offer
him sympathy.
And it would be stupid to place final blame on the managers or
the publishers, or to think that such things as drama leagues can
furnish a fundamental remedy for the apathy of the public. The
whole structure of society must be altered, and the quality of the
individual human spirit must be quickened, before our leaders can
find any adequate reaction in the crowds. We have denied ourselves
the artistic stimulus of a cohesive aristocracy. How shall we vitalize
our democracy?

If they (men) were books, I would not read them.—Goethe.

Some people term a book poor and unreal because it happens to be outside the
reality with which they themselves happen to be acquainted—a reality which is to
actual reality what a duck-pond is to the ocean.—George Brandes.
 The Theatre

Forbes-Robertson’s Hamlet
(Blackstone Theatre)

O ne of the noblest things I have ever seen on the stage—or ever

expect to see—is the Hamlet of Forbes-Robertson. The poet,
the scholar, the philosopher, the great gentleman, the lover, the
brilliant talker, the anguished boy—they are all there in the tall man
in black with the graven face and the wonderful hands and the voice
of surpassing richnesses—the tall, graceful, impetuous, humorous,
agonized man in black who reads Shakespeare as if he were
improvising and makes a true and charming human being out of a
character that has had the misfortune to become a problem. “And
please observe,” writes Bernard Shaw, “that this is not a cold
Hamlet. He is none of your logicians who reason their way through
the world because they cannot feel their way through it; his intellect
is the organ of his passion; his eternal self-criticism is as alive and
thrilling as it can possibly be.” His moment of expiation, alone at the
back of the stage, with his arms raised to the vaulted heavens; and
his gallant last moment on the throne with its single silver sentence,
“The rest is silence”—these things are too moving to be articulate
about. Richard Le Gallienne has expressed it all as well as it can be
done: “All my life I seem to have been asking my friends, those I
loved best, those who valued the dearest, the kindest, the greatest,
and the strongest, in our strange human life, to come with me and
see Forbes-Robertson die in Hamlet. I asked them because, as that
strange young dead king sat upon his throne, there was something,
whatever it meant—death, life, immortality, what you will—of a
surpassing loneliness, something transfiguring the poor passing
moment of trivial, brutal murder into a beauty to which it was quite
natural that that stern Northern warrior, with his winged helmet,
should bend the knee. I would not exchange anything I have ever
read or seen for Forbes-Robertson as he sits there so still and starlit
upon the throne of Denmark.”
M. C. A.

“The Yellow Ticket”

(Powers’ Theatre)

A bleeding chunk of reality is not art, but it is a bleeding chunk of

reality; your aesthetic emotions may sleep at the sight of a tortured
animal, but your humane emotions will roll up to your throat when
you witness the simple tragedy of a Jewish girl in St. Petersburg,
presented in Michael Morton’s play, The Yellow Ticket. To me such a
realistic play in such a realistic presentation has as little to do with
dramatic art as a reporter’s story has to do with literature; but I
brushed aside my memories of Rheinhardt and Komissarzhevskaya
when I went to see a piece of Russian life at Powers’. And I saw it
indeed—real, nude, appalling.
Some of my acquaintances have asked me whether the tragedy
could be true, whether a Jewish girl has no right to live in St.
Petersburg, unless she has bought her protection from the police by
selling her reputation—that is by procuring a yellow ticket, the trade-
licence of a prostitute. Yes, it is true. A Jew is forbidden to abide
outside the Pale of Settlement, with the exception of certain
merchants and persons of a university education, and prostitutes.
The latter form the most desirable element in the eyes of
government officials, since their occupation does not generally
presuppose any predilections for revolutionary ideas or free thought.
I have known instances where women involved in the Revolution,
gentiles as well as Jewesses, obtained yellow tickets which served
them the rôle of a carte blanche from the molestations of the police.
There are many anecdotic facts in Russian life that seem incredible
to the outsider, and Mr. Morton has produced in his play a mass of
such facts with photographic verisimilitude. It must be said to the
credit of the actors that they have escaped the slippery path of
melodramatic overdoing.
K.

“Jael”
(The Little Theatre)

“Hosanna!” I felt like shouting, when the curtains slowly concealed

the mysterious stage. I am still under the spell of the oriental
atmosphere, not yet cooled off for objective criticism. What Florence
Kiper Frank has done with the biblical subject may terrify the
orthodox student of the Bible, but I greeted her daring heresy and
free manipulation of epochs and styles. She has skilfully blended the
bloodthirsty, gloating outcries of Deborah’s Song with the idyllic
lyrics of Solomon’s Songs, and has presented in Jael a composite
type, a mixture of the savage tent-woman, of the passionate yet
gentle Shulamite, and of the eternal jealous female. The result, as
far as the creation of an atmosphere goes, is a positive success.
A word about the staging. Maurice Browne, on the privilege of a
pioneer, may be congratulated on the progress he has made in
leaving behind mouldy conventions and approaching the state where
he can produce pure aesthetic emotions. The three one-act plays on
the present bill, regardless of their merits or demerits, demonstrate
the great possibilities of an artistic stage manager, who can do away
with elaborate accessories and produce suggestive illusions with the
aid of an ultramarine background and calico apple blossoms. Yet, as
in all pioneering, there are signs of hesitation and of half-measures.
I am sure that the effect of Jael would not in the least diminish (it
would rather be intensified), if we were spared the inevitable storm-
pyrotechnics. The verses in themselves imply the idea of battle and
tempest, and Miss Kiper in the title rôle has the voice and diction to
serve the purpose.
K.
 Harold Bauer in Chicago
Herman Schuchert

T here yet remain certain pianists and other opinionated craftsmen

in music who will say, when approached on the subject of Harold
Bauer’s piano playing: “Oh, yes; but you know Bauer is—well,
shall we say?—a monotonist. His playing is all of one style—beautiful
tone, to be sure; but, oh, such a sameness! He shades beautifully—
yes, surely, but it’s all too colorless.” And it probably never occurs to
these critics that a pianist who uses an entirely beautiful tone, who
shades delicately, and who is definitely individual in his playing,
might not seem monotonous to the admirers of true piano-artistry.
And it is quite certain that these carpers failed to attend Bauer’s last
Sunday afternoon recital in Orchestra Hall, when and where the
above composite quotation was put to shame.
The program was headed by that most unequal set of little pieces
—interesting, dull, graceful, and often clumsy:—Brahm’s Waltzes.
The Brahms faddists may sacrifice all the credit to their idol, but he
deserves only a part of it; for Bauer made these waltzes float as
lightly and pleasantly as the material permitted, and invested them
with all possible contrast and pulse. There was no lack of what
pianists call “point,” either in this opening number or in the
remainder of the program; and it is this quality of “point,” which is
the season more in evidence in Bauer’s work than ever before, which
makes the carpers appear rather uninformed. “Point” is nothing
mysterious; it means definite and crisp rhythm, brightness of tone
designed to contrast with richness and warmth of tone, sharp
shadings artistically brought out, and a deeply satisfying precision in
tempi. This man’s work deserves this inclusive term. Whatever lack
there might have been in seasons past (there has been a fragile
foundation for the criticism mentioned at the beginning of this
appreciation, when, as late as three years ago, his tonal ideals
apparently did not include great brilliance), this Sunday recital went
far to establish the fact that Bauer has a happy variety of tone-colors
at his command, which variety includes no little brilliance. Sheer
facility and digital expertness have never seemed to occupy the
attention of this master-pianist, except insofar as such facility and
expertness would give expression to purely musical content; and
now if the carpers continue to shrug their shoulders at the praise of
Bauer, it will be because they miss the usual bombast and key-
swatting of esteemed mediocrity, and certainly not because of any
inadequacy of technic for musical purposes, or lack of pianistic
lustre. No mediocrity of a technic-worshipper or piano-eater ever
gave a performance of Beethoven’s Opus 3 that could compare with
that of Bauer on Sunday afternoon; for he then projected a deeply
significant art, particularly in the first movement of the sonata,
which must be inexplicable in words. Schumann’s Scenes from
Childhood were given a highly imaginative treatment—a treatment
which penetrated even the academics. And Schumann’s Toccata—
that battered veteran of many an ivory struggle—ceased for once to
be an endurance stunt, and hummed forth (as the composer hoped
and indicated) as a strangely beautiful bit of music. Bauer’s playing
of this will remain long in the awakened music-receptacles. So will
his interpretation of his own arrangement of César Franck’s Prelude,
Choral, and Fugue—which are three movements vieing with each
other for supreme religious solidity—and his nonchalant handling of
the tricky D-flat Study of Liszt. The Chopin Scherzo in C-sharp minor
closed a program which would surely have been sombre and sleepy
under the fingers of any less than a pianistic musician. In certain
splendid moments Bauer seems like a high priest performing a tonal
miracle, or like a potent magician weaving curious and impossible
dream-fabrics. And, with all pleasant fancies put aside, he is an
exponent of modern pianism at its best.
In music a light blue is like a flute, a darker blue a ’cello, a still darker a
thunderous double bass; and the darkest blue of all—an organ.—Kandinsky.

Color is a power which directly influences the soul. Color is the key-board, the
eyes are the hammers, the soul is the piano with many strings. The artist is the
hand which plays, touching one key or another, to cause vibrations of the soul.—
Kandinsky.
 A Ferrer School in Chicago
Dr. Rudolf von Liebich

T he Havasupai Indian mother says: “I must not beat my boy. If I

do, I will break his will.” Unlike her pale-faced friends, she is not
obsessed with the mania for governing. We, in our insane
subservience to traditions, continue to train our children to obey.
Slaves they shall be; that is the slogan. We no longer whip men; we
whip children only because they are weaker than we are. So, a child
is the slave in successive stages of home, church, school,
government, and either boss or “superior officer.” Could Europe be at
war unless its men were made molluscous by discipline and their
mental paralysis completed through respectability?
Children are born materialists, poets, and joy-worshippers. We
tame them and they grow up philistines, supernaturalists, and
respectable believers in the disinterested love of dullness. Instead of
teaching them theories and superstitions, we should tell them that
they are parts of the universe; that the carbon, iron, sulphur,
phosphorus, oxygen, hydrogen, nitrogen, the zero-gases, and the
dozen other elements of which our bodies are made are also the
main elements of sun, moon, and stars,—of the whole material
universe. The next step might be to show the child, through actual
experiments, the known physical and chemical properties of these
elements, thus preparing its mind for the greatest of all poetries—
the poetry of evolution. These things need but be shown, not
laboriously learned by rote; they need only to be told, not to be
taught; and if the child’s healthy inquisitiveness has not been ruined
by repression, it will delight in feeling the pull of the magnet; in
watching the electric spark that unites oxygen and hydrogen into
water; in drawing the marvelous beauties of snow flakes and other
crystal formations; in watching and aiding the growth of birds,
beasts, flowers, or fruits; in the thrill of blended voices or in other
forms of voluntary co-operation. All these things, all the realities
need but be shown to delight the untainted mind of childhood; while
daily free association with other children will soon give to each child
a practical working knowledge of ethics (quite impossible to attain
under the boss-system of the government schoolmistress) from
which, as a basis, the errors of our economic and social systems can
be pointed out and discussed. In the minds and hearts of these free
children, ideals can then be formulated which will tend toward their
development into the free society of the future, whose coming their
own efforts will hasten. For it is only through the successive
enslavement of each succeeding generation that governments can
retain their powers.
Such should be some of the activities of a Ferrer or Modern
School, free from the noxious taint of authority, superstition, or
respectability. If we cannot do better let us begin, at least, with a
Sunday school. However that may be, and whatever the future of
such a school, all those interested in establishing it are cordially
invited to communicate with the editor of The Little Review, with
William Thurston Brown, 1125 N. Hoyne Ave., with Anthony Udell,
817½ N. Clark St., or with the writer, 1240 Morse Ave.
 The Old Spirit and the New Ways in
Art
William Saphier

F ull of visions and ideals and eager to express them in their own
way, a group of striving young painters and sculptors in this city
is working industriously without regard for applause from either
the crowd or the few. Just as there are religious and social rebels—
people who refuse to accept the old dogmas and habits merely
because they were successful at a certain time and fit for a certain
period in human history—these young artists refuse to adopt
methods and views of the past for the purpose of expressing their
views on modern subjects.
In striving to realize the new idea in form and color they are of
necessity passing through that period in which the intellect discerns
and style is chosen—the period of experiment. And if they do not
achieve as great a success as the old masters, they certainly work in
the spirit of a Monet or a Rembrandt. We print this month
reproductions of work done by four of these artists. They have
nothing in common except that they are all trying to express
themselves in their own way.
Jerome S. Blum, the oldest and best known of the group, is an
extraordinary painter of the usual. He does not rely on a dramatic
subject, or on a sensational technic, to arouse interest in his work. It
is his unusual way of looking at people and nature, and his vigorous
and interesting color schemes, that have made his paintings notable.
Mr. Blum is far too imaginative to be natural, far too poetic to be
“real.” All his work strikes one as a spontaneous expression of almost
childish delight in color.