Introduction to Deep Learning Using R: A Step-by-Step Guide to Learning and Implementing Deep Learning Models Using R Taweh Beysolow Ii - The latest ebook is available for instant download now
Introduction to Deep Learning Using R: A Step-by-Step Guide to Learning and Implementing Deep Learning Models Using R Taweh Beysolow Ii - The latest ebook is available for instant download now
com
https://textbookfull.com/product/introduction-to-deep-
learning-using-r-a-step-by-step-guide-to-learning-and-
implementing-deep-learning-models-using-r-taweh-beysolow-ii/
OR CLICK HERE
DOWLOAD EBOOK
https://textbookfull.com/product/deep-learning-with-python-develop-
deep-learning-models-on-theano-and-tensorflow-using-keras-jason-
brownlee/
textbookfull.com
https://textbookfull.com/product/r-deep-learning-essentials-1st-
edition-wiley/
textbookfull.com
https://textbookfull.com/product/deep-learning-with-r-1st-edition-
francois-chollet/
textbookfull.com
Learning Tensorflow A Guide to Building Deep Learning
Systems 1st Edition Tom Hope
https://textbookfull.com/product/learning-tensorflow-a-guide-to-
building-deep-learning-systems-1st-edition-tom-hope/
textbookfull.com
https://textbookfull.com/product/advanced-data-analytics-using-python-
with-machine-learning-deep-learning-and-nlp-examples-mukhopadhyay/
textbookfull.com
https://textbookfull.com/product/webpack-for-beginners-your-step-by-
step-guide-to-learning-webpack-4-1st-edition-mohamed-bouzid/
textbookfull.com
Taweh Beysolow II
Introduction to Deep Learning Using R
Taweh Beysolow II
San Francisco, California, USA
ISBN-13 (pbk): 978-1-4842-2733-6 ISBN-13 (electronic): 978-1-4842-2734-3
DOI 10.1007/978-1-4842-2734-3
Library of Congress Control Number: 2017947908
Copyright © 2017 by Taweh Beysolow II
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole
or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical
way, and transmission or information storage and retrieval, electronic adaptation, computer
software, or by similar or dissimilar methodology now known or hereafter developed.
Trademarked names, logos, and images may appear in this book. Rather than use a trademark
symbol with every occurrence of a trademarked name, logo, or image we use the names, logos,
and images only in an editorial fashion and to the benefit of the trademark owner, with no
intention of infringement of the trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if
they are not identified as such, is not to be taken as an expression of opinion as to whether or not
they are subject to proprietary rights.
While the advice and information in this book are believed to be true and accurate at the
date of publication, neither the authors nor the editors nor the publisher can accept any legal
responsibility for any errors or omissions that may be made. The publisher makes no warranty,
express or implied, with respect to the material contained herein.
Cover image designed by Freepik
Managing Director: Welmoed Spahr
Editorial Director: Todd Green
Acquisitions Editor: Celestin Suresh John
Development Editor: Laura Berendson
Technical Reviewer: Somil Asthana
Coordinating Editor: Sanchita Mandal
Copy Editor: Corbin Collins
Compositor: SPi Global
Indexer: SPi Global
Artist: SPi Global
Distributed to the book trade worldwide by Springer Science+Business Media New York,
233 Spring Street, 6th Floor, New York, NY 10013. Phone 1-800-SPRINGER, fax (201)
348-4505, e-mail [email protected], or visit www.springeronline.com. Apress
Media, LLC is a California LLC and the sole member (owner) is Springer Science + Business
Media Finance Inc (SSBM Finance Inc). SSBM Finance Inc is a Delaware corporation.
For information on translations, please e-mail [email protected], or visit
http://www.apress.com/rights-permissions.
Apress titles may be purchased in bulk for academic, corporate, or promotional use.
eBook versions and licenses are also available for most titles. For more information, reference
our Print and eBook Bulk Sales web page at http://www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this book is
available to readers on GitHub via the book's product page, located at the following link:
https://github.com/TawehBeysolowII/AnIntroductionToDeepLearning.
For more detailed information, please visit http://www.apress.com/source-code.
Printed on acid-free paper
Contents at a Glance
■
■Chapter 1: Introduction to Deep Learning�������������������������������������� 1
■
■Chapter 2: Mathematical Review������������������������������������������������� 11
■
■Chapter 3: A Review of Optimization and Machine Learning������� 45
■
■Chapter 4: Single and Multilayer Perceptron Models������������������� 89
■
■Chapter 5: Convolutional Neural Networks (CNNs)��������������������� 101
■
■Chapter 6: Recurrent Neural Networks (RNNs)�������������������������� 113
■■Chapter 7: Autoencoders, Restricted Boltzmann Machines,
and Deep Belief Networks���������������������������������������������������������� 125
■
■Chapter 8: Experimental Design and Heuristics������������������������� 137
■
■Chapter 9: Hardware and Software Suggestions������������������������ 167
■
■Chapter 10: Machine Learning Example Problems��������������������� 171
■
■Chapter 11: Deep Learning and Other Example Problems���������� 195
■
■Chapter 12: Closing Statements������������������������������������������������� 219
Index���������������������������������������������������������������������������������������������� 221
iii
Contents
■
■Chapter 1: Introduction to Deep Learning�������������������������������������� 1
Deep Learning Models���������������������������������������������������������������������������� 3
Single Layer Perceptron Model (SLP)����������������������������������������������������������������������� 3
Multilayer Perceptron Model (MLP)�������������������������������������������������������������������������� 4
Convolutional Neural Networks (CNNs)�������������������������������������������������������������������� 5
Recurrent Neural Networks (RNNs)�������������������������������������������������������������������������� 5
Restricted Boltzmann Machines (RBMs)������������������������������������������������������������������ 6
Deep Belief Networks (DBNs)����������������������������������������������������������������������������������� 6
Other Topics Discussed��������������������������������������������������������������������������� 7
Experimental Design������������������������������������������������������������������������������������������������� 7
Feature Selection������������������������������������������������������������������������������������������������������ 7
Applied Machine Learning and Deep Learning��������������������������������������������������������� 7
History of Deep Learning������������������������������������������������������������������������������������������ 7
Summary������������������������������������������������������������������������������������������������� 9
■
■Chapter 2: Mathematical Review������������������������������������������������� 11
Statistical Concepts������������������������������������������������������������������������������ 11
Probability��������������������������������������������������������������������������������������������������������������� 11
And vs. Or��������������������������������������������������������������������������������������������������������������� 12
v
■ Contents
Bayes’ Theorem������������������������������������������������������������������������������������������������������ 14
Random Variables��������������������������������������������������������������������������������������������������� 14
Variance������������������������������������������������������������������������������������������������������������������ 15
Standard Deviation������������������������������������������������������������������������������������������������� 16
Coefficient of Determination (R Squared)��������������������������������������������������������������� 17
Mean Squared Error (MSE)������������������������������������������������������������������������������������� 17
Linear Algebra��������������������������������������������������������������������������������������� 17
Scalars and Vectors������������������������������������������������������������������������������������������������ 17
Properties of Vectors���������������������������������������������������������������������������������������������� 18
Axioms�������������������������������������������������������������������������������������������������������������������� 19
Subspaces�������������������������������������������������������������������������������������������������������������� 20
Matrices������������������������������������������������������������������������������������������������������������������ 20
Summary����������������������������������������������������������������������������������������������� 43
■
■Chapter 3: A Review of Optimization and Machine Learning������� 45
Unconstrained Optimization������������������������������������������������������������������ 45
Local Minimizers���������������������������������������������������������������������������������������������������� 47
Global Minimizers��������������������������������������������������������������������������������������������������� 47
Conditions for Local Minimizers����������������������������������������������������������������������������� 48
Neighborhoods�������������������������������������������������������������������������������������� 49
Interior and Boundary Points���������������������������������������������������������������������������������� 50
Regression Models�������������������������������������������������������������������������������� 51
Linear Regression��������������������������������������������������������������������������������������������������� 51
vi
■ Contents
What Is Multicollinearity?���������������������������������������������������������������������� 62
Testing for Multicollinearity������������������������������������������������������������������� 62
Variance Inflation Factor (VIF)��������������������������������������������������������������������������������� 62
Ridge Regression���������������������������������������������������������������������������������������������������� 62
Least Absolute Shrinkage and Selection Operator (LASSO)����������������������������������� 63
Comparing Ridge Regression and LASSO��������������������������������������������������������������� 64
Evaluating Regression Models������������������������������������������������������������������������������� 64
Receiver Operating Characteristic (ROC) Curve������������������������������������������������������ 67
Confusion Matrix���������������������������������������������������������������������������������������������������� 68
Limitations to Logistic Regression�������������������������������������������������������������������������� 69
Support Vector Machine (SVM)������������������������������������������������������������������������������� 70
Sub-Gradient Method Applied to SVMs������������������������������������������������������������������� 72
Extensions of Support Vector Machines����������������������������������������������������������������� 73
Limitations Associated with SVMs�������������������������������������������������������������������������� 73
vii
■ Contents
Random Forest������������������������������������������������������������������������������������������������������� 83
Limitations to Random Forests������������������������������������������������������������������������������� 83
Bayesian Learning��������������������������������������������������������������������������������� 83
Naïve Bayes Classifier�������������������������������������������������������������������������������������������� 84
Limitations Associated with Bayesian Classifiers��������������������������������������������������� 84
Final Comments on Tuning Machine Learning Algorithms�������������������������������������� 85
Reinforcement Learning������������������������������������������������������������������������ 86
Summary����������������������������������������������������������������������������������������������� 87
■
■Chapter 4: Single and Multilayer Perceptron Models������������������� 89
Single Layer Perceptron (SLP) Model���������������������������������������������������� 89
Training the Perceptron Model������������������������������������������������������������������������������� 90
Widrow-Hoff (WH) Algorithm���������������������������������������������������������������������������������� 90
Limitations of Single Perceptron Models���������������������������������������������������������������� 91
Summary Statistics������������������������������������������������������������������������������������������������ 94
Summary��������������������������������������������������������������������������������������������� 100
■
■Chapter 5: Convolutional Neural Networks (CNNs)��������������������� 101
Structure and Properties of CNNs������������������������������������������������������� 101
Components of CNN Architectures������������������������������������������������������ 103
Convolutional Layer���������������������������������������������������������������������������������������������� 103
Pooling Layer�������������������������������������������������������������������������������������������������������� 105
Rectified Linear Units (ReLU) Layer���������������������������������������������������������������������� 106
Fully Connected (FC) Layer����������������������������������������������������������������������������������� 106
Loss Layer������������������������������������������������������������������������������������������������������������ 107
viii
■ Contents
ix
■ Contents
Summary��������������������������������������������������������������������������������������������� 136
■
■Chapter 8: Experimental Design and Heuristics������������������������� 137
Analysis of Variance (ANOVA)�������������������������������������������������������������� 137
One-Way ANOVA��������������������������������������������������������������������������������������������������� 137
Two-Way (Multiple-Way) ANOVA��������������������������������������������������������������������������� 137
Mixed-Design ANOVA�������������������������������������������������������������������������������������������� 138
Multivariate ANOVA (MANOVA)������������������������������������������������������������������������������ 138
x
■ Contents
Summary��������������������������������������������������������������������������������������������� 166
■
■Chapter 9: Hardware and Software Suggestions������������������������ 167
Processing Data with Standard Hardware������������������������������������������ 167
Solid State Drives and Hard Drive Disks (HDD)����������������������������������� 167
Graphics Processing Unit (GPU)����������������������������������������������������������� 168
Central Processing Unit (CPU)������������������������������������������������������������� 169
Random Access Memory (RAM)���������������������������������������������������������� 169
Motherboard���������������������������������������������������������������������������������������� 169
Power Supply Unit (PSU)��������������������������������������������������������������������� 170
Optimizing Machine Learning Software���������������������������������������������� 170
Summary��������������������������������������������������������������������������������������������� 170
xi
■ Contents
■
■Chapter 10: Machine Learning Example Problems��������������������� 171
Problem 1: Asset Price Prediction������������������������������������������������������� 171
Problem Type: Supervised Learning—Regression����������������������������������������������� 172
Description of the Experiment������������������������������������������������������������������������������ 173
Feature Selection�������������������������������������������������������������������������������������������������� 175
Model Evaluation��������������������������������������������������������������������������������� 176
Ridge Regression�������������������������������������������������������������������������������������������������� 176
Support Vector Regression (SVR)�������������������������������������������������������������������������� 178
Problem 2: Speed Dating�������������������������������������������������������������������������������������� 180
Problem Type: Classification��������������������������������������������������������������������������������� 181
Preprocessing: Data Cleaning and Imputation������������������������������������������������������ 182
Summary��������������������������������������������������������������������������������������������� 194
■
■Chapter 11: Deep Learning and Other Example Problems���������� 195
Autoencoders�������������������������������������������������������������������������������������� 195
Convolutional Neural Networks����������������������������������������������������������� 202
Preprocessing������������������������������������������������������������������������������������������������������� 204
Summary��������������������������������������������������������������������������������������������� 218
■
■Chapter 12: Closing Statements������������������������������������������������� 219
Index���������������������������������������������������������������������������������������������� 221
xii
About the Author
xiii
About the Technical
Reviewer
xv
Acknowledgments
To my family, who I am never grateful enough for. To my grandmother, from whom much
was received and to whom much is owed. To my editors and other professionals who
supported me through this process, no matter how small the assistance seemed. To my
professors, who continue to inspire the curiosity that makes research worth pursuing.
To my friends, new and old, who make life worth living and memories worth keeping. To
my late friend Michael Giangrasso, who I intended on researching Deep Learning with.
And finally, to my late mentor and friend Lawrence Sobol. I am forever grateful for your
friendship and guidance, and continue to carry your teachings throughout my daily life.
xvii
Introduction
xix
CHAPTER 1
Introduction to Deep
Learning
With advances in hardware and the emergence of big data, more advanced computing
methods have become increasingly popular. Increasing consumer demand for better
products and companies seeking to leverage their resources more efficiently have
also been leading this push. In response to these market forces, we have recently seen
a renewed and widely spoken about interest in the field of machine learning. At the
cross-section of statistics, mathematics, and computer science, machine learning refers
to the science of creating and studying algorithms that improve their own behavior in
an iterative manner by design. Originally, the field was devoted to developing artificial
intelligence, but due to the limitations of the theory and technology that were present
at the time, it became more logical to focus these algorithms on specific tasks. Most
machine learning algorithms as they exist now focus on function optimization, and the
solutions yielded don’t always explain the underlying trends within the data nor give
the inferential power that artificial intelligence was trying to get close to. As such, using
machine learning algorithms often becomes a repetitive trial and error process, in which
the choice of algorithm across problems yields different performance results. This is fine
in some contexts, but in the case of language modeling and computer vision, it becomes
problematic.
In response to some of the shortcomings of machine learning, and the significant
advance in the theoretical and technological capabilities at our disposal today, deep
learning has emerged and is rapidly expanding as one of the most exciting fields of
science. It is being used in technologies such as self-driving cars, image recognition on
social media platforms, and translation of text from one language to others. Deep learning
is the subfield of machine learning that is devoted to building algorithms that explain
and learn a high and low level of abstractions of data that traditional machine learning
algorithms often cannot. The models in deep learning are often inspired by many sources
of knowledge, such as game theory and neuroscience, and many of the models often
mimic the basic structure of a human nervous system. As the field advances, many
researchers envision a world where software isn’t nearly as hard coded as it often needs to
be today, allowing for a more robust, generalized solution to solving problems.
Deep neural networks are distinguished by having many hidden layers, which
are called “hidden” because we don’t necessarily see what the inputs and outputs of
these neurons are explicitly beyond knowing they are the output of the preceding layer.
The addition of layers, and the functions inside the neurons of these layers, are what
distinguish an individual architecture from another and establish the different use cases
of a given model.
More specifically, lower levels of these models explain the “how,” and the higher-levels
of neural networks process the “why.” The functions used in these layers are dependent
on the use case, but often are customizable by the user, making them significantly more
robust than the average machine learning models that are often used for classification and
regression, for example. The assumption in deep learning models on a fundamental level is
that the data being interpreted is generated by the interactions of different factors organized
2
Chapter 1 ■ Introduction to Deep Learning
in layers. As such, having multiple layers allows the model to process the data such that it
builds an understanding from simple aspects to larger constructs. The objective of these
models is to perform tasks without the same degree of explicit instruction that many
machine learning algorithms need. With respect to how these models are used, one of the
main benefits is the promise they show when applied to unsupervised learning problems,
or problems where we don’t know prior to performing the experiment that the response
variable y should be given a set of explanatory variables x. An example would be image
recognition, particularly after a model has been trained against a given set of data. Let’s say
we input an image of a dog in the testing phase, implying that we don’t tell the model what
the picture is of. The neural network will start by recognizing eyelashes prior to a snout,
prior to the shape of the dog’s head, and so on until it classifies the image as that of a dog.
3
Chapter 1 ■ Introduction to Deep Learning
4
Chapter 1 ■ Introduction to Deep Learning
5
Chapter 1 ■ Introduction to Deep Learning
6
Chapter 1 ■ Introduction to Deep Learning
Experimental Design
The emphasis of this text ultimately is to give the reader a theoretical understanding of
the deep learning models such that they feel comfortable enough to apply them. As such,
it is important to discuss elements of experimental design to help the reader understand
proper ways to structure their research so it leads to actionable insights and not a waste
of time and/or energy. Largely, I will draw upon Fisher’s principles in addition to defining
best practices given the problems often utilized by deep learning.
Feature Selection
A component of experimental design, but ultimately entirely a subtopic of research unto
itself, I will cover the concept of variable selection and multiple methods used often by
data scientists to handle high dimensional data sets. Specifically, I will speak in depth
about principal components analysis as well as genetic algorithms. All the algorithms
discussed are available in the R statistical language in open source packages. For those
who want to research this area of research further, I’ll reference papers relevant to this
topic. From a deep learning perspective, we will discuss in depth how each model
performs its own specific methods of feature selection by design of the layer architecture
in addition to addressing recent discoveries in the field.
7
Chapter 1 ■ Introduction to Deep Learning
The first working learning algorithm that is often associated with deep learning
models was developed by Ivakhenenko and Lapa. They published their findings in a
paper entitled “Networks Trained by the Group Method of Data Handling (GMDH)” in
1965. These were among the first deep learning systems of the feed-forward multilayer
perceptron type. Feed-forward networks describe models where the connections between
the units don’t form a cycle, as they would be in a recurrent neural network. This model
featured polynomial activation functions, and the layers were incrementally grown and
trained by regression analysis. They were subsequently pruned with the help of a separate
validation set, where regularization was used to weed out superfluous units.
In the 1980s, the neocognitron was introduced by Kunihio Fukushima. It is a
multilayered artificial neural network and has primarily been used for handwritten
character recognition and similar tasks that require pattern recognition. Its pattern
recognition abilities gave inspiration to the convolutional neural network. Regardless,
the neocognitron was inspired by a model proposed by the neurophysiologists Hubel
and Wiesel. Also during this decade, Yann LeCun et al. applied the back-propagation
algorithm to a deep neural network. The original purpose of this was for AT&T to
recognize handwritten zip codes on mail. The advantages of this technology were
significant, particularly right before the Internet and its commercialization were to occur
in the late 1990s and early 2000s.
In the 1990s, the field of deep learning saw the development of a recurrent neural
network that required more than 1,000 layers in an RNN unfolded in time, and the
discovery that it is possible to train a network containing six fully connected layers and
several hundred hidden units using what is called a wake-sleep algorithm. A heuristic,
or an algorithm that we apply over another single or group of algorithms, a wake-sleep
algorithm is a unsupervised method that allows the algorithm to adjust parameters in
such a way that an optimal density estimator is outputted. The “wake” phase describes
the process of the neurons firing from input to output. The connections from the inputs
and outputs are modified to increase the likelihood that they replicate the correct activity
in the layer below the current one. The “sleep” phase is the reverse of the wake phase,
such that neurons are fired by the connections while the recognitions are modified.
As rapidly as the advancements in this field came during the early 2000s and the
2010s, the current period moving forward is being described as the watershed moment
for deep learning. It is now that we are seeing the application of deep learning to a
multitude of industries and fields as well as the very devoted improvement of the
hardware used for these models. In the future, it is expected that the advances covered in
deep learning will help to allow technology to make actions in contexts where humans
often do today and where traditional machine learning algorithms have performed
miserably. Although there is certainly still progress to be made, the investment made
by many firms and universities to accelerate the progress is noticeable and making a
significant impact on the world.
8
Chapter 1 ■ Introduction to Deep Learning
Summary
It is important for the reader to ultimately understand that no matter how sophisticated
any model is that we describe here, and whatever interesting and powerful uses it may
provide, there is no substitute for adequate domain knowledge in the field in which these
models are being used. It is easy to fall into the trap, for both advanced and introductory
practitioners, of having full faith in the outputs of the deep learning models without
heavily evaluating the context in which they are used. Although seemingly self-evident,
it is important to underscore the importance of carefully examining results and, more
importantly, making actionable inferences where the risk of being incorrect is most
limited. I hope to impress upon the reader not only the knowledge of where they can
apply these models, but the reasonable limitations of the technology and research as it
exists today.
This is particularly important in machine learning and deep learning because
although many of these models are powerful and reach proper solutions that would be
nearly impossible to do by hand, we have not determined why this is the case always. For
example, we understand how the back-propagation algorithm works, but we can’t see it
operating and we don’t have an understanding of what exactly happened to reach such
a conclusion. The main problem that arises from this situation is that when a process
breaks, we don’t necessarily always have an idea as to why. Although there have been
methods created to try and track the neurons and the order in which they are activated,
the decision-making process for a neural network isn’t always consistent, particularly
across differing problems. It is my hope that the reader keeps this in mind when moving
forward and evaluates this concern appropriately when necessary.
9
CHAPTER 2
Mathematical Review
Statistical Concepts
No discussion about statistics or machine learning would be appropriate without initially
discussing the concept of probability.
Probability
Probability is the measure of the likelihood of an event. Although many machine learning
models tend to be deterministic (based off of algorithmic rules) rather than probabilistic,
the concept of probability is referenced specifically in algorithms such as the expectation
maximization algorithm in addition to more complex deep learning architectures such
a recurrent neural networks and convolutional neural networks. Mathematically, this
algorithm is defined as the following:
P( AÇB)
P( A| B) = ,
P (B)
Provided P ( B ) > 0.
P ( A | B ) = P ( A)
P (B | A) = P (B )
P ( A Ç B ) = P ( A) P (B )
In Figure 2-1, we can envision events A and B as two sets, with the union of A and B
as the intersection of the circles:
Should this equation not hold in a given circumstance, the events A and B are said to
be dependent.
And vs. Or
Typically when speaking about probability—for instance, when evaluating two events
A and B—probability is often in discussed in the context of “the probability of A and B” or
“the probability of A or B.” Intuitively, we define these probabilities as being two different
events and therefore their mathematical derivations are difference. Simply stated, or
denotes the addition of probabilities events, whereas and implies the multiplication of
probabilities of event. The following are the equations needed:
12
Chapter 2 ■ Mathematical Review
P ( A Ç B ) = P ( A) P (B | A)
= P (B)P ( A | B)
P ( A Ç B ) = P ( A) P (B )
Or (additive law of probability) is the probability of the union of two events A and B:
P ( A È B ) = P ( A) + P (B ) - P ( A Ç B )
The probabilities of A and B exclusively are the section of their respective spheres
which do not intersect, whereas the probability of A or B would be the addition of these
two sections plus the intersection. We define S as the sum of all sets that we would
consider in a given problem plus the space outside of these sets. The probability of S is
therefore always 1.
With this being said, the space outside of A and B represents the opposite of these
events. For example, say that A and B represent the probabilities of a mother coming
home at 5 p.m. and a father coming home at 5 p.m. respectively. The white space
represents the probability that neither of them comes home at 5 p.m.
13
Chapter 2 ■ Mathematical Review
Bayes’ Theorem
As mentioned, Bayesian statistics is continually gaining appreciation within the fields
of machine learning and deep learning. Although these techniques can often require
considerable amounts of hard coding, their power comes from the relatively simple
theoretical underpinning while being powerful and applicable in a variety of contexts.
Built upon the concept of conditional probability, Bayes’ theorem is the concept that the
probability of an event A is related to the probability of other similar events:
P ( A | Bj )P (Bj )
P (B j | A) =
åik P ( A | Bi ) P ( Bi )
Referenced in later chapters, Bayesian classifiers are built upon this formula as well
as the expectation maximization algorithm.
Random Variables
Typically, when analyzing the probabilities of events, we do so within a set of random
variables. We define a random variable as a quantity whose value depends on a set of
possible random events, each with an associated probability. Its value is known prior to it
being drawn, but it also can be defined as a function that maps from a probability space.
Typically, we draw these random variables via a method know as random sampling.
Random sampling from a population is said to be random when each observation is
chosen in such a way that it is just as likely to be selected as the other observations within
the population.
Broadly speaking, the reader can expect to encounter two types of random variables:
discrete random variables and continuous random variables. The former refers to
variables that can only assume a finite number of distinct values, whereas the latter are
variables that have an infinite number of possible variables. An example is the number of
cars in a garage versus the theoretical change in percentage change of a stock price. When
analyzing these random variables, we typically rely on a variety of statistics that readers
can expect to see frequently. But these statistics often are used directly in the algorithms
either during the various steps or in the process of evaluating a given machine learning or
deep learning model.
As an example, arithmetic means are directly used in algorithms such as K-means
clustering while also being a theoretical underpinning of the model evaluation statistics
such as mean squared error (referenced later in this chapter). Intuitively, we define the
arithmetic mean as the central tendency of a discrete set of numbers—specifically it is the
sum of the values divided by the number of the values. Mathematically, this equation is
given by the following:
1 N
x= åxi
N i =1
14
Discovering Diverse Content Through
Random Scribd Documents
“When her note came—that note, you know—what would you
have said in her place? No, I don’t mean that. You’d have said: ‘Well,
I’m damned!’ But what would you have expected her to say?”
“‘Great God!’ or perhaps ‘Good gracious!’” Sir Paget suggested
doubtfully.
“She’s gone—gone!” I ventured to submit.
“Just so—just what I should have said,” Aunt Bertha agreed.
“Something like that. What our friend Mrs. Knyvett did say to me
was, ‘Miss Fleming, she’s done it!’”
“What did you say?” Sir Paget as nearly snapped this out as a
man of his urbanity could snap.
“I don’t think I said anything. There seemed nothing to——”
“Then you knew what she meant?”
Aunt Bertha pouted her lips and looked, as it might be,
apprehensively, at Sir Paget.
“Yes, I suppose I must have,” she concluded—with an obvious air
of genuine surprise.
“We sometimes find that we have known—in a way—things that
we never realized that we knew,” said Sir Paget—“much what I said
before. But—well, you and Mrs. Knyvett both seem to have had
somewhere in your minds the idea—the speculation—that Lucinda
might possibly do what she has done. Can you tell us at all why?
Because that sort of thing doesn’t generally happen.”
“By God, no!” Waldo grunted out. “And I don’t see much good in
all this jaw about it.”
A slight, still pretty, flush showed itself on Aunt Bertha’s wrinkled
cheeks—hers seemed happy wrinkles, folds that smiles had turned,
not furrows plowed by sorrow—“I’ve never been married,” she said,
“and I was only once in love. He was killed in the Zulu war—when
you were no more than a boy, Paget. So perhaps I’m no judge. But
—darling Waldo, can you forgive me? She’s never of late looked like
—like a girl waiting for her lover. That’s all I’ve got to go upon,
Paget, absolutely all.”
I saw Waldo’s hands clench; he sat where he was, but seemed to
do it with an effort.
“And Mrs. Knyvett?”
“Nothing to be got out of her just now. But, of course, if she
really had the idea, it must have been because of Arsenio Valdez!”
The name seemed a spur-prick to Waldo; he almost jumped to
his feet. “Oh, we sit here talking while——!” he mumbled. Then he
raised his voice, giving his words a clearer, a more decisive
articulation. “I’ve told you what I’m going to do. Julius can come
with me or not, as he likes.”
“No, Waldo, you’re not going to do it. I love—I have loved—
Lucinda. I held my arms open to her. I thought I was to have what I
have never had, what I have envied many men for having—a
daughter. Well, now——” his voice, which had broken into
tenderness, grew firm and indeed harsh again. “But now—what is
she now?”
“Monkey Valdez’s woman!”
These words, from Waldo’s lips, were to me almost incredible.
Not for their cruelty—I knew that he could be cruel in his rage—but
for their coarse vulgarity. I did not understand how he could use
them. A second later he so far repented—so far recovered his
manners—as to say, “I beg your pardon for that, Aunt Bertha.”
“My poor boy!” was all the old lady said.
“Whatever she may be—even if she were really all that up to to-
day you thought—you mustn’t go after her now, Waldo—neither you
nor Julius with you.” He paused a moment, and then went on slowly.
“In my deliberate judgment, based on certain facts which have
reached me, and reënforced by my knowledge of certain persons in
high positions, all Europe will be at war in a week, and this country
will be in it—in a war to the death. You fellows will be wanted; we
shall all be wanted. Is that the moment to find you two traipsing
over the Continent on the track of a runaway couple, getting
yourselves into prison, perhaps; anyhow quite uncertain of being
able to get home and do your duty as gentlemen? And you, Waldo,
are a soldier!”
Waldo sat down again; his eyes were set on his father’s face.
“You can’t suspect me of a trick—or a subterfuge. You know that
I believe what I’m telling you, and you know that I shouldn’t believe
it without weighty reasons?”
“Yes,” Waldo agreed in a low tone. His passion seemed to have
left him; but his face and voice were full of despair. “This is pretty
well a matter of life and death to me—to say nothing of honor.”
“Where does your honor really lie?” He threw away his cigarette,
walked across to his son, and laid a hand on his shoulder. But he
spoke first to me. “As I told you, I am breaking my word in
mentioning this knowledge of mine. It is desirable to confine that
breach of confidence to the narrowest possible limits. If I convince
Waldo, will you, Julius, accept his decision?”
“Of course, Sir Paget. Besides, why should I go without him?
Indeed, how could I—well, unless Mrs. Knyvett——”
“Mrs. Knyvett has nothing to do with our side of the matter.
Waldo, will you come out with me for an hour?”
Waldo rose slowly. “Yes. I should like to change first.” He still
wore his frock coat and still had a white flower in his buttonhole.
Receiving a nod of assent from Sir Paget, he left the room. Sir Paget
returned to the fireplace and lit a fresh cigarette.
“He will do what’s right,” he pronounced. “And I think we’d better
get him to Cragsfoot to-morrow. You come too, Julius. We’ll wait
developments there. I have done and said what I could in quarters
to which I have access. There’s nothing to do now but wait for the
storm.”
He broke away from the subject with an abrupt turn to Aunt
Bertha. “It’s a damned queer affair. Have you any views?”
“The mother’s weak and foolish, and keeps some rather second-
rate company,” said the old lady. “Surroundings of that sort have
their effect even on a good girl. And she’s very charming—isn’t she?”
“You know her yourself,” Sir Paget observed with a smile.
“To men, I mean. In that particular way, Paget?”
“Well, Julius?”
“Oh, without a doubt of it. Just born to make trouble!”
“Well, she’s made it! We shall meet again at tea, Aunt Bertha? I’ll
pick up Waldo at his room along the passage. And I’d better get rid
of my wedding ornament too.” He took the rose out of the lapel of
his coat, flung it into the fireplace, and went out of the room,
leaving me with Aunt Bertha.
“On the face of it, she has just suddenly and very tardily changed
her mind, hadn’t the courage to face it and own up, and so has
made a bolt of it?” I suggested.
“From love—sudden love, apparently—of Arsenio Valdez, or just
to avoid Waldo? For there seems no real doubt that Arsenio’s taken
her. He’s only once been to the flat, but the girl’s been going out for
walks every day—all alone; a thing that I understand from her
mother she very seldom did before.”
“Oh, it’s the Monkey all right. But that only tells us the fact—it
doesn’t explain it.”
“Very often there aren’t any explanations in love affairs—no
reasonable ones, Julius. Waldo takes it very hard, I’m afraid.”
“She’s made an ass of him before all London. It can’t really be
hushed up, you know.”
“Well,” Aunt Bertha admitted candidly, “if such an affair happened
in any other family, I should certainly make it my business to find
out all I could about it.” She gave a little sigh. “It’s a shock to me.
I’ve seen a lot, and known a lot of people in my day. But when you
grow old, your world narrows. It grows so small that a small thing
can smash it. You Rillington men had become my world; and I had
just opened it wide enough to let in Lucinda. Now it seems that I
might just as well have let in a high explosive. In getting out again
herself, she’s blown the whole thing—the whole little thing—to bits.”
“Love’s a mad and fierce master,” I said—with a reminiscence of
my classics, I think. “He doesn’t care whom or what he breaks.”
“No! Poor Lucinda! I wish she’d a nice woman with her!”
I laughed at that. “The nice woman would feel singularly de trop,
I think.”
“She could make her tea, and tell her that in the circumstances
she could hardly be held responsible for what she did. Those are the
two ways of comforting women, Julius.”
“As it is, she’s probably gone to some beastly foreign place where
there isn’t any tea fit to drink, and Monkey Valdez is picturesquely,
but not tactfully, insisting that her wonderful way has caused all the
trouble!”
“Poor Lucinda!” sighed Aunt Bertha again.
And on that note—of commiseration, if not actually of excuse—
our conversation ended; rather contrary to what might have been
expected, perhaps, from two people so closely allied to the deserted
and outraged lover, but because somehow Aunt Bertha enticed me
into her mood, and she—who loved men and their company as much
as any woman whom I have known—never, I believe, thought of
them en masse in any other way than as the enemy-sex. If and
where they did not positively desire that lovely women should stoop
to folly, they were always consciously or unconsciously, by the law of
their masculine being, inciting them to that lamentable course. Who
then (as the nice woman would have asked Lucinda as she handed
her the cup of tea) were really responsible when such things came
about? This attitude of mind was much commoner with Aunt
Bertha’s contemporaries than it is to-day. Aunt Bertha herself,
however, always praised Injured Innocence with a spice of malice.
There was just a spice of it in her pity for Lucinda and in the
remedies proposed for her consolation.
My own feeling about the girl at this juncture was much what
one may have about a case of suicide. She had ended her life as we
had known her life in recent years; that seemed at once the object
and the effect of her action. What sort of a new life lay before her
now was a matter of conjecture, and we had slender data on which
to base it. What did seem permissible—in charity to her and without
disloyalty to Waldo—was some sympathy for the struggle which she
must have gone through before her shattering resolve was reached.
CHAPTER IV
A
S Sir Paget had suggested, we—we three Rillington men and Aunt
Bertha—spent the Twelve Days, the ever-famous Twelve Days before
the war, at Cragsfoot. On the public side of that period I need say
nothing—or only just one thing. If we differed at all from the public
at large in our feelings, it was in one point only. For us, under Sir
Paget’s lead, it was less a time of hope, fear, and suspense than of mere
waiting. We other three took his word for what was going to happen;
his certainty became ours—though, as I believe (it is a matter of belief
only, for he never told me what he told Waldo on that walk of theirs on
the afternoon of the wedding day—which was not the day of a
wedding), his certainty was based not so much on actual information as
on a sort of instinct which long and intimate familiarity with
international affairs had given him. But, whatever was his rock of
conviction, it never shook. Even Waldo did not question it. He accepted
it—with all its implications, public and private.
Yes, and private. There his acceptance was not only absolute; it was
final and—a thing which I found it difficult to understand—it was
absolutely silent. He never referred to his project of pursuit—and of
rescue, or revenge, or whatever else it had been going to be. He never
mentioned Lucinda’s name; we were at pains never to pronounce it in
his presence. It was extraordinary self-control on the part of a man
whom self-control could, on occasion, utterly forsake. So many people
are not proof against gossiping even about their own fallen idols,
though it would be generally admitted that silence is more gracious;
pedestal-makers should be sure that they build on a sound foundation.
However, Waldo’s silence was not due to delicacy or to a recognition of
his own mistake; that, at least, was not how I explained it. He
recognized the result of his own decision. The event that was to raise
for all the civilized world a wall of division between past and future—
whom has it not touched as human being and as citizen?—erected a
barrier between Lucinda and himself, which no deed could pass, which
no word need describe. Only memory could essay to wing over it a blind
and baffled flight.
In spite of the overwhelming preoccupation of that national crisis—
Sir Paget remained in close touch with well-informed people in town,
and his postbag gave rise to talk that lasted most of the morning—my
memory, too, was often busy with those bygone days at Cragsfoot,
when the runaways had been of the party. Tall, slim, and fair, a girl on
the verge of womanhood, ingenuous, open, and gay though she was,
the Lucinda of those days had something remote about her, something
aloof. The veil of virginity draped her; the shadow of it seemed to fall
over her eyes which looked at you, as it were, from out of the depths of
feelings and speculations to which you were a stranger and she herself
but newly initiated. The world faced her with its wonders, but the
greatest, the most alluring and seductive wonder was herself. The
texture of her skin, peculiarly rich and smooth—young Valdez once,
sitting on a patch of short close moss, had jokingly compared it to
Lucinda’s cheeks—somehow aided this impression of her; it looked so
fresh, so untouched, as though a breath might ruffle it. Fancy might find
something of the same quality in her voice and in her laughter, a
caressing softness of intonation, a mellow gentleness.
What were her origins? We were much in the dark as to that; even
Aunt Bertha, who knew everything of that sort about everybody, here
knew nothing. The boys, Waldo and Valdez, had met mother and
daughter at a Commem’ Ball; they came as guests of the wife of one of
their dons—a lady who enjoyed poor health and wintered in “the
South.” There, “in the South,” she had made friends with the Knyvetts
and, when they came to England, invited them to stay. Mrs. Knyvett
appeared from her conversation (which was copious) to be one of those
widows who have just sufficient means to cling to the outskirts of
society at home and abroad; she frequently told us that she could not
afford to do the things which she did do; that “a cottage in the country
somewhere” was all she wanted for herself, but that Lucinda must “have
her chance, mustn’t she?” The late Mr. Knyvett had been an architect;
but I believe that Lucinda was by far the greatest artistic achievement in
which he could claim any share.
So—quite naturally, since Waldo always invited any friends he chose
—the pair found themselves at Cragsfoot in the summer of 1912. And
the play began. A pleasant little comedy it promised to be, played
before the indulgent eyes of the seniors, among whom I, with only a
faint twinge of regret, was compelled to rank myself; to be in the
thirties was to be old at Cragsfoot that summer; and certain private
circumstances made one less reluctant to accept the status of an elder.
Valdez paid homage in the gay, the embroidered, the Continental
fashion; Waldo’s was the English style. Lucinda seemed pleased with
both, not much moved by either, more interested in her own power to
evoke these strange manifestations than in the meaning of the
manifestations themselves. Then suddenly the squall came—and, as
suddenly, passed; the quarrel, the “row,” between Waldo and Valdez;
over (of all things in the world) the Legitimist principle! The last time I
had seen Waldo in a rage—until the day that was to have brought his
wedding with Lucinda! It had been a rage too; and Valdez, a fellow not
lacking in spirit as I had judged him, took it with a curious meekness;
he protested indeed, and with some vigor, but with a propitiatory air,
with an obvious desire to appease his assailant. We elders discussed
this, and approved it. Waldo was the host, he the guest; for Aunt
Bertha’s and Sir Paget’s sake he strove to end the quarrel, to end the
unpleasantness of which he was the unfortunate, if innocent, cause. He
behaved very well indeed; that was the conclusion we arrived at. And
poor dear Waldo—oh, badly, badly! He quite frightened poor Lucinda.
Her eyes looked bright—with alarm; her cheeks were unwontedly,
brilliantly red—with excited alarm. The girl was all of a quiver! It was
inexcusable in Waldo; it was generous of Valdez to accept his apologies
—as we were given to understand that he had when the two young
men appeared, rather stiff to one another but good friends, at the
breakfast table the next morning.
How did this view look now—in the light of recent events? Was there
any reason to associate the old quarrel of 1912 with the catastrophe
which had now befallen Waldo? I had an impulse to put these questions
to Aunt Bertha, perhaps to Sir Paget too. But, on reflection, I kept my
thoughts to myself. Silence was the mot d’ordre; Waldo himself had set
the example.
It was on the Saturday—the day on which the question of Belgian
neutrality defined itself, according to my uncle’s information, as the vital
point—that, wearied by a long talk about it and oppressed by Waldo’s
melancholy silence, I set out for a walk by myself. Cragsfoot, our family
home, lies by the sea, on the north coast of Devon; a cleft in the high
cliffs just leaves room for the old gray stone house and its modest
demesne; a steep road leads up to the main highway that runs along
the top of the cliff from east to west. I walked up briskly, not pausing till
I reached the top, and turned to look at the sea. I stood there, taking in
the scene and snuffing in the breeze. A sudden wave of impatient
protest swept over my mind. Wars and rumors of wars—love and its
tragedies—troubles public and private! My holiday was being completely
spoilt. A very small and selfish point of view, no doubt, but human, after
all.
“Oh, damn the whole thing!” I exclaimed aloud.
It must have been aloud—though I was not conscious that it was—
for another audible voice spoke in response.
“That’s just what Father said this morning!”
“It’s just what everybody’s saying,” I groaned. “But—well, how are
you after all this time, Miss Frost?”
For it was Nina Frost who stood beside me and I felt oddly surprised
that, in my retrospect of that earlier summer at Cragsfoot, I had never
thought of her; because she had been a good deal with us in our sports
and excursions. But the plain fact is that there had been little about her
in those days that would catch a mature man’s attention or dwell in his
memory. She was a chit of a girl, a couple of years or so younger than
Lucinda, much more the school-girl, pretty enough but rather
insignificant, attaching herself to the other three rather by her own
perseverance than thanks to any urgent pressing on their part. Lucinda
had altogether outshone her in the eyes of us all; she had been “little
Nina Frost from Briarmount.”
But now—she was different. A first glance showed that. She was not
only taller, with more presence; she had acquired not merely an ease of
manner; it was a composure which was quite mature, and might almost
be called commanding.
“You’ve changed!” I found myself exclaiming.
“Girls do—between sixteen and eighteen—or nearly nineteen!
Haven’t you noticed it, Mr. Rillington?” She smiled. “Hasn’t Lucinda
changed too? I expect so! Oh, but you’ve been abroad, haven’t you?
And since she didn’t—I mean, since the wedding didn’t—Oh, well,
anyhow, perhaps you haven’t seen her?”
“No, I haven’t seen her.” I had not—officially. “Are you going towards
Briarmount? May I walk with you?”
“Yes, do. And perhaps I haven’t changed so much, after all. You see,
you never took much notice of me. Like the others, you were dazzled by
Lucinda. Are you at liberty to tell me anything, Mr. Rillington? If you
aren’t, I won’t ask.”
She implied that she was not much changed. But would any child of
sixteen put it like that? I thought it precocious for eighteen; for it
cornered me. I had to lie, or admit practically the whole thing. I tried to
fence.
“But didn’t you go to the wedding yourself?” I asked. “If you did
——”
“No, I didn’t. Father wasn’t very well, and I had to stay down with
him.”
As we walked, I had been slyly studying her face: she had grown
handsome in a style that was bold and challenging, yet in no way
coarse; in fact, she was very handsome. As she gave me her most
respectable reason for not having attended—or attempted to attend—
Waldo’s wedding, she grew just a little red. Well, she was still only
eighteen; her education, though I remained of opinion that it had
progressed wonderfully, was not complete. She was still liable to grow
red when she told fibs. But why was she telling a fib?
She recovered her composure quickly and turned to me with a rather
sharp but not unpleasant little laugh. “As it turned out, I’m glad. It must
have been a very uncomfortable occasion.” She laughed again—
obviously at me. “Come, Mr. Rillington, be sensible. There are servants
at Cragsfoot. And there are servants at Briarmount. Do you suppose
that I haven’t heard all the gossip through my maid? Of course I have!
And can’t I put two and two together?”
I had never—we had never—thought of this obvious thing. We had
thought that we could play the ostrich with its head in the sand! Our
faithful retainers were too keen-sighted for that!
“Besides,” she pursued, “when smart society weddings have to be
put off, because the bride doesn’t turn up at the last moment, some
explanation is put in the papers—if there is an explanation. And she
gets better or worse! She doesn’t just vanish, does she, Mr. Rillington?”
I made no reply; I had not one ready.
“Oh, it’s no business of mine. Only—I’m sorry for Waldo, and dear
Miss Fleming.” A gesture of her neatly gloved and shapely hands
seemed to dismiss the topic with a sigh. “Have you seen anything of
Don Arsenio lately?” she asked the next moment. “Is he in England?”
“Yes. He was at the wedding—well, at the church, I mean.”
She came to a stop, turning her face full round to me; her lips were
parted in surprise, her white teeth just showing; her eyes seemed full of
questions. If she had “scored off” me, at least I had startled her that
time. “Was he?” she murmured.
At the point to which our walk had now brought us, the cliffs take a
great bulge outwards, forming a bold rounded bluff. Here, seeming to
dominate, to domineer over, a submissive Bristol Channel, Mr. Jonathan
Frost (as he then was—that is, I think, the formula) had built his
country seat; and “Briarmount” he had called it.
“Good Heavens,” said I, “what’s happened to the place? It’s grown!
It’s grown as much as you have!”
“We’ve built on a bit—a few more bedrooms, and bathrooms. And
garages, you know. Oh, and a ballroom!”
“No more than that?”
“Not at present. Come in and have a look—and some tea. Or are you
in too deep mourning?”
I found myself not exactly liking the girl, but interested in her, in her
composure—and her impudence. I accepted her invitation.
Since he could very well afford it, no blame need rest on Mr. Frost
for building himself a large house and equipping it sumptuously. The
only thing was that, when he had got it, he did not seem to care a bit
about it. Probably he built it to please Nina—or to enshrine Nina; no
doubt he found in his daughter a partial and agreeable solution of the
difficulty of how to spend the money which he could not help making.
He himself was a man of the simplest ways and tastes—almost of no
tastes at all. He did not even drink tea; while we took ours, he
consumed a small bowlful of one of those stuffs which, I believe, they
call cereals—this is a large domed hall of glass—conservatory, winter-
garden, whatever it should be called—full of exotic plants and opening
on a haughty terrace with a view of the sea. He was small, slight,
shabby, simple, and rather nervous. Still I gazed on him with some awe;
he was portentously rich; Mother Earth labored, and her children
sweated, at his bidding; he waved wands, and wildernesses became—
no, not quite paradises perhaps, but at all events garden-cities; he
moved mountains and where the ocean had been he made dry land.
Surely it beseems us to look with some awe on a man like that? I, at
least, being more or less in the same line of business, recognized in him
a master.
He greeted me very kindly, though I think that it had cost him an
effort to “place” me, to remember who I was. He spoke warmly of the
kindness which my uncle and Miss Fleming had shown to his motherless
girl. “They’ve made you quite at home at Cragsfoot, haven’t they, Nina?
And your cousin Waldo—Mr. Waldo taught you billiards, didn’t he?”
(There was no billiard room at Cragsfoot; these lessons presumably
took place at Briarmount.) “And he made company for your rides, too! I
hope he’s very well, Mr. Rillington? Oh, but didn’t you tell me that he
was engaged to be married, my dear?”
One must allow for preoccupation with important affairs. Still, this
was Saturday; as recently as the preceding Tuesday week, Mr. Frost
would have attended Waldo’s wedding, but for his own indisposition. I
stole a glance at Nina; she was just a little red again. I was not far from
embarrassment myself—on Waldo’s account; I gave a weak laugh and
said: “I’m afraid it’s not quite certain that the event will come off.”
“Oh, I’m sorry, I’m sorry,” he murmured apologetically. “It was the
pretty girl who came here with him once or twice—Miss—Miss—yes,
Miss Knyvett?”
“Yes, it was, Mr. Frost. But the—well, the arrangement is sort of—of
suspended.” With that distinctly lame explanation I rose to take my
leave.
I rather thought that Nina, being by now pretty plainly convicted of
fibbing, would stay where she was, and thus avoid being left alone with
me. However, she escorted me back through Briarmount’s spacious hall
—furnished as a sitting-room and very comfortable. She even came out
into the drive with me and, as she gave me her hand in farewell, she
said, with a little jerk of her head back towards the scene of my talk
with her father, “After that, I suppose you’re wondering what was the
real reason for my not coming to the wedding?”
“Perhaps I am. Because you seem to have kept up the old friendship
since I’ve been away.”
“Sometimes people don’t go to functions because they’re not
invited.”
“What, you mean to say——”
“I should have been the skeleton at the feast!” She looked me in the
face, smiling, but in a rather set, forced fashion. Then, as she turned
away, she added with a laugh, “Only, as it turned out, there was no
feast, was there, Mr. Rillington?”
When I got back to Cragsfoot, I met Waldo in the garden, walking
up and down in a moody fashion and smoking his pipe. “Been for a
walk?” he asked.
“I started on one, but I met Nina Frost and she took me in to tea.”
He stood still, smoking and staring out to sea. “Did she say anything
about me?” he asked.
“Hardly about you yourself. She referred to—the affair. The servants
have been chattering, it seems. Well, they would, of course!”
He gave a nod of assent. Then he suddenly burst out in a vehement
exclamation: “She wasn’t there to see it, anyhow, thank God!” With that
he walked quickly away from me and was soon hidden in the shrubbery
at the end of the walk.
How did he know that she had not come to the church? He had not
been in the body of the church himself—only in the vestry. Many people
had actually gone in—early arrivals; Sir Paget had told me so. Many
more had been turned away from the doors. Waldo could not have
known from his own observation that Nina Frost was not there. Possibly
somebody had told him. More probably he had known beforehand that
she would not be there, because she had not been invited. But why
should he thank God that she was not at the church?
So there was the coil—unexplained, nay, further complicated by the
intrusion of a fourth party, Miss Nina Frost. Unexplained I had to leave
it. The next morning—Sunday though it was—Sir Paget carried me off to
town, by motor and rail, to interview some bigwig to whom he had
mentioned me and who commanded my attendance. I had not even a
chance of a private talk with Aunt Bertha, whose silence about Nina
now struck me as rather odd.
The war was upon us. It had many results for many people. One
result of it was that, instead of the start of hours for which they had
schemed, our runaway couple secured a start of years. That made a
great difference.
CHAPTER V
I
DO not want to say more about the war or my doings during it
than is strictly necessary to my purpose. The great man to
whom I have referred took a note of my qualifications. Nothing
came of this for a good many months, during which I obtained
a commission, went through my training, and was for three months
fighting in France. Then I was called back, and assigned to non-
combatant service (it was not always strictly that, as a nasty scar on
my forehead, the result of a midnight “scrap” in a South American
seaport where I happened to be on business, remains to testify). My
knowledge of various parts of the world and my command of
languages made me of value for the quasi-diplomatic, quasi-
detective job with which I was entrusted, and I continued to be
employed on it throughout the war. It entailed a great deal of
traveling by sea and land, and a lot of roughing it; it was interesting
and sometimes amusing; there was, of course, no glory in it. I was a
mole, working underground; there were a lot of us. For the best part
of a year I was out of Europe; I was often out of reach of letters,
though now and then I got one from Aunt Bertha, giving me such
home news as there was, and copying out extracts from what she
described as “Waldo’s miserable letters” from France—meaning
thereby not unhappy—he wrote very cheerfully—but few, short, and
scrappy. Sir Paget, it appeared, had found some sort of advisory job
—a committee of some kind—in connection with the Foreign Office.
It was when I came back to Europe, in the spring of 1916, and
was staying for a few days at a small town in the South of France—I
was at the time covering my tracks, pending the receipt of certain
instructions for which I was waiting, but there is no harm in saying
now that the town was Ste. Maxime—that I ran into Lucinda
Knyvett. That is almost literal. I came round a sharp corner of the
street from one direction, she from another. A collision was so
narrowly avoided that I exclaimed, “Pardon!” as I came to an abrupt
stop and raised my hat. She stopped short too; the next moment
she flung out both her hands to me, crying, “You, Julius!” Then she
tried to draw her hands back, murmuring, “Perhaps you won’t——!”
But I had caught her hands in mine and was pressing them. “Yes!
And it’s you, Lucinda!”
It was about midday, and she readily accepted my suggestion
that we should lunch together. I took her to a pleasant little
restaurant on the sea-front. It was bright, warm, calm weather; we
ate our meal out of doors, in the sunshine. In reply to her inquiries—
made without any embarrassment,—I told her what Cragsfoot news
I had. She, in return, told me that Arsenio—he also was mentioned
without embarrassment—had gone to Italy when that country
entered the war, and was at this moment on the staff of some
General of Division; he wrote very seldom, she added, and, with
that, fell into silence, as she sipped a glass of wine.
She had changed from a girl into a woman; yet I did not divine in
her anything like the development I had marked in Nina Frost. In
appearance, air, and manner she was the Lucinda whom I had
known at Cragsfoot; her eyes still remotely pondering, looking
inwards as well as outwards, the contour of her face unchanged, her
skin with all its soft beauty. But she was thinner, and looked rather
tired.
“Arsenio told me that you saw me in the taxi that day,” she said
suddenly.
“He must have been very much amused, wasn’t he? He certainly
made a pretty fool of me! And put the cap on it by coming to the—to
the church, didn’t he?”
“I suppose, when once he’d met you, he was bound to go there,
or you’d have suspected.”
“He could have made some excuse to leave me, and not turned
up again.”
She did not pursue her little effort to defend Valdez; she let it go
with a curious smile, half-amused, half-apologetic. I smiled back.
“Monkey Valdez, I think!” said I. She would not answer that, but her
smile persisted. “You were looking very happy and bonny,” I added.
“I was happy that day. I had at last done right.”
“The deuce you had!” That was to myself. To her I said, rather
dryly, “It certainly was at the last, Lucinda.”
“It was as soon as I knew—as soon as I really knew.”
The waiter brought coffee. She took a cigarette from me, and we
both began to smoke.
“And it’s true that I didn’t dare to face Waldo. I was physically
afraid. He’d have struck me.”
“Never!” I exclaimed, indignant at the aspersion on my kinsman.
“Oh, but yes!—I thought that he would fight Arsenio that night at
Cragsfoot—the night Arsenio first kissed me.” She let her cigarette
drop to the ground, and leant back in her chair. Her eyes were on
mine, but the shadow of the veil was thick. “It all began then—at
least, I realized the beginning of it. It all began then, and it never
stopped till that day when I ran away. Shall I tell you about it?”
“We were all very fond of you—all of us. I wish you would.”
She laid her hand on my arm for a moment. “I couldn’t have told
then—perhaps I can now. But, dear Julius, perhaps not quite plainly.
There’s shame in it. Some, I think, for all of us—most, I suppose, for
me.”
At this point a vision of Aunt Bertha’s “nice woman” flitted before
my mind’s eye; it was a moment for her ministrations—or ought to
have been, perhaps. Lucinda was rather ruminative than distressed.
“We were very happy that summer. I had never had anything
quite like it. Mother and I went to lunches and teas—and I’d just
begun to go to a few dances. But people didn’t ask us to stay in
country houses. Three days’ visit to Mrs. Wiseman at Oxford was an
event—till Cragsfoot came! I love that old house—and I shall never
see it again!—Oh, well——! The boys were great friends; all three of
us were. If anything, Waldo and I took sides against Arsenio,
chaffing him about his little foreign ways, and so on, you know.
Waldo called him Monkey; I called him ‘Don’—sometimes ‘Don
Arsenio.’ I called Waldo just ‘Waldo’—and I should have called
Arsenio just by his name, only that once, when we were alone, he
asked me to, rather sentimentally—something about how his name
would sound on my lips! So I wouldn’t—to tease him. I thought him
rather ridiculous. I’ve always thought him ridiculous at times. Well,
then, Nina Frost took to coming a good deal; Miss Fleming had pity
on her, as she told me—her mother wasn’t long dead, you know, and
she was all alone at Briarmount with a governess. Do you remember
Fräulein Borasch? No? I believe you hardly remember Nina? You
hardly ever came on excursions, and so on, with us. The boys told
me all that sort of thing bored ‘old Julius.’ Nina rather broke up our
trio; we fell into couples—you know how that happens? The path’s
too narrow, or the boat’s too small, or you take sides at tennis. And
so on. For the first time then the boys squabbled a little—for me. I
enjoyed that—even though I didn’t think victory over little Nina
anything to boast about. Well, then came that day.”
Lucinda leant forward towards me, resting her arms on the table
between us; she was more animated now; she spoke faster; a slight
flush came on her cheeks; I likened it to an afterglow.
“Nina had been there all the afternoon, but she went home after
tea. We’d been quite jolly, though. But after dinner Waldo whispered
to me to come out into the garden. I went—it was a beautiful
evening—and we walked up and down together for a few minutes.
Waldo didn’t say anything at all, but somehow I felt something new
in him. I became a little nervous—rather excited. We were at the
end of the walk, just where it goes into the shrubbery. He said,
‘Lucinda!’—and then stopped. I turned sharp round—towards the
house, suddenly somehow afraid to go into the shrubbery with him;
his voice had sounded curious. And there—he must have come up as
silently as a cat—was Arsenio, looking so impishly triumphant! Waldo
had turned with me; I heard him say ‘Damn!’ half under his breath.
‘Do I intrude?’ Arsenio asked. Waldo didn’t answer. The moon was
bright; I could see their faces. I felt my cheeks hot; Waldo looked so
fierce, Arsenio so mischievous. I felt funnily triumphant. I laughed,
cried, ‘Catch who catch can!’ turned, and ran down the winding path
through the shrubbery. I ran quite a long way. You know how the
path twists? I looked back once, and saw Arsenio running after me,
laughing: I didn’t see Waldo, but I could hear his footsteps. I ran
round another turn. By then Arsenio was quite close. I was out of
breath and stopped under a big tree. I put my back against it, and
faced Arsenio; I think I put out my hands to keep him off—in fun,
you know. But he came and took hold of my hands, and pulled me
to him and kissed me on my lips. ‘Caught!’ he said as he let me go.
Then I saw Waldo just a few yards off, watching us. I was trembling
all over. I ran away from them, back towards the house; but I didn’t
dare to go straight in; I felt that I shouldn’t be able to answer, if
anybody spoke to me. I sat down on the bench that stands close by
the door, but is hidden from it by the yew hedge. Presently I heard
them coming; I heard Waldo speaking angrily, but as they got nearer
the house, he stopped talking, so I didn’t hear anything that he said.
But Arsenio told me—later on—that he said that English gentlemen
didn’t do things like that, though dirty Spaniards might—and so on. I
sat where I was, and let them go in. But presently I felt that I must
see what was happening. So I went in, and found them quarreling:
at least, Waldo was abusing Arsenio—but you know about that; you
were there. I thought they’d fight—they would have if you and Sir
Paget hadn’t been there—but somehow, by now, I didn’t mind if
they did. I wasn’t frightened any more; I was excited. You know how
it ended. I didn’t then, because after a good deal of it Sir Paget sent
me to bed—don’t you remember? I went to bed, but I didn’t go to
sleep for ever so long. I felt that something great had happened to
me. Men had tried to kiss me a few times before; one or two had
managed just to kiss my cheek in a laughing kind of way. This was
different to me. And there was Waldo too! I was very young. I
suddenly seemed to myself immensely important. I wondered—oh,
how I wondered—what they would do the next morning—and what I
should do. I imagined conversations—how I should be very stiff and
dignified—and Arsenio very penitent, but protesting his devotion. But
I couldn’t imagine how Waldo would behave. Anyhow, I felt that the
next morning would be the most awfully exciting moment in my life,
that anything might happen.”
Lucinda paused, looking at me with a smile that mocked the girl
whose feelings she had been describing. “Nothing did!”
After another pause she went on: “Later on, of course, I heard
how that was. I’ve heard it from both of them! Arsenio didn’t really
care for me at that time, though Waldo did. And Arsenio was very
fond of Waldo; he felt he’d behaved rather badly, and he didn’t bear
malice against Waldo for abusing him. Arsenio is malicious in a way;
it’s fun to him to make people look and feel silly; but he doesn’t
harbor malice. He’s not rancorous. He went to Waldo’s room early in
the morning—while Waldo was still in bed—and apologized. He said
he must have had a glass too much of champagne, that he hadn’t
meant anything, and that if he’d had the least notion how Waldo
would feel about it—and so on! In fact, he made light of the whole
thing, so far as I was concerned. Waldo listened to it all in silence,
and at the end just said, ‘All right, old chap. There’s an end of it.’ But
he didn’t really forgive Arsenio—and he didn’t forgive me, though it
hadn’t been my fault—had it? In the first-place, between us we’d
made him give himself away; he’s very proud, and he hates that. In
the second, he’s much better than you’d suppose at seeing into
things; he has a sort of instinct; and from that day, right on, he was
instinctively afraid of Arsenio; he felt that, if Arsenio chose, he could
be dangerous—about me. I know it, from the way he used to speak
of him later on—when we were engaged—always trying to probe
me, to find out my feelings about Arsenio, whether I was thinking
about him, whether I ever heard from him, and things like that. All
the time he never had Arsenio out of his mind. Well—he was right.
“But I knew nothing of all that at the time. To me they seemed
just a little sulky to one another, and to me, too. Otherwise they
ignored what had happened, made nothing of it, never referred to it
in any way. I was most frightfully hurt and—and let down. To me it
had been a great beginning—of something, though I didn’t know of
what. I couldn’t understand how Arsenio could treat it as nothing—
that he shouldn’t apologize and abase himself if he’d meant nothing
serious, that he shouldn’t speak to me again if he really cared for
me. I felt utterly bewildered. Only I had a strange feeling that
somehow, in some way, Arsenio had acquired a right over me by
kissing my lips. Of that feeling I never got rid.”
From a frown she broke into a smile again, as she went on. “It
was a miserable week—till we went. Both the boys avoided me
whenever they could. Both have told me why since, but I don’t
believe that either of them told me the truth. Arsenio said it was
because he couldn’t trust himself not to make love to me, and he
had practically promised that he wouldn’t. I think it was because he
thought I would expect to be made love to (I did!), and he didn’t
want to; he wasn’t in love with me then; besides he was afraid of
Waldo. Waldo said it was because he was ashamed of himself. I
daresay he was ashamed, but it was much more because he was in
love with me, but was too proud to seem to compete with Arsenio.
Whatever the reasons, the result was—triumph for Nina! She was
invited over every day and all day. Both of them tried to keep with
her—in order to avoid me. I wasn’t exactly jealous, because I knew
that they really wanted to be with me—but for the complications.
But I was exasperated to see that she thought—as, of course, she
must—that she had cut me out. How her manner changed! Before
this she had adored me—as younger girls do older ones sometimes;
‘Darling Lucinda!’ and so on! I’d noticed her trying to imitate me,
and she used to ask where I got such pretty frocks. Now she
patronized me, told me how I must wish I had a nice home (she
knew I hadn’t) like Cragsfoot or Briarmount, and said what a pity it
was my mother couldn’t give me more chances of riding, so that I
could improve! She did ride much better than I—which made it
worse.”
Here I looked at Lucinda, asking leave to laugh. She gave it in
her own low-murmuring laughter at herself. “So it ended. We went
away, and I was very glad when we did. I went away without either
Arsenio or Waldo having said to me a single word that mattered.”
“I must have been very dull to have noticed nothing—except just
the quarrel; well, the quarrel itself, and how you looked while it was
going on—till you were sent to bed.”
“How did I look?”
“Just as you did when I saw you in the taxi at the corner by
Marlborough House.”
“I’m very glad I didn’t see you! You’d have brought back what I’d
managed to put out of my mind. As though I could put it out of my
life!”
Suddenly and abruptly she pushed her chair back from the table.
“Aren’t we staying here a frightfully long time? That waiter’s staring
at us.”
“But surely I haven’t heard all the story yet?”
“All the story? No. Only the prologue. And the prologue’s a
comedy, isn’t it? A children’s comedy! The rest isn’t quite like that.
Pay the bill and let’s go. For a walk, if you like—and have time.”
“I ought just to call at my hotel—the Méditerranée—and see if
there’s anything for me—any telegrams. If there aren’t, I should like
to sit by the sea, and smoke, and hear the next chapter.”
At the moment Lucinda merely nodded. But as we walked away,
she put her arm in mine and said, “The next chapter is called
‘Venice,’ and it’s rather a difficult one for me to tell.”
“I hope I’m not a person who has to have all the t’s crossed and
all the i’s dotted. Arsenio has—or had—a ‘palazzo’ at Venice?”
“Yes. We stayed there.”
CHAPTER VI
VENICE
T
HE instructions for which I was waiting did not reach me for
three days: I found reason to suspect, later on, that bribery
had been at work; they had almost certainly been delayed,
copied, and communicated to enemy quarters. The bulk of
these enforcedly idle hours I spent with Lucinda—at the restaurant,
on the sea-front, once or twice at my hotel, but never in the little
house where she had a room: I often escorted her to the door, but
she never asked me in. But we grew intimate; she told, I think, all,
or almost all, the story, though often still with the air of examining
herself, or of rendering an account to herself, rather than of being
anxious to tell me: sometimes she would seem even to forget my
presence. At other points, however, she would appeal directly to me,
even urgently, as though she hung on my verdict. These changes
gave variety and life to her story; one saw her living again through
all her moods and experiences: on the other hand, it cannot be
denied that they lengthened the narrative.
In the spring of 1913—the spring after their visit to Cragsfoot—
her mother and Lucinda went to stay on the top floor but one in
Arsenio Valdez’s palazzo at Venice, Valdez himself inhabiting the
attics immediately above them. Poverty, the satirist remarked long
ago, has no harsher incident than that of making people ridiculous; it
may have worse moral effects. Mrs. Knyvett had not so much
accepted Valdez’s invitation as intrigued and cadged for it; and they
stayed rent free, though even then Valdez was by no means a well-
to-do man. And Mrs. Knyvett could not receive favors in the grand
manner. She took, but she took cringingly; she over-acknowledged,
constantly by manner and even by word, reminding the donor and
herself of the gift, reminding her daughter also. She did not, it is
true, know about the kiss in the garden at Cragsfoot; Lucinda kept
that to herself; her view was that in her mother’s hands it would
have been another lever. “Arsenio lodged us free as it was; if mother
had known that, she’d have made him board us too!” Even as it was,
he seemed to have entertained them a good deal (as was only
natural) while he played cicerone, showing them the sights and
pleasures of the place.
It was by no means Mrs. Knyvett’s intention or desire that her
daughter should marry Arsenio. Her ambition flew higher. Cragsfoot
was to her still the most eligible prospect or project which had so far
presented itself; she kept in touch with it by letters to Aunt Bertha;
in them she angled for another invitation there, just as she had
cadged for Arsenio’s invitation to the palazzo. How many invitations
does a charming daughter “make” in the arithmetic of genteel
poverty? Arsenio was quite aware of her attitude towards him, but it
pleased his monkeyish humor to pretend to believe that she favored
a suit which he had himself no intention of pressing. Arsenio could
not afford to marry a poor girl, and probably did not want to marry
at all. His taste was for a bachelor life, and his affairs were in a
precarious state. He could hardly be said to live by gambling; he
existed in spite of it—in a seesaw between prosperity and penury; as
such men do, he splashed his lire about when he had them; when he
was “cleaned out,” he would disappear from the ken of the Knyvetts
for a day or two, engaged in “milking” sundry old and aristocratic
friends of his father, who still resided at Venice in a stately and
gloomy seclusion, and could be persuaded to open their not very fat
purses to help a gentleman of Spain who upheld the Legitimist
principle, as we know—from past events—that Arsenio did! No, he
certainly did not intend at the beginning of their visit to mate poverty
to poverty.
But—there was Lucinda! Lucinda under blue skies by day and soft
moonlight by night. There was that secret memory between them,
the meeting of their lips; for him an incentive to gallantry, almost an
obligation, according to his code; for her, more subtly, a tie, a union
that she could not lightly nor wholly disown. He did not speak of it
directly, but he would circle round it in talk, and smile in an impish
exchange of the unspoken memory; he would laugh at Waldo, while
with feigned sincerity he praised his sterling qualities. “Oh, his
reliability, his English steadiness—dear, good, old Waldo! You’d trust
him—even in a gondola, Lucinda!”
The gondola! Let it stand for the whole of Venice’s romantic
paraphernalia; an old theme, a picture painted a thousand times. No
need to expatiate on it here. To him it was all very familiar—the
nearest thing he had to a home; to her, of course, it was a
revelation. They were both susceptible to impressions, to beauty. He
retained his sensibility, she developed hers. She saw new things
through his eyes; he saw old ones newly reflected in the light of
hers. His feelings regained freshness, while hers grew to maturity—a
warm ripeness in which the man and the place were fused together
in one glowing whole. “Oh, I lived then!” she cried, clasping her
hands together and beating them upon her knee.
Yet it must still have been with her own aloofness, delicacy,
difficulty of approach; the fires gleamed through the veil, but the veil
was round them. He complained, it appeared, of her coldness, of the
distance at which she kept him, at relapses into formality after hours
of unreserved merriment. Mrs. Knyvett chid her; was he not the
friend, the host, the benefactor? Within prudent bounds he should
be handsomely encouraged—and rewarded. “Mother told me that
well-bred girls knew how to make themselves respected without
being prudish.” Maternal philosophy of an affectionately utilitarian
order—one eye on present amenities, the other on grander prospects
in the future!
But was there no fear also in that maternal breast? Did the
situation and the actors raise no apprehension? To some people—to
how many? Some have maintained to all!—morality is not a master,
but a good and ever vigilant servant. It preserves the things that are
of real value, the marketable stuff. And it dignifies its watch and
ward with such high names, such sacred and binding traditions, that
—well, really, what between the august sanctions on the one hand
and the enormous material advantages on the other, can it be
dreamt of that any reasonable girl will forget herself? So one may
suppose that Mrs. Knyvett reasoned. For what, after all, is the
“leading article” in a girl’s stock-in-trade? Who, properly instructed,
would sell that under market price, and so stand bankrupt?
So much may be said in apology for Mrs. Knyvett’s blindness to
her daughter’s peril; for in peril she was. Then an apology is needed
for Arsenio? It would show a lack of humor to tender it; it is the last
thing which those who have known and liked Monkey Valdez would
think of doing. He was a “good Catholic” by tradition, and a
gentleman by breeding; but he was an honest man only by fits and
starts—when honesty appealed to his histrionic sense, when it
afforded him the chance of a beau geste, when he felt himself under
the eyes of the men with whom he had been brought up, who
expect honesty even in dealings with women—at all events, with
girls of their own caste; who draw a broad distinction between an
intrigue and a seduction; who are, in fact (not to labor the subject),
born and trained adepts in the niceties, some of them curious, of the
code of honor, which is certainly not a religious rule or an ethical
system, but may be considered to embody the laws of sex warfare,
to be a Hague Convention between the sexes.
Yet there is no need to picture the poor Monkey as the deliberate
villain of the stage. Your true villain must be deliberate and must
rejoice in his villainy, or all the salt is out of him. Arsenio was
certainly not deliberate, and in no way realized himself as a villain.
The event—the course of affairs afterwards—proves that. He
probably let his boat drift pleasantly, delightfully, down the river, till
the swirl of rapids caught it; it is likely that he was himself surprised;
the under-nature stormed the hesitating consciousness.
She gave me no particulars; I asked for none. She shrank from
them, as I did. It was after a delightful evening alone together, on
the water, that it came. Mrs. Knyvett had gone to bed; they were
alone, full of the attraction of each other—and of “it all.” So Lucinda
summed up the notoriously amatory influences of the Adriatic’s
Queen. She appealed to me—woman now, to a man of middle age—
to understand how it happened. As she told me—well, she hardly
told me, she let me see—she laid her hand in mine, her eyes sought
mine, straight, in question—yet hardly to me—rather to some
tribunal which she blindly sought, to which she made a puzzled but
not despairing, not altogether too tragic, appeal: “At Cragsfoot he
had kissed my lips, you know; and I wasn’t angry. That meant I liked
him, didn’t it? That meant——? That meant—the same?”
That seemed to me to record—as she, saying it, still seemed to
retain—a wonderful freedom from the flesh. She judged things by
the spirit. A terribly dangerous criterion; anybody can distort it;
anybody may snigger at it—though I think that it offers more
resistance to an honest laugh. There is a sort of pathos about it.
Meant the same! Poor dear! The gulf between the two things!
Immeasurable! Let speak religion (though there perhaps the voices
have varied), morality, prudence, the rest of them! And virgin
modesty? Shall we lay its fall most essentially in the less or the
greater—in the parley or in the surrender? That’s what she seemed
to ask. But what answer could a plain man of the world give her?
She had a few—a very few days of happiness, of forgetfulness of
everything except their love. Then the clouds gathered. She waited
for a word from him that did not come—not the first time that he
had kept her thus waiting—yet how different! Arsenio grew fretful,
disconsolate, and sometimes sullen. One of his disappearances
occurred; he was raising the wind among his long-suffering
aristocrats; he was scraping together every coin he could and
throwing them all on the gaming table. If fortune smiled, he would
do the right thing, and do it handsomely; if she frowned—and there
could be no doubt that she was frowning now—what lay before him,
before them? A scamped and mean ménage à trois, existence eked
out with the aid of Mrs. Knyvett’s scanty resources, and soured by
her laments! No money for gayety, for play, to cut a figure with! He
shrank from the prospect. He could not trust his love with it;
probably he did not trust hers either. He began to draw away from
her; she would not reproach or beseech. “I had taken the chances; I
had gambled too,” she said.
Unless something had happened which put Arsenio under an
even more imperative obligation—one which, as I would fain believe,
he must have honored—it seems probable that the affair would in
any case have ended as it did; but the actual manner of its ending
was shaped by an external incident.
The two were sitting together one morning in the Knyvett salon,
Lucinda mending her gloves, Arsenio doing nothing and saying
nothing, melancholy and fagged after a bout of gambling the night
before. Mrs. Knyvett came in, with an air of triumph, holding a letter
in her hand. She was still ignorant of the situation; still sure that her
daughter was making herself respected—though surely less
apprehensive of her prudishness? And, while they had been pursuing
their devices, she had had hers also to pursue. Success had crowned
her efforts. The letter was from “dearest Miss Fleming”; it invited
mother and daughter to pay another visit to herself and Sir Paget as
soon as they returned to England; that is, in about six weeks; for
they had a stay with friends in Paris arranged in the immediate
future—a thing that had already begun to trouble Lucinda.
“It’s delightful!” said Mrs. Knyvett. “Won’t it help us splendidly
through the summer! Any chance of your being there too, Don
Arsenio? That would make it perfect!”
The good lady did not stay for an answer. She had her hat on,
and was going out to do her marketing. She laid the letter down on
the table between them, and bustled out, her face still radiant with
the joy of successful maneuver.
So Cragsfoot, completely forgotten of recent days, made its
reëntry on the scene.
For a few moments they sat silent still, with the letter between
them. Then Lucinda said, “What are we to do, Arsenio?” She raised
her eyes from her sewing and looked across at him. He did not
return her glance; he was scowling. The invitation to Cragsfoot (he
did not know about the French visit, which Mrs. Knyvett could readily
have put off if she had preferred to stay on at Venice) brought him
up short; it presented him with an issue. It forced Lucinda’s hand
also. No mere excuse, no mere plea of disinclination, would prevent
Mrs. Knyvett from going to Cragsfoot and taking her daughter with
her. To stay there was not only a saving and a luxury, in her eyes it
was also prestige—and a great possibility!
“Damn Cragsfoot!” she heard him mutter. And then he laid his
head between his hands on the table and began positively to sob.
How much for unsuccessful gambling, how much for too successful
love, Heaven knows! But Monkey Valdez sobbed.
She put down her work, went round to the back of his chair, and
put her arms about his neck. “I know, I know, Arsenio. Don’t be so
miserable, dear. I understand. And—and there’s no harm done. You
only loved me too much—and if you can’t do what—what I know you
want to do——”
He raised his head and said (in what she called “a dead voice”),
“I’m what he called me, that’s the truth. He called me a dirty
Spaniard; he said no English gentleman would do what I did. The
night I kissed you at Cragsfoot! Waldo!”
“He said that to you? He told you that? Waldo? Oh, I knew he
was very angry; but you’ve never told me that he said that.”
“Then,” said Lucinda, as she told her story to me, “I did
something, or said something, that seemed to make him suddenly
angry. What he repeated—what Waldo had said—somehow struck
me with a queer sense of puzzle. It seemed to put him and Waldo
back into the same sort of conflict—or, at least, contrast—that I had
seen them in at Cragsfoot. I didn’t, of course, accept the ‘dirty
Spaniard’ part; Waldo was just angry when he said that. But the
words did bring Waldo back to my mind—over against Arsenio, so to
speak. I don’t know whether you’ve ever noticed that I sometimes
fall into what they call a brown study? I get thinking things over, and
rather forget that I’m talking to people. I wasn’t angry with Arsenio;
I was feeling sorry for him; I loved him and wanted to comfort him.
But I had to think over what he had told me—not only (perhaps not
so much) as it bore on Arsenio, but as it bore on myself—on what I
had done and felt, and—and allowed, you know. Well, Arsenio
suddenly called out, quite angrily, ‘You needn’t pull your arms away
like that!’ I had done it, but I hadn’t been conscious of doing it; I
didn’t think about it even then. I was thinking of him—and Waldo.
And I know that I was smiling, as the old Cragsfoot days came back
to me. I wasn’t thinking in the least about where my arms were! ‘Of
course you and Waldo are curiously different,’ I said.
“He jumped to his feet as if I had struck him, and broke out in a
torrent of accusation against me. A few minutes before he had
himself said that Waldo had told the truth about him. Now he
declared that it was I who had said it. I hadn’t said anything of the
sort—at all events, meant anything of the sort. I suppose I was sore
in my heart, but I should never have said a word. But he would have
it that I had meant it. He talked very fast, he never stopped. And—I
must tell you the truth, Julius—it all seemed rather ridiculous to me,
rather childish. I believe that I listened to most of it smiling—oh, not
a merry smile, but a smile all the same. I was waiting for him to
work himself out, to run down; it was no good trying to interrupt.
And all the time the contrast was in my mind—between him and
Waldo, between Waldo’s anger and—this! I felt as I suppose a
woman feels towards her naughty child; I wanted to scold and to
kiss him both at once. I even thought of that wicked nickname that
Waldo has for him! At last—after a great deal of it—he dashed one
hand through his hair, thumped the table with the other, and flung
out at me, ‘Then go to him! Go to your English gentleman! Leave me
in the gutter, where I belong!’ And he rushed out of the room. I
heard his steps pattering up the stone stairs to his own floor.”
“You must have been terribly distressed,” I said—or something
formal of that kind.
“No. I didn’t believe that anything had really happened. I waited
half an hour to let him cool down. But Mother might be back every
minute; there was still that question about Cragsfoot! I had to have
some answer! I went up to his apartment and knocked. I got no
answer. I went down to Amedeo the portière, and he told me that
Arsenio had gone out ten minutes before—I hadn’t heard his
footsteps coming down again, he must have stolen down softly; he
was carrying a bag, had a gondola called, and went off in the
direction of the station, saying that he would be back in a few days.
That was the end of—Venice!”
She came to a stop, gently strumming her fingers on the arm of
her chair. On an impulse I leant forward and asked her a question:
“Are you Madame Valdez now, Lucinda?”
“Donna Lucinda Valdez, at your service, sir! Since the day after
you saw me in the taxi.”
“Then he must have explained—Venice?”
“Never. From the first day that we met again, we have never
mentioned Venice.” She touched my arm for a moment. “I rather like
that. It seems to me rather a tactful apology, Julius. He began
courting me all afresh when he came to England. At least he took it
up from where it had stopped at Cragsfoot.”
“It may be tactful; it’s also rather convenient,” I commented
gruffly. “It avoids explanations.”
A gleam of amusement lit up her eyes. “Poor Arsenio! He was in
a difficulty—in a corner. And he’d been losing, his nerves were
terribly wrong. There was the question of—me! And the question of
Cragsfoot! And then Waldo came into it—oh, I’m sure of that. Those
two men—it’s very odd. They seem fated to—to cross one another—
to affect one another sometimes. I wonder whether——!” She broke
off, knitting her brow. “He sounded most genuine in that outbreak of
his when he mentioned Waldo. I think he was somehow realizing
what Waldo would think and say, if he knew about Venice. Perhaps
so, perhaps not! As for the rest of it——”
“You think he wasn’t quite as angry as he pretended to be?”
She seemed to reflect for a moment. “I didn’t say his anger was
unreal, did I? I said it was childish. When a child runs heedlessly into
something and hurts himself, he kicks the thing and tells his mother
that it’s horrid. I was the thing, you see. Arsenio’s half a child.” Again
she paused. “He’s also an actor. And he contrived, on the whole, a
pretty effective exit!”
“That you ever let him come back again is the wonder!” I cried.
“No. It’s what happened before he came back that puzzles me,”
she said.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
textbookfull.com