Download ebooks file Ethics Of Artificial Intelligence S. Matthew Liao all chapters

Download the full version of the textbook now at textbookfull.
com
Ethics Of Artificial Intelligence S. Matthew

Liao
https://textbookfull.com/product/ethics-of-
artificial-intelligence-s-matthew-liao/
Explore and download more textbook at https://textbookfull.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
AI for the Good: Artificial Intelligence and Ethics

(Management for Professionals) Stefan Vieweg
https://textbookfull.com/product/ai-for-the-good-artificial-
intelligence-and-ethics-management-for-professionals-stefan-vieweg/
textbookfull.com
Artificial Intelligence By Example Develop machine

intelligence from scratch using real artificial
intelligence use cases Denis Rothman
https://textbookfull.com/product/artificial-intelligence-by-example-
develop-machine-intelligence-from-scratch-using-real-artificial-
intelligence-use-cases-denis-rothman/
textbookfull.com
Distributed Artificial Intelligence Second International

Conference DAI 2020 Nanjing China October 24 27 2020
Proceedings Matthew E. Taylor
https://textbookfull.com/product/distributed-artificial-intelligence-
second-international-conference-dai-2020-nanjing-china-
october-24-27-2020-proceedings-matthew-e-taylor/
textbookfull.com
Children and Sustainable Development Ecological Education

in a Globalized World 1st Edition Antonio M. Battro
https://textbookfull.com/product/children-and-sustainable-development-
ecological-education-in-a-globalized-world-1st-edition-antonio-m-
battro/
textbookfull.com
Emerging Practices in Scholarship of Learning and Teaching
in a Digital Era 1st Edition Siu Cheung Kong
https://textbookfull.com/product/emerging-practices-in-scholarship-of-
learning-and-teaching-in-a-digital-era-1st-edition-siu-cheung-kong/
textbookfull.com
NASCLA Contractors Guide to Business, Law and Project

Management (13th Edition) National Association Of State
Contractors Licensing Agencies (Nascla)
https://textbookfull.com/product/nascla-contractors-guide-to-business-
law-and-project-management-13th-edition-national-association-of-state-
contractors-licensing-agencies-nascla/
textbookfull.com
Who Let a Demon Lord Into the Mage Tower A Fantasy Light
Novel 1st Edition Atlas Kane Kane Atlas
https://textbookfull.com/product/who-let-a-demon-lord-into-the-mage-
tower-a-fantasy-light-novel-1st-edition-atlas-kane-kane-atlas/
textbookfull.com
Sunnis And Shi'a: A Political History Of Discord 1st

Edition Laurence Louër
https://textbookfull.com/product/sunnis-and-shia-a-political-history-
of-discord-1st-edition-laurence-louer/
textbookfull.com
The History of Sociology in Britain: New Research and

Revaluation Plamena Panayotova
https://textbookfull.com/product/the-history-of-sociology-in-britain-
new-research-and-revaluation-plamena-panayotova/
textbookfull.com
Thermal and Rheological Measurement Techniques for
Nanomaterials Characterization. A volume in Micro and Nano
Technologies Sabu Thomas
https://textbookfull.com/product/thermal-and-rheological-measurement-
techniques-for-nanomaterials-characterization-a-volume-in-micro-and-
nano-technologies-sabu-thomas/
textbookfull.com
Ethics of Artificial Intelligence
Ethics of Artificial
Intelligence
E D I T E D B Y S . M AT T H EW L IAO
1
3
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press

198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2020
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by license, or under terms agreed with the appropriate reproduction
rights organization. Inquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

Names: Liao, S. Matthew, editor.
Title: Ethics of artificial intelligence / edited by S. Matthew Liao.
Description: New York, NY, United States of America : Oxford University
Publication, 2020. | Includes bibliographical references and index.
Identifiers: LCCN 2020004473 (print) | LCCN 2020004474 (ebook) |
ISBN 9780190905040 (paperback) | ISBN 9780190905033 (hardcover) |
ISBN 9780190905064 (epub)
Subjects: LCSH: Artificial intelligence—Moral and ethical aspects.
Classification: LCC Q334 .7 .E84 2020 (print) | LCC Q334 .7 (ebook) |
DDC 174/.90063—dc23
LC record available at https://lccn.loc.gov/2020004473
LC ebook record available at https://lccn.loc.gov/2020004474
1 3 5 7 9 8 6 4 2
Paperback printed by LSC Communications, United States of America
Hardback printed by Bridgeport National Bindery, Inc., United States of America
For Wibke, Caitlin, and Connor
Acknowledgments
This volume emerged in part from a conference in October 2016 on the ethics
of artificial intelligence at New York University hosted by the NYU Center for
Mind, Brain and Consciousness and the NYU Center for Bioethics. I would
like to thank my co-organizers, Ned Block and David Chalmers, for making the
conference such a tremendous success. Thanks are also due to Jonathan Simon,
Cassandra Coste, and Leigh Bond, who provided logistical and organization
support for all facets of the conference. Tom Carew, the former Dean of the
Faculty of Arts and Science at NYU, deserves special thanks for providing the
opening remarks at the conference. I would also like to express my appreciation
to all the speakers and panelists for their insightful talks and presentations: Peter
Asaro, John Basl, Nick Bostrom, Meia Chita-Tegmark, Kate Devlin, Vasant Dhar,
Virginia Dignum, Mara Garza, Daniel Kahneman, Adam Kolber, Yann LeCun,
Gary Marcus, Steve Petersen, Francesca Rossi, Stuart Russell, Ronald Sandler,
Jürgen Schmidhuber, Susan Schneider, Eric Schwitzgebel, Frans Svensson,
Jaan Tallinn, Max Tegmark, Wendell Wallach, Stephen Wolfram, and Eliezer
Yudkowsky.
I am particularly grateful to David Chalmers, who provided intellectual
support and guidance throughout this process. The thirty contributors to this
volume merit a special gratitude for their excellent intellectual work. I would
like to thank Peter Ohlin, my editor at Oxford University Press, for suggesting
the idea of producing this volume and for his enthusiasm and encouragement
throughout its production.
Colin Allen, David Chalmers, Mala Chatterjee, Felipe De Brigard, Sarah
Gokhale, Robin Hanson, Athmeya Jayaram, Ryan Jenkins, Andrew Lee, Robert
Long, Neil McArthur, Rune Nyrup, Huw Price, Duncan Purves, Jonathan Simon,
Daniel Viehoff, and Roman V. Yampolskiy read and provided astute comments
on various chapters in this book, for which I am very grateful. I have also been
helped greatly by Sarah Gokhale, Yulia Gamper, and Nicholas Tilmes, who pro-
vided invaluable research assistance.
Finally, I would like to thank my family, Wibke, Caitlin, and Connor, for their
love and support as I worked to complete this book.
Contributors
Peter Asaro is an Associate Professor and the Director of the Graduate Program in Media
Studies at The New School, and an Affiliate Scholar of Stanford Law School's Center
for Internet and Society. He is also the Co-founder and Vice-Chair of the International
Committee for Robot Arms Control, and a founding member of the Campaign to Stop
Killer Robots.
Jean-François Bonnefon is a Research Director at the French Centre National de la

Recherche Scientifique and is affiliated with the Toulouse School of Economics, the
Toulouse School of Management, and the Institute for Advanced Study in Toulouse. He is
the Moral AI Chair at the Artificial and Natural Intelligence Toulouse Institute.
Nick Bostrom is a Professor at Oxford University, where he leads the Future of

Humanity Institute as its founding director. He is the author of over 200 publications,
including Anthropic Bias, Global Catastrophic Risks, Human Enhancement, and
Superintelligence: Paths, Dangers, Strategies, a New York Times bestseller which helped
spark a global conversation about artificial intelligence.
Andrew Critch is a Research Scientist at the Center for Human-Compatible AI in the

Department of Electrical Engineering and Computer Sciences at the University of
California, Berkeley. He was a Research Fellow at the Machine Intelligence Research
Institute.
Allan Dafoe is an Associate Professor and Senior Research Fellow in the International
Politics of Artificial Intelligence at the University of Oxford. He is also the Director
of the Centre for the Governance of AI at University of Oxford’s Future of Humanity
Institute.
Kate Devlin is a Senior Lecturer (Associate Professor) in Social and Cultural Artificial
Intelligence at King's College London. She is the author of Turned On: Science, Sex and
Robots, which examines the ethical and social implications of technology and intimacy.
Carrick Flynn is a Research Fellow at Georgetown University’s Center for Security and
Emerging Technology focused on national security, technology law, and AI policy. He was
the founding Assistant Director of the Center for the Governance of AI at the University
of Oxford.
Mara Garza is a former philosophy PhD student at the University of California, Riverside
and is currently an artist.
xii List of Contributors
Hanna Gunn is an Assistant Professor in Cognitive and Information Sciences at the

University of California, Merced. Her research focuses on themes in social epistemology
including agency, community, new media, and the Internet.
Aaron James is a Professor of Philosophy at the University of California, Irvine. He is

the author of several books, including Fairness in Practice: A Social Contract for a Global
Economy.
Frances M. Kamm is the Henry Rutgers University Professor of Philosophy and

Distinguished Professor of Philosophy at Rutgers University. She is the author of nu-
merous works in normative ethical theory and applied ethics, including The Trolley
Problem Mysteries.
Patrick LaVictoire is a machine learning engineer, most recently for Lyft. After a postdoc-
toral position at the University of Wisconsin, he worked on AI alignment theory at the
Machine Intelligence Research Institute.
S. Matthew Liao is the Arthur Zitrin Chair of Bioethics, the Director of the Center for
Bioethics, and an Affiliated Professor in the Department of Philosophy at New York
University. He is the author or editor of numerous books including The Right to Be Loved
and Moral Brains: The Neuroscience of Morality.
Andrea Loreggia is a Research Associate at the European University Institute. His re-
search interests in artificial intelligence span from knowledge representation to deep
learning.
Nicholas Mattei is an Assistant Professor of Computer Science at Tulane University. His

research focuses on the theory and practice of artificial intelligence, machine learning,
data science, and the impact of these technologies on society.
Cathy O’Neil is the author of Weapons of Math Destruction: How Big Data Increases
Inequality and Threatens Democracy. She is the founder of O'Neil Risk Consulting &
Algorithmic Auditing.
Steve Petersen is an Associate Professor of Philosophy at Niagara University. In addition

to the ethics of AI, he works on algorithmic approaches to epistemology.
Iyad Rahwan is a Director of the Max Planck Institute for Human Development, where he
founded and directs the Center for Humans & Machines. He is also an Associate Professor
of Media Arts & Sciences at the Massachusetts Institute of Technology
Peter Railton is the Gregory S. Kavka Distinguished University Professor and John
Stephenson Perrin Professor of Philosophy at the University of Michigan, where he
specializes in ethics and philosophy of science. He is also a Fellow of the American
Academy of Arts and Sciences.
Francesca Rossi is based at the T.J. Watson IBM Research Lab in New York. Before joining
IBM, she was a Professor of Computer Science at the University of Padova.
List of Contributors xiii
Stuart Russell is a Professor of Computer Science at the University of California, Berkeley,

an Honorary Fellow of Wadham College, Oxford, and an Andrew Carnegie Fellow. He
co-authored (with Peter Norvig) Artificial Intelligence: A Modern Approach, the standard
text in the field. He has been active in the arms control community for nuclear and auton-
omous weapons.
Susan Schneider is the NASA/Baruch Blumberg Chair of Astrobiology and Technological

Innovation at NASA and the Library of Congress and the Director of the AI, Mind and
Society Group at the University of Connecticut. She writes about the nature of the self and
mind, from the perspectives of philosophy, AI, cognitive science and astrobiology.
Eric Schwitzgebel is a Professor of Philosophy at University of California, Riverside, and

a member of its program in Speculative Fiction and Cultures of Science. His most recent
book is A Theory of Jerks and Other Philosophical Misadventures.
Azim Shariff is an Associate Professor of Psychology at the University of British

Columbia where he directs the Centre for Applied Moral Psychology. His research
addresses ethical issues related to religion, free will, economics, and technological trends.
Jessica Taylor is a technical philosopher who has researched topics including logical prob-
ability, decision theory, ontology, and distributed consensus algorithms. She works at the
Median Group, and she was a Research Fellow at the Machine Intelligence Research Institute.
Shannon Vallor is the Baillie Gifford Chair in the Ethics of Data and Artificial Intelligence
at the Edinburgh Futures Institute at the University of Edinburgh, where she is also ap-
pointed in Philosophy. She is the author of Technology and the Virtues: A Philosophical
Guide to a Future Worth Wanting.
K. Brent Venable is a Senior Research Scientist at the Florida Institute of Human and
Machine Cognition in joint appointment with University of West Florida, where she is a
Professor of Computer Science. She has co-authored two books and published over 100
academic papers.
Wendell Wallach chaired the Technology and Ethics Research Group for the past eleven
years at Yale University's Interdisciplinary Center for Bioethics. His latest book, a primer
on emerging technologies, is entitled, A Dangerous Master: How to Keep Technology
from Slipping Beyond Our Control, and he co-authored (with Colin Allen) Moral
Machines: Teaching Robots Right from Wrong.
Stephen Wolfram is the creator of Mathematica, Wolfram|Alpha and the Wolfram

Language; the author of A New Kind of Science; and the founder and CEO of Wolfram
Research. Over the course of nearly four decades, he has pioneered the development and
application of computational thinking and has been responsible for many discoveries,
inventions, and innovations in science, technology and business.
Eliezer Yudkowsky has been a Research Fellow at the Machine Intelligence Research
Institute since 2000, working on decision theory and on AI alignment.
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
A Short Introduction to the Ethics
of Artificial Intelligence
S. Matthew Liao
I.1. Overview
Artificial intelligence (AI) is progressing rapidly. AI can now recognize objects in

images and videos; transcribe speech; translate between languages; beat humans
at Jeopardy,1 at Go,2 and at poker;3 paint in the style of van Gogh;4 write Beatles-
like music;5 help prepare legal documents; trade stocks; drive cars; fly drones;
write its own encryption language;6 identify cancer in tissues;7 and solve the
quantum state of many particles at once.8 In the coming years, it is expected that
AI will reach and exceed human performance on many more, and increasingly
complex, tasks.
As AI technologies continue to advance, questions about the ethics of AI be-
come more pressing than ever. Complex ethical issues surround current and
near-future AI systems. For instance, leading technology companies are building
self-driving cars, which promise to increase personal mobility for elderly and
disabled people and to save lives by reducing driver error; however, in an emer-
gency, should a self-driving car prioritize the lives of the passengers or the lives
of pedestrians?9
Many countries are developing autonomous weapon systems capable of iden-
tifying and attacking a target without human intervention. Autonomous weapon
systems offer the potential to decrease risks to military personnel and civilians
by being better than stressed-out soldiers in the heat of a battle at distinguishing
civilians from combatants and at making reasonable trade-offs between military
gains and risk or harm to civilians. But should we as a society give machines this
kind of moral decision-making power? Who is at fault if an autonomous weapon
system attacks a hospital or a school? Does the exercise of deadly force always re-
quire “meaningful human control” to be legitimate?10
AI has been found to be particularly useful for revealing otherwise unrec-
ognizable patterns in complex processes. As a result, police departments are
investigating ways to use AI to identify likely criminals among the general pop-
ulation.11 Likewise, some judges and prison officials are interested in using AI to
S. Matthew Liao, A Short Introduction to the Ethics of Artificial Intelligence In: Ethics of Artificial Intelligence. Edited by:
S. Matthew Liao, Oxford University Press (2020). © Oxford University Press.
DOI: 10.1093/oso/9780190905040.003.0001
2 A Short Introduction to the Ethics of AI
develop “risk prediction” tools to assist with decisions on criminal sentencing,

bail hearings, and parole.12 However, among other things, machine learning
requires good data because incomplete or unrepresentative data can exacerbate
problems of bias.13 At the same time, advanced AI systems tend to process large
volumes of data, and their inner workings tend not to be transparent. Given this,
how can we create AI systems that are fair and that do not inadvertently produce
biased results?
At the same time, Japan, Italy, and some other countries are looking into using
robots to care for and provide companionship to their elderly population.14 Many
companies are racing to build sex robots with sophisticated AI.15 What ethical
issues are raised by companion robots and sex robots? Will humans be able to
marry robots?16 How will this impact our relationships with other humans?
It is estimated that 47% of American jobs could be lost to AI and automation
in the next twenty years.17 How should we help people adjust to this level of un-
employment? Should taxes be levied on robots that replace human workers, as
Bill Gates has suggested?18 Should we offer everyone a universal basic income?19
Current AI is what is known as narrow AI20 because it is designed to per-
form a narrowly defined task such as driving a car or identifying a hostile
target. In the long term, a number of AI researchers hope to create artificial
general intelligence (AGI), which would be capable of performing any intellec-
tual task that a human being can.21 On one understanding, such AI, sometimes
referred to as strong AI, would be capable of actual thought and reasoning and
would possess sentience and consciousness.22 Some writers speculate that once
a sufficiently intelligent AI is developed, it could develop even more intelligent
systems, which in turn could develop systems with even greater intelligence,
resulting in an “intelligence explosion” or “singularity,” whereby superintelli-
gent machines would come to possess capacities that greatly exceed human
capacities.23
The prospect of superintelligent AIs raises at least two kinds of ethical issues.
One issue pertains to the impact of these AIs on humans. In particular, can we
prevent them from causing harm to humans or even human extinction? To illus-
trate, consider an example from Nick Bostrom in which a machine designed to
make as many paper clips as possible becomes incredibly intelligent.24 Given its
goal, the machine could decide to convert everything in the universe, including
humans, to paper clips. Is it possible to shape the development of AI now and
align the eventual values and goals of superintelligent AIs with those of humans
so that these AIs will not end up harming or destroying humanity?25 A second
issue is how we should treat these AIs. Would such AIs be conscious? Would they
have the moral status that humans have, that is, would they be rightsholders?26
What kind of rights and responsibilities would they have?27 Could they have
moral status greater than that of humans?
A Short Introduction to the Ethics of AI 3
In recent years, the ethical implications of near-term and long-term AI have

received considerable attention in both the popular media and academia.28 In
this introduction, I aim to outline some of the key issues in the study of the ethics
of AI, identify some of the core claims that have been made, and propose some
ways of taking these discussions further. With respect to near-term AI, it will
be useful to distinguish between (a) ethical issues that arise because of limita-
tions to current machine learning systems, what might be called “vulnerabilities
in machine learning,” and (b) ethical issues that arise because current machine
learning systems may be working too well and humans can be vulnerable in the
presence of or when interacting with these intelligent systems, what might be
called “human vulnerabilities.” The chapters in this volume will then continue
this discussion by offering a variety of new perspectives on the ethics of AI.
I.2. Key Concepts in Machine Learning
To begin, let me say something about what AI is.29 There is no agreed-upon

definition of AI. In 1956 at a conference in Dartmouth, John McCarthy coined
the term “artificial intelligence” and defined it as “the science and engineering
of making intelligent machines.”30 Stuart Russell and Peter Norvig, authors of
one of the most popular textbooks on AI, propose that there are four ways of
defining AI: as (a) acting humanly, (b) thinking humanly, (c) thinking ration-
ally, and (d) acting rationally. They are interested in AI concerned with (d), that
is, rational action.31 For our purpose, we can broadly understand AI as getting
machines to do things that require cognitive functions such as thinking, learning,
and problem-solving when done in intelligent beings such as humans.
On this understanding, AI can take different forms.32 One form is symbolic
AI, or good-old-fashioned artificial intelligence (GOFAI), which dominated
the field of AI research from the 1950s to the 1980s.33 Symbolic AI attempts to
represent cognitive functions such as thinking, learning, and problem-solving
through symbolic reasoning and logic. In particular, these systems use a series
of explicitly programmed if-then rules and statements to establish the relations
between inputs and outputs. Examples of symbolic AI include rules engines such
as expert systems (where the rule set is a representation of an expert’s know-
ledge) and knowledge graphs (where a database stores information in a graphical
format). A limitation of symbolic AI is that it is difficult to revise the rules once
they are encoded into such a system.
Another form of AI is machine learning, which uses algorithms to learn from
data without being explicitly programmed. Within machine learning, one can
distinguish between supervised learning, unsupervised learning, and reinforce-
ment learning. In supervised learning, an algorithm aims to learn a function that
best approximates the relationship between input and output in the data. To do
so, the algorithm is trained on a training data set in which the correct answers for
certain data are known and the data are labeled accordingly. In this way, the al-
gorithm can use the labeled information to learn the relationship between inputs
and outputs. Once the algorithm is properly trained, it is then able to apply what
it has learned to predict the correct answer in different (target) data sets.
There are two main types of supervised learning algorithms: classification and re-
gression.34 In classification algorithms, the desired output is a discrete label with a
finite set of possible outcomes. Cases of binary classification have only two possible
outcomes; for example, either something is a car or it is not a car. Cases of multilabel
classification have more than two possible outcomes, for instance, text categorization
of news articles.35 In regression algorithms, the desired outputs are continuous.36
To give an example of supervised learning, suppose that there is a box
containing images of cars and motorcycles. A classification algorithm can be
trained by repeatedly manually telling the algorithm which images depict a
car and which images depict a motorcycle. Over time, a properly trained algo-
rithm will learn the features that are more likely to make something a car rather
than a motorcycle, and vice versa. After that, suppose that there is another box
containing new images of cars and motorcycles. The properly trained algorithm
should be able to distinguish images of cars from those of motorcycles on its own.
In unsupervised learning, a given data set is not labeled, and the algorithm
aims to sort the data on its own. One type of unsupervised learning algorithm
is clustering, where the algorithm aims to group data that are more similar to
each other than they are to data in other groups.37 Another type of unsupervised
learning is association rule learning, where the algorithm tries to discover rules
that describe large portions of the data. To illustrate unsupervised learning, sup-
pose that there is a box with images of cars and motorcycles that have not been
labeled or sorted. An unsupervised learning algorithm would attempt to sort and
categorize these images based on their similarities and differences.
In reinforcement learning, the algorithm attempts to learn through expe-
rience.38 A reinforcement learning algorithm learns by being rewarded if it
succeeds in a task and/or punished if it fails. Through trial and error, the algo-
rithm strives to maximize the long-term reward.
Currently, the driver for many breakthroughs in AI is deep learning. Deep
learning uses artificial neural networks, which loosely emulate the activity of
neurons in the brain. Simpler versions of such networks existed as far back as
the 1950s,39 but were not taken seriously by the mainstream AI community for
decades. Recently, advances in computing power and data storage, coupled with
the development of more sophisticated neural networks by researchers such as
Geoffrey Hinton,40 Yann LeCun,41 and Yoshua Bengio,42 have unleashed deep
learning’s potential.43
A deep neural network typically has an input layer, an output layer that acts
as the final decision-maker, and multiple hidden layers of nodes in between. The
“deep” in “deep learning” refers to the number of hidden layers of nodes in a
neural network. Each layer is responsible for performing a specific function that
contributes to solving the problem at hand. Each node has weights associated
with each of its connections. Adjusting the weights on these connections causes a
node to produce certain outputs, which are then fed through the following layers.
Deep learning uses these layers of nodes to detect increasingly abstract
features of a data set that capture the most information while minimizing losses
in accuracy. To do this, it uses the method of back-propagation,44 which takes the
error between the expected result and the actual result of a neural network and
adjusts the neural network’s weights in the direction of less error. In doing so, the
entire network progressively gets better at predicting the correct answer.
As an example, consider again a classification task involving images of cars
and motorcycles. The input layer of a deep neural network might take in external
data such as pixels in an image. It would then feed this information forward to
some or all of the connections in the next layer. Each connection in that layer
would integrate the inputs from the first layer and pass the results forward to
subsequent layers. Eventually the output layer would offer a prediction, for ex-
ample, that an image is a car or a motorcycle.
Different types of neural networks are used for different purposes. For ex-
ample, recurrent neural networks (RNNs),45 which have a sort of built-in
memory, are particularly useful for solving temporal data problems such as
predicting sequences of words. Convolution neural networks (CNNs),46 which
are good at recognizing patterns across space, are particularly useful for image
classifications and computer vision tasks. Deep learning can be combined with
other machine learning methods such as reinforcement learning to create deep
reinforcement learning algorithms. A deep reinforcement learning algorithm
was used to beat the world champion Lee Sedol at the game of Go.47
I.3. Vulnerabilities in Machine Learning
As impressive as machine learning is, it also suffers certain limitations. As we

shall now see, these limitations can give rise to a host of ethical issues.
I.3.1. Machine Learning Is Data Hungry
First, machine learning needs a lot of data to work well. For example, supervised
learning algorithms can fine-tune themselves and achieve great predictive power
when they have access to a vast amount of data. Consequently, this incentivizes
companies and organizations to harvest or buy data, including sensitive, per-
sonal data, even when doing so might involve violating an individual’s right to
privacy. For example, Cambridge Analytica collected millions of Facebook users’
data without their knowledge for political purposes.48 A borderline case might
be when the drug maker GlaxoSmithKline bought the exclusive rights to mine
the genetic data of customers of the DNA testing service 23andMe for drug
discovery.49
I.3.2. Garbage In/Garbage Out
Second, machine learning is only as good as the data from which it learns. If a
machine learning algorithm is trained on inadequate or inaccurate data, then the
algorithm will make bad predictions even if it is itself well designed. For instance,
in 2015 Google Photo autolabeled Jacky Alcine, a black software developer, and
his friend as “gorillas”50 because, in all likelihood, the data used to train the algo-
rithm did not include sufficient images of people from different racial and ethnic
backgrounds.51
I.3.3. Faulty Algorithms
Third, even if a machine learning algorithm receives adequate and accurate data,
if the algorithm itself is bad, it will also make bad predictions. For instance, a
bad machine learning algorithm may identify a pattern even if there isn’t one,
a problem known as “overfitting,”52 or may fail to identify a pattern even when
there is one, a problem known as “underfitting.”53 A machine learning algorithm
may also give too much or too little weight to certain features or fail to include
certain relevant features altogether. Faulty algorithms can have serious ethical
implications. For example, in 2016 the Arkansas Department of Human Services
began to use an algorithmic tool developed by interRAI to determine how many
hours of home care some people with disabilities should receive.54 The depart-
ment implemented the algorithm’s recommendation to reduce drastically the
number of home care hours for many beneficiaries, which caused several people
to be hospitalized. After a lawsuit was filed, an investigation revealed that, among
other things, the algorithm had incorrectly coded conditions such as cerebral
palsy and had not accounted for conditions such as diabetes, which led it to rec-
ommend reduced home care hours for hundreds of people. Ultimately a judge
ruled that the department had insufficiently implemented the interRAI algo-
rithm and ordered that its use be terminated.
I.3.4. Deep Learning Is a Black Box
Fourth, deep learning is a black box that raises issues such as interpreta-
bility, explainability, and trust.55 Deep learning is impenetrable even to its
programmers because it typically employs thousands or millions of connections
that interact with one another in complex ways. As a result, it is difficult to inter-
pret how these connections are interacting with each other and why they make
certain predictions. The issue of explainability arises because humans often need
to know how a decision is reached. However, deep learning announces its predic-
tion without explaining (in human terms) how it arrived at that prediction. To
see why this could be a problem, consider the following example. Suppose that a
deep learning algorithm predicts that there is a 74% chance that Kay will commit
another crime in the future, and a judge denies Kay parole on this basis. The
deep learning algorithm does not, for example, say, “There is a 74% chance that
Kay will commit a crime in the future because she has committed such crimes
in the past, and the last time Kay was on parole, she recommitted a crime within
two weeks.” Without such an explanation, the judge would not be able to ex-
plain, and justify, to Kay why she was denied parole. Beyond explainability, this
also raises the issue of trust in the deep learning system, since we do not know
whether it makes its predictions on reasonable and reliable grounds. For high-
stakes decisions such as those concerning parole, not being able to trust the deep
learning system is especially problematic.
Are there ways to address or mitigate deep learning’s black box problem?
Some AI researchers are currently exploring technical fixes such as “interpret-
able machine learning.”56 One interpretable machine learning method involves
adding an additional layer to deep learning models after the hidden layer(s) of
nodes and before the output.57 The added layer would provide information such
as which features were the most important for arriving at a particular prediction,
which features could have had an even greater impact on the prediction, how
each feature in the data bears on a particular prediction, and how each feature
would affect different possible predictions.58 The hope is that this information
would make the deep learning system more interpretable.
While interpretable machine learning is a promising idea, there are reasons to
question whether it can alleviate the problems of interpretability, explainability,
and trust in deep learning. One concern is that, since it is placed outside the
black box, the additional layer provides a post hoc explanation of the black box
after the deep learning system has already made its predictions. One might
wonder whether such post hoc explanations can identify the actual reasons
why the deep learning system gave the predictions that it did. The concern can
be put in the form of a dilemma: either the predictions are based on these post
hoc explanations or they are not. If the predictions are not based on these post
hoc explanations, what is their value? They might just be a post hoc rationaliza-
tion that does not correspond to how the black box arrived at its predictions. Of
course, post hoc explanations could be useful in some contexts even if the orig-
inal process was not based on them, as long as they adequately captured some
aspects of the process.59 However, given the black box nature of deep learning, it
is unclear how we could know whether these post hoc explanations (adequately)
captured some of what was going on in the process. Suppose instead that the
predictions are based on the post hoc explanations. If so, it should be possible to
design a new model using just these post hoc explanations. But if this is the case,
it implies that the black box is dispensable. Indeed it suggests a way to test the
value of these interpretable machine learning systems. If the black box remained
indispensable for making predictions, this would seem to suggest that the post
hoc explanations do not completely explain why a black box gave the predictions
that it did.
Other people have attempted to address the black box problem by arguing
that the importance of interpretability and explainability may be overstated.60
According to this line of thought, there is a trade-off between accuracy and
explainability in deep learning. If a deep learning system can make accu-
rate predictions, so the thought goes, it may not matter in certain cases if it
is not interpretable and explainable. For instance, consider medicine. One
would think that interpretability and explainability are especially important
in medicine, given its high-stakes nature. However, some argue that clinicians
often prescribe medications without fully understanding why they work.61
For example, clinicians frequently prescribe aspirin as an analgesic and lithium
as a mood stabilizer despite persistent uncertainty about the mechanisms
through which they work.62
There are reasons to be skeptical of this particular argument. While it may be
correct that we do not fully understand how some medications work, we do have
some ideas regarding the causal mechanisms through which they work. For in-
stance, people knew that something from willow trees causes fevers and pain to
be reduced, even if they did not know about salicylic acid, an active ingredient
in aspirin.63 This is distinct from a deep learning system that works through
associations and is, at least for now, unable to track causal relations. Likewise,
it is true that for a long time we did not know exactly how lithium stabilized an
individual’s mood. (The current hypothesis is that lithium moderates glutamate
levels in the brain.)64 Still, we know that lithium causes moods to be stabilized.
Again, we cannot say the same about a deep learning system that cannot track
causal relations.65
It might be asked why it matters whether a deep learning system can or cannot
track causal relations. To answer this question, it is worth noting that deep
learning is vulnerable to certain kinds of adversarial attacks, which are inputs
that are designed to cause a machine learning model to make a mistake.66 For in-
stance, deep neural networks are vulnerable to the so-called one-pixel attacks.67
In one study, by changing just one pixel in an image, researchers were able to get
a deep learning algorithm to classify an image of a car as a dog. The researchers
found that one-pixel attacks are successful on nearly three-quarters of standard
training images and that altering more pixels made this type of attack even more
effective.68 In another study, researchers modified 0.04% of the pixels in an
image, that is, about four hundred pixels out of a million. These changes were
imperceptible to the human eye. Nevertheless the deep neural network classi-
fied a panda as a gibbon “with 99.3% confidence.”69 Recently researchers found
that unmodified real-world images can also be used in adversarial attacks.70
The fact that deep learning networks are vulnerable to these types of attacks
suggests that these networks are not learning “real” features of the world such as
causal relations or what a macro-level object like a panda really is; instead these
deep learning networks are learning only superficial features. For our purpose,
if a deep learning network can be tricked in these ways, issues of interpreta-
bility, explainability, and trust remain highly relevant, especially in high-stakes
domains such as medicine and law where human beings could be harmed.
I.3.5. Machine Learning Is Weak AI
Fifth, in addition to being narrow AI, current machine learning systems are
also weak AI in that they do not have self-awareness or consciousness and they
cannot think for themselves.71 As we have seen, they lack understanding of “real”
features of the world such as causal relations. Importantly, they also lack a moral
sense, that is, the capacity to assess and determine whether an action is right or
wrong. Yet machine learning systems are being deployed in situations in which
they may have to make moral decisions without human oversight. Can morality
and ethical decision-making be built into such AI? If so, how? A whole literature
called “machine ethics” or “machine morality” is devoted to addressing these
questions.72 Owing to space, I will not attempt to give a detailed overview of that
literature, but here are some main takeaways.
Some people have proposed designing machines that behave ethically by
building moral rules and principles into them using overarching ethical the-
ories such as deontology and utilitarianism. For instance, one might create a
utilitarian machine that would aim to secure the greatest good for the greatest
number. Or one might create a deontological machine that would follow Kant’s
categorical imperative.73 A problem with this top-down approach to machine
ethics is that it can be difficult to know when to apply a moral rule or principle
and when the rule or principle has been satisfied. For instance, it is a well-known
problem for utilitarianism that it is difficult to calculate when acting would se-
cure the greatest good for the greatest number. This calculation problem would
also apply to machine learning systems under consideration. A more pressing
issue is that not everyone believes that deontology and utilitarianism are correct
moral theories. Given this, there are questions about whether it is appropriate to
build such moral theories into systems that are likely to affect everyone.
Others have proposed building virtue ethics into machines.74 According to
this line of thought, one should model a machine on what a virtuous agent would
do in a particular situation. A concern about this approach is that it is not alto-
gether clear who is a virtuous agent and what a virtuous agent would do in a par-
ticular situation. For instance, would Mother Teresa be a virtuous agent? Suppose
that she is. How would one know what Mother Teresa would do in a particular
situation? Also, some people may not regard Mother Teresa as a moral exem-
plar. Should we instead let each person decide who the virtuous agent should be?
What if someone’s idea of a virtuous agent is a racist, malevolent dictator?
Instead of overarching theories, perhaps we could use case-driven approaches
from moral philosophy to build ethics into machines. Since the famous trolley
dilemmas are often mentioned in this context, it is worthwhile providing some
background on them. Philosophers initially used trolley dilemmas to question
utilitarianism.75 Consider the following two cases.
Sidetrack: A runaway trolley is headed toward five people who will be killed.
You can hit a switch that will turn the trolley onto a sidetrack where an-
other person sits, thereby killing him instead of the five.
Footbridge: As before, a runaway trolley is threatening to kill five people. You
are standing next to a large man on a footbridge and you can push the large
man off the bridge. The large man will die, but his body will stop the trolley
from reaching the five people, thereby saving the five.
In both cases, the choice is between killing one person and letting five others die.
It seems that utilitarianism would say that the actions in both cases are morally
on a par and that we should kill the one and save the five. However, many people
believe that while it is morally permissible to hit the switch in Sidetrack and kill
the one, it is impermissible to push the large man in Footbridge to save the five.
Suppose that people’s judgments about these cases are correct. This would seem
to call utilitarianism into question.
It might be asked, how does one explain the difference in people’s judgments
about Sidetrack and Footbridge, supposing that there is a moral difference between
these two cases? According to Judith Jarvis Thomson, who coined the term, the
“Trolley Problem” is the problem of explaining why our judgments differ between
these two cases.76 One explanation for the Trolley Problem appeals to the Doctrine
Visit https://textbookfull.com
now to explore a rich
collection of eBooks, textbook
and enjoy exciting offers!
of Double Effect (DDE), which relies on a distinction between intending harm and
merely foreseeing harm. According to one interpretation of the DDE, there is a
moral constraint on acting with the intention of doing harm, even when the harm
is used as a means to a greater good.77 However, it is permissible to act with the
intention of employing neutral or good means to promote a greater good, even
though one foresees the same harmful side effects, if (a) the good is proportionate
to the harm, and (b) there is no better way to achieve this good. Using the DDE,
one can explain the permissibility of hitting the switch in Sidetrack on the ground
that you merely foresee the innocent bystander’s death but you do not intend him
to be hit as a means to saving the five. In contrast, in Footbridge, because it seems
that you intend the innocent bystander to be hit by the trolley as a means to stop-
ping the trolley from hitting the five, it is not permissible for you to push him off the
footbridge. The DDE is not without its critics.78 Other philosophers have offered
alternative explanations of the Trolley Problem.79
For our purpose, it might be thought that one could use trolley-like scenarios
to figure out people’s judgments about various cases.80 One might then be able
to program people’s judgments about those cases into a machine. There is some-
thing to this idea. Consider the following:
Empty Track: A runaway trolley is headed toward five people who will be
killed. You can hit a switch that will turn the trolley onto an empty side-
track where there is no one.
In this case, there should be no disagreement that one should hit the switch and
turn the trolley toward the empty track. If so, perhaps one could use clear cases
such as Empty Track to create some clear ethical boundaries for machines.
Still, this approach also has problems. For one thing, not every case will be
as easy as Empty Track. In more difficult cases, people’s judgments are likely to
differ. Indeed, even with respect to the original trolley dilemma, some people
believe that it is impermissible to hit the switch in Sidetrack, while others be-
lieve that it is permissible to push the large man off the bridge in Footbridge.81
Of course, we could still think that there is a fact of the matter with respect to
these cases even though they are difficult. History is replete with examples such
as slavery where there were strong disagreements that have been resolved. Even
so, at the very least, more philosophical work will need to be done before we can
program a machine to act one way or another with respect to these difficult cases.
In addition, there are many difficult real-world cases that would need to be re-
solved before one could build these cases into a machine. Hence, even if we could
use the case-based approach to program some easier cases into machines, we are
a long way from giving them any sort of complete ethical decision procedure.
Perhaps recognizing that this is a limitation of current AIs, companies typi-

cally strive for the less ambitious goal of making sure that their AI systems are
safe, akin to having safety protocols for aircrafts. But even meeting this aim can
be challenging. Consider an example concerning self-driving cars. To ensure
that self-driving cars are safe for everyone involved, manufacturers have to con-
sider, among other things, how the cars should behave when accidents are about
to occur. For instance, suppose that a self-driving car’s brakes malfunction and
the car can either hit five people on the road or swerve off the road and thereby
kill the passenger. Suppose that it is not the passenger’s fault that the brakes
malfunctioned. What should the self-driving car do? Some manufacturers
have suggested that the self-driving car should prioritize the passenger in such
circumstances.82 This means that the car should hit the five people instead of
endangering the passenger. There is some plausibility to this line of thinking. In
ethics, it is commonly held that under ordinary circumstances, we do not have to
sacrifice our lives in order to prevent others from being harmed, especially if we
are not at fault. Also, practically speaking, most people are unlikely to purchase
a car that would sacrifice them for the greater good. Moreover, always priori-
tizing the passenger seems like a protocol that could be technically implemented
in machines in a straightforward manner. The problem, though, is that it is not
clear that we should always prioritize the passenger in an accident. Consider the
following:
Child: A runaway self-driving car is headed toward a child who will be killed.
The self-driving car can swerve slightly to avoid hitting the child. Swerving
the car slightly to avoid hitting the child has a low (but not zero) chance of
harming the passenger in the car.
In this case, if the passenger should have absolute priority, then the car should
not swerve, since swerving introduces some risk to the passenger. Yet it seems
that the car should swerve in this case because the risk to the passenger is low and
the benefit of not hitting the child is great. This case suggests that it will not be
straightforward to devise and implement protocols that would ensure the safety
of everyone involved.
Here are some incremental proposals that might help to make self-driving cars
safer. First, whatever safety protocols we end up adopting for self-driving cars, it
seems that the testing of self-driving cars in the “wild,” that is, in actual streets,
should be much more regulated. Just as pharmaceutical companies are not per-
mitted to test new drugs by giving them randomly to people on the street, per-
haps some oversight body should determine when and where self-driving cars
can be tested in consultation with members of the community. Second, given
that self-driving cars are weak AIs, and given the difficulty of anticipating every
possible scenario on the road, at least for now we should consider having dedi-
cated lanes for self-driving cars.83 This would free self-driving cars from having
to cope with the unpredictability of human driving, among other things. Third,
we might consider equipping self-driving cars with devices that would enable
them to communicate with each other so that they could coordinate their actions
and reduce the number of accidents. Of course, we would also need to protect
such devices from hackers and ensure that they would not undermine users’
privacy.
I.4. Human Vulnerabilities
In the previous section, we discussed ethical issues that can arise because current
machine learning systems are limited in certain ways. In this section, we shall
consider ethical issues that can arise because current machine learning systems
may be working too well and humans can be vulnerable in their presence. I shall
give four examples of such human vulnerabilities, although there are certainly
others. I shall also consider how we should address these vulnerabilities.
First, facial recognition technologies can already detect faces in a crowd with
great accuracy.84 In the near future these technologies will likely be able to track
constantly anyone who enters a public space at any hour of the day. On the pos-
itive side, these technologies can help police find criminals more quickly and
identify missing or kidnapped children. On the negative side, a government
could use this technology to monitor its citizens or to profile and discriminate
against minorities. For instance, a controversial study from Stanford University
allegedly found that a machine learning system could correctly distinguish be-
tween gay and straight sexual orientation 81% of the time for men and 74% of
the time for women just by examining photos of their faces.85 A government
that criminalizes homosexuality could use such facial recognition technology to
identify and discriminate against homosexuals. For our purpose, this is an ex-
ample where machine learning may be working too well and ethical issues arise
because people may be tempted to use it for ill.
Second, we are on the cusp of being able to use machine learning to fabri-
cate videos so realistic that humans cannot tell that they are fake. These so-called
“deepfakes” use generative adversarial networks to produce new types of data
out of existing data sets.86 One can use this technique to create videos of a person
saying or doing things that he or she has never said or done. For instance, the
director Jordan Peele and his brother-in-law, BuzzFeed CEO Jonah Peretti, used
this technology to produce a video in which President Barack Obama declared
that the villain Killmonger in the film Black Panther was “right” about his plan
for world domination.87 While amateur hobbyists might use deepfakes to make
people appear to say or do funny things for entertainment, bad actors could use
these digital forgeries to conduct smear campaigns against politicians or private
citizens and to spread fake news that erodes trust in our institutions.88 It should
be noted that these deepfakes are not yet good enough to fool us completely.
However, there is evidence that digital forgeries need not be very convincing for
them to be believed and cause significant damage. For instance, someone manip-
ulated a video of the current U.S. House speaker, Nancy Pelosi, by slowing the
speed of the video to 75%.89 This was enough for the video to go viral and for
people to accuse Pelosi of slurring her words. In any case, deepfakes will likely
advance to a point where it will be difficult for us to detect whether or not they
are fake. For our purpose, deepfakes serve as another example where machine
learning is working too well and can be used to exploit our tendency to believe
what we see, which is reasonable in ordinary contexts but less so as deepfakes
proliferate.
Third, given that robots can perform certain tasks better and faster than
humans, do not need sleep, can be duplicated and replaced, and so on, many
people are worried that robots and automation will replace a significant portion
of current human labor in the near future. A study from McKinsey suggested
that by 2030, 30% of human labor could be replaced by automation.90 Another
study from Oxford University predicted that 47% of jobs in the United States will
be under threat from intelligent machines in the next two decades.91 Yet some
people believe that while automation will make some jobs obsolete, it will also
create new ones.92 After all, there were similar concerns during the Industrial
Revolution about machines taking over human jobs, but as it turned out, while
the spinning jenny and the steam engine did displace some workers, they also
created many new jobs in textiles and manufacturing. Also, automation is likely
to replace tasks that are more repetitive and undesirable, which means that at
least in the short term, jobs that are more creative will still require humans.93
Nevertheless it seems certain that some people will lose their jobs as a result of
increased automation and that a subset of these people will not be able to transi-
tion to new jobs.
Fourth, as robots become more and more sophisticated, some people have
begun to regard them as companions. For instance, in 2018 Akihiko Kondo, a
Japanese school administrator, married the hologram of the popular anime
character Hatsune Miku.94 In Love and Sex with Robots, David Levy predicts
that some people will come to prefer robot companions over humans in the fu-
ture.95 Similarly companies are developing robot caregivers for the elderly that
can bring drinks to them, remind them to take medication, and play games with
them.96 There is also evidence that some elderly people are becoming attached
to their robots.97 However sophisticated, at least given the current state of ma-
chine learning, these robots are still weak AI and are not full agents. As such,
among other things, there is a concern that these are not genuine, reciprocal
relationships.
How should we address issues that arise because machine learning systems
are working too well and humans are vulnerable in their presence? I would like
to suggest that we adopt a theoretical human rights framework. To see why, let
me first briefly say what human rights are and offer a theoretical account of what
human rights we have. I shall then suggest that a theoretical framework enables
us to explain why certain rights claims are indeed genuine human rights.
Human rights, as A. J. Simmons states, are “rights possessed by all human
beings (at all times and in all places), simply in virtue of their humanity.”98 But
which features of humanity ground human rights? Elsewhere I have defended
what I call a Fundamental Conditions Approach to human rights, which says that
human rights protect the fundamental conditions for pursuing a good life.99 All
too briefly, the fundamental conditions are various goods, capacities, and options
that human beings qua human beings need, whatever else they qua individuals
might need, in order to pursue certain basic activities. Some basic activities in-
clude deep personal relationships with one’s partner, friends, parents, children;
knowledge of the workings of the world, of oneself, of others; active pleasures
such as creative work and play; and passive pleasures such as appreciating beauty.
The fundamental goods are resources that human beings qua human beings
need in order to sustain themselves corporeally, including food, water, and air.
The fundamental capacities are powers and abilities that human beings qua
human beings require in order to pursue the basic activities. These capacities
include the capacity to think, to be motivated by facts, to know, to choose an act
freely (liberty), to appreciate the worth of something, to develop interpersonal
relationships, and to have control of the direction of one’s life (autonomy). The
fundamental options are those social forms and institutions that human beings
qua human beings require if they are to be able to exercise their essential capaci-
ties to engage in the basic activities. These social forms and institutions include
the options to have social interaction, to acquire further knowledge, to evaluate
and appreciate things, and to determine the direction of one’s life.
The Fundamental Conditions Approach can explain why many of the rights
in the Universal Declaration of Human Rights are genuine human rights. For
instance, consider the right to life, liberty, and security of person (Article 3).
Whatever else human beings (qua individuals) need, they (qua human beings)
need life, liberty, and security of person in order to pursue the basic activities.
If they are not alive, if they cannot freely choose to act to some degree, or if the
security of their person is not guaranteed, then they cannot pursue the basic ac-
tivities. Given this, on the Fundamental Conditions Approach, human beings
would have human rights to life, liberty, and security of person. Or consider
the right to freedom of thought, conscience, and religion (Article 18), the right
Random documents with unrelated
content Scribd suggests to you:
people, and—yes, it had better be said,—a slight tendency to the
habit of officious lying—(the cardinal went purple)—there, it is said
and done with: you have had your lesson, and you know better now:
—after those things, the only reason why your episcopate has not
been a very brilliant one is that you started with the false idea of the
necessity of speaking that corrupt and obsolete dialect."
"But does not Your Holiness think that a foreigner——"
"No: England is the dominant race: her language is the language of
all her colonies. Why a triplet of little conquered countries should
refuse to learn English—should be permitted to insist on their
barbarous and unliterary languages, We never could understand.
They are conquered countries, annexed to their conqueror. They
have lost their national existence for centuries. They have no
national existence, or any kind of existence apart from England. No.
Nationality does not come into the question of your successor at all.
That is where the Church of Christ differs from all religions. Rome
can do, and does do, what no other ecclesiastical power durst do.
Our predecessors sent an Italian to Canterbury, and even a Greek,
Theodore; and We are sending a Kelt to Pimlico. As for Caerleon—do
you remember John Jennifer, the priest of Selce? You do:—he was a
white man at Mary vale:—and since? Good. He is Bishop of
Caerleon."
"He speaks the language, Holy Father," said Talacryn, laughing.
"The merest accident. We selected him for his steadfast sturdy
goodness under great difficulty at Maryvale. Oh, we remember——"
And the Pope's gaze went far away into the past.
Cardinal Talacryn mentioned that the Secretary of State desired to
know whether His Holiness would require the services of the
Patriarch of Byzantion at the present juncture.
"The Patriarch of Byzantion?"
"It was thought that as he had negotiated with England during the
reign of Your Holiness's predecessors——"
"Oh. Then, no. The services of the Patriarch of Byzantion are not
required. When His Grace is not smirking in 'black' drawing-rooms,
or writing defamatory letters to duchesses——"
"Defamatory letters, Holy Father!"
"Yes: defamatory letters, such as this one which he wrote in 1890."
The Pope got up, took off His episcopal ring, unlocked and dived into
an alphabetical letter-case, and handed a most ingeniously
suggestive and lethific note to the cardinal. "Well, when His Grace is
not engaged in these disedifying pastimes, he has his patriarchate to
attend-to. In fact unless he can see his way to become a resident
patriarch in Byzantion within the month, he may look for a decree of
deposition." The Supreme Pontiff's aspect was austere. "Your
Eminency will convey that response to Cardinal Ragna's obliging
suggestion."
Talacryn made haste to kneel. "Give me a blessing, Holy Father, and
I will immediately proceed to my new see, whatever."
Hadrian smiled. "God bless you, son. But do not go yet. Pimlico has
been in the hands of the Vicar-General and the Coadjutor for years;
and the Vicar-Capitular can manage for the present. Stay here a little
while. We shall need you. We shall not need you long."
And Talacryn went out from the Presence, glad, yet grave.
During a few days, questions and answers incessantly passed
between the Vatican and Windsor Castle. Hadrian consulted
sovereigns: discussed difficulties with statesmen. Baron de Boucert
expressed the opinion that it would be futile to oppose the inevitable
expansion of Germany. Signor Barconi himself officiated at an
instrument installed in the apostolic antechamber, until he was
carried away in nervous collapse. Hadrian envied him: and forced
Himself to resist temptation. He had much to do yet. Messages,
messages, study of maps, collation of ms. notes, filled a score of
each twenty-four hours. There was need of profound thought, so
that the clairvoyant undazzled eye like a diver might reach the
bottom of deep-preserving thought. The four hours which remained
chiefly were spent at the tomb of St. Peter in the basilica. The
Arbitrator slept not at all in these days. He ate while at work; and
only sought refreshment under the ice-cold tap in the bath-room. A
squadron of English cruisers escorted a procession of royal yachts
and battleships, which conveyed the Congress of Windsor to Golden
and immortal Rome.
Then came the issue of the Epistle to the Princes, in which the
Apostle reiterated the evangelic counsels, predicating a scheme of
utter self-sacrifice and non-resistance in imitation of the "sweet
reasonableness of Christ." This would mean, said He, the deliberate
loosening and casting away of all conventions which bound society
together. It was right: it was straight: it was the most direct road to
heaven. But it was not in accordance with the human will: it would
be called utopian, and unconventional; and it would be derided more
than followed: it would cause confusion inconceivable if it were
attempted on the grand scale. Truth more quickly emerges from
error than from confusion. Men, being what they are, i.e. bound to
err, would be better for having their errancy guided. They would
diverge from the road: but they should not leave it out of sight; and,
properly guided, their movement at least could be made to tend
towards the Point Desirable. Individuality so long had been
suppressed, that its efforts required administration. Therefore the
Pontiff shewed, as well as an unconventional, a conventional way of
approaching that Point Desirable. He maintained the aristocratic and
monarchic principle in strict integrity. A rebel was worse than the
worst prince, and rebellion was worse than the worst government of
the worst prince that hitherto had been. He proclaimed the anarchy
of France and Russia to be a manifestation of diabolic ebullience,
which ought to be restrained and stamped out by all right means,
even the most stringent. France and Russia, having forfeited the
right of being deemed capable of ruling themselves, henceforth must
submit to be ruled. Satan finds mischief for idle hands to do.
Occupation, and scope for occupation, alone will enable individuals
and nations to work out their own salvation humanly speaking. Men
must use themselves:—for good or ill. Most human ills were caused
by the lack of scope for energy. Sitting on, or screwing down, the
safety valve invariably was fatal:—a doctrine which He enforced on
the attention and obedience of the clergy. These principles involved
a re-arrangement of various spheres of influence. The Ruler of the
World, Peter, the Supreme Arbitrator, decreed that the only nations,
in which the "facultas regendi" survived in undiminished energy,
were England, America, Japan, Germany, Italy. Some of the old
monarchies, however, had not yet reached that point of decay when
their extinction would become desirable: they were Norway, Sweden,
Denmark, the German kingdoms and principalities and duchies,
Spain, Portugal, Greece, Roumania, Albania, Montenegro, the
republics of Switzerland and San Marino. These were to be
maintained as sovereign states and to preserve their national
characters. Some also of the old monarchies, which had tolerated
unmerited suppression, were to be given an opportunity of proving
themselves worthy of corporate existence. These were Hungary,
Bohemia, and Russian and German Poland. They were revived as
kingdoms; and required to provide themselves with constitutions
(after the manner of England), and to elect their respective
monarchical dynasties. Switzerland and San Marino were confirmed
as republics. The Sultan at the instigation of England, his ally, would
move his capital to Damascus, in order to concentrate the main force
of Islam in Asia. Servia was added to the Principality of Montenegro.
Turkey-in-Europe and Bulgaria would become merged in the
kingdom of Greece. So far for particulars.
Hadrian denounced, as bad and idle dreams, the plans of recent
political schemers who had adumbrated ideas of a federation of the
English-speaking and the Teutonic races. He dwelled upon the
essential differences which divided Germany from America, and both
from England. No blend was possible between the English and the
Germans; and Americans were not qualified for bonds. Each one of
the three was unique; and each would stand alone. Three such
enormous powers must have each its own separate and singular
existence and sphere of action. Three such spheres must be found,
in which the three nations independently might thrive. It was room
for independent development which must be sought out, and
assigned.
He stated the case of the continent of Europe. Belgium had 228
inhabitants to the square-kilometre: Holland, 160: Germany, 104:
Austria, 87: France, 72: Russia was so sparsely populated that only a
migration of 109,000,000 people from the rest of Europe would raise
her to the European average. Hence, the Pope proclaimed the
instauration of the Roman Empire, under two Emperors, a Northern
Emperor and a Southern Emperor; and confirmed the same to the
King of Prussia and the King of Italy as representatives of the
dynasties of Hohenzollern and Savoy respectively. He ordained that
this instauration should not be deemed 'the ghost of the dead
Roman Empire sitting crowned upon the grave thereof, but its
legitimate heir and successor, justified by the ancient virtues of the
Romans, the beneficence of their rule,' and the vigorous aspiration
to well-doing which characterized their present representatives. The
Northern Emperor William would nominate sovereign dynasties for
Belgium and Holland. He might replace the present exiled monarchs
on their respective thrones: or he might depose them and substitute
members of his Imperial family. He then would extend the borders of
Germany, eastward to the Ural Mountains by the inclusion of Russia,
westward to the English Channel and Bay of Biscay by the inclusion
of France, southward to the Danube by the inclusion of Austria. At
the same time, he would federate the constitutional monarchies of
Norway and Sweden, Denmark, Holland, Belgium, Hungary,
Bohemia, Poland, Roumania, and the republic of Switzerland with
the other sovereign states already under his suzerainty: while the
Southern Emperor Victor Emanuel would federate the constitutional
monarchies of Portugal, Spain, the extended kingdom of Greece, the
principalities of Montenegro and Albania, and the republic of San
Marino, with the kingdom of Italy, which last now was to include
Italia Redenta. The frontier dividing the Northern Empire from the
Southern was to be formed by the Pyrenees, Alps, Danube, and
Black Sea.
The case of America was defined. The United States were to be
increased by the inclusion of all the states and republics of the two
Americas from the present northern frontier of the United States to
Cape Horn.
The Japanese Empire was authorized to annex Siberia.
All Asia (except Siberia), Africa, Canada, Australia, New Zealand and
All Islands, were erected into five constitutional kingdoms, and
added to the dominions of the King of England, Ireland, Wales, and
Scotland. The title "Emperor" being antipathetic to the English Race
(on account of its primary significance "War-Lord"), the official style
of the Majesty of England, Ireland, Wales, Scotland, Asia, Africa,
Canada, Australia, New Zealand and All Islands, henceforth would be
"The Ninefold King."
Thus the Supreme Arbitrator provided the human race with scope
and opportunity for energy. The provisions of the Epistle to the
Princes were drawn up in the form of Treaty dividing the world, till
midnight (G.T.) of December 31st (N.S.) of the year 2000 of the
Fructiferous Incarnation of the Son of God, into the Ninefold
Kingdom, the American Republic, the Japanese Empire and the
Roman Empire. This Treaty was signed, in the Square of St. Peter's
at Rome, by the Pontiff, the Sovereigns and the Presidents, on the
Festival of the Annunciation of Our Lady the Virgin; and the armies
and navies of the signatories instantly set about the pacification of
France and Russia by martial law.
CHAPTER XXI
April brought to Hadrian an experience of one of those periods of
psychical disturbance which are incidental to the weakness of
humanity, and inevitable by a man of His particular temper. Things
lost their significance to Him, persons lost their personality, events
their importance; and time was not. He kept a straight face, and
forced Himself to courteous demeanour: but He was living in a world
in which He felt Himself to be just off the floor and floating, a world
in which everything was strange and everybody was quite strange, a
world where nobody and nothing mattered the least little bit. He had
the sense at the beginning to include Himself in secret behind
guarded doors; and also to hold His tongue when His attendants
were in the Presence. He simply sat and wondered—wondered who
He was, how He came there, who dressed Him like that, and when;
—and decided that it did not matter. He nursed His cat, cooing and
mewing and talking cat-language in a most enjoyable manner. When
the creature went away,—it did not matter. He used to gaze at His
cross by the hour together, planning combinations of lights and
shades and backgrounds of book-backs: placing the golden symbol
there, and revelling in the supple splendour of the Form, its dignity,
its grace, the majestic youth of the Face, noble and grave. He would
close His eyes and learn the lovely planes and contours with delicate
reverent touch. It pleased Him to think that He had created a type of
incarnate divinity, which neither was the Orpheys of the catacombs,
nor the Tragic Mask of the Vernicle, nor the gross sexless
indecencies wherewith pious Catholics in their churches insult the
One among ten thousand, the Altogether Lovely. That thought
brought Him back to Space and Time. Indignation at images at least
eleven heads long, proportioned like female fashion-plates, visaged
like emasculate noodles whom you would slap in the face on sight,
simply for their tepid attenuate silliness, if you met them in the flesh
—this drew down Hadrian to realities and life.—He felt utterly
exhausted. An exposition of sleep seized Him. He was always
drowsy; and would fall asleep in the day-time over the writing and
reading which He put Himself to do, in His armchair by the window,
in His favourite seat by the old wall in the garden where He spent
the vivid afternoons of spring. Only toward night-fall, was He able to
write that beautiful clear script of His, to bring any of His usual
alertness to bear upon affairs: even then that alertness was
extraordinarily diluted. His intellect was nebulous, uncertain. He
could not select saliencies, could not concentrate his thoughts: His
constructive faculty was in abeyance: His imagination was in chains.
He spent a long time over His scanty meals, chewing, chewing,
reading, reading, and remembering nothing which He read. In an
inert perfunctory way, He blamed Himself for waste of time; and
continued to waste it. No doubt it was divine nature's will. Let it be
understood that He was not slothful in the confessional sense of the
word. He was merely lethargic, dulled, blunted, listless, eager for
nothing, except to flee away and be at rest—at rest.
From this stupor, He awoke in panic, as though nympholeptose,
lymphatic, driven to phrenzy by some unknown external agency. He
became inspired with an appalling consciousness of the absolute
necessity for instant active continuous exertion,—if He were to
continue alive upon this earth. He felt that, if He were to permit
Himself to relax for one instant, if for one instant He were to
abdicate command of His physical forces, to let Himself go,—that
instant would be His last. With this in His mind, He prepared for
momentary unconscious lapses from violent activity. He posed with
care, so that, if Death should seize Him unawares, He might not
present a disedifying or untidy spectacle to the finders of His corpse.
He carefully avoided postures from which, when He should be reft
from the body, His form would fall indecorously. He did not trouble
His confessor more often than twice a week as usual: but His one
prayer, His incantation, always was on His lips, "Dear Jesus, be not
to me a Judge, but a Saviour." He was losing hold of the world.
Continually, through every hour of the day and night, His head rang
with the reverberating boom—boom—boom—boom of His strong
heart's beating. The rhythm was maddening. He used to count the
pulsations, wondering, after "fourteen," whether He would be able
to say "fifteen": after "ninety-seven," whether He would be in Rome
to say "ninety-eight": expecting the sudden wrench of self from
body: conjecturing the nature of that unique experience. Once, He
put Himself to the question "Was He afraid?" He answered, No,
because He dared to hope; and, Yes, because He had not been there
before. But Sokrates had said that death was our greatest
possession on earth; and Seneca said that death was the best of the
inventions of life; and Seneca's friend Saint Paul said "to die is gain."
On the whole, He was not afraid, afraid, of death. But, He did not
dare to go—to go—to sleep now. At night, He used to lie in bed, first
on His right side, then at full length on His back with the pillow
under His neck, and His hands crossed on the breast which had
been tattoed with a cross when He was a boy, and His ankles
crossed like a crusader, rigid, as He wished to lie in His coffin,—and
His brain active, active, counting physical pulsations, meditating on
the future, scheming, planning, counting each breath, and waiting
for the last—and death.
Sometimes He wondered whether it was all worth while: whether it
was in accordance with God's Will that He should be so will-full. He
decided to risk an affirmative to that, on the ground of the existence
of His will. He knew that He tried rightly to use it. He hoped for
mercy on account of lapses. One point He determined. With all due
respect to Sokrates and Seneca, Death came by Sin, and Sin was
God's enemy, and God's friends must fight God's enemies to the
bitter end. To relax was suicide, and suicide was sin; and, tired with
conflict as He was, eager for rest and peace as He was, it certainly
was not worth while to add to His tale of sin: it was not worth while
to exchange tiresome earth for untiring hell: to lose, what Petrarch
calls 'the splendour of the angelic smile.' He had no steel in His
possession except safety-razors: knives and scissors He had
abolished long ago; and now He had light strong gratings fixed to all
His windows. He would not go into temptation. 'I am fawned upon
by hope. Ah, would that she had a voice which I could understand, a
voice like that of a herald, that I might not be agitated by distracting
thought,' He said to Himself in the words of Elektra at the tomb of
Agamemnōn. Had He been trained in boyhood at a public-school, in
adolescence at an university, had His lines been cast in service, He
would not have had to put so severe restraint upon Himself. The
occasion would not have arisen. A simple and perhaps a stolid
character would have been formed of His temper, potent and brilliant
enough to distinguish Him from the mob, but incapable of
hypersensation. Instead, His frightfully self-concentrated and lonely
life, denied the ordinary opportunities of action, had developed this
heart-rending complexity: had trained him in mental gymnastics to a
degree of excellence which was inhuman, abominable, (in the first
intention of the words), in its facile flexible solert dexterity. He was
not restrained by any sense whatever of modesty or of decorum. He
had no sense of those things. He knew it; and regretted it. He was
Himself. He distrusted that self, rejoiced in it, and determined to deal
well and righteously with it. Dr Guido Cabelli, at length summoned,
found Him positively furious with the pain of physical and intellectual
struggles. The physician exhibited Pot. Brom., Tinct. Valerian. Am.,
Tinct. Zinzil., Sp. Chlorof., Aq. Menth. Pip., once every three hours. It
made the Pontiff conscious that He stank like a male cat in early
summer: but He heard no more boom-booming in his ears. It
strung-up His nervous system for the time. He put on His pontifical
mask; and addressed Himself from the ideal to the real.
He put the affairs of nations on one side. They, the nations all were
tumbling over one another in their eagerness to re-arrange
themselves upon the pattern which He had devised for them. If He
adopted the Pythagorean rôle of an uninterested spectator, either He
would be annoyed by something ugly or something silly, or He would
have a chance of glorifying Himself on account of some success. And
He wished to do otherwise than that. "In this world, God and His
angels only may be spectators."
The affairs of religion, as far as He could see, amounted to the
service of others and the cultivation of personal holiness, the
correspondence with Divine Love. Someone had told Him that—yes,
Talacryn in confession, of course,—that the key to all His difficulties,
present and to come, was Love. That was all very pretty and
theological on the part of the bishop, the cardinal-archbishop: but it
was the baby who had taught Him the secret of the method. He
would, He really would keep His troubles to Himself. His office was
the office of leader and exemplar. Nothing must interfere. He put
Himself to review the first year of His pontificate: and a black
enough tale it seemed to Him. Without surprize, without emotion,
He noted the blurs of impatience, pride,—pride,—humanity.—
Retrospection was the most wearisome most fatuous banality.
Onward!
Leader and exemplar! One thing was clear. He must come down
among the led and following. He must be seen of men. And He was
not seen. No. Peculiar personal preference kept Him apart,
mysterious. He rather enjoyed (not the being misunderstood but)
the not being understood; and, at the same time, He had been
doing a lot of people the gross injustice of crediting them with the
possession of intelligence similar to His Own, of perspicacity equal to
His Own, of the ability to keep up with His rapid pace and abrupt
manœuvres. That was unrighteous. No doubt it had been all very
fine and noble and so forth to sit down silent under calumny, for
example. One could afford to do that when one was innocent. But,
when millions of people (to give the devils their due) actually wanted
to believe one innocent, and would be grieved and perhaps injured
because the opportunity to believe innocence was withheld, was it
righteous to refuse to condescend? No, such a pose was mere pride.
The Servant of the servants of God must not fear to soil the
whiteness of His robe in any kind of ordure. Also, to save others was
the best way of retrieving oneself.
He sent for the nearest cardinals. Ragna, Saviolli, Semphill, Sterling,
Talacryn, Carvale, Van Kristen, Gentilotto, Leighton, Whitehead,
responded to the summons. Hadrian received them in the throne-
room, but without formality; and contrived to give them an easy and
genial greeting. They thought Him to be looking seriously ill. There
was the dead whiteness of a gardenia in the hue of His face and
hands: His reddish-brown hair was going grey over the left ear: His
intense and rigid mask was the sign of pain. His whole aspect also
was diaphanous, wasted. But His manner was vivid: He was not
inaccessible. Their Eminencies gave Him their attention; and
wondered what He was going to bring-out of the dispatch box by His
side. He was extremely glad to see the Secretary of State: for He
knew how antipathetic He was to that one; and now He was going
to try to give him satisfaction. At least it should not be His fault if
Ragna's ordinary attitude of discreet and convulsive brutality
remained unmitigated.
"Lord Cardinals," the Supreme Pontiff said, "it has occurred to Us
that ye have many things to say: that there be many things which ye
desire to know. We, on Our part, are ready to hear; and We are
willing to respond to questions."
Questions instantly were born in each man's brain. Ragna was the
first to deliver Himself of his. "Holiness, will You answer a question
about the Epistle to the Princes?"
"Yes."
Ragna collected himself. "I am curious to know why the rights of
France in Egypt were not even named. I can see that the very
nature of Your Holiness's counsels demanded that Africa as a whole
should pass to England: but I cannot understand why Germany, in
taking over France, should not also have taken-over the
condominium of Egypt. Why did that fall to England; and why did
Germany consent to its falling to England?"
Hadrian made an effort to conquer His natural incapacity for coming
near a subject at the first attempt; and put Himself to be concise.
"Your Eminency knows that since—We forget the exact date—but
since a very short time ago, no international obligations have existed
which could restrain Egypt from legitimate attempts at emancipation.
Nothing but Ottoman firmans held her. Very well. We discovered that
when the King of England and the Sultan, last October, made
alliance, the latter issued a firman in which England was named
Protector of Egypt. Then (the speaker slightly smiled), when the task
of arbitration was submitted to Us, We found that the German
colonies in Africa, not only did not pay their way but, required a
yearly subsidy of £1,500,000; and therefore, taking one thing with
another, We arranged to give Germany sufficient employment for a
century nearer home. She promptly recognized that 'megli' è
fringuello in man' che tordo in frasca.' The fact is that she was only
too glad to be rid of her own parasitic colonies, which had severed
their connection from the parent stem, and derived their nutriment
from other states: while the colonies of France which were epiphytic,
having no existence apart from the source from which they sprang,
were wiped out (as French colonies) when France was wiped out."
"And no doubt Germany, in her pretty Gothic way, was in such a
desperate hurry to grab France, that she forgot all about Egypt. D'ye
know they say she's going to call her conquest Gallia again?"
Semphill put in with a sniff. "And now I'll ask a question. Holy Father,
may I smoke?"
"But smoke!" Hadrian assented with pleasure; and held-out His Own
hand for a cigarette. Some of the others did likewise; and the gear
began to run much more easily. Van Kristen expressed joy that the
Germans were not to have chances of doing more monkey business
on the Erechtheion and the Akropolis at Athens.
"Yes," Ragna meditatively continued: "I suppose I ought to have
understood all that. But now, Holiness, there's another thing: why
did the Sultan consent to evacuate Europe?"
"Simply because, with all the examples which he has had lately, he
goes in mortal terror of assassination. He has managed to persuade
himself that he only can be warranted against that, as long as he is
under the ægis of England. Well: seeing England and Turkey allied,
We moved England and England moved Ismail. The former had
sense: the latter, sentiment. But Ismail really is not half bad: in fact
he's rather decent. If We only had another dear charming child-like
naked Christian like Blessed Brother Francis——"
"What?" said Carvale with animation. He happened to have noted
that, when Hadrian rioted in superlatives, it meant no more than
positives: but, when He negligently drawled comparatives, "not half
bad" or "rather decent," the ultimate of praise was signified. "What?"
the cardinal repeated.
"We would send him to give points to Ismail's mollahs and
dervishes."
"St. Francis has innumerable sons, Holiness," Saviolli put in.
"And We only know one who in the slightest degree resembles his
father," the Pope responded, waving away the subject.
"One would like to know," said Sterling, "whether Your Holiness is
not really of the opinion that the Epistle to the Princes was perhaps
a trifle too sentimental and——"
"Sentimental? Yes. The Ruler, who rules sentiment out of his
calculations, ignores one of the most potent forces in human affairs.
Too sentimental? No. And what else was Your Eminency about to say
—a trifle too sentimental and——"
"One would have said perhaps a trifle too arbitrary."
"Dear man——" the Pope gleefully began.
But Ragna interrupted "Nothing of the kind. That particular Epistle
was replete with pontifical dignity: it was the finest thing——"
Hadrian stopped him "We were about to remind Cardinal Sterling
that when the Ruler of the World geographically rules the world, He
is accustomed to do His ruling with a ruler. Our predecessor
Alexander VI. used a ruler on a celebrated occasion on the Atlantic
Ocean."
Everybody burst out laughing: laughed for a few moments; and
returned to a serious demeanour. There was a question, an
important question, which sat upon all tongues, wing-preened, ready
to fly. But His Holiness already had refused to discuss it. Those, who
had tried to persuade, so seriously had been hurt by His icy
reticence or by His blunt aloofness, that no one now was
temerarious enough to attempt the re-opening of so unsavoury and
so personal a matter, except upon explicit invitation. Knowing what
he did of men, Hadrian had expected hesitation: but, seeing that His
purpose was likely to fail of completion; and, being determined that
it should not fail, He slowly and significantly drew-off the pontifical
ring from His first finger, and put it in His pocket. "Gentlemen," He
said with quite a change of manner, "some of you would like to put
George Arthur Rose to the question?"
They would indeed. They would whatever. They would like it so
much that they all spoke in unison. The sum of their words
amounted to a request that George Arthur Rose would give them
some sort of statement concerning newspaper calumnies, some sort
of statement by way of support to their contention that he had been
grossly wronged and mispresented.
It was George the Digladiator who responded. He seemed to step
down into the arena, naked, lithe, agile, with bright open eyes, and
ready to fight for life. "Very well," he said—"I will give that
statement to you: but understand that I will not defend myself in the
newspapers. If I were a layman, I should have whipped in a writ for
libel, and have given my damages to Nazareth House. I should have
preferred to trust my reputation rather to an English judge and jury,
than to the nameless editors of Erse or Radical newspapers. Fancy
having one's letters edited by the Catholic Hour, for example: fancy
having one's letters, which are one's defence, nefariously garbled by
a nameless creature who is one's prosecutor, and one's judge, and
one's jury, all in one! However, not being a layman, I cannot go to
law; and I will not condescend to have dealings with those
newspapers. Understand also, that I tell you what I am about to tell
you, not because I have been provoked, abused, calumniated,
traduced, assailed with insinuation, innuendo, mispresentation, lies:
not because my life has been held up to ridicule, and to most inferior
contempt: not because the most preposterous stories to my
detriment have been invented, hawked about, believed. No. Please
understand that I am not going to speak in my own defence, even to
you. I personally and of predilection, can be indifferent to opinions.
But officially I must correct error. So I will give you some
information. You may take it, or leave it: believe it or disbelieve it.
You shall have as photographic a picture as I can give you of my life,
and of the majestic immobility by which you clergy tire out—
assassinate a man's body—perhaps his soul. You are free to use it or
abuse it. When I shall have finished speaking, I never will return to
this subject."
"Of course we shall believe what you say," Semphill rather nervously
intercalated. "I'm sure we believe it unsaid. We take it as said, you
know. But if you could see your way to give us details, say on half a
dozen points, that would be quite enough."
"The Daily Anagraph has not apologized for its latest slander,"
Carvale put in.
"Why should it?" George inquired.
"Well, I sent an authenticated account of what happened in the last
consistory. The other papers printed it; and I should have thought
the least the Daily Anagraph could have done would be——"
"Carvale, you're making a mistake. The Daily Anagraph has no
personal grudge against me: although the last editor had, because I
once innocently asked him whether historical accuracy came within
the scope of a Radical periodical. That was years ago, at the time of
the second Dreyfus case. I know that he was furious; because
Bertram Blighter, the novel-man, told me that that editor in revenge
was going to put me on the newspaper black-list, whatever that may
be. No, it is not a personal matter, a matter in which an apology is
customary. It's simply an example of the ethics of commercial
journalism. The man wanted to increase the sale of his paper. I
happened by chance to be before the world just then. And he took
the liberty of increasing his circulation at my expense. Actually that
is all. You can't (at least I don't), expect an editor, who is capable of
doing such a thing, to apologize for doing it. The case of the other
papers is verisimilar: except of course the Catholic Hour. That simply
exists on sycophanty by sycophants for sycophantophagists, as
Semphill knows."
"Yes I know," said Semphill. "And I don't allow the thing to enter my
house."
"But the others—in their case it's not lurid malignance, but legal
malfeasance. Did you say that they apologized?"
"No. None of those, which printed the calumnies, apologized. They
just kept silence. But all the respectable papers, which had not
calumniated you, printed my refutation of the Daily Anagraph."
George made a gesture of scorn, of satisfaction, of dismissal. "Then
the Pope is clear;" he said. "Now I will try to tell you, as briefly as
possible, what you want to know about the other person." He
produced a sheaf of newspaper-cuts. He was in such a white rage at
having to do what he was about to do, that he wreaked his anger on
those who listened to him, piercingly eyeing them, speaking with
swift fury as one would speak to foes. "The Catholic Hour states that
in 1886 I was under an under-master at Grandholme School: that I
had to leave my master-ship because I became Catholic. That is true
in substance and absolutely false in connotation. I was an under-
master: but as I also had charge of the school-house, I was called
the house-master. You also perhaps may be aware that there is only
one head-master in a school; and that all the rest are under-
masters. But, when slander is your object, 'under-master' is a nice
disgraceful dab of mud to sling at your victim for a beginning. Well: I
resigned my house-mastership of my own free and unaided will for
the reason alleged; and I have yet to learn that the becoming
Catholic is an extraordinarily slimy deed. Further, note this, far from
my resignation being the dishonourable affair which the Catholic
Hour implies, the head-master of Grandholme School remained my
dear and intimate and honoured friend through thick and thin, for
more than twenty years, and is my only dear and intimate friend at
this moment."
Semphill and Carvale looked up, and then down. Sterling looked
down, down. Van Kristen looked up. The others, anywhere. Talacryn
looked annoyed. The taunt was flung out; and the flying voice went-
on. "The Catholic Hour thus casts its diatribe in a key of
depreciation. Next, I am said to have gone to a school for outcasts,
to have quarrelled with the two priest-chaplains; and presently to
have been 'again out.' The idea being to infer evil, it is rather
cleverly done in that statement of the case. But here are the facts.
The school perhaps might be called a school for outcasts. But I, a
young inexperienced Catholic of six months, was lured by
innumerable false pretences, on the part of the eccentric party who
offered me the post, to accept what he called the Head-mastership
of a Cathedral Choir School. He did not tell me that he was forcing
the establishment on the bishop of the diocese, nor that the Head-
mastership had been refused by several distinguished priests simply
on account of the impossible conditions. I bought my experience.
That I quarrelled with the chaplains is quite true. I did not quarrel
effectually though. They were a Belgian and a Frenchman. They
drank themselves drunk on beer, out of decanters, chased each
other round the refectory tables in a tipsy fight, defied my authority
and compelled the ragamuffins of the school to do the same. I
naturally resigned that post as quickly as possible. Then follows a
pseudo-history of the beginning of my ecclesiastical career at
Maryvale. Talacryn knows all about that; and can tell you at your
leisure. Afterwards, I came across, (I am quoting), 'came across a
certain Pictish lairdie, and was maintained by him for three or four
months——'"
"And I know all about that," Semphill interrupted: "You gave a great
deal more than you got."
"The fallacies connected with my career at and expulsion from St.
Andrew's College are known?"
"Thoroughly," assented Semphill, Talacryn, and Carvale in a breath.
"The statement that I contracted large debts there——"
"What about those debts?" Ragna asked.
Carvale told him. "They all were contracted under the personal
supervision of the Vice-Rector. They were quite insignificant. Besides
that, they would not have been contracted but for the promise of
Archbishop Smithson and the advice of Canon Dugdale——"
"And the advice of me," Semphill added in a low tone.
"Oh, you at length acknowledge it?" George fiercely thrust at him.
"Yes, I acknowledge it."
"Well then, we're quits now:" George quietly and mysteriously
mewed.
"One confesses that the question of the pseudonym interests one,"
Sterling judicially said.
"I had half-a-dozen. You see when I was kicked out from college,
without a farthing or a friend at hand, I literally became an
adventurer. Thank God Who gave me the pluck to face my
adventures. I was obliged to live by my wits. Thank God again Who
gave me wits to live by."
Cardinal Leighton was standing-up, blinking and blushing with
indignation which distorted his honest placid features. "Holy Father,
don't say another word." He twitched round towards his fellow-
collegians. "How can you torture the man so!" he cried. "Can't you
see what you're doing, wracking the poor soul like this, pulling him
in little pieces all over again? Shame on ye!—Holy Father don't say
another word."
"Oh if I had only known!" cried Van Kristen.
"You did! I told you myself; and you didn't believe me!" George
fulminated.
The youngest cardinal wept into his handkerchief, shaking with sobs.
George neither saw nor noted anyone. He was glaring like a python.
Demurrers to Leighton's remarks arose. No one wanted to wrack
anybody. Questions had been invited. Of course no one believed. But
it would be so much more satisfactory—Ragna added. George sat
violently still in his chair while they talked: let them talk; and
prepared to resume.
"If Your Holiness would condescend——" Carvale began.
"There is no Holiness here," George interrupted, in that cold white
candent voice which was more caustic than silver nitrate and more
thrilling than a scream.
"If you would do us the favour of just noticing a few heads."
"As you please," George chucked at him: "agree among yourselves
as to those heads; and you shall have bodies and limbs and finger-
nails and teeth to fit them."
Their Eminencies began agreeing. George meanwhile went into the
secret chamber for ten minutes or so: and returned with his cat on
his neck, and his own tobacco-pouch. He was beginning a cigarette;
and his gait was the gait of a challenged lion. Sterling presented him
with a pencilled slip of paper. He read aloud "Pseudonym: begging
letters: debts: luxurious living: idleness: false pretences as to means
and position."
"I think it right to say that I myself am perfectly satisfied on all those
points," said Semphill. "I've read the calumnies—and I call them
dastardly calumnies—in the light of my own knowledge of the facts;
and I can only say that the worst thing which they've alleged against
you is that you've been used to go-about bilking landlords. All the
rest is excusable, not to say harmless."
"Gracious Heavens!" George exclaimed in a rictus of rage. "Do you
suppose that a man of my description goes-about bilking landlords
for the sake of the fun of the thing? It's no such deliriously jolly
work, I can tell you. However, I've never bilked any landlords if that's
what you want to know. Never. They saw that I worked like nineteen
galley-slaves; and they offered to trust me. I voluminously explained
my exact position and prospects to them. I was foolish enough to
believe that you Catholics would keep your promises and pay me for
the work which I did at your orders. So I accepted credit. I wish I
had died. When at length I was defrauded, legally, mind!—for, as my
employers were Catholics and sometimes priests, I trusted to their
honour, and obtained no stamped agreement:—when I was
defrauded of my wages, my landlords lost patience (poor things—I
don't blame them,) harried me, reproached me, at length turned me
out, and so prevented me from paying them. I dug myself out of the
gutter with these bare hands again and again; and started anew to
earn enough to pay my debts. Debts! They never were off my chest
for twenty years, no matter what these vile liars say. Debts! They
say that I incurred them for luxurious living, unjustifiably——"
His passionate voice subsided: he became frightfully cool and tense
and terse, analytical, quite merciless to himself. Their Eminencies
never before had seen a surgical knife at work in a human heart and
brain. They sat all vigilant and attentive, as self-dissection
proceeded. "They say that I gorged myself with sumptuous
banquets at grand hotels. Once, after several days' absolute
starvation, I got a long earned guinea; and I went and had an
omelette and a bed at a place which called itself a grand hotel. It
wasn't particularly grand in the ordinary sense of the term; and my
entertainment there cost me no more than it would have cost me
elsewhere, and it was infinitely cleaner and tastier. They say that I
ate daintily, and had elaborate dishes made from a cookery book of
my own. The recipes, (there may have been a score of them,) were
cut-out of a penny weekly, current among the working classes. The
dishes were lentils, carrots, anything that was cheapest, cleanest,
easiest, and most filling—nourishing—at the price. Each dish cost
something under a penny; and I sometimes had one each day. As I
was living on credit, I tried to injure no one but myself. That's the
story of my luxurious living. Let me add though that I was
extravagant, in proportion to my means, in one thing. Whenever I
earned a little bit, I reserved some of it for apparatus conducing to
personal cleanliness, soap, baths, tooth-things, and so on. I'm not a
bit ashamed of that. Why did I use credit? Because it was offered:
because I hoped: because—— That I did not abuse it you may see,
actually see, by my style of living,—here are the receipted bills;—and
by the number and quantity and quality of the works of my hands. I
never was idle. I worked at one thing after another. The Catholic
Hour admits my skill; and mispresents that as a crime. At the same
time, I myself don't claim my indefatigability as a virtue. Nothing of
the kind. It's something lower than that. It's comical to say it: but
my indefatigability was nothing but a purely selfish pose, put-on
solely to make philanthropists look unspeakably silly, to give the lie
direct to all their idiotic iniquitous shibboleths. It wasn't that I
couldn't stop working: but that I wouldn't. The fact is that I long, I
burn, I yearn, I thirst, I most earnestly desire, to do absolutely
nothing. I am so tired. I have such a genius for elaborate repose.
But convention always alleges idleness, or drunkenness, or lechery,
or luxury, to be the causa causans of scoundrelism and of poverty.
That's a specimen of the 'Eidola Specus,' the systematizing spirit
which damns half the world. People never stop to think that there
may be other causes—that men of parts become rakes, or
scoundrels, or paupers, for lack of opportunity to live decently and
cleanly. Look at François Villon, and Christopher Marlowe, and Sir
Richard Steele, and Leo di Giovanni, and heaps of others. Well: I
resolutely determined that you never righteously should allege those
things of me. Simply to deprive you of that excuse for your failure to
do your duty to your neighbour—simply to deprive you of the chance
of classifying me among the ruck which your neglect has made—I
courted semi-starvation and starvation, I scrupulously avoided drink,
I hardly ever even spoke civilly to a woman; and I laboured like a
driven slave. No: I never was idle. But I was a most abject fool. I
used to think that this diligent ascetic life eventually would pay me
best. I made the mistake of omitting to give its due importance to
the word 'own' in the adage 'Virtue is its own reward.' I had no other
reward, except my unwillingly cultivated but altogether undeniable
virtue. A diabolic brute once said to me 'If I had your brains I would
be earning a thousand a year.' I replied 'Take them: tell me what to
do: give me orders, and I implicitly will obey you. Then, take that
thousand a year, and give me two hundred; and I'll bless you all my
days.' He said nothing; and he did nothing. He was just a fatuous
liar. I mocked him: caught him stealing my correspondence—there is
his written confession;—and, he wrote these anonymous calumnies
in long cherished revenge." The dreadful lambent voice flickered for
a moment;—and more rapidly flashed-on. "I repeat, I never was
idle. I did work after work. I designed furniture, and fire-irons. I
delineated saints and seraphim, and sinners, chiefly the former: a
series of rather interesting and polyonomous devils in a period of
desperate revolt. I slaved as a professional photographer, making
(from French prints) a set of negatives for lantern-slides of the Holy
Land which were advertised as being 'from original
negatives'—'messing about' the Catholic Hour elegantly denominates
that portion of my purgatory. Well I admit it was messy, and
insanitary within the meaning of the act too—but then you see I was
working for a Catholic. I did journalism, reported inquests for
eighteen pence. I wrote for magazines. I wrote books. I invented a
score of things. Experts used to tell me that there was a fortune
waiting for me in these inventions: that any capitalist would help me
to exploit them. They were small people themselves, these experts,
—small, in that they were not obliged to pay income tax: they had
no capital to invest: but they recommended me, and advised me, to
apply to lots of people who had:—gave me their names and
addresses, dictated the letters of application which I wrote. I trusted
them, for they were 'business men' and I knew that I was not of that
species. I quieted my repugnance; and I laid invention after
invention, scheme after scheme, work after work, before capitalist
after capitalist. I was assured that it was correct to do so. I despised
and detested myself for doing it. I scoured the round world for a
'patron.' These were my 'begging letters.'—At that time I was totally
ignorant of the fact that there are thousands of people who live by
inviting patronage; and that most of them really have nothing to be
patronized: while the rest are cranks. I knew that I had done such
and such a new thing: that I had exhausted myself and my
resources in doing it: that my deed was approved by specialists who
thoroughly knew the subject. I was very ashamed to ask for help to
make my invention profitable: but I was quite honest—generous: I
always offered a share in the profits—always. I did not ask for, and I
did not expect, something for nothing. I had done so much; and I
wanted so little: but I did want that little,—for my creditors,—for
giving ease to some slaves of my acquaintance. I was a fool, a
sanguine ignorant abject fool! I never learned by experience. I still
kept on. A haggard shabby shy priestly-visaged individual, such as I
was, could not hope to win the confidence of men who daily were
approached by splendid plausible cadgers. My requests were too
diffident, too modest. I made the mistake of appealing to brains
rather than to bowels, to reason rather than to sentiment. I wanted
hundreds, or thousands—say two: others wanted and got tens and
hundreds of thousands. A cotton-waste merchant could not risk
fifteen-hundred on my work, although he liked me personally and
said that he believed in the value of my inventions: but, at the same
time, he cheerfully lost twelve-thousand in a scheme for 'ventilated
boots.' I myself was wearing ventilated boots, then: but the
ventilated-boot man wore resplendent patent leather Cardinals'
secretaries could live at the rate of two-thousand-two-hundred-and-
ninety pounds a year and borrow three-thousand-and-sixty pounds,
on a salary of two-hundred pounds a year; and they could become
bankrupt for four-thousand-one-hundred-and-twenty pounds with
one-hundred-and-eighty pounds worth of assets. But I,—I could not
get my due from that man, one of whose secretaries wrote his
business to me on the franked note-paper of the late Queen of
England's Treasury: while the other, the bankrupt, gave me a winter
of starvation, because his lord had altered his mind, quoth he, about
the job on which I was working, and had determined to put his
money into a cathedral. No. I never accomplished the whole art and
mystery of mendicity. I perfectly could see what was required of him
who would be a successful swindler. I was not that one. I was
playing another kind of game—unfortunately an honest one. Take
that 'unfortunately' for irony, please. I mean—but you perfectly know
what I mean.—I made nothing of my inventions. By degrees, I had
the mortification of seeing others arrive at the discovery which I had
made years before. They contrived to turn it into gold and fame.
That way, one after another of my inventions became nulled to me. I
think I am right in saying that there are only four remaining at the
present moment. Finance them now? Engage in trade like a monk or
a nun? No. No. I shall give them to—that doesn't matter. It shall be
done to-day.—Idle? Idle? When I think of all the violently fatuous
frantic excellent things I've done in the course of my struggles for an
honest living—ouf! It makes me sick! Oh yes, I have been helped.
God forgive me for bedaubing myself with that indelible blur. I had
not the courage to sit-down and fold my hands and die. A brute
once said that he supposed that I looked upon the world as mine
oyster. I did not. I worked; and I wanted my wages. When they
were withheld, people encouraged me to hope on; and offered me a
guinea for the present. I took the filthy guinea. God forgive me for
becoming so degraded. Not because I wanted to take it: but
because they said that they would be so pained at my refusal. But
one can't pay all one's debts, and lead a godly righteous sober life
for ever after on a guinea. I was offered help: but help in
teaspoonfuls: just enough to keep me alive and chained in the mire:
never enough to enable me to raise myself out of it. I asked for
work, and they gave me a guinea,—and a tacit request to go and
agonize elsewhere. My weakness, my fault was that I did not die
murdered at Maryvale, at St. Andrew's College. The normal man,
treated as I was ill-treated, would have made no bones whatever
about doing so. But I was abnormal. I took help, when it was offered
gently. I'm thankful to say that I flung it back when it was offered
charitably, as the Bishop of Claughton offered it, and Monsignor—
you know whom I mean, Talacryn,—and John Newcastle of the
Weekly Tabule. I'll tell you about the last. He said that, being
anxious to do me a good turn, he had deposited ten pounds with a
printer-man, who would be a kind friend to me, and would consult
me as to how that sum could be expended in procuring permanent
employment for me. I took seven specimens of my handicraft to that
printer-man. He admired them: offered me a loan of five pounds on
their security. With that, I fulfilled a temporary engagement. Then I
consulted the printer-man, the 'kind friend.' He proposed to give me
a new suit of clothes, (I was to do without shirts or socks), to accept
my services at no salary, and to teach me the business of a printer's
reader for three months; and, then, to recommend me for a
situation as reader to some other printer. But, I said, why waste
three months in learning a new trade when I already had four trades
at my fingers' ends? But, I said, what was I to live on during those
three months? But, I said, what certainty was there at the end of
those three months? But, he said, that he would 'have none of' my
'lip, for' he 'knew all' my 'capers'; and he bade me begone and take
away my drawings. Those were ruined: he had let them lie on his
dirty office floor for months. Oh I admit that I have been helped—
quite brutally and quite uselessly. Helped? Yes. Once, when they told
me at the hospital that I was on the verge of a nervous collapse, a
Jesuit offered to help me. He would procure my admission to a
certain House of Rest, if I would consent to go there. By the Mercy
of God I remembered that it was a licensed madhouse, where they
imprisoned you by force and tortured you. Fact! There had been a
fearful disclosure of their methods in the P.M.G. Well: I refused to
go. Rather than add that brand to what I had incurred through being
Catholic, I made an effort of will; and contrived to escape that
danger: contrived to recover my nerves; and I continued my battle.
—Regarding my pseudonyms—my numerous pseudonyms—think of
this: I was a tonsured clerk, intending to persist in my Divine
Vocation, but forced for a time, to engage in secular pursuits both to
earn my living and to pay my debts. I had a shuddering repugnance
from associating my name, the name by which I certainly some day
should be known in the priesthood, with these secular pursuits. I
think that was rather absurd: but I am quite sure that it was not
dishonourable. However, for that reason I adopted pseudonyms. I
took advice about adopting them: for, in those days, I used to take
advice about everything, not being man enough to act upon my own
responsibility. Also, the idea of using pseudonyms was suggested to
me; and the first one was selected for me. As time went on, and
Catholic malfeasance drove me from one trade to another—for you
know—Talacryn—Carvale—Semphill—Sterling—that two excellent
priests declared in so many words that they would prevent me from
ever earning a living—legal assassination, you see definitely was
contemplated—I say as Catholic malfeasance drove me from one
trade, I invented another, and another; and I carried on each of
these under a separate pseudonym. In fact I split up my personality.
As Rose I was a tonsured clerk: as King Clement, I wrote and
painted and photographed: as Austin White, I designed decorations:
as Francis Engle, I did journalism. There were four of me at least. I
always have thought it so inexplicable that none of the authorities—
you, Talacryn, with your pretended confidence in me and your
majestic immobility towards me,—that none of you ever realized the
tremendous amount of energy which was being expended,
misdirected, if you like. Certainly no one of you ever made a
practical attempt to direct that energy. I was a like a wild colt
careering round and round a large meadow. You all looked on and
sneered 'Erratic!' Of course I was erratic, for you all did your very
best, by stolidity, hints, insinuations, commands, to create obstacles
over which I had to jump, through which I had to tear a way; and
there was no one to bit and bridle me, to ride me, and to share his
couch with me. And of course my pseudonymity has been
misunderstood by the stupid, as well as mispresented by the
invidious. Most people have only half developed their single
personalities. That a man should split his into four and more; and
should develop each separately and perfectly, was so abnormal that
many normals failed to understand it. So when 'false pretences' and
similar shibboleths were shrieked, they also took alarm and howled.
But, there were no false pretences. I told my name to everyone
whom it concerned. I am not the only person who has traded under
pseudonyms or technikryms. Take, for example, the man whose
shop I am said to have offered to buy. He himself used a trade-
name. He begged for my acquaintance when I was openly living as a
tonsured clerk, about a couple of years before my first pseudonym
even was thought of. Take, for another example, those priests, Fr.
Aleck of Beal, and the Order of Divine Love, who are alleged to have
'charitably maintained' me. By the way, they never did that. They
always were paid for my entertainment, in hard coin, and their own
price—always. And the Fathers of Divine Love refused me shelter for
one night in 1892 at the very time when they are said to have
'charitably maintained' me. They did suggest a common lodging-
house at fourpence, though; and I flung back the suggestion in their
faces and walked the streets all night. But all these people knew all
about me and my pseudonyms. In fact, the very priest who
suggested the common lodging-house, was the man on whose
advice I adopted my first pseudonym. It was invented by an old lady
who chose to call herself my grandmother: she was that priest's
patron and penitent. It was approved by him and adopted by me.
And there you have the blind and naked truth on that point. It now
is pretended that 'King Clement' was a jesuitical machiavellian device
of mine, implying royalty, dominions, wealth, and interminable
nonsense. I think that the pretension is due to malice and imbecillity.
It is malignant now: but I firmly believe that it began by being
imbecile. I confess that the name, taken together with my
domineering manner, my pedantic diction, my austere and (shall I
say) exclusive habit, was liable to misconstruction by the low coarse
half-educated uncultured boors among whom I lived. It's an example
of the 'Eidola Fori,' the strange power of words and phrases over the
mind. I think it really was believed, in some vague way, that I was
an exiled sovereign or some rot of that sort. I believe that I
perceived it; and laughed to myself about it. But I did my best to
disabuse the fools of their foolery. That made things worse. Liars
themselves, they could not conceive of a man speaking truth to his
own detriment. My disclaimer was taken for a lie; and they honoured
me the more for it; and chuckled at the thought of their own
perspicacity:—that is to say, when what I said was intelligible to
them. You see I used to be a great talker. I have had many
experiences; and I used freely to talk of them. It amused and
instructed; and I like to amuse and to instruct. You will understand
that my voice and my manner of speech did not resemble the voice
and the manner of speech of the ruffians with whom I worked and
lived. Live as poorly as I would, dress as shabbily as I would, the
moment I opened my mouth I was discovered to be different to
those people. They perceived it; and I never could disguise my
speech. Also, I'm quite sure that they could not understand my
speech—follow my argument. I used words which were strange to
them to express ideas unimagined by them, while their half-
developed minds were more than half occupied, not in listening to
me but, in contemplating me, and in trying to form their particular
idea of me by the aid of the 'Vulgi sensus imperiti,' the imperfection
of undisciplined senses, at their disposal. I called that Imbecillity.
Perhaps Ignorance is the apter term. The Malice is to be found
among people who ought to know better: people to whom I have
told the exact truth about myself, exact at the time of telling:
people, who being possessed by a desire to think evil, think evil:
people who read between, instead of on, the lines: people, prone to
folly, whom I have not helped to avoid their predilection. I tried to
be simple and plain, to sulk (if you like) in my own corner by myself.
It was no good. Anyhow, I told no tales of realms or wealth as mine.
I made no false pretences. I myself was grossly deceived:
barbarously man-woman-and-priest-handled. I was foolish to try to
explain myself. I was foolish to try to work with, to live with, to
equal myself in every respect with, verminous persons within the
meaning of the act. I ought to have died. But I did not die. That is
all. It is not half. Now you know. Make what you please of it."
"Tell me," Gentilotto instantly said: "Why did you never go to the
Trappists?"
"Because I went to something worse, to something infinitely terribly
more ghastly. Trappists live in beautiful silent solitude; they have
clean water, beds, regular meals, and peace. I went to live in
intellectual silence and solitude in an ugly obscene mob, where clean
water was a difficulty, food and a bed an uncertainty, and where I
had the inevitable certainty of ceaseless and furious conflict."
He hurled the words like javelins, and drew back in his chair. The old
bitter feeling of disgust with himself inspired him. He feared lest
perhaps he might have seemed to be pleading for sympathy. So he
angrily watched to detect any signs of a wish to insult him with
sympathy. But he really had gone far, far beyond the realm of human
sympathy. There was not a man on the earth who would have dared
to risk rebuff, to persist against rebuff, to soar to him with that
blessed salve of human sympathy—for which,—underneath his
armour,—and behind his warlike mien,—he yearned. Pity perhaps,
horror perhaps, dislike perhaps, might have met him. But he only
had emphasized his own fastidious aloofness. He had cleared-off the
mire: but he had disclosed the cold of marble, not the warmth of
human flesh.
The cardinals remained silent for a minute. Then Ragna said "'An
enemy hath done this!' Who is it?"
George blazed with vigorous candid delight. "That is the first
genuine word which I have had from the heart of Your Eminency!"—
He returned to his repellent manner. "I gave the names of my
calumniators to Cardinal Leighton."
"Jerry Sant the Liblab, aided by the woman and a clot of worms who
had turned;" Leighton said to Ragna.
"Let them be smothered in the dung-hill. Anathema sint." Ragna
growled.
Again there was an exposition of silence in the throne-room. George
was frozen hard and white. Ragna and Leighton continued to look at
each other. Carvale's eyes had the blue brilliance of wet stars.
Saviolli, Semphill, Talacryn, Whitehead, were as though they had
seen the saxificous head of the Medoysa. Stirling gazed straight
before him, in the manner of the sphinx carven of black basalt.
George was watching them with half-shut eyes from the illimitable
distance of his psychic altitude. Presently, the pure pale old face of
Gentilotto and the pure pale young face of Van Kristen
simultaneously were lifted; and their eyes met His. He blushed:
slowly drew out the pontifical ring: and put it on His finger.
"Lord Cardinals, it is Our will to be alone:" the Supreme Pontiff said.
They came one by one and kissed His ring; and retired in silence.
CHAPTER XXII
When the door was shut, Hadrian remained quite motionless on the
throne; and set Himself to review what He had said. He wondered
whether He for once had got-down to and laid-bare the root of the
matter: whether He for once had made His argument clear and
convincing.—Good God! Who even could hope to be convincing?—He
flung the thing away from Him; and for ever closed that volume of
the book of His life.

Download ebooks file Ethics Of Artificial Intelligence S. Matthew Liao all chapters

Uploaded by

Copyright:

Available Formats

Download ebooks file Ethics Of Artificial Intelligence S. Matthew Liao all chapters

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Download ebooks file Ethics Of Artificial Intelligence S. Matthew Liao all chapters

Uploaded by

Copyright:

Available Formats

Download the full version of the textbook now at textbookfull.

Ethics Of Artificial Intelligence S. Matthew

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

AI for the Good: Artificial Intelligence and Ethics

Artificial Intelligence By Example Develop machine

Distributed Artificial Intelligence Second International

Children and Sustainable Development Ecological Education

NASCLA Contractors Guide to Business, Law and Project

Sunnis And Shi'a: A Political History Of Discord 1st

The History of Sociology in Britain: New Research and

Published in the United States of America by Oxford University Press

© Oxford University Press 2020

All rights reserved. No part of this publication may be reproduced, stored in

You must not circulate this work in any other form

Library of Congress Cataloging-in-Publication Data

Jean-François Bonnefon is a Research Director at the French Centre National de la

Nick Bostrom is a Professor at Oxford University, where he leads the Future of

Andrew Critch is a Research Scientist at the Center for Human-Compatible AI in the

Hanna Gunn is an Assistant Professor in Cognitive and Information Sciences at the

Aaron James is a Professor of Philosophy at the University of California, Irvine. He is

Frances M. Kamm is the Henry Rutgers University Professor of Philosophy and

Nicholas Mattei is an Assistant Professor of Computer Science at Tulane University. His

Steve Petersen is an Associate Professor of Philosophy at Niagara University. In addition

Stuart Russell is a Professor of Computer Science at the University of California, Berkeley,

Susan Schneider is the NASA/Baruch Blumberg Chair of Astrobiology and Technological

Eric Schwitzgebel is a Professor of Philosophy at University of California, Riverside, and

Azim Shariff is an Associate Professor of Psychology at the University of British

Stephen Wolfram is the creator of Mathematica, Wolfram|Alpha and the Wolfram

Artificial intelligence (AI) is progressing rapidly. AI can now recognize objects in

develop “risk prediction” tools to assist with decisions on criminal sentencing,

In recent years, the ethical implications of near-term and long-term AI have

I.2. Key Concepts in Machine Learning

To begin, let me say something about what AI is.29 There is no agreed-upon

I.3. Vulnerabilities in Machine Learning

As impressive as machine learning is, it also suffers certain limitations. As we

I.3.1. Machine Learning Is Data Hungry

I.3.2. Garbage In/Garbage Out

I.3.3. Faulty Algorithms

I.3.4. Deep Learning Is a Black Box

I.3.5. Machine Learning Is Weak AI

Perhaps recognizing that this is a limitation of current AIs, companies typi-

I.4. Human Vulnerabilities

You might also like