Download the Full Version of textbook for Fast Typing at textbookfull.

com

Fuzziness in Information Systems How to Deal with

Crisp and Fuzzy Data in Selection Classification
and Summarization 1st Edition Miroslav Hudec
(Auth.)
https://textbookfull.com/product/fuzziness-in-information-
systems-how-to-deal-with-crisp-and-fuzzy-data-in-selection-
classification-and-summarization-1st-edition-miroslav-hudec-
auth/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Classification and Data Science in the Digital Age 1st

Edition Paula Brito

https://textbookfull.com/product/classification-and-data-science-in-
the-digital-age-1st-edition-paula-brito/

textboxfull.com

Decision Support Systems VII. Data, Information and

Knowledge Visualization in Decision Support Systems: Third
International Conference, ICDSST 2017, Namur, Belgium, May
29-31, 2017, Proceedings 1st Edition Isabelle Linden
https://textbookfull.com/product/decision-support-systems-vii-data-
information-and-knowledge-visualization-in-decision-support-systems-
third-international-conference-icdsst-2017-namur-belgium-
may-29-31-2017-proceedings-1st-e/
textboxfull.com

A First Course in Fuzzy Logic Fuzzy Dynamical Systems and

Biomathematics Theory and Applications 1st Edition Laécio
Carvalho De Barros
https://textbookfull.com/product/a-first-course-in-fuzzy-logic-fuzzy-
dynamical-systems-and-biomathematics-theory-and-applications-1st-
edition-laecio-carvalho-de-barros/
textboxfull.com

Statistical and Fuzzy Approaches to Data Processing, with

Applications to Econometrics and Other Areas: In Honor of
Hung T. Nguyen's 75th Birthday Vladik Kreinovich
https://textbookfull.com/product/statistical-and-fuzzy-approaches-to-
data-processing-with-applications-to-econometrics-and-other-areas-in-
honor-of-hung-t-nguyens-75th-birthday-vladik-kreinovich/
textboxfull.com
Recent Advances in Intuitionistic Fuzzy Logic Systems and
Mathematics Said Melliani

https://textbookfull.com/product/recent-advances-in-intuitionistic-
fuzzy-logic-systems-and-mathematics-said-melliani/

textboxfull.com

Fuzzy Machine Learning Algorithms for Remote Sensing Image

Classification 1st Edition Anil Kumar

https://textbookfull.com/product/fuzzy-machine-learning-algorithms-
for-remote-sensing-image-classification-1st-edition-anil-kumar/

textboxfull.com

Modern Data Mining Algorithms in C++ and CUDA C: Recent

Developments in Feature Extraction and Selection
Algorithms for Data Science 1st Edition Timothy Masters
https://textbookfull.com/product/modern-data-mining-algorithms-in-c-
and-cuda-c-recent-developments-in-feature-extraction-and-selection-
algorithms-for-data-science-1st-edition-timothy-masters/
textboxfull.com

Records Classification: Concepts, Principles and Methods.

Information, Systems, Context Umi Asma Mokhtar And Zawiyah
Mohammad Yusof (Auth.)
https://textbookfull.com/product/records-classification-concepts-
principles-and-methods-information-systems-context-umi-asma-mokhtar-
and-zawiyah-mohammad-yusof-auth/
textboxfull.com

Information Technology and Data in Healthcare Using and

Understanding Data 1st Edition David Hartzband (Author)

https://textbookfull.com/product/information-technology-and-data-in-
healthcare-using-and-understanding-data-1st-edition-david-hartzband-
author/
textboxfull.com
Miroslav Hudec

Fuzziness in
Information
Systems
How to Deal with Crisp and Fuzzy
Data in Selection, Classification, and
Summarization
Fuzziness in Information Systems
Miroslav Hudec

Fuzziness in Information
Systems
How to Deal with Crisp and Fuzzy Data
in Selection, Classification,
and Summarization

123
Miroslav Hudec
Faculty of Economic Informatics
University of Economics in Bratislava
Bratislava
Slovakia

ISBN 978-3-319-42516-0 ISBN 978-3-319-42518-4 (eBook)

DOI 10.1007/978-3-319-42518-4

Library of Congress Control Number: 2016948117

© Springer International Publishing Switzerland 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG Switzerland
I dedicate this book to Martina. I would like
to skip writing reasons, because this part
should be shorter than a book chapter.
Foreword

It is a pleasure to write this foreword because this book is an important contribution

to the literature on applications of fuzzy models. There are many books dealing with
fuzzy sets in a general way but this work is an essential contribution to the
description of fuzziness in information systems.
Usually in statistical information systems data are stored as numbers which
pretend a precision which is not justified, because real data are frequently not
available as precise numbers but they are more or less non-precise. This imprecision
is different from errors and it is best modelled by the so-called fuzzy numbers,
which are special fuzzy subsets of the set of real numbers. To describe fuzziness
in quantitative mathematical terms, is an important innovation in science and
management.
When Karl Menger introduced the generalization of classical sets in the year
1951 by generalizing the indicator function of classical sets, this was a theoretical
concept and it took many years until practical applications of these generalized sets
came up. An important step was the paper by Lotfi Zadeh in 1965 when he
introduced the name fuzzy set and defined generalized set operations based on the
defining functions of fuzzy sets. These defining functions were called membership
functions and in the last decades of the twentieth century enormous research
activities developed the theory and applications of fuzzy sets. The concept of
fuzziness was extended to generalize logics, and the calculus of fuzzy logics was
created. In the meantime fuzzy concepts are applied in many scientific fields, for
example in civil engineering for risk analysis of structures, in medical science for
diagnostic systems, in measurement science to describe results of precision mea-
surement, in statistics for the description and analysis of real data, in Bayesian
analysis to model fuzzy a priori information, in information science to describe
fuzzy information and to formulate fuzzy questions.
Chapter 1 of the book gives an introduction to fuzzy sets and fuzzy logic,
linguistic variables, fuzzy quantifiers, and related references.
Chapter 2, Fuzzy Queries, considers the way from crisp to fuzzy queries, the
construction of fuzzy sets for flexible conditions, the conversion of fuzzy conditions

vii
viii Foreword

to SQL ones, the calculation of matching degrees, empty and overabundant

answers, and some issues related to practical realization.
Chapter 3, Linguistic Summaries, explains the benefits of linguistic summa-
rization (LS), the basic structure of LS, relative quantifiers, quality measures of LS,
applicability of LS, and building summaries.
Chapter 4, Fuzzy Inference, is devoted to fuzzy models for control systems.
After introducing fuzzy inference engines, the chapter presents a section on fuzzy
classification, and concludes with remarks to applications.
The next chapter, Fuzzy Data in Relational Databases, is central to the book. It
starts with the classical relational databases, has a section on fuzziness in the real
world, explains fuzzy databases and their basic model, fuzzy data in traditional
relational databases, aggregation functions in queries, and linguistic summaries on
fuzzy data.
Chapter 6, Perspectives, Synergies and Conclusion, briefly explains the rela-
tionship between fuzzy inference and fuzzy databases and linguistic summaries as
well as fuzzy queries.
The references at the end of each chapter are helpful for further reading.
Appendixes, Illustrative Interfaces and Applications for Fuzzy Queries and
Illustrative Interfaces and Applications for Linguistic Summaries, and an Index are
helpful for the reader.
This book explains important applications of fuzzy logic in information systems.
Congratulations to the author for this valuable and up-to-date contribution to
information science.

Wien Reinhard Viertl

April 2016
Preface

The increasing use of information systems by governmental agencies and businesses

has created mountains of data that contain potentially valuable knowledge.
Admittedly, these data do constitute “golden mines” which should be swiftly and
efficiently processed and interpreted to be useful. Users (e.g. decision-makers) would
like to efficiently reveal relevant data. Moreover, users are often not interested in
large sheets of figures, but in knowledge that is usually overshadowed by large
amount of data.
People can relatively easily answer imprecise questions like, is it true that most
of tall persons in the room wear blue or green shirts? Different hues of these
colours as well as the meaning of the vague term tall people are not limitations for
solving this task. However, if we want to know, which of these two sentences: most
of young commuters commute short distances; most of medium aged commuters
commute short distances better explains the commuting behaviour, then we have to
adapt this query to mine the truth value form the data. The same holds for querying
cheap hotel with good references and if possible near to the city centre and
common-sense reasoning: if customer buys products very often, then provide high
discount.
The initial research in the theory of fuzzy sets and fuzzy logic was motivated by
the perception that traditional computing techniques are not effective in dealing
with problems, in which vagueness, imprecision and subjectivity are immanent, and
therefore should not be neglected. These types of uncertainty are commonly called
fuzziness.
According to Prof. Zadeh, four principal rationales for handling fuzziness exist.
Two of them, which are relevant for this book, are: “don’t need rationale” and
“don’t know rationale”. In the former, the tolerance for imprecision is in accord
with the remarkable human capability to solve variety of tasks without precise
calculations. For example, summarizing data by short questions of natural
language; creating queries with flexible conditions and approximate inference. In
the latter, the values of attributes are not known with sufficient precision to justify
the use of traditional databases for storing these data. Many data cannot be ade-
quately expressed as precise numbers or as one linguistic term, due to non-sharp

ix
x Preface

boundaries of observations, tendency of people to estimate or guess answers in

surveys and tolerance intervals of measurement instruments. Therefore, the data are
often vague and include both quantitative and qualitative elements. Storing these
data as crisp values might cause loss of valuable information.
Keeping in mind the aforementioned facts, fuzzy queries, fuzzy inference pro-
cesses, linguistic summaries and managing fuzzy data in information systems could
be the option. We have chosen these areas, because businesses of all sizes and
governmental agencies cope with them in their work. The motivation for this book
has arisen from the author’s experience in teaching courses of fuzzy logic for
business informatics and database design and in research and development of
information systems and data mining applications mainly for the official statistics
purposes. Furthermore, many small- and medium-sized enterprises cannot afford
sophisticated tools or experts for information systems and data mining, even though
they are aware of limitation of sharp boundaries in data analyses. Many tasks can be
solved in a classical way, but their complexity becomes high. The complexity of the
problem can be reduced by including the intensity of the examined property. This
permits us to discern elements with the same property, based on the intensity
matching it.
Roughly, the intent of the book could be depicted in Fig. 1. The usual scenario is
that user wants to retrieve data or summarized information from a database.
Furthermore, user might be interested to classify data. Often user is not aware of the
nature of collected data or cannot determine sharp criteria. In addition, all data
including vague ones are usually stored as crisp values.
In the book we examine these approaches theoretically as well as on the
municipal statistics data. The latter is illustrated in appendixes. These data are
suitable source, due to larger number of municipalities, which are often very similar
in several attributes. Second reason is that some of attributes are fuzzy in their
nature, but are limited to crisp values.
We should not expect that domain experts are familiar with the fuzzy logic
theory. Therefore, the book demonstrates developing user-friendly interfaces to
allow users exploring advantages of fuzzy logic in their tasks. Furthermore,

Fig. 1 Intent of the book

Preface xi

companies usually keep data in relational databases. We should keep this in mind
during adapting database to cope with the fuzzy data.
The main target audience of the book are students, researchers and practitioners
working in fields of data analysis, database design and business intelligence. This
book does not go deeply into the foundation and the mathematical theory of fuzzy
logic and relational algebra (e.g. theorems’ proofs). Hence, intermediate knowledge
of fuzzy logic and relational databases is recommended.
The book is divided into six chapters in the following way. Chapter 1 is focused
on the theory of fuzzy sets and fuzzy logic to a level, which is advisable to know in
order to proceed to next chapters. Readers skilled in fuzzy logic theory can skip this
chapter.
Chapter 2 is devoted to flexible queries. The following aspects of flexible queries
are examined: constructing fuzzy sets for query conditions; aggregation operators
for commutative and non-commutative conditions with and without priorities;
dealing with empty and overabundant answer problems and issues related to
practical realizations.
Chapter 3 is dedicated to linguistic summaries. We start with the basic linguistic
summary and build more complex ones. To meet this goal, selecting appropriate
aggregations, implications for preferences and issues related to construction of
membership functions are examined. Quality measures of created summaries are
also considered. Finally, several possible applicabilities are discussed.
Chapter 4 presents fuzzy logic control architecture adjusted to the aims of
business and governmental agencies. It shows fuzzy rules, construction of fuzzy
sets and procedures for solving inference tasks by generalized modus ponens. In the
first part we explain reasoning procedures. In the second part fuzzy expert systems
are discussed. In the last part classification by IF-THEN rules is examined.
Chapter 5 covers fuzzification of classical relational databases. We briefly
review classical relational databases and fuzzy database models. The emphasis is on
storing fuzzy data in classical relational databases in a way that existing data and
normal forms are not affected. Furthermore, practical aspects of user-friendly
interfaces for storing, updating, querying and summarizing are examined.
Chapter 6 shortly discusses possible integration of fuzzy queries, summarization
and inference related to crisp and fuzzy databases. Use of these approaches in a
complementary, rather than competitive way, can support variety of tasks.
Finally, we suppose that the book will provoke at least some interest to continue
research and also will be of support for developing tailored applications commu-
nicating with users by easy-to-use interfaces. Maybe the next generations of rela-
tional database management systems and applications will include many fuzzy
characteristics and users will enjoy easy-to-use interfaces for fuzzy queries, fuzzy
inferences, fuzzy summarization, fuzzy recommending and so on, without the
necessity of knowing mathematics of fuzzy logic. We hope that the book will
contribute to this field with a membership degree greater than 0.25.

Bratislava Miroslav Hudec

April 2016
Acknowledgements

Advice and suggestions touching theoretical as well as practical implications related

to my work were valuable source for this research. In this direction, I express my
deep appreciation to Tomáš Bacigál, Simona Balbi, Mojca Bavdaž, Piet Daas, Boris
Delibašić, Erich Peter Klement, Gabriela Kristová, Andreas Meier, Radko Mesiar,
Dušan Praženka, Dragan Radojević, Agnieszka Stawinoga, František Sudzina,
Matthias Templ, Valentin Todorov, Vanessa Torres van Grinsven, Steven Vale,
Enrique Herrera-Viedma, Reinhard Viertl, Miljan Vučetić and Mirko Vujošević.
I also address my gratitude to the authors of publications listed in the references.
Their contributions were inspiration for my work and are reflected in the book.
Concerning data required for experiments I thank Statistical Office of the Slovak
Republic for providing the data. I am grateful to Eleonora Vallová for proofreading
and editing my manuscript.
I am also grateful to Springer for offering me opportunity to publish my humble
view and opinion about the fuzziness in their publishing house and especially Ralf
Gerstner for the cooperation during the whole process.
My final thanks go to the heads of the Faculty of Economic Informatics and
especially Department of Applied Informatics for providing me with space for
research which enabled me to spend long hours in researching and writing the book.

xiii
Contents

1 Fuzzy Set and Fuzzy Logic Theory in Brief . . . . . . . . . . . . . . . . . . . . 1

1.1 From Crispness to Fuzziness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Fuzzy Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Properties of Fuzzy Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.2 Types of Fuzzy Sets (Membership Functions). . . . . . . . . . . 7
1.2.3 Operations on Fuzzy Sets . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.2.4 Fuzzy Numbers and Fuzzy Arithmetic . . . . . . . . . . . . . . . . 13
1.2.5 Measures of Fuzziness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.2.6 Fuzzy Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.3 Fuzzy Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.3.1 Fuzzy Conjunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.3.2 Fuzzy Negation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.3.3 Fuzzy Disjunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.3.4 Fuzzy Implication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4 Linguistic Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.5 Fuzzy Quantifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.6 Some Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2 Fuzzy Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.1 From Crisp to Fuzzy Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2 Construction of Fuzzy Sets for Flexible Conditions . . . . . . . . . . . . 37
2.3 Converting Fuzzy Conditions to SQL Ones . . . . . . . . . . . . . . . . . . 40
2.4 Calculation of Matching Degrees . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4.1 Independent Conditions Aggregated
by the “And” Operator . . . . . . . . . . . . . . . . . . . . . . . . .... 42
2.4.2 Fuzzy Preferences Among Atomic Query Conditions . .... 44
2.4.3 Answer to the Second Atomic Condition Depends
on the Answer to the First One . . . . . . . . . . . . . . . . . . .... 47
2.4.4 Constraints and Wishes . . . . . . . . . . . . . . . . . . . . . . . . .... 50
2.4.5 Quantified Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 55

xv
xvi Contents

2.4.6 Querying Changes of Attributes over Time . . . . . . . . . . . . . 56

2.5 Empty and Overabundant Answers . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5.1 Empty Answer Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.5.2 Overabundant Answer Problem . . . . . . . . . . . . . . . . . . . . . . 61
2.6 Some Issues Related to Practical Realization . . . . . . . . . . . . . . . . . 62
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3 Linguistic Summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.1 Benefits and Protoforms of Linguistic Summarization . . . . . . . . . . 67
3.2 The Basic Structure of LS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.3 Relative Quantifiers in Summaries . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.4 LS with Restriction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.5 Mining Linguistic Summaries of Interest . . . . . . . . . . . . . . . . . . . . 75
3.6 Quality Measures of LSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.6.1 Quality Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.6.2 Aggregation of Quality Measures . . . . . . . . . . . . . . . . . . . . 83
3.6.3 Influence of Constructed Fuzzy Sets and T-Norms
on Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.7 Some Applicability of LS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.7.1 Quantified Queries (LS as a Nested Condition) . . . . . . . . . . 87
3.7.2 Generating IF-THEN Rules . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.7.3 Estimation of Missing Values . . . . . . . . . . . . . . . . . . . . . . . 93
3.8 Building Summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4 Fuzzy Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.1 From Classical to Fuzzy Inference . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2 Fuzzy Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.2.1 Inference Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.2.2 Defuzzification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.2.3 Illustrative Examples and Issues . . . . . . . . . . . . . . . . . . . . . 114
4.3 Fuzzy Inference Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.3.1 Mamdani Model (Logical Model) . . . . . . . . . . . . . . . . . . . . 120
4.3.2 Sugeno Model (Functional Model) . . . . . . . . . . . . . . . . . . . 122
4.4 Fuzzy Rule-Based System Design . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.5 Fuzzy Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.5.1 A View on Crisp Classification . . . . . . . . . . . . . . . . . . . . . . 126
4.5.2 Managing Fuzzy Classification . . . . . . . . . . . . . . . . . . . . . . 127
4.5.3 Fuzzy Classification by Fuzzy Queries . . . . . . . . . . . . . . . . 130
4.6 Remarks to Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5 Fuzzy Data in Relational Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.1 Classical Relational Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.2 Fuzziness in the Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.3 Fuzzy Databases: An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Contents xvii

5.4 Basic Model of Fuzzy Database . . . . . . . . . . . . . . . . . . ......... 143

5.4.1 Structure of Basic Model . . . . . . . . . . . . . . . . . ......... 144
5.4.2 Querying Basic Model . . . . . . . . . . . . . . . . . . . ......... 145
5.5 Fuzzy Data in Traditional Relational Databases
Managed by Fuzzy Meta Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.5.1 Creating Fuzzy Meta Model . . . . . . . . . . . . . . . . . . . . . . . . 148
5.5.2 Storing and Representing Tuples . . . . . . . . . . . . . . . . . . . . . 151
5.5.3 Inserting Fuzziness into Existing Databases . . . . . . . . . . . . 155
5.5.4 Managing Fuzziness in Data and in Inference
Rules by the Same Database . . . . . . . . . . . . . . . . . . . . . . . . 157
5.6 Querying Fuzzy Relational Databases . . . . . . . . . . . . . . . . . . . . . . . 160
5.6.1 Aggregation Functions in Queries . . . . . . . . . . . . . . . . . . . . 160
5.6.2 Query Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
5.7 Linguistic Summaries on Fuzzy Data . . . . . . . . . . . . . . . . . . . . . . . 171
5.8 Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
6 Perspectives, Synergies and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 177
6.1 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
6.2 Synergy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
6.2.1 Fuzzy Inference and Fuzzy Databases . . . . . . . . . . . . . . . . . 178
6.2.2 Linguistic Summaries and Fuzzy Inference . . . . . . . . . . . . . 180
6.2.3 Fuzzy Queries for Fuzzy Classification and Summaries . . . 181
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Appendix A: Illustrative Interfaces and Applications
for Fuzzy Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Appendix B: Illustrative Interfaces and Applications
for Linguistic Summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Acronyms

1NF First Normal Form

2NF Second Normal Form
3NF Third Normal Form
COG Center of Gravity
COM Center of Maxima
CRM Customer Relationship Management
DBMS Database Management System
DDL Data Definition Language
DNF Disjunctive Normal Form
FIS Fuzzy Inference System
FMB Fuzzy Meta-knowledge Base
FMM Fuzzy Meta Model
GLC Generalized Logical Condition
GMP Generalized Modus Ponens
HDM Height Defuzzification Method
HOF Half of Field
IBA Interpolative Realization of Boolean Algebra
LOM Left of Maxima
LSs Linguistic Summaries
MOM Mean of Maxima
NFE Necessarily Fuzzy Equal to
NFL Necessarily Fuzzy Less than
OWA Ordered Weighted Averaging Operator
PFE Possibly Fuzzy Equal to
PFL Possibly Fuzzy Less than
RDBMS Relational Database Management System
ROM Right of Maxima
SQL Structured Query Language

xix
Notations

In each field variables, sets, properties, functions, etc. are either marked by letters
using an (informal) agreement, or by letters the authors decided to use in their
seminal papers. In this textbook we have decided to keep usual notation of most
used terms and adjust the notation of other terms in order to avoid misinterpretation.
As it is not always possible, some letters are not used for a single term only, but the
explanation on their usage avoids misinterpretation of used letters and indexes.

A Attribute, fuzzy set, answer to query


A Complement of fuzzy set
AðaÞ a-cut of fuzzy set
Ac Accuracy
a; m; b Parameters of fuzzy sets
B Fuzzy set
core(A) Core of fuzzy set
card(A) Cardinality of fuzzy set, jAj
c Negation
cs Standard negation
cg Gödel negation
cdg Dual Gödel negation
C Coverage
iC Coverage index
D Domain, distance
d Measure of fuzziness
Ec Specificity
F Fuzzification operator, fuzzy set
h Height of fuzzy set
H Highest value of attribute in a database
isKD Kleene–Dienes implication
(continued)

xxi
xxii Notations

(continued)
isL Łukasiewicz implication
iqZ Zadeh implication
irGd Gödel implication
irGg Goguen implication
L Lowest value of attribute in a database
N Negative preference
O Outlier measure
P Predicate, positive preference
Q Quantifier
Qc Quality of summary
r Database tuple
R Relation, rule, restriction
s t-conorm or s-norm
sm Maximum s-norm
sa Algebraic sum
sL Łukasiewicz s-norm
sd Drastic s-norm
supp(A) Support of fuzzy set
S Summarizer, simplicity
t t-norm
tm Minimum t-norm
tp Product t-norm
tL Łukasiewicz t-norm
td Drastic product
tnM Nilpotent minimum t-norm
T Transformation
U Usefulness
v Validity of linguistic summary
w Weight
X Universe of disclosure, universal set
a And if possible operator
b Or else operator, threshold
d Firing degree of rule
e Length of slope of fuzzy set
l Membership function
u Characteristic function
h Length of flat segment of fuzzy set
N Set of natural numbers
R Set of real numbers
Chapter 1
Fuzzy Set and Fuzzy Logic Theory in Brief

Abstract A set consists of elements sharing the same property. This property
is essential for setting set boundaries. Hence, the following question appears:
Can we always unambiguously define these boundaries? The answer is, no. We
can unambiguously define a set containing all municipalities belonging to the
district D. Municipality either belongs to the district D (from administrative point
of view), or does not belong. However, for the set expressing high distance we
cannot clearly define sharp boundary to distinguish high from non-high distance.
This section begins with the classical sets in order to smoothly continue to fuzzy
sets. Next, relevant properties and operations of fuzzy sets are discussed. Further, the
concept of fuzzy number, as a subcategory of fuzzy sets, is explained. Fuzzy sets and
many-valued logics are basis for fuzzy logic. Fuzzy logic facilitates commonsense
reasoning with imprecise predicates expressed as fuzzy sets. In the second part
fuzzy conjunction, negation, disjunction, implication and quantifiers are examined.
Mentioned concepts are used throughout the book.

1.1 From Crispness to Fuzziness

The crisp set is a collection of elements which share the same property. The principal
concept in the set theory is belonging or membership to a set. If an element of the
universal set X belongs to the set A, we simply write x ∈ A. If x is not a member
of A, we write x ∈ / A. It means that belonging to a set should be clear [47].
A crisp set can be described by several methods. The listing method lists all
elements by putting them into the braces: A = {strongly agree, agree, do not know,
disagree, strongly disagree}, where A denotes all possible answers in a questionnaire,
for example. The order of elements is irrelevant. This method is feasible only if a
set contains finite number of elements. Otherwise, a set should be described by the
membership rule (property or predicate which has to be satisfied)

A = {x ∈ X | x satisfies property P} (1.1)

© Springer International Publishing Switzerland 2016 1

M. Hudec, Fuzziness in Information Systems,
DOI 10.1007/978-3-319-42518-4_1
2 1 Fuzzy Set and Fuzzy Logic Theory in Brief

where A denotes a set of all x such that x satisfies property P e.g. A = {x ∈ R | x >
350}, where R is a set of real numbers.
Finally, crisp set A can be defined by the characteristic function ϕ A that matches
each element of the universal set X to the set A in the following way:

ϕ A (x) : X → {0, 1} (1.2)

Example 1.1 For illustrative example, a governmental agency decided to financially

support highly polluted municipalities. In order to discern a set of highly polluted
municipalities (H P), the agency should define property explaining the H P set. If
high pollution means more or equal than 20 mg of the measured pollutant, then the
crisp set H P is defined as: H P = {x ∈ X | x ≥ 20}, where X is the universal set of
all possible pollutions (real values greater or equal zero). Membership to a set H P
can be expressed as a characteristic function ϕ(x) in the following way:

0 for x < 20
ϕ H P (x) = .
1 for x ≥ 20
The set expressing high pollution is shown in Fig. 1.1. At the first glance we see
the drawback of crisp sets. A municipality having the value of 19.93 mg will not
receive any financial support (it is treated in the same way as municipality having
value of e.g. 0.2 mg), whereas municipality having pollution of 20 mg will receive
full financial support as well as municipality having pollution of 60 mg. Furthermore,
measured pollution is expected to be a crisp real number. But in reality, either it is
not possible to realize extremely precise measuring [31], or values are fuzzy in their
nature. It especially holds for the environmental data [44]. 

Vagueness concerning the description of the semantic meaning of events or phe-

nomena is called fuzziness [57]. Hence, uncertainty is not based on randomness or
probability; it cannot be presented as a crisp value. Vague terms such as high, cheap,
medium, around m (m ∈ R), heap share three interrelated features of vagueness
[18]: admit borderline cases, lack sharp boundaries and are susceptible to sorties
paradoxes.

Fig. 1.1 Crisp set high

pollution
1.1 From Crispness to Fuzziness 3

In Fig. 1.1, two kinds of fuzziness are neglected: fuzziness in data and fuzziness
in belonging to a set [16]. If managing pollution by sharp sets remains, several small
intervals e.g. [0, 15)—no support; [15, 18)—30 % of full support; [18, 22)—65 % of
full support; [22, 25)—85 % of full support; etc., to ensure that similar municipalities
receive similar support can be employed.
However, when we include additional attributes, such as unemployment or number
of respiratory diseases, then managing rules by crisp sets become more complex.
The interpretation of fuzzy sets [56] has arisen from the generalization of the
classical sets to embrace the vague notions and unclear boundaries. It may not be
always clear, if an element x belongs to a set A, or not. Thus, its membership may be
measured by a degree, commonly known as the membership degree taking a value
from the unit interval by agreement.
Consequently, a fuzzy set A over the universe of discourse X is defined by function
μ A that matches each element of the universe of discourse with its membership degree
to the set A

μ A (x) : X → [0, 1] (1.3)

where μ A (x) = 0 says that an element x definitely does not belong to a fuzzy set
A, μ A (x) = 1 says that x without any doubt is member of fuzzy set A. Higher
value of μ A (x) indicates the higher degree of membership of an element x to a fuzzy
set A. Each fuzzy set is defined by one membership function. A membership function
maps each element of the universal set X into real numbers from the [0, 1] interval.
We should emphasize that the universal set X is always a crisp set [21].
A fuzzy set can be defined as a set of ordered pairs

A = {(x, μ A (x)) | x ∈ X ∧ μ A (x) ∈ (0, 1]} (1.4)

When the universal set is finite, fuzzy set constructed on this universal set can be
expressed by counting the elements and their respective membership degrees

μ A (x1 ) μ A (x2 ) μ A (xn )

A= + + ··· + (1.5)
x1 x2 xn

Example 1.2 Let us consider highly polluted municipalities (Example 1.1) from the
fuzzy sets point of view. The crisp set H P from Fig. 1.1 could be converted into the
fuzzy set F H P in the following way (Fig. 1.2):
⎧
⎪ 0 for x ≤ 15
⎨
x − 15
μ F H P (x) = for 15 < x < 20 (1.6)
⎪
⎩ 5
1 for x ≥ 20

In this way the soft transition between belonging and non-belonging to a set is
ensured. 
4 1 Fuzzy Set and Fuzzy Logic Theory in Brief

Fig. 1.2 Fuzzy set high

pollution

Example 1.3 Let us consider respondent’s consent to a question in a questionnaire.

The fuzzy set consent (A) can be expressed as A={(strongly agree, 1), (agree, 0.65),
(do not know, 0.15)}. Ordered pairs (disagree, 0) and (strongly disagree, 0) are not
included into fuzzy set A. 

1.2 Fuzzy Sets

The first publications of fuzzy set theory by Zadeh [56] and Goguen [11] were
focused on the generalization of the classical notion of sets and propositions in order
to mathematically cover fuzziness. Although the needs as well as importance of
managing fuzziness were recognized earlier (e.g. [39]), the acceptance by scientific
and practitioners communities was not high, especially at the beginning. Reasons for
accepting and non-accepting fuzzy sets and related topics were summarized in [52].

1.2.1 Properties of Fuzzy Sets

In this section properties relevant for the next sections are examined.
Scalar and relative scalar cardinality
For any fuzzy set A defined on a finite universal set X we define its scalar cardinality
by the formula

card(A) = |A| = μ A (x) (1.7)
x∈X

The scalar cardinality of fuzzy set (1.7) is a generalization of the classical cardi-
nality. Elements of universal set belong to the fuzzy sets with different membership
degrees and therefore we cannot count elements of a set A, but their respective mem-
bership degrees should be summed. Some authors refer to |A| as the sigma count of
A [21].
1.2 Fuzzy Sets 5

The relative scalar cardinality is defined by the formula:

card(A) |A|
||A|| = = (1.8)
card(X ) |X |

where card(A) is defined in (1.7) and card(X ) represents the number of elements
in X . These cardinalities are broadly used in e.g. linguistic summaries.
The third type of cardinality is fuzzy cardinality expressed as ordered pair: number
of elements belonging to a particular α-cut and α-cut [22] when universal set is a
finite one. Cardinalities are closely examined in e.g. [45].
Scalar cardinality of a fuzzy set can be expressed as the area bounded by the
membership function of fuzzy set and the x-axis [35]. This approach is demonstrated
on the trapezoidal fuzzy set in Sect. 1.2.2.
Support
The support of a fuzzy set A is the crisp set with the following property:

supp(A) = {(x ∈ X | μ A (x) > 0} (1.9)

This property is broadly used in flexible queries, among others.

Core
The core of a fuzzy set A is the crisp set with the following property:

core(A) = {(x ∈ X | μ A (x) = 1} (1.10)

In the fuzzy sets literature the term kernel is used as a synonym for the core.
Height
The height is the highest value of membership degree of all elements in the considered
fuzzy set A, i.e.

h(A) = sup μ A (x) (1.11)

x∈X

From (1.10) and (1.11) we can infer that if the core is not an empty set, then the
height is equal to the value of 1. The opposite does not always hold.

Example 1.4 A heap of maize grains obviously contains large number of grains.
Because by crisp sets we cannot unambiguously discern the two sets, heap and non-
heap, a fuzzy set should be applied. For example, one could agree that 2 000 grains
is a large quantity (heap), and between 1 and 2 000, the belonging to a heap grows.
Thus, the membership function of the set heap could be:

0 for x = 0
μheap (x) = 2 (1.12)
π
arctan(0.09 · x) for x > 0
6 1 Fuzzy Set and Fuzzy Logic Theory in Brief

Fig. 1.3 Heap of maize grains explained by fuzzy set

The graphical interpretation of the resulting membership function is shown

in Fig. 1.3. This figure explains the core and height properties of fuzzy set. In
this case the support (1.9) is unlimited. The height (1.11) is equal to 1, because
lim x→∞ π2 arctan(0.09 · x) = 1. The core (1.10) is an empty set. Although,
h(heap) = 1, this value is not reached by any element.

Normalized fuzzy set

Fuzzy set A is normalized, if the membership degree of at least one element is equal
to 1, i.e.:

∃x ∈ X, μ A (x) = h(x) = 1 (1.13)

Crossover point
The element xcp of a fuzzy set A that has a membership degree equal to 0.5 represents
the crossover point, i.e.:

xcp = {x ∈ X |μ A (x) = 0.5} (1.14)

α-cut and strong α-cut

One of the important concepts used in fuzzy sets is the α-cut. The α-cut A(α) and its
restrictive variant strong α-cut A(α+) are defined in the following way:

A(α) = {x ∈ X | μ A (x) ≥ α} (1.15)

A(α+) = {(x ∈ X | μ A (x) > α} (1.16)

where α ∈ [0, 1].

The α-cut of a fuzzy set A is a crisp set containing all the elements of the X whose
membership degrees in A are greater than or equal than the specified value of α. This
1.2 Fuzzy Sets 7

Fig. 1.4 Convex and non-convex fuzzy sets

property is used in many directions, e.g. working with elements which significantly
belong to a fuzzy set.
Convexity of fuzzy sets
A fuzzy set is convex, if and only if [56]:

μ A (λx + (1 − λ)y) ≥ min(μ A (x), μ A (y)) (1.17)

for all x and y ∈ X and all λ ∈ [0, 1]. Convex and non-convex fuzzy sets are plotted
in Fig. 1.4.

1.2.2 Types of Fuzzy Sets (Membership Functions)

Membership functions are classified into two main groups [10]: linear and Gaussian
or curved. All membership functions explained in this section are normalized fuzzy
sets.
Triangular fuzzy set (Fig. 1.5) is defined by its lower limit a, its upper limit b
and the modal (highest) value m as

Fig. 1.5 Triangular fuzzy

set (membership function)
8 1 Fuzzy Set and Fuzzy Logic Theory in Brief

Fig. 1.6 Gaussian fuzzy set

(membership function)

⎧
⎪
⎪ 1 for x = m
⎪
⎪ x −a
⎪
⎨ for a < x < m
m
μ A (x) = b − −a (1.18)
⎪ x
⎪
⎪ for m < x < b
⎪
⎩b−m
⎪
0 for x ≤ a ∨ x ≥ b

Gaussian fuzzy set (Fig. 1.6) is defined by the modal value (centre) m and width
k as

μ A (x) = e−k(x−m)
2
(1.19)

The bell of the Gaussian function depends on the value k. If the value k is lower,
then the bell is narrower.
Trapezoidal fuzzy set (Fig. 1.7) is defined by its lower limit a, its upper limit b,
and the flat segment [m 1 , m 2 ] representing the highest value of height (1.11) as
⎧
⎪
⎪ 1 for m 1 ≤ x ≤ m 2
⎪ x −a
⎪
⎪
⎨ for a < x < m 1
μ A (x) = mb1−−xa (1.20)
⎪
⎪ < <
⎪
⎪ for m 2 x b
⎩ b − m2
⎪
0 for x ≤ a ∨ x ≥ b

Fig. 1.7 Trapezoidal fuzzy

set
1.2 Fuzzy Sets 9

Fig. 1.8 L fuzzy set

The scalar cardinality (1.7) of trapezoidal fuzzy set is calculated by its area in the
following way [35]:
(m 1 − a) + (b − m 2 )
card(A) = area(A) = (m 2 − m 1 ) + (1.21)
2
where parameters are the same as in (1.20). In case of triangular fuzzy set the left
part is not used.
Trapezoidal, triangular and Gaussian fuzzy sets are suitable for modelling con-
cepts such as medium value or approximate m, where m is a real number. The sup-
port (1.9) of the Gaussian fuzzy set is spread over the whole universe of disclosure,
although with values close to 0 near the edges of the universe of disclosure (or far
from the value of m). This could be a problem in fuzzy relational databases, which
is discussed later on.
L fuzzy set (Fig. 1.8) is defined by two parameters, m and b, in the following
way:
⎧
⎪ 1 for x ≤ m
⎨ b−x
μ A (x) = for m < x < b (1.22)
⎪
⎩b−m
0 for x ≥ b

L fuzzy set is suitable for defining sets expressing small values of the analysed
concepts such as small pollution. These concepts can be defined by nonlinear func-
tions as well. Concerning practical applications examined in the next chapters and
the simplicity for end users, nonlinear functions are not further considered. Anyway,
approaches examined in the book are valid for nonlinear functions. The difference
is in calculated values of membership degrees.
R fuzzy set (linear gamma) (Fig. 1.9) is defined by two parameters, a and m, in
the following way:
⎧
⎨ 0x − a for x ≤ a
⎪
μ A (x) = for a < x < m (1.23)
⎪
⎩m−a
1 for x ≥ m
10 1 Fuzzy Set and Fuzzy Logic Theory in Brief

Fig. 1.9 R (linear gamma)

fuzzy set

Fig. 1.10 Singleton fuzzy

set

R fuzzy set is suitable for defining sets expressing high values of the analysed
concepts, such as high turnover. The same comment for nonlinearity of L fuzzy sets
holds for the R fuzzy sets.
Singleton fuzzy set (Fig. 1.10) takes the value zero in all the x ∈ X except in the
point x = m, where it takes the value 1 in the following way:

0 for x = m
μ A (x) = (1.24)
1 for x = m

At the first glance, the singleton is an usual crisp number and there is no special
need to express crisp number in this way. But, in the tasks of approximate reasoning
and in managing fuzziness by relational databases, singletons are indispensable.
According to (1.17) all aforementioned fuzzy sets (1.18)–(1.24) are convex. Fur-
thermore, applying α-cut (1.15) we can say that a fuzzy set A is convex, if and only
if all A(α) intervals are convex for ∀α ∈ [0, 1].
Summarizing this part, we could say that fuzzy sets allow users to express the
uncertainty of the analysed problem. On the other hand, the analysed system will not
work properly, if membership functions are badly defined. Hence, these functions
have to be carefully defined [10].

1.2.3 Operations on Fuzzy Sets

The operations with fuzzy sets A and B are defined via operations on their respective
membership functions.
1.2 Fuzzy Sets 11

Equality
The fuzzy sets A and B are equal (A = B), if for ∀x ∈ X :

μ A (x) = μ B (x) (1.25)

This operation is the generalization of equality from classical set theory. However,
sets might be more or less equal. For this purpose distance measures and equality
indexes (optimistic, medium, pessimistic) are employed [10], as well as generalized
equality [34].
Generalized equality generalizes the operator =. One of the ways is the general-
ization of a well known equality relation from the crisp set theory

(A ⊆ B ∧ B ⊆ A) ⇔ A = B (1.26)

Straightforwardly, for the fuzzy equality holds: if (A ⊆ F B ∧ B ⊆ F A), then

A = F B.
Possibility measure
The possibility that the fuzzy value B belongs to a fuzzy concept A is defined as
(Fig. 1.11)

Poss(B, A) = sup[t (μ A (x), μ B (x))] (1.27)

x∈X

where t stands for t-norm. T-norms are examined in Sect. 1.3.1. Usually minimum
t-norm is used and therefore this equation is known as:

Poss(B, A) = sup[min(μ A (x), μ B (x))] (1.28)

x∈X

The possibility measure gets value of 0 when intersection of two fuzzy sets is
empty. This consequence is used in fuzzy queries over fuzzy values in the relational
databases, among others.

Fig. 1.11 The possibility

that the fuzzy value belongs
to a fuzzy concept
12 1 Fuzzy Set and Fuzzy Logic Theory in Brief

Fig. 1.12 Intersection of

two fuzzy sets

Inclusion
The fuzzy set A is included in the fuzzy set B, if for every x ∈ X holds

μ A (x) ≤ μ B (x) (1.29)

Hence, fuzzy set A is a subset of fuzzy set B.

Intersection
The intersection operation of fuzzy sets A and B is defined as

μ A  B (x) = min(μ A (x), μ B (x)) (1.30)

Fuzzy sets and their intersection (marked as a bold line) are plotted in Fig. 1.12.
If membership degrees are reduced to values of 0 and 1, this function meets the
definition of intersection in the classical set theory. Furthermore, this operation is
often subnormalized fuzzy set, that is, its height (1.11) is lower than the value of 1,
except when A(1) B (1) = ∅. Furthermore, if fuzzy sets A and B are convex, so is
their intersection [56].

Union
The union operation of fuzzy sets A and B is defined as

μA B (x) = max(μ A (x), μ B (x)) (1.31)

Fuzzy sets and their union (marked as a bold line) are shown in Fig. 1.13. If mem-
bership degrees are reduced to values of 0 and 1, this function meets the definition of
union in the classical set theory. The union of two fuzzy sets is mainly non-convex,
except when the intersection of cores (1.10) is not empty and both sets are convex.

Complement
The fuzzy sets A and A are complements if

μ A (x) = 1 − μ A (x) (1.32)

Random documents with unrelated
content Scribd suggests to you:
accusation by appeal to the emperor, after the example of the
blessed Paul the Apostle. But Paul, when accused by his own nation
before the princes of Judaea, but not as yet judged, appealed to
Caesar, and by the princes he was sent to Caesar to be judged. That
does not at all coincide with the present case. For this cleric of evil
repute was accused, and judged, and sent to prison, and thence
escaped, and contrary to law entered the basilica, which he ought
not to have entered till after he had done penance, and still—it is
said—ceases not to live perversely; this man you say has appealed
to Caesar in the same manner as Paul. But he certainly is not
coming to Caesar as Paul did.
“We have given orders to Bishop Theodulf, by whom he was
judged and sent to prison, and from whose custody he escaped, that
he be brought back; and the bishop must bring him to our audience,
whether he speaks truth or falsehood; for it consists not with our
dignity that for such a man as this there should be any change of our
original order.
“We greatly wonder that to you alone it should seem fit to go
against our authoritative sanction and decree, when it is quite clear,
both from ancient custom and from the constitution, that the decrees
of enactments ought to be unalterable, and that to no one is it
permitted to disregard their edicts and statutes. And herein we can
not sufficiently marvel that you have preferred to yield to the
entreaties of that wretch, rather than to our authoritative commands.
“Now you yourselves, who are called the congregation of this
monastery and the servants of God, yea the true God, know how
your life is now frequently evil spoken of by many, and not without
cause. You declare yourselves sometimes to be monks, sometimes
canons, sometimes neither. And we, acting for your good and to
remove your evil repute, looked out a suitable master and rector for
you and invited him to come from a distant province. He by his words
and admonitions, and—for that he is a religious man—by his
example of good conversation, could have amended the manner of
your life. But—ah, the grief of it—all has turned out the other way.
The devil has found you as his ministers for sowing discord exactly
in the wrong place, namely, between wise men and doctors of the
church. And those who ought to correct and chastise sinners you
drive into the sin of envy and wrath. But they, by God’s mercy, will
not lend an ear to your evil suggestions.
“And you, who stand out as contemners of our command, whether
you be called canons or monks,[218] know that at our pleasure, as
our present messenger will indicate to you, you must appear before
us; and although a letter sent to us here excuses you of actual
sedition, you must come and wipe out your unjust crime by condign
amends.”
Alcuin’s reply was more than twice as long.
“To the lord most excellent, and of all honour Ep. 184.
most worthy, Charles, king, emperor, and most
victorious most great most good and most serene Augustus, Albinus
his servitor wishes the welfare of present prosperity, and of future
beatitude, eternal in Christ the Lord God.
“On the first face of this letter I see that thanks from my whole
heart must be given by me to our Lord God for your safety and
welfare, not to me only but to all Christians most necessary. Next,
with prostrate body, contrite heart, tearful voice, mercy must be
begged of the piety of your goodness for the brethren of St. Martin,
to whose service your goodness delegated me however little worthy.
I call God as the witness of my conscience that never have I
understood the brethren to be such as I hear that they are called by
some who are more ready to accuse than to save. As far as can be
seen and known, they worthily perform the office in the churches of
Christ, and I most truly bear witness that never any where have I
seen other men celebrating more perfectly or more diligently, in daily
course interceding for your safety and the stability of the Christian
empire. Of their life and conversation you can learn from a perfect
man, an incorrupt judge, and a faithful messenger, Wido [Count of
the shore of Britany]. He has looked into all their affairs and knows
what they have done and how they have lived.
“I have not been slow to admonish them concerning the strictness
of the monastic life, as they themselves will testify, if any one will
accept their testimony. And I do not know what faults they have
committed against their accusers, that they should pursue them with
such hatred.
“It is a matter of wonder why they[219] wish to push themselves,
contrary to the edict of the law, into another’s harvest. The illustrious
doctor forbids this where he says[220] Who art thou that judgest
another man’s servant? To his own master he standeth or falleth.
Yea he shall stand, for God is able to make him stand. For the city of
Tours has a pastor [Joseph, the Archbishop], in his life elect, in
preaching devout, who knows how best to give to the family of Christ
their portion of meat. Let each shepherd watch over his own flock,
that no member of it lack the grace of God; that when the shepherd
of all shall come He may find them worthy of eternal reward.
“With regard to the concourse and tumult which arose in the
church of St. Martin, or without in the atrium, I testify in the sight of
Him that knows the heart of each that it took place without any
incitement or foreknowledge or even wish of mine. And I confess that
never was I in greater trouble for other men’s offences than then.
Nor, as far as I have been able to understand or to hear, was any
thing done by design of the brethren. I have not even been able to
learn that they wished it; and there can be no doubt that no one who
fears God and cares for his own salvation, should—I will not say do
such a thing but—even think of it.
“Did not the venerable man Teotbert, sent by your authority, spend
nineteen days among them for the purpose of this enquiry? Whom
he would, he flogged; whom he would, he put in chains; whom he
would, he put on oath; whom he pleased, he summoned to your
presence.
“In vain have I so long time served my Lord Jesus Christ if His
mercy and providence have so forsaken me that I should fall into this
impious wickedness in the days of my old age....
“The true cause of this tumult, as far as I have been able to
understand, I am not ashamed to lay before your excellency, sparing
no one, so that I may produce testimony to the truth.
 “It appears to me that in the doing of this impious deed no one has
offended more gravely than the guard of this wretch, from whose
negligence so many evils came. If I may say so to those who hear
this letter read, I think it would be more just that he by whose
negligence the accused man escaped from his bonds should suffer
the same bonds, than that the fugitive to the protection of Christ our
God and of His saints, should be sent back from the church into the
same bonds. I will not put this on my own opinion, I am supported by
the word of God who bade[221] the prophet say to the king of Israel
who had let go out of his hand the king of Syria, Thus saith the Lord,
Because thou hast let go a man worthy of death, thy life shall be for
his life.
“In the second place, I take it that the men were the cause of the
tumult who came armed in larger number than was necessary from
Orleans to Tours; especially because the report ran through the
populace that they had come to carry off with violence a man who
had fled to the protection of the Church of Christ and St. Martin. For
all men everywhere take it ill that their holy ones are dishonoured.
Perhaps, too, the miserable man had called upon the rustics who
came to his dwelling in their cups to defend the church of St. Martin
and not allow him to be snatched from it.
“There was a third cause of the tumult. Our holy father and pontiff
[Archbishop Joseph] inopportunely, the people being present,
entered the church along with the men who were supposed to have
come to drag away the man. He may have done this in the simplicity
of his heart, not imagining that any harm could come.[222] When the
ignorant people, always doing thoughtlessly inconvenient things, saw
this, they cried out, they took to their clubs; some energetic men ran
out when they heard the bells sound. They were rung by unskilled
hands; your own judges ascertained that, and our accusers
themselves allowed that it was so, for in their presence the holy
Gospel was brought; there was laid upon it the wood of the holy
Cross; they made such of the brethren as they chose, swear by that.
When the brethren heard the bells, they rushed out of the refectory
to learn why they were being rung. As I am informed, they did what
they could to allay the tumult; only some youths, who were found
and sent to your presence, were the offenders in the concourse.
From them it can be learned what they did; they have sworn that
they acted on the prompting of no man, only on the impulse of their
own folly. Not one of the servants of St. Martin was there, except a
man called Amalgarius, who was with me at the moment. Him I sent
at once with the other brethren to appease the tumult, and to
extricate the men of the venerable bishop from the hands of the
people, so that no harm should be done them. As soon as the tumult
was appeased, they were brought into the monastery, where they
were safe. These men were so burning with wrath against me that
they turned a kindness I had ordered to be done to them into evil,
saying that it was in insult that I had sent them some food.[223] This
was absolutely false. They did not know that I was imbued with the
Lord’s command, Do good unto them that hate you.
“Let your holy piety, most pious lord, consider these facts and
recognize the truth. Be favourable to thy servants in the love of God
omnipotent and in the honour of the holy Martin your intercessor,
who always has been honoured in the kingdom and by the kings of
the Franks.
“We are wont to say in confessing our sins, If thou, Lord, wilt be
extreme to mark what is done amiss, O Lord, who may abide it? And
to thee we may say, forasmuch as we know thee to be a member of
that same Head, if thou wilt be extreme to mark what is done amiss,
who, lord, may abide it? Above all, because the special virtue,
goodness, and praise of emperors has always been their clemency
towards their subjects; in so much that the most noble emperor Titus
said that no one should leave the presence of the emperor sad.
Rejoice the minds of thy servants by the highest gift of thy mercy; let
mercy rejoice against judgement. Men who have been guilty of the
greatest crimes of perfidy against your authority you have been able
to pardon with laudable piety; overlook our infelicity, in accordance
with the most pious nobility of your most holy disposition, which I
have always known to abound in a marvellous degree in the mind of
your wisdom. We read how David, the ancestor of Christ, was
praised in the greatness of his mercy and the justness of his
judgements. In like manner we know that your blessedness is, by the
gift of Christ, always worthy of all laudation and praise for these two
great merits.
“May the omnipotent God the Father, by His only Son our Lord
Jesus Christ, illumine, fill, and rejoice the heart of your blessedness
with all blessing and wisdom in the Spirit the Comforter, and deign to
grant to your most noble offspring, for the welfare of a Christian
people, perpetual prosperity, most dearly loved lord, best and most
august father of the fatherland.”
We know no more than this. There appears to be no possibility of
carrying the investigation further. Reading between the lines we
seem to see signs of ecclesiastical tension between the archbishop,
seated at his cathedral church of St. Gatian, and Abbat Alcuin of St.
Martin’s. Until the time of Alcuin’s penultimate predecessor, the
abbat of St. Martin’s had been the archbishop of Tours, and, as we
have seen, there are curious references to a claim of St. Martin’s to
have bishops of its own. This may have caused tension, beyond that
which was not very improbable under the ordinary conditions.
Theodulf of Orleans was an old friend of Alcuin, and an admirer.
He gives to Alcuin a large place in his description of the court of
Charlemagne. Theodulf was a laudatory poet, and his poem was
very properly meant to please those whom he described. Of the king
himself he says—

O face, face more shining than gold thrice refined,

Happy he who always is with thee.
The head illustrious, the chin, the neck so beautiful,
The hands of gold, that banish poverty.
The breast, the legs, the feet, all laudable,[224]
All shining forth in beauty and in strength.

The latest wife of the king, Luitgard, has eight pretty lines devoted to
her, after an inauspicious opening address to “the fair virago,
Luitgard”. This dates the poem before 801, in which year Luitgard
died at Tours. The tower of St. Martin’s, now called the tower of
Charlemagne, was raised over her tomb.[225]
 Alcuin was evidently a very prominent figure at court, keeping
things alive by his knowledge and wit and subtleties.

And Flaccus too is there, the glory of our poets,

Who pours forth many things in lyric foot.
An able sophist, a poet, too, melodious,
Able in mind and able in practice alike.
He brings forth pious lessons from Holy Writ,
And solves the puzzles of numbers with favouring jest.
He puts an easy question now, and then a hard;
Of this world now, then of the world above.
The king alone, of many that fain would,
Can solve the skilful puzzles Flaccus sets.

There was evidently no standing ill-feeling against the Abbat of St.

Martin’s on the part of the Bishop of Orleans.
 CHAPTER XV
Alcuin’s letters to Charlemagne’s sons.—Recension of the Bible.—The “Alcuin
Bible” at the British Museum.—Other supposed “Alcuin Bibles”.—Anglo-Saxon
Forms of Coronation used at the coronations of French kings.

There is in the Bibliothèque Nationale in Paris a letter headed “In

nomine Dei summi incipit scriptum Albini magistri ad Karolum
imperatorem”. It is, however, held to be uncertain whether the letter
is addressed to the emperor or to his son Charles, who died some
three years before his father. The internal evidence appears to be
decidedly against its having been addressed to the emperor. Alcuin
could not have denied himself the pleasure of referring to the
emperor when he mentions king David as the authority for his
advice, and we have no letter of Alcuin to the emperor so completely
free from honorific titles and phrases, with nothing but the simple vos
throughout. It is to be said on the other hand that the author of the
Life of the blessed Alchuin the Abbat, with which we dealt fully in
Chapters I and II, refers[226] to a libellus which Alcuin wrote for
Charlemagne, setting forth the psalms which he was to use
according as penitence, tribulation, or joy, was his theme.
The interest of the letter in question fortunately lies in its advice,
not in the person to whom the advice is given. This is the letter, with
its ordinary heading:—
“Alcuin dedicates to Charles the Emperor a Ep. 244.
breviary[227] of prayer to God.
“The blessed David, the great king and servant of God most high,
gave us the rule of singing, how man should pour forth prayers to
God at certain stated hours. ‘Seven times a day,’ he says, ‘do I
praise Thee,’—that is, at the first hour of the day, the second, third,
sixth, ninth, the evening hour, and the twelfth. David the king, then,
gave praise to God at these seven hours. The holy Daniel, the
prophet, at the third, sixth, and ninth hour of the day, went into his
chamber to pray to the Lord, and with hands stretched upward to
Heaven entreated God for himself and for the people of Israel. The
same David said[228] further, ‘I will make mention of Thy
righteousness only.’ And again, ‘At midnight I will rise to give thanks
unto thee,’ that is, at the hour of night. And again he says, ‘I have
thought upon Thy name in the night season,’ that is, at cock-crow.
And, ‘Have I not remembered Thee in my bed, and thought upon
Thee when I was waking?’ Here are three courses of the office
during the night, and seven by day, making the ten courses which we
sing, following the number of the ten laws of Moses. But you have
asked me to write to you in a net form the order in which a layman in
active life should pray to God at the stated hours. You live after a
Christian fashion, and you desire to do Christian deeds; you are not
ignorant how prayer should be made to the Lord; but at your request
I will briefly state my opinion. When you have risen from your bed,
say first ‘O Lord Jesu Christ, son of the living God, in Thy name will I
lift up my hands, make haste to deliver me.’ Say this thrice, with the
psalm ‘Ponder my words, O Lord, consider my meditation. O
hearken thou unto the voice of my calling, my king and my God, for
unto Thee will I make my prayer. My voice shalt thou hear betimes,
O Lord, early in the morning will I direct my prayer unto Thee.’ Then,
‘Our Father,’ and the prayers, ‘Vouchsafe O Lord to keep us this
day,’ ‘Perfect my steps,’ ‘Praised be the Lord daily,’ ‘Direct and
sanctify,’ ‘O Lord let Thy mercy lighten upon us.’ Then, rising, begin
the verse ‘Thou shalt open my lips, O Lord’. When that is ended, with
the Gloria, begin the psalm ‘Lord how are they increased’. Then
follows ‘God be merciful unto me’. Then ‘O come, let us sing unto the
Lord’. Then psalms, as many as you will.”
We have two letters of Alcuin which were certainly written to
Charles the king, the eldest son of Charlemagne. The first was
written in 801 to congratulate Charles on his anointment as king by
Leo III on the same day (Christmas Day, 800) that saw his father
crowned as emperor.
“I have heard from the lord apostolic [Leo III] that Ep. 162.
with the consent of the most excellent Lord David
[Charlemagne] the title of king and the crown of kingly dignity have
been conferred upon you. I greatly rejoice in the honour both of the
title and of the power. I pray that your dignity and nobleness may be
for the safety of many peoples, nations, and churches of Christ; may
be glorious in the world and terrible to the adversaries of the
Christian religion; may be vigorous and strong through a long season
of prosperity; and with the blessing of God may always follow after
better things, ascend to higher, and grow even unto the perfect day
of eternal blessedness.
“Do justice, my best-loved son, and mercy, among Christian
people, for it is these, as Solomon testifies, that exalt the throne of a
kingdom and render the kingly power laudable and pleasing to God.
Have as counsellors men good, pious, prudent, and god-fearing;
men in whom truth reigns, not covetousness, for the gift blindeth the
wise and perverteth the words of the righteous.[229] Never allow the
dishonesty of others to sully the name of your dignity, nor permit
others to do with wicked mind in covetousness that which you would
not yourself do; the fault of the subject is often imputed to the ruler.
Let not the impious will of some, under the name of thy beatitude, fill
their money-bags with the mammon of unrighteousness.
“Good examples are not far to seek. In the home in which you
were brought up you have the best examples of all goodness. You
may have perfect confidence that you will by the gift of God attain to
the blessing of that most excellent and in all honour most noble
father of thine, ruler and emperor of a Christian people, if you strive
to imitate the manner of his nobility and piety and complete
discretion; and will most fully obtain the mercy of God, which is
better than all the glory of the world.
“Wheresoever your way may lead, may the footsteps of piety ever
follow thee, that you may have praise of men and eternal reward with
God.”
Alcuin must needs end a congratulatory letter to a royalty with
hexameter and pentameter:—

Prosperous even for ever be thou great hope of the nations.

 Be to thee Christ as love, light, way, and safety, and life.

The next letter to King Charles was probably later. It seems to

indicate some anxiety on the part of Alcuin, and, indeed, Charles
was not as fine a character as his brother Louis, who is mentioned in
this letter. Alcuin would appear to have kept a copy of the former
letter, and to have made a good deal of it do service a second time.
“I rejoice, my dearest son, in the devotion of your Ep. 245.
good will which Osulf your attendant has narrated
to me, whether as regards the largeness of your alms-giving, or as
regards the gentleness of your rule. Know of a surety that all this is
greatly pleasing to God, and deserves at the hand of His mercy
perpetual blessing. Do thou, my son, my dearest son, always to the
utmost of your power work for the honour of God Almighty in all
goodness and piety; following the example of your most excellent
father in all honesty and sobriety, that the divine clemency of Christ
the God may grant to thee to possess his blessing by right of
inheritance.
“Be a pious hearer of the wretched, and judge their cause with the
utmost justness. Do not permit the judges who are under you to
judge for presents and gifts; for gifts, it is said in Holy Scripture, blind
the hearts of the wise, and subvert the words of the just. Hold in
honour the servants of Christ, those who are true servants of God,
for some come in sheeps’ clothing but inwardly are ravening wolves.
The Truth says, By their fruits ye shall know them. Have as
counsellors wise men, who fear God; not flatterers, for a flatterer, as
it is said, is a bland enemy and often seduces those who consent
unto him. Be prudent in thought and cautious in speech; always
setting your hope on God, for He never faileth them whose hope is
set on Him.
“Would that it were allowed me more frequently to address a letter
of advice to thy benignity, as the most noble youth Louis your brother
has asked me to do frequently for him. This I have done, and, if God
will, I shall continue to do; he reads my letters with great humility.
 “My greatest joy is when I hear—as, indeed, it is right that I should
hear—of a good manner of life on your part. For this is the gift of
God, the prosperity of a kingdom, that the rulers of a Christian
people live most strict lives, and have their conversation among men
in a way pleasing to God. Thus a blessing from heaven is certain to
come on the nation and kingdom, which may God vouchsafe to grant
eternally to your nobility.
“May you flourish, grow, and be strong, advancing in all that is
good and prosperous, to the exaltation of His Holy Church, my
dearest son.”
We have only one of Alcuin’s letters to King Pepin, who died
young, leaving a son Bernard who became king on his father’s
death.
“To the most noble and beloved son Pippin Ep. 77.
Albinus sends greeting in the love of Christ.
“We give thanks to thy benevolence and to the piety of the lord
King who has piously consented to our petition concerning the
redemption of captives. I know that in such works of piety you earn
blessing and a long and prosperous reign.
“And do thou, most excellent youth, study to adorn nobility of birth
by nobility of conduct. Strive with all thy power to fulfil the will and the
honour of the omnipotent God, that His ineffable piety may exalt the
throne of thy kingdom and extend its bounds, and subject the nations
to thy power. Be liberal to the wretched, good to foreigners, devout in
the service of Christ, treating honourably His servants and His
churches that their sedulous prayer may aid thee. Be clean in
conversation, chaste in body. Rejoice with the wife of thy youth and
let not other women have any part in thee, that the blessing granted
unto thee may lead to a long posterity of descendants.
“Be strong against adversaries, faithful to friends, humble to
Christians, terrible to pagans, affable to the wretched, provident in
council. Use the advice of the old men, the service of the young. Let
equity be the judgement in thy kingdom. Let the praise of God
everywhere resound at the fitting hours, and especially in the
presence of thy piety. This kind of devotion to the offices of the
church will render thee loveable to God and honoured among men.
Let thoughts of sobriety be in your heart, words of truth in your
mouth, examples of honour in your conduct, that the divine clemency
may in all ways exalt and preserve thee.
“I pray you let this letter go with you as a testimony of my love.
Though it be not worthy to be hung at the girdle of thy veneration, yet
let its admonition be worthy to be stored in the mind of thy wisdom.”
We must now say something on the part which Alcuin played in
connexion with the revision of the manuscripts of the Bible.
Alcuin is credited with a revision of the whole of the Latin Bible,
both the Old Testament and the New. We have a letter of his in
which he states in precise terms that he had been commissioned by
Karl to correct the corrupted text. The letter is addressed to Gisla,
Abbess of Chelles, Karl’s sister, and Rotruda, Karl’s daughter, whom
he addresses as Columba, the Dove.
“I have sent for the solace of your sanctity a Ep. 136. a.d. 800.
small book, written in short sections, that you may
use it during these days[230] for your holy devotion. In such study
you best spend these most holy days, and especially in the Gospel
of the blessed John, wherein are the deeper mysteries of divinity,
and the most holy words of our Lord Jesus Christ which He spoke on
that night when He willed to be betrayed for the salvation of the
world.
“I might have sent you an exposition of the whole Gospel, if I had
not been occupied, by the command of the lord king, in the
emendation of the Old and the New Testament. But if life last and
God help, I will, when occasion serves, finish the task now begun,
and dedicate the completed work to your name.”
Gisla and Rotruda sent him a delightfully Ep. 137.
affectionate and bright letter in reply. They liken
Alcuin to Jerome sending the Scriptures from his cave in Bethlehem
to Rome; and in begging him to send the rest of the commentary on
St. John they remind him that the shallow Loire is crossed with less
danger than the Tuscan Sea, and that a messenger gets more easily
from Tours to Paris than from Bethlehem to Rome.
It is certain from the dedicatory verses of Alcuin’s which have been
preserved, that at least four complete copies of the whole Bible had
been corrected by him or under his direction, and sent to the
emperor. Of these, not one is known to be still in existence. Of one of
them Alcuin makes definite mention in the following letter:—
“To the most desired and entirely loveable David Ep. 205. a.d. 801-
the king Albinus wishes present prosperity and 3.
eternal beatitude in Christ.
“I have long deliberated upon the question what could the devotion
of my mind think of as worthy to be given towards the splendour of
your imperial power and the increase of your most rich treasury. I
feared lest the poor intelligence of my mind should remain torpid in
empty idleness, while others were offering various rich gifts, and the
messenger of my littleness should come before the presence of your
beatitude with empty hands. I have at length, by the inspiration of the
Holy Spirit, found something which it is fitting that I should send and
it may be agreeable to your prudence to receive.
“In the most sacred solicitude of your piety it is clear beyond doubt
what the Holy Spirit works through you for the safety of the whole
Church, and how earnestly all faithful people should pray that your
empire be extended to full glory, and be loved at home by all God’s
people, and terrible abroad to all the enemies of His Son. To my
questioning and desiring mind, nothing seemed more worthy of your
most peace-giving honour than a present of the divine books, which
by the dictation of the Holy Spirit and the ministration of Christ God
have been written by the pen of divine grace for the salvation of the
whole race of man. These, brought together into the sanctity of one
most clear body, and diligently emended, I have sent to your most
lofty authority by this dearest son of ours and faithful servant of
yours, that with full hands he may with most joyous service stand
before your dignity. He has been ill for a long time, but now that by
God’s mercy he has to some extent recovered, he has with the
greatest satisfaction hastened to approach your piety.
 “The small gifts of my tears I send by faithful promise in prayer to
St. Martin for the ardently desired prosperity of your authority. Let my
messenger serve the most pious lord as is fitting; I will pray for the
most loved lord as the visitation of the Holy Spirit shall deign to
illumine my heart. If the devotion of my mind could have found
anything better, I would with ready will offer it towards the increase of
your honour.”
The messenger was Nathanael, that is, Fredegisus. We learn this
from Letter 206, which commences “Albinus greets Nathanael”, and
after addressing him as though he were the real Nathanael who was
seen under the fig-tree by Jesus, proceeds thus:—
“Salute Lucia my sister and Columba our Ep. 206.
daughter.[231] Pray them to be mindful of my old
age in sacred prayers and of their own salvation in good works. And
hide not from them the beauty of your wisdom, but irrigate the flower-
beds of good will in them. What is more beautiful than the flowers of
wisdom, which never fade? What is richer than the wealth of
knowledge, which is never exhausted? To this exhort them. Let them
live day and night in meditation on the law of God, that they may find
Him of whom Moses in the law wrote, and the prophets. Bid them
hold Him and not let Him go till they are led into the chambers of the
Kings glory to be supported by flowers of eternal blessedness, the
Bridegroom’s left hand of present prosperity under their head, and
the right hand of eternal bliss embracing them.[232]
“Convey the letter of my littleness, with the most holy gift of divine
Scripture and peaceful words of salutation, to my lord David. To him
we owe as many thanks and praises for all his goodness to me and
to my sons as this Book has syllables; to him may God give as many
blessings as in this Book there are letters.”
The natural supposition is that Alcuin brought—or had sent—from
York accurate copies of the Scriptures, from which he corrected the
faulty manuscripts of France and Germany, to use modern names.
Errors were due, probably, at least as much to mispronunciation on
the part of the person who dictated to the writers, or to mis-hearing
on their part, as to carelessness in transcribing. We have to
remember that the practice was for one monk to read out word by
word the sentence which the writers in the scriptorium were to take
down, so that in this way twenty or thirty—it is said as many as two
hundred—copies of a poem or a book could be written at the same
time. This practice gave many opportunities for error.
We have at the British Museum a magnificent Bible, one of the
largest manuscripts in existence, called Alcuin’s Bible. It contains
449 sheets of very fine parchment, 20 by 14½ inches. It was
purchased for the Museum in 1836 for £750, the price asked at first
being £12,000, reduced to £6,500 as “an immense sacrifice”. The
story of its acquisition, and the question of its date and its connexion
with Alcuin, were stated and discussed by Sir F. Madden in the
Gentleman’s Magazine for 1836, pages 358 to 363, 468 to 477, 580
to 587. That able archaeologist believed it to be of Alcuin’s own time,
and, indeed, to be the very copy which Alcuin presented to
Charlemagne in 801, on the completion of the recension which Karl
had entrusted to him. The evidence in favour of this view is found on
the last page of the MS., in some elegiac verses composed by
Alcuin. The verses begin with an appeal from the book itself to its
readers that it may be called a Pandect, and not a Bibliotheca,[233]
and after eight more verses, in which it is called a Codex, they end
as follows:—

Mercedes habeat, Christo donante, per aevum

Is Carolus qui iam scribere iussit eum.
Haec dator aeternus cunctorum, Christe, bonorum
Munera de donis accipe sancta tuis,
Quae Pater Albinus, devoto pectore supplex,
Nominis ad laudes obtulit ecce tui;
Quem tua perpetuis conservet dextra diebus,
Ut felix tecum vivat in arce poli.
Pro me quisque legas versus orare memento,
Alchuine dicor ego. Tu sine fine vale.

“May Charles, who bade this book be written, receive eternal

rewards. May the giver of all good accept this offering of His own
gifts, which Father Albinus has made, whom may Thy hand preserve
to live with Thee. Thou who readest these verses, remember to pray
for me; my name is Alchuine; mayest thou for ever fare well.”
That these verses were written in the great Pandect of Alcuin’s
recension, which Alcuin presented to Charlemagne, we may take to
be certain. But we may also take it as certain that they would be
written also in copies made from that special Pandect; and it has
been decided by the most competent modern critics that the Bible in
the Museum was not written till a generation had passed away after
Alcuin’s death.
That the verses were entered in other copies also is certain. The
Fathers of the Oratory della Vallicella at Rome had a copy of this
recension, which was believed to be written by Alcuin’s own hand
and presented to Charlemagne. In it there is a long copy of verses,
including those in the Museum Bible, but with curious alterations and
additions, which make it probable that the Vallicella Bible was written
for Charlemagne’s grandson, Charles le Chauve. Quae Pater
Albinus devoto pectore supplex is altered into Quae tibi devoto
Carolus rex pectore supplex, and verses are added, stating that the
Bible was written for a new church which Charles had just built. The
alteration cuts out the personal note of Alcuin, and the addition cuts
out Charlemagne and points to another Charles. This is far from
being the only case in which confusion is caused by the fact that
Charlemagne was himself for many years Charles the king; that his
oldest son was Charles the king; that his grandson was Charles the
king; as also two great grandsons, a great great grandson, and even
two generations further still.
Others besides Alcuin and the royal family were interested in the
various versions of Scripture. For example, his contemporary
Theodulf, the learned bishop of Orleans, sent to his own daughter
Gisla a psalter, radiant with silver and gold, with both the earlier and
the later versions of Jerome.
Our use of the word Graduale for the book containing the words
and the music sung by the choir at the service of the Mass is an
evidence of the large part played by the Gallican Church in the
arrangement and improvement of the early mediaeval service books.
Rome spoke of the Antiphonale Missarum and Antiphonale Horarum,
while Gaul spoke of the Graduale for Mass Music and Antiphonale
for the Music of the Hours. Under Alcuin’s guiding hand, the
influence of Charlemagne and his times upon the services was wide
and deep. In the document described as Ep. 31, a.d. 794, Karl has a
good deal to say about the success of his own efforts to put down
irregular methods of singing the services, and to bring all into
general accord with the Roman method.[234] Alcuin’s work re-acted
upon the Roman use itself, and is understood to have been the
operating cause of the mark left upon it.
Alcuin had strong opinions as to the best manner Ep. 72. a.d. 796.
of singing the services. In a letter to Eanbald II, he
writes thus, for the benefit of the Church of York:—“Let the clergy
chant with moderated voice, striving to please God rather than men.
An immoderate exaltation of voice is a sign of boastfulness. And let
them not be above learning the Roman Orders of Service, that they
may have eternal benediction from the blessed Peter, chief of the
Apostles, whom Our Lord Jesus Christ made the head of His elect
flock.”
Alcuin was versed in secular music also. We learn from Ep. 100
that Karl had asked him to write peaceful and soothing songs, both
words and music, for soldiers to sing when engaged in the horrors of
war, and that he complied with the request.
We have some very interesting evidences of the borrowing of
Anglo-Saxon manuscripts for use in France, and of the influence of
Anglo-Saxon forms on French services. There are two Anglo-Saxon
forms for the coronation of a king. One of these is found in the
Pontifical of Ecgbert, the Bishop and later the Archbishop of York, to
which a date of about 745 may be given. It is merely the supplement
to the Mass on the occasion of a coronation, and accordingly it does
not give the details of the ceremony. The other is a later form, and it
gives at length the details of the ceremony, one of the longest
prayers describing the king as raised to the royal throne of the
Angles and Saxons. But, curiously enough, we learn the most
interesting parts of the ceremony of crowning an Anglo-Saxon king,
not from this manuscript, but from three manuscripts of the form for
the coronation of a king of the French. The first of these to be
mentioned is a manuscript form of an Abbat of Corbie. In it we find
the prayer for “This thy servant whom with suppliant devotion we
elect equally to the kingdom of the whole of Albion, that is to say, of
the Franks ... That he may nourish and teach the Church of the
whole of Albion, with the peoples committed to his charge”. Here it
would appear that a marginal note had been added to the Anglo-
Saxon form at the first mention of “Albion”, “that is to say, of the
Franks,” and has afterwards been incorporated in one place and not
in the other. The “elect equally” indicates that the form was used for
an Anglo-Saxon king who claimed to be king of the whole land, while
yet the old division into three main nations was fresh in mind.[235] It
is a further evidence in favour of this being an Anglo-Saxon form,
that the only saint mentioned besides the Blessed Virgin and St.
Peter is “Holy Gregory, Apostolic of the Angles”. In the preparation of
the Sens Order, to be mentioned later, this flaw had been
discovered, and St. Denys and St. Remy put in the place of St.
Gregory.
In a manuscript in the National Library of Paris, we have a second
Order for the Coronation of a King of the Franks, which is indubitably
an Anglo-Saxon Order. The following phrases occur: “This thy
servant whom with suppliant devotion we elect equally ... That the
sceptre desert not the royal throne, that is to say, of the Saxons,
Mercians, and Northumbrians (Nordanchimbrorum) ... That
supported by the due subjection of both of these peoples...”
In a third Order for the Coronation of French Kings, from the
Pontifical of the illustrious Church of Sens, we find the prayer “that
the sceptre desert not the royal throne, that is to say, of the Saxons,
Mercians, and Northumbrians (Nordan Cymbrorum)”, and “that the
king, supported by the due subjection of both these peoples....”
It may be added that the French Benedictional of Archbishop
Robert, now at Rouen, has the form “Angles and Saxons”. So late as
1364 Charles V of France was crowned with a form which named the
throne as that of the Saxons, Mercians, and Northchimbrians; while
at the same time the peers of Guienne swore to protect him against
the king of England, his people, and allies.[236]
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com