Get the full ebook with Bonus Features for a Better Reading Experience on ebookmeta.

com

Making Presentation Math Computable A Context

Sensitive Approach for Translating LaTeX to
Computer Algebra Systems 1st Edition André
Greiner-Petter
https://ebookmeta.com/product/making-presentation-math-
computable-a-context-sensitive-approach-for-translating-
latex-to-computer-algebra-systems-1st-edition-andre-greiner-
petter/

OR CLICK HERE

DOWLOAD NOW

Download more ebook instantly today at https://ebookmeta.com

André Greiner-Petter

Making
Presentation
Math Computable
A Context-Sensitive Approach
for Translating LaTeX to Computer
Algebra Systems
Making Presentation Math Computable
André Greiner-Petter

Making Presentation
Math Computable
A Context-Sensitive Approach for
Translating LaTeX to Computer
Algebra Systems
André Greiner-Petter
Berlin, Germany

ISBN 978-3-658-40472-7 ISBN 978-3-658-40473-4 (eBook)

https://doi.org/10.1007/978-3-658-40473-4

© The Editor(s) (if applicable) and The Author(s) 2023. This book is an open access publication.
Open Access This book is licensed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give appropriate
credit to the original author(s) and the source, provide a link to the Creative Commons license and
indicate if changes were made.
The images or other third party material in this book are included in the book’s Creative Commons
license, unless indicated otherwise in a credit line to the material. If material is not included in the
book’s Creative Commons license and your intended use is not permitted by statutory regulation or
exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with
regard to jurisdictional claims in published maps and institutional affiliations.

This Springer Vieweg imprint is published by the registered company Springer Fachmedien
Wiesbaden GmbH, part of Springer Nature.
The registered company address is: Abraham-Lincoln-Str. 46, 65189 Wiesbaden, Germany
Front Matter

Contents

FRONT MATTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Zusammenfassung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CHAPTER 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Motivation & Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Research Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Research Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.1 Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.2 Research Path . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

CHAPTER 2
Mathematical Information Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1 Background and Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2 Mathematical Formats and Their Conversions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.1 Web Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.2 Word Processor Formats. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.3 Computable Formats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2.4 Images and Tree Representations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.2.5 Math Embeddings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.3 From Presentation to Content Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.3.2 Benchmarking MathML . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.3.3 Evaluation of Context-Agnostic Conversion Tools . . . . . . . . . . . . . . . . . . . . . . . 48
2.3.4 Summary of MathML Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4 Mathematical Information Retrieval for LaTeX Translations . . . . . . . . . . . . . . . . . . . . . 51

v
 CHAPTER 3
Semantification of Mathematical LaTeX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.1 Semantification via Math-Word Embeddings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1.1 Foundations and Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1.2 Semantic Knowledge Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.1.3 On Overcoming the Issues of Knowledge Extraction Approaches . . . . . . . 68
3.1.4 The Future of Math Embeddings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.2 Semantification with Mathematical Objects of Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.2.1 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.2.2 Data Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.2.3 Frequency Distributions of Mathematical Formulae . . . . . . . . . . . . . . . . . . . . . 76
3.2.4 Relevance Ranking for Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.2.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.2.6 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.3 Semantification with Textual Context Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.3.1 Semantification, Translation & Evaluation Pipeline . . . . . . . . . . . . . . . . . . . . . . 91

CHAPTER 4
From LaTeX to Computer Algebra Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.1 Context-Agnostic Neural Machine Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.1.1 Training Datasets & Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.1.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.1.3 Evaluation of the Convolutional Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.2 Context-Sensitive Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.2.3 Formal Mathematical Language Translations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.2.4 Document Pre-Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.2.5 Annotated Dependency Graph Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.2.6 Semantic Macro Replacement Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

CHAPTER 5
Qualitative and Quantitative Evaluations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.1 Evaluations on the Digital Library of Mathematical Functions . . . . . . . . . . . . . . . . . . . 114
5.1.1 The DLMF dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.1.2 Semantic LaTeX to CAS translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.1.3 Evaluation of the DLMF using CAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.1.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.1.5 Conclude Quantitative Evaluations on the DLMF . . . . . . . . . . . . . . . . . . . . . . . . 131
5.2 Evaluations on Wikipedia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.2.1 Symbolic and Numeric Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.2.2 Benchmark Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.2.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.2.4 Error Analysis & Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.2.5 Conclude Qualitative Evaluations on Wikipedia . . . . . . . . . . . . . . . . . . . . . . . . . 139

vi Contents
CHAPTER 6
Conclusion and Future Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.2 Contributions and Impact of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.3 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.3.1 Improved Translation Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.3.2 Improve LaTeX to MathML Converters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
6.3.3 Enhanced Formulae in Wikipedia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
6.3.4 Language Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

BACK MATTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Bibliography of Publications, Submissions & Talks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

Contents vii
Front Matter

List of Figures

2.1 Reference map of mathematical formats and translations between them. . . . 20

2.2 The math template editor of Microsoft’s Word [395]. . . . . . . . . . . . . . . . 32
2.3 An expression tree representation of the explicit Jacobi polynomial definition
in terms of the hypergeometric function. . . . . . . . . . . . . . . . . . . . . . 35
2.4 GUI to support the creation of our gold standard MathMLben. . . . . . . . . . 46
2.5 Overview of the MathML tree edit distances to the gold standard. . . . . . . . 50
2.6 Runtime performances of LATEX to MathML conversion tools. . . . . . . . . . . 51
2.7 Four different layers of math objects in a single mathematical expression. . . . 53

3.1 t-SNE plot of top-1000 closest vectors of the identifier f . . . . . . . . . . . . . 68

3.2 Unique subexpressions for each complexity in arXiv and zbMATH. . . . . . . 77
3.3 Frequency and Complexity Distributions of Math Expressions. . . . . . . . . . 78
3.4 Most frequent math expressions in arXiv. . . . . . . . . . . . . . . . . . . . . . 80
3.5 Comparison plot of most frequent expressions in arXiv and zbMATH. . . . . . 82
3.6 Top-20 search results from topic-specific subsets. . . . . . . . . . . . . . . . . . 85
3.7 Search results for the query ‘Jacobi polynomial.’ . . . . . . . . . . . . . . . . . 90
3.8 Pipeline of the proposed context-sensitive conversion process. . . . . . . . . . 92

4.1 Mathematical semantic annotation in Wikipedia. . . . . . . . . . . . . . . . . . 102

4.2 The workflow of our context-sensitive translation pipeline. . . . . . . . . . . . 106

5.1 Example argument identifications for sums. . . . . . . . . . . . . . . . . . . . 121

5.2 The workflow of the evaluation engine and the overall results. . . . . . . . . . 123
5.3 The numeric test values and global constraints. . . . . . . . . . . . . . . . . . . 126

6.1 Layers of a mathematical expression with mathematical objects (MOI). . . . . 143

6.2 The annotated defining formula of Jacobi polynomials in the English Wikipedia
article. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.3 Translation information for equation (6.3). . . . . . . . . . . . . . . . . . . . . 149
6.4 Proposed pipeline to improve existing LATEX to MathML converters. . . . . . . 155
6.5 Semantic enhancement of the formula E = mc2 . . . . . . . . . . . . . . . . . 157

ix
Front Matter

List of Tables

1.1 Different representations of a Jacobi polynomial. . . . . . . . . . . . . . . . . . 2

1.2 Examples of Mathematica’s LATEX import function. . . . . . . . . . . . . . . . . 3
1.3 The results of importing π(x + y) in different CAS. . . . . . . . . . . . . . . . 4
1.4 Different computation results for arccot(−1) (inspired by [84]). . . . . . . . . 5
1.5 Overview of the primary publications. . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Overview of secondary publications. . . . . . . . . . . . . . . . . . . . . . . . 9

2.1 Overview table of available mathematical format translations. . . . . . . . . . 21

2.2 LATEX to CAS translation comparison. . . . . . . . . . . . . . . . . . . . . . . . 28
2.3 Special content symbols added to LATExml. . . . . . . . . . . . . . . . . . . . . . 44

3.1 Find the Term where Term is a word that is to X what Y is to Z. . . . . . . . . . 65

3.2 The cosine distances of f regarding to the hits in Table 3.1. . . . . . . . . . . . 66
3.3 Descriptive terms for f in a given context. . . . . . . . . . . . . . . . . . . . . 67
3.4 Mathematical Objects of Interests Dataset Overview. . . . . . . . . . . . . . . 77
3.5 Settings for the retrieval experiments. . . . . . . . . . . . . . . . . . . . . . . . 84
3.6 Top s(t, D) scored expressions in zbMATH. . . . . . . . . . . . . . . . . . . . . 86
3.7 Most frequent expressions on topic-specific subsets of zbMATH. . . . . . . . . 88
3.8 Suggested autocompleted math expressions. . . . . . . . . . . . . . . . . . . . 89

4.1 Results of our neural machine translations. . . . . . . . . . . . . . . . . . . . . 98

4.2 Comparison between Mathematica and our machine translation. . . . . . . . . 99
4.3 Machine translations on 100 random DLMF samples. . . . . . . . . . . . . . . 99
4.4 Examples of our machine translations from LATEX to Mathematica. . . . . . . . 100
4.5 Mappings and likelihoods for a semantic LATEX macro. . . . . . . . . . . . . . . 111

5.1 Examples blueprints for subscripts of sums and products. . . . . . . . . . . . . 120

5.2 Translations for the prime derivative of the Hurwitz zeta function. . . . . . . . 122
5.3 The symbolic and numeric evaluations on Wikipedia. . . . . . . . . . . . . . . 134
5.4 Performance of description extractions via MLP. . . . . . . . . . . . . . . . . . 136
5.5 Performance of semantification from LATEX to semantic LATEX. . . . . . . . . . . 137
5.6 Performance comparison for translating LATEX to Mathematica. . . . . . . . . . 138

xi
FRONT MATTER

Abstract

This thesis addresses the issue of translating mathematical expressions from LATEX to the syntax
of Computer Algebra Systems (CAS). Over the past decades, especially in the domain of Science,
Technology, Engineering, and Mathematics (STEM), LATEX has become the de-facto standard
to typeset mathematical formulae in publications. Since scientists are generally required to
publish their work, LATEX has become an integral part of today’s publishing workflow. On the
other hand, modern research increasingly relies on CAS to simplify, manipulate, compute, and
visualize mathematics. However, existing LATEX import functions in CAS are limited to simple
arithmetic expressions and are, therefore, insufficient for most use cases. Consequently, the
workflow of experimenting and publishing in the Sciences often includes time-consuming and
error-prone manual conversions between presentational LATEX and computational CAS formats.
To address the lack of a reliable and comprehensive translation tool between LATEX and CAS,
this thesis makes the following three contributions.
First, it provides an approach to semantically enhance LATEX expressions with sufficient semantic
information for translations into CAS syntaxes. This, so called, semantification process analyzes
the structure of the formula and its textual context to conclude semantic information. The
research for this semantification process additionally contributes towards related Mathematical
Information Retrieval (MathIR) tasks, such as mathematical education assistance, math recom-
mendation and question answering systems, search engines, automatic plagiarism detection,
and math type assistance systems.
Second, this thesis demonstrates the first context-aware LATEX to CAS translation framework
LACAST. LACAST uses the developed semantification approach to transform LATEX expressions
into an intermediate semantic LATEX format, which is then further translated to CAS based
on translation patterns. These patterns were manually crafted by mathematicians to assure
accurate and reliable translations. In comparison, this thesis additionally elaborates a non-
context aware neural machine translation approach trained on a mathematical library generated
by Mathematica.
Third, the thesis provides a novel approach to evaluate the performance for LATEX to CAS
translations on large-scaled datasets with an automatic verification of equations in digital math-
ematical libraries. This evaluation approach is based on the assumption that equations in digital
mathematical libraries can be computationally verified by CAS, if a translation between both
systems exists. In addition, the thesis provides an in-depth manual evaluation on mathematical
articles from the English Wikipedia.
The presented context-aware translation framework LACAST increases the efficiency and reliability
of translations to CAS. Via LACAST, we strengthened the Digital Library of Mathematical Functions
(DLMF) by identifying numerous of issues, from missing or wrong semantic annotations to sign
errors. Further, via LACAST, we were able to discover several issues with the commercial CAS
Maple and Mathematica. The fundamental approaches to semantically enhance mathematics
developed in this thesis additionally contributed towards several related MathIR tasks. For

xiii
 instance, the large-scale analysis of mathematical notations and the studies on math-embeddings
motivated new approaches for math plagiarism detection systems, search engines, and allow
typing assistance for mathematical inputs. Finally, LACAST translations will have a direct real-
world impact, as they are scheduled to be integrated into upcoming versions of the DLMF and
Wikipedia.

xiv Abstract
FRONT MATTER

Zusammenfassung

Diese Dissertation befasst sich mit der Problematik von Übersetzungen mathematischer For-
meln zwischen LATEX und Computeralgebrasystemen (CAS). Im Laufe des digitalen Zeitalters
wurde LATEX zum Quasistandard für das Schreiben mathematischer Formeln auf dem Computer,
insbesondere in den Disziplinen Mathematik, Informatik, Naturwissenschaften und Technik
(MINT). Da Wissenschaftler gemeinhin ihre Arbeit publizieren, ist LATEX zu einem integralen
Bestandteil moderner Forschung geworden. Gleichermaßen verlassen sich Wissenschaftler
immer mehr auf die Möglichkeiten moderner CAS, um effektiv mit mathematischen Formeln
zu arbeiten, zum Beispiel, indem sie diese umformen, lösen oder auch visualisieren. Die mo-
mentanen Ansätze, welche eine Übersetzung von LATEX zu CAS erlauben, wie beispielsweise
interne Import-Funktionen einiger CAS, sind jedoch häufig auf einfache arithmetische Aus-
drücke beschränkt und daher nur wenig hilfreich im realen Arbeitsalltag. Infolgedessen ist die
Arbeit moderner Wissenschaftler in den MINT Disziplinen häufig geprägt von zeitraubenden
und fehleranfälligen manuellen Übersetzungen zwischen LATEX und CAS.
Die vorliegende Dissertation leistet die folgenden Beiträge, um das Problem des Übersetzens
von mathematischen Ausdrücken zwischen LATEX und CAS zu lösen.
Zunächst ist LATEX ein Format, welches lediglich die visuelle Präsentation mathematischer Aus-
drücke kodiert, nicht jedoch deren semantische Informationen. Die semantischen Informationen
sind jedoch notwendig für CAS, welche keine mehrdeutigen Eingaben erlauben. Daher führt
die vorliegende Arbeit als ersten Schritt für eine Übersetzung eine sogenannte Semantifizierung
mathematischer Ausdrücke ein. Diese Semantifizierung extrahiert semantische Informationen
aus dem Kontext und den Bestandteilen der Formel, um Rückschlüsse auf ihre Bedeutung zu
ziehen. Da die Semantifizierung eine klassische Aufgabe auf dem Gebiet der mathematischen
Informationsgewinnung darstellt, leistet dieser Teil der Dissertation auch Beiträge zu verwand-
ten Themengebieten. So sind die hier vorgestellten Ansätze auch nützlich für pädagogische
Programme, Frage-Antwort Systeme, Suchmaschinen und die digitale Plagiatserkennung.
Als zweiten Beitrag, stellt die vorliegende Dissertation das erste kontextbezogene LATEX zu
CAS Übersetzungsprogramm vor, genannt LACAST. LACAST nutzt die zuvor eingeführte Seman-
tifizierung, um LATEX in ein Zwischenformat zu transformieren, welches die semantischen
Informationen explizit darstellt. Dieses Format wird semantisches LATEX genannt, da es eine
technische Erweiterung von LATEX ist. Die weitere Übersetzung zu CAS wird durch heuristi-
sche Übersetzungsmuster für mathematische Funktionen realisiert. Diese Übersetzungsmuster
wurden in Zusammenarbeit mit Mathematikern definiert, um eine korrekte Übersetzung in
diesem letzten Schritt zu gewährleisten. Um die Vorzüge einer kontextbezogenen Übersetzung
besser zu verstehen, stellt diese Arbeit zum Vergleich auch eine Maschinenübersetzung auf
neuronalen Netzen vor, welche den Kontext einer Formel nicht berücksichtigt.
Der dritte Beitrag dieser Dissertation führt eine neue Methode zur Evaluierung von mathe-
matischen Übersetzungen ein, welche es erlaubt, auch eine große Anzahl an Übersetzungen
auf ihre Korrektheit hin zu überprüfen. Diese Methode folgt dem Ansatz, dass Gleichungen

xv
 in mathematischen Bibliotheken auch nach der Übersetzung in ein CAS noch korrekt sein
müssten. Ist dies nicht der Fall, ist entweder die Ausgangsgleichung, die Übersetzung, oder
das CAS fehlerhaft. Hierbei ist zu beachten, dass jede Fehlerquelle einen Mehrwert für das
jeweilige System darstellt. Zusätzlich zu dieser automatischen Evaluierung, erfolgt noch eine
manuelle Analyse von Übersetzungen auf Basis englischer Wikipedia Artikel.
Zusammenfassend ermöglicht das kontextbezogene Übersetzungsprogramm LACAST eine effizi-
entere Arbeitsweise mit CAS. Mit Hilfe dieser Übersetzungen konnten auch mehrere Probleme,
wie falsche Informationen oder Vorzeichenfehler, in der Digital Library of Mathematical Func-
tions (DLMF) sowie Fehler in den kommerziell vertriebenen CAS Maple und Mathematica
automatisch aufgedeckt und behoben werden.
Die hier vorgestellte Grundlagenforschung zum semantischen Anreichern mathematischer
Ausdrücke, hat zudem etliche Beiträge zu verwandten Forschungsthemen geleistet. Zum Bei-
spiel hat die Analyse der Verteilung von mathematischen Notationen in großen Datensätzen
neue Ansätze in der digitalen Plagiatserkennung ermöglicht. Des Weiteren wird zurzeit daran
gearbeitet, die Übersetzungen von LACAST in kommende Versionen von Wikipedia und der DLMF
zu integrieren.

xvi Zusammenfassung
FRONT MATTER

Acknowledgements

This thesis would not have been possible without the tremendous help and support from nu-
merous family members, friends, colleagues, supervisors, and several international institutions.
In the following, I want to take the opportunity to thank all the individuals and organizations
that helped me along the way to make this work possible.
My first sincere wishes go to my prodigious doctoral advisers Bela Gipp and Akiko Aizawa.
Their continuous support and counsel enabled me to realize this thesis at marvelous places and
together with numerous wonderful people from all over the world. Their enduring encourage-
ment and assistance, Bela’s abiding and infectious positivity, and Akiko’s steadfast and kind
endorsement empowered my personal and professional life. Both of their competent and sincere
guidance helped me to find my way in the intricate maze of research and career decisions and
turned my often onerous time into a joyful and memorable experience.
Moreover, I am very grateful to my adviser and friend Moritz Schubotz, who supported and
guided me throughout the entire time of my doctoral thesis and even beyond. Our fruitful and
always engaging discussions, even when exhausting, enriched and positively affected most, if
not all, of my work. It is not an exaggeration to admit that my career, including my Master’s
thesis and this doctoral thesis, would not have been possible and nearly as successful and joyful
as it has been without his continuous and sincere support over the years. I am wholeheartedly
thankful for all the years we worked together.
I further wish to gratefully acknowledge my friends, colleagues, and advisers Howard Cohl,
Abdou Youssef, and Bruce Miller at the National Institute of Standards and Technology (NIST) for
their valuable advice, continuous drive to perfection, and our rewarding collaborations. I thank
Jürgen Gerhard at Maplesoft, who kindly provided me access and support for Maple on several
occasions. I am just as thankful for the assistance and support from Norman Meuschke, who
always helped me to overcome governmental and organizational hurdles, Corinna Breitinger,
who never failed to refit my gibberish, and my colleagues and friends Terry Lima Ruas and
Philipp Scharpf for many visionary discussions. I also thank all my collaborators and colleagues
with whom I had the distinct opportunity to work together, including Takuto Asakura, Fabian
Müller, Olaf Teschke, William Grosky, Marjorie McClain, Yusuke Miyao, Malte Ostendorff,
Bonita Saunders, Kenichi Iwatsuki, Takuma Udagawa, Anastasia Zhukova, and Felix Hamborg.
I further want to thank the students I worked with, including Avi Trost, Rajen Dey, Joon Bang,
Kevin Chen, and Felix Petersen. I especially appreciate the help and assistance from people
at the National Institute of Informatics (NII) to overcome governmental and daily life issues. I
wish to especially thank Rie Ayuzawa, Noriko Katsu, Akiko Takenaka, and Goran Topic.
My genuine gratitude also goes to my host organizations and those that provided financial
support for my research. I am thankful for the German Academic Exchange Program (DAAD)
for enabling two research stays at the NII in Tokyo, the NII for providing me a wonderful
work environment, the German Research Foundation (DFG) to financially support many of
my projects, the NIST for hosting me as a guest researcher, and Maplesoft for offering me

xvii
 an internship during my preliminary research project on the Digital Library of Mathematical
Functions (DLMF). I finally thank the ACM Special Interest Group on Information Retrieval
(SIGIR), the University of Konstanz, the University of Wuppertal, and Maplesoft for supporting
several conference participations.
My last and most crucial gratitude goes to my family and friends, who always cheered me in
good and bad times and constantly backed and supported me so that I could selfishly pursue my
dreams. I am deeply grateful for my lovely parents Rolf & Regina, who have always been on
my side and make all this possible behind the scenes. I am also tremendously thankful for the
enduring personal support from my dear friends Kevin, Lena, Vici, Dong, Peter, Vitor, Ayuko,
and uncountably more. Finally, I thank my lovely partner Aimi for brightening even the darkest
times and pushing every possible obstacle aside. I dedicate this thesis to my lovely parents, my
dear friends, and my enchanting girlfriend.

xviii Acknowledgements
Other documents randomly have
different content
feeling of a cavern wall will grow upon him, of a cavern deep, below
roaring seas, in which the waves are there, though they do not enter
in upon him; or rather not the waves, but the very bowels of the
ocean. He will feel as though the floods surrounded him, coming and
going with their wild sounds, and he will hardly recognize that
though among them he is not in them. And they, as they fall with a
continual roar, not hurting the ear, but musical withal, will seem to
move as the vast ocean waters may perhaps move in their internal
currents. He will lose the sense of one continued descent, and think
that they are passing round him in their appointed courses. The
broken spray that rises from the depth below, rises so strongly, so
palpably, so rapidly, that the motion in every direction will seem
equal. And, as he looks on, strange colours will show themselves
through the mist; the shades of grey will become green or blue, with
ever and anon a flash of white; and then, when some gust of wind
blows in with greater violence, the sea-girt cavern will become all
dark and black. Oh, my friend, let there be no one there to speak to
thee then; no, not even a brother. As you stand there speak only to
the waters.
North America (London, 1862).
 NIAGARA FALLS
(NORTH AMERICA)
CHARLES DICKENS
We called at the town of Erie, at eight o’clock that night, and lay
there an hour. Between five and six next morning, we arrived at
Buffalo, where we breakfasted; and being too near the Great Falls to
wait patiently anywhere else, we set off by the train, the same
morning at nine o’clock, to Niagara.
It was a miserable day; chilly and raw; a damp mist falling; and the
trees in that northern region quite bare and wintry. Whenever the
train halted, I listened for the roar; and was constantly straining my
eyes in the direction where I knew the Falls must be, from seeing
the river rolling on towards them; every moment expecting to behold
the spray. Within a few moments of our stopping, not before, I saw
two great white clouds rising up slowly and majestically from the
depths of the earth. That was all. At length we alighted; and then
for the first time, I heard the mighty rush of water, and felt the
ground tremble underneath my feet.
The bank is very steep, and was slippery with rain, and half-melted
ice. I hardly know how I got down, but I was soon at the bottom,
and climbing, with two English officers who were crossing and had
joined me, over some broken rocks, deafened by the noise, half-
blinded by the spray, and wet to the skin. We were at the foot of the
American Fall. I could see an immense torrent of water tearing
headlong down from some great height, but had no idea of shape,
or situation, or anything but vague immensity.
When we were seated in the little ferry-boat, and were crossing the
swollen river immediately before both cataracts, I began to feel what
it was—but I was in a manner stunned, and unable to comprehend
the vastness of the scene. It was not until I came on Table Rock,
and looked—Great Heaven, on what a fall of bright green water!—
that it came upon me in its full might and majesty.
Then, when I felt how near to my Creator I was standing, the first
effect, and the enduring one—instant and lasting—of the
tremendous spectacle, was Peace. Peace of Mind, tranquillity, calm
recollections of the Dead, great thoughts of Eternal Rest and
Happiness: nothing of gloom or terror. Niagara was at once stamped
upon my heart, an Image of Beauty; to remain there, changeless
and indelible, until its pulses cease to beat, for ever.
Oh, how the strife and trouble of daily life receded from my view,
and lessened in the distance, during the ten memorable days we
passed on that Enchanted Ground! What voices spoke from out the
thundering water; what faces, faded from the earth, looked out
upon me from its gleaming depths; what Heavenly promise glistened
in those angels’ tears, the drops of many hues, that showered
around, and twined themselves about the gorgeous arches which
the changing rainbows made!
I never stirred in all that time from the Canadian side, whither I had
gone at first. I never crossed the river again; for I knew there were
people on the other shore, and in such a place it is natural to shun
strange company. To wander to and fro all day, and see the cataracts
from all points of view; to stand upon the edge of the great Horse-
Shoe Fall, marking the hurried water gathering strength as it
approached the verge, yet seeming too, to pause before it shot into
the gulf below; to gaze from the river’s level up at the torrent as it
came streaming down; to climb the neighbouring heights and watch
it through the trees, and see the wreathing water in the rapids
hurrying on to take its fearful plunge; to linger in the shadow of the
solemn rocks three miles below; watching the river as, stirred by no
visible cause, it heaved and eddied and awoke the echoes, being
troubled yet, far down beneath its surface, by its giant leap; to have
Niagara before me, lighted by the sun and by the moon, red in the
day’s decline, and grey as evening slowly fell upon it; to look upon it
every day, and wake up in the night and hear its ceaseless voice:
this was enough.

NIAGARA FALLS IN WINTER.

I think in every quiet season now, still do those waters roll and leap,
and roar and tumble, all day long; still are the rainbows spanning
them, a hundred feet below. Still, when the sun is on them, do they
shine and glow like molten gold. Still, when the day is gloomy, do
they fall like snow, or seem to crumble away like the front of a great
chalk cliff, or roll down the rock like dense white smoke. But always
does the mighty stream appear to die as it comes down, and always
from its unfathomable grave arises that tremendous ghost of spray
and mist which is never laid: which has haunted this place with the
same dread solemnity since Darkness brooded on the deep, and that
first flood before the Deluge—Light—came rushing on Creation at
the word of God.
American Notes for General Circulation (London, 1842).
 FUJI-SAN
(JAPAN)
SIR EDWIN ARNOLD
I have just made in the company of Captain John Ingles, R. N.,
Naval Adviser to the Imperial Government of this country, and a
young Japanese gentleman—Mr. Asso—a very fortunate and
delightful ascent of Fuji-San, the famous mountain—you would not
wonder, residing here, that everybody in Japan talks about Fuji, and
thinks about her; paints her on fans, and limns her with gold on
lacquer; carves her on temple-gates and house-fronts, and draws
her for curtains of shops and signboards of inns, rest-houses and
public institutions. Living in Tokio or Yokohama, or anywhere along
this Tokaidô—the Southern road of Japan—you would soon perceive
how the great volcano dominates every landscape, asserts
perpetually her sovereignty over all other hills and mountains, and
becomes in reality as well as imagination, an indispensable element
in the national scenery. Far away at sea, when approaching Japan, if
the weather be clear, long before the faintest blue line of coast is
discernible from the deck, there is seen hanging in the air a dim
white symmetrical cone, too constant for a cloud, which is Fuji-San.
After you have landed and taken up your residence in Yokohama,
Tokio, or any point of the southeastern littoral, you will be always
seeing Fuji-Yama from some garden-nook, some tea-house gallery,
some grove of cryptomerias, or thicket of bamboo, or even from the
railway-carriage window. In the spring and autumn, as frequently as
not, she will, indeed, be shrouded in the dense masses of white or
grey cumulus which her crest collects, and seems to create in the
mists of the Pacific. But during summer, when the snows are all
melted from the vast cone, and again in winter, when she is covered
with snow half-way down her colossal sides, but the air is clear, the
superb mountain stands forth, dawn after dawn, and evening after
evening—like no other eminence in the world for beauty, majesty,
and perfectness of outline. There are loftier peaks, of course, for
Fuji-San is not much higher than Mont Blanc, but there is none—not
even Etna—which rises so proudly alone, isolated, distinct, from the
very brink of the sea—with nothing to hide or diminish the dignity of
the splendid and immense curves sweeping up from where the
broad foot rests, planted on the Suruga Gulf, to where the imperial
head soars, lifted high above the clouds into the blue of the
firmament. By many and many a picture or photograph you must
know well those almost perfectly matched flanks, that massive base,
the towering lines of that mighty cone, slightly truncated and
dentated at the summit. But no picture gives, and no artist could
ever reproduce, the variety and charm of the aspect which Fuji-San
puts on from day to day and hour to hour under the differing
influences of air and weather. Sometimes it is as a white cloud that
you see her, among the white clouds, changeless among the
changeful shapes from which she emerges. Sometimes there will
break forth, high above all clouds, a patch of deep grey against the
blue, the broad head of Fuji. Sometimes you will only know where
she sits by the immense collection of cirrus and cirro-cumulus there
alone gathered in the sky; and sometimes—principally at dawn and
nightfall—she will suddenly manifest herself, from her foot, jewelled
with rich harvests, to her brow, bare and lonely as a desert—all
violet against the gold of the setting sun, or else all gold and green
against the rose and silver of the daybreak....
As late as the Fourteenth Century Fuji was constantly smoking, and
fire is spoken of with the eruptions, the last of which took place in
December, 1707, and continued for nearly forty days. The Ho-Yei-
san, or hump in the south face, was probably then formed. In this,
her final outbreak, Fuji covered Tokio itself, sixty miles away, with six
inches of ash, and sent rivers of lava far and wide. Since then she
has slept, and only one little spot underneath the Kwan-nom-Gatake,
on the lip of the crater, where steam exhales, and the red pumice-
cracks are hot, shows that the heart of this huge volcano yet glows,
and that she is capable of destroying again her own beauty and the
forests and rich regions of fertility which clothe her knees and feet.

FUJI SAN.
It is a circuit of 120 miles to go all round the base of Fuji-San. If you
could cut a tunnel through her from Yoshiwara to Kawaguchi, it
would be forty miles long. Generally speaking, the lower portion of
the mountain is cultivated to a height of 1,500 feet, and it is a whole
province which thus climbs round her. From the border of the farms
there begins a rough and wild, but flowery moorland, which
stretches round the hill to an elevation of 4,000 feet, where there
the thick forest-belt commences. This girdles the volcano up to 7,000
feet on the Subashiri side and 8,000 on the Murayama fall, but is
lower to the eastward. Above the forest extends a narrow zone of
thicket and bush, chiefly dwarfed larch, juniper and a vaccinium;
after which comes the bare, burnt, and terribly majestic peak itself,
where the only living thing is a little yellow lichen which grows in the
fissures of the lava blocks, for no eagle or hawk ventures so high,
and the boldest or most bewildered butterfly will not be seen above
the bushes half-way down.
The best—indeed, the only—time for the ascent of the mountain is
between July 15th and September 5th. During this brief season the
snow will be melted from the cone, the huts upon the path will be
opened for pilgrims, and there will be only the danger of getting
caught by a typhoon, or reaching the summit to find it swathed day
after day in clouds, and no view obtainable. Our party of three
started for the ascent on August 25th, taking that one of the many
roads by which Fuji is approached that goes by Subashiri. Such an
expedition may be divided into a series of stages. You have first to
approach the foot of the mountain by train or otherwise, then to ride
through the long slope of cultivated region. Then, abandoning
horses or vehicles, to traverse on foot the sharper slopes of the
forest belt. At the confines of this you will reach the first station,
called Sho or Go; for Japanese fancy has likened the mountain to a
heap of dry rice and the stations are named by rice-measure. From
the first station to the ninth, whatever road you take, all will be
hard, hot, continuous climbing. You must go by narrow, bad paths,
such as a goat might make, in loose volcanic dust, gritty pumice, or
over the sharp edges of lava dykes, which cut boots and sandals to
shreds....
At daybreak the horses are brought, and the six coolies, two by two,
bind upon their backs the futons and the food. We start, a long
procession, through a broad avenue in the forest, riding for five
miles, under a lovely dawn, the sun shining gloriously on the
forehead of Fuji, who seems further off and more immensely lofty
the nearer we approach. The woodland is full of wild strawberries
and flowers; including tiger-lilies, clematis, Canterbury bells, and the
blue hotari-no hana, or fire-fly blossom. At 6:30 a. m., we reach
Uma-Gayeshi, or “turn-the-horses-back”; and hence to the mountain
top there is nothing for it but to walk every step of the long, steep,
and difficult path. Two of the men with the lightest loads led the way
along the narrow path, in a wood so thick that we shall not see Fuji
again till we have passed through it. It takes us every now and then
through the gates and precincts of little Shinto temples, where the
priests offer us tea or mountain water. In one of them, at Ko-mitake,
we are invited to ring the brass gong in order that the Deity may
make our limbs strong for the task before us. And this is solemnly
done by all hands, the ninsoku slapping their brown thighs piously
after sounding the bell....
The shortest time in which the ascent has been made is six hours
and a half. We, taking it more easily, made no attempt to beat the
record, and stopped frequently to botanize, geologize, etc. The
rarefaction of the air gave our Japanese companion, Takaji San, a
slight headache, which soon passed as the circulation became
accustomed to the atmosphere; but Captain Ingles and I, being I
suppose, both in excellent health and strength, experienced no
inconvenience worth mentioning.
At half-past four next morning, while I was dreaming under my thick
coverings, a hand touched me and a voice said softly, “Danna Sama,
hi no de!” “Master, here is the sun!” The shoji at my feet were
thrown open. I looked out, almost as you might from the moon, over
a prodigious abyss of space, beyond which the eastern rim of all the
world seemed to be on fire with flaming light. A belt of splendid rose
and gold illumined all the horizon, darting long spears of glory into
the dark sky overhead, gilding the tops of a thousand hills, scattered
over the purple plains below, and casting on the unbroken
background of clouds beyond an enormous shadow of Fuji. The
spectacle was of unparalleled splendour, recalling Lord Tennyson’s
line—

“And, in the East,

God made himself an awful Rose of Dawn.”

Moment by moment it grew more wonderful in loveliness of colour

and brilliant birth of day; and then, suddenly, just when the sun
rolled into sight—an orb of gleaming gold, flooding the world
beneath with almost insufferable radiance—a vast mass of dense
white clouds swept before the north wind over the view, completely
blotting out the sun, the belt of rose and gold, the lighted mountains
and plains, and the lower regions of Fuji-San. It was day again, but
misty, white, and doubtful; and when we started to climb the last
two stages of the cone the flags of the stations were invisible, and
we could not know whether we should find the summit clear, or
wrapped in enveloping clouds.
All was to be fortunate, however, on this happy day; and after a
hard clambering of the remaining 2,000 feet we planted our staffs
victoriously on the level ground of the crater’s lip and gazed north,
south, east, and west through clear and cloudless atmosphere over a
prodigious prospect, whose diameter could not be less than 300
miles. It was one of the few days when O-ana-mochi, the Lord of
the Great Hole, was wholly propitious! Behind the long row of little
black huts standing on the edge of the mountain, gaped that awful,
deadly Cup of the Volcano—an immense pit half a mile wide and six
or seven hundred feet deep, its sides black, yellow, red, white, and
grey, with the varying hues of the lava and scoriæ. In one spot
where a perpetual shadow lay, from the ridge-peaks of Ken-ga-mine
and the Shaka-no-wari-ishi, or “Cleft Rock of Buddha,” gleamed a
large patch of unmelted snow, and there was dust-covered snow at
the bottom of the crater. We skirted part of the crater, passed by the
dangerous path which is styled “Oya-shirazu, Ko-shirazu,” “The place
where you must forget parents and children, to take care of
yourself;” saw the issue of the Kim-mei-sai or “Golden famous
water,” and of the Gim-mei-sai, or “Silver famous water”; and came
back to breakfast at our hut silent with the delight and glory, the
beauty and terror of the scene. Enormous flocks of fleecy clouds and
cloudlets wandered in the lower air, many thousand feet beneath,
but nowhere concealed the lakes, peaks, rivers, towns, villages,
valleys, sea-coasts, islands, and distant provinces spreading out all
round. Imagine the prospect obtainable at 13,000 feet of elevation
through the silvery air of Japan on a summer’s morning with not a
cloud, except shifting, thin, and transitory ones, to veil the view!...
At the temple with the bell we were duly stamped—shirts, sticks,
and clothing—with the sacred mark of the mountain, and having
made the hearts of our faithful and patient ninsoku glad with extra
pay, turned our backs on the great extinct volcano, whose crest,
glowing again in the morning sunlight, had no longer any secrets for
Captain Ingles, or Takaji San, or myself.
Seas and Lands (New York, 1891).
 THE CEDARS OF LEBANON
(SYRIA)
ALPHONSE DE LAMARTINE
The Sheik of Eden, the last inhabited village towards the summit of
Lebanon, was the maternal uncle of M. Mazoyer, my interpreter.
Informed by his nephew of our arrival in Tripoli, the venerable sheik
descended the mountain with his eldest son and a portion of his
retinue; he came to visit me at the convent of the Franciscans, and
offered me hospitality at his home in Eden. From Eden to the Cedars
of Solomon it is only a three hours’ march; and if the snows that
cover the mountains will permit us, we can visit these ancient trees
that have spread their glory over all Lebanon and that are
contemporaries of the great king; we accepted, and the start was
arranged for the following day.
At five o’clock in the morning we were on horseback. The caravan,
more numerous than usual, was preceded by the Sheik of Eden, an
admirable old man whose elegance of manner, noble and easy
politeness, and magnificent costume were far from suggesting an
Arab chieftain; one would have called him a patriarch marching at
the head of his tribe; he rode upon a mare of the desert whose
golden-bay skin and floating mane would have made a worthy
mount for a hero of Jerusalem; his son and his principal attendants
caracoled upon magnificent stallions, a few paces before him; we
came next, and then the long file of our moukres and our Saïs....
The sheik has sent three Arabs over the route to the Cedars to learn
if the snow will permit us to approach those trees; the Arabs
returning say that access is impracticable; there are fourteen feet of
snow in a narrow valley which must be crossed before reaching the
trees;—wishing to get as near as possible, I entreat the sheik to give
me his son and several horsemen; I leave my wife and my caravan
at Eden; I mount the strongest of my horses, Scham, and we are en
route at break of day;—a march of three hours over the crests of the
mountains, or in the fields softened with melting snow. I arrive at
the edge of the valley of the Saints, a deep gorge where the glance
sweeps from the rocky height to a valley more confined, more
sombre and more solemn even than that of Hamana; at the top of
this valley, at the place where, after continually rising, it reaches the
snows, a superb sheet of water falls, a hundred feet high and two or
three toises wide; the entire valley resounds with this waterfall and
the leaping torrents that it feeds; on every side the rocky flanks of
the mountain stream with foam; we see almost beyond our vision, in
the depths of the valley, two large villages the houses of which can
scarcely be distinguished from the rocks rolled down by the torrent;
the tops of the poplars and the mulberries from here look like tufts
of reed or grass; we descend to the village of Beschieraï by paths
cut in the rock, and so abrupt that one can hardly imagine that men
will risk themselves upon them; people do perish sometimes; a
stone thrown from the crest where we stand would fall upon the
roofs of these villages where we shall arrive after an hour’s descent;
above the cascade and the snows, enormous fields of ice extend,
undulating like vapours in tints greenish and blue by turns; in about
a quarter of an hour towards the left in a half circular valley formed
by the last mounts of Lebanon, we see a large, black blot upon the
snow,—the famous group of cedars; they crown the brow of the
mountain like a diadem; they mark the branching off of numerous
and large valleys that descend from there; the sea and the sky are
their horizon.
We put our horses to a gallop over the snow to get as near as
possible to the forest; but on arriving five or six hundred steps from
the trees, we plunge our horses up to their shoulders; we realize
that the report of the Arabs is correct, and we must renounce the
hope of touching these relics of the centuries and of nature; we
alight and sit upon a rock to contemplate them.
 THE CEDARS OF LEBANON.
These trees are the most celebrated natural monuments in the
whole universe. Religion, poetry, and history have equally
consecrated them. Holy Writ celebrates them in several places. They
are one of the favourite images which the prophets employ. Solomon
wished to consecrate them—doubtless on account of the renown of
magnificence and sanctity that these prodigies of vegetation enjoyed
at this epoch—to the ornamentation of the temple that he was the
first to elevate to the one God. These were certainly the trees; for
Ezekiel speaks of the cedars of Eden as the most beautiful of
Lebanon. The Arabs of all sects have a traditional veneration for
them. They attribute to these trees, not only a vegetative force that
gives them eternal life, but even a soul that makes them give signs
of wisdom, of foresight, similar to those of instinct in animals and
intelligence in men. They know the seasons in advance; they move
their enormous branches like human limbs, they spread or contract
their boughs, they raise their branches towards the sky or incline
them to the earth, according as the snow is preparing to fall or to
melt. They are divine beings under the form of trees. They grow on
this single spot of the mounts of Lebanon; they take root far beyond
the region where all prolific vegetation dies. All this strikes the
imagination of the Oriental people with astonishment, and I do not
know that science is not even more astonished. Alas! however,
Basan languishes and Carmel and the flower of Lebanon fade.—
These trees diminish every century. Travellers formerly counted
thirty or forty, later seventeen, and still later, about a dozen.—There
are now only seven of those whose massive forms can presume to
be contemporaneous with Biblical times. Around these old memorials
of past ages, which know the history of the ground better than
history herself, and which could tell us, if they could speak, of many
empires, religions, and vanished human races, there remains still a
little forest of cedars more yellow it appears to me than a group of
four or five hundred trees or shrubs. Each year in the month of June
the population of Beschieraï, Eden, and Kanobin, and all the villages
of the neighbouring valleys, ascend to the cedars and celebrate
mass at their feet. How many prayers have resounded beneath their
branches! And what more beautiful temple, what nearer altar than
the sky! What more majestic and holier daïs than the highest
plateau of Lebanon, the trunks of the cedars and the sacred boughs
that have shaded and that will still shade so many human
generations pronouncing differently the name of God, but who
recognize him everywhere in his works and adore him in his
manifestations of nature! And I, I also prayed in the presence of
those trees. The harmonious wind that resounded through their
sonorous branches played in my hair and froze upon my eyelids
those tears of sorrow and adoration.
Voyage en Orient (Paris, 1843).
 THE GIANT’S CAUSEWAY
(IRELAND)
WILLIAM MAKEPEACE THACKERAY
The road to the Causeway is bleak, wild, and hilly.
The cabins along the road are scarcely better than those of Kerry,
the inmates as ragged, and more fierce and dark-looking. I never
was so pestered by juvenile beggars as in the dismal village of
Ballintoy. A crowd of them rushed after the car, calling for money in
a fierce manner, as if it was their right; dogs as fierce as the children
came yelling after the vehicle; and the faces which scowled out of
the black cabins were not a whit more good-humoured. We passed
by one or two more clumps of cabins, with their turf and corn-stacks
lying together at the foot of the hills; placed there for the
convenience of the children, doubtless, who can thus accompany the
car either way, and shriek out their “Bonny gantleman, gi’e us a
ha’p’ny.” A couple of churches, one with a pair of its pinnacles blown
off, stood in the dismal open country, and a gentleman’s house here
and there: there were no trees about them, but a brown grass round
about—hills rising and falling in front, and the sea beyond. The
occasional view of the coast was noble; wild Bengore towering
eastwards as we went along; Raghery Island before us, in the steep
rocks and caves of which Bruce took shelter when driven from
yonder Scottish coast, that one sees stretching blue in the northeast.
I think this wild gloomy tract through which one passes is a good
prelude for what is to be the great sight of the day, and got my mind
to a proper state of awe by the time we were near the journey’s
end. Turning away shorewards by the fine house of Sir Francis
Macnaghten, I went towards a lone handsome inn, that stands close
to the Causeway. The landlord at Ballycastle had lent me Hamilton’s
book to read on the road; but I had not time then to read more than
half-a-dozen pages of it. They described how the author, a
clergyman distinguished as a man of science, had been thrust out of
a friend’s house by the frightened servants one wild night, and
butchered by some Whiteboys who were waiting outside and called
for his blood. I had been told at Belfast that there was a corpse in
the inn: was it there now? It had driven off, the car-boy said, “in a
handsome hearse and four to Dublin the whole way.” It was gone,
but I thought the house looked as if the ghost was there. See,
yonder are the black rocks stretching to Portrush: how leaden and
grey the sea looks! how grey and leaden the sky! You hear the
waters rushing evermore, as they have done since the beginning of
the world. The car drives us with a dismal grinding noise of the
wheels to the big lone house: there’s no smoke in the chimneys; the
doors are locked. Three savage-looking men rush after the car: are
they the men who took out Mr. Hamilton—took him out and
butchered him in the moonlight? Is everybody, I wonder, dead in
that big house? Will they let us in before those men are up? Out
comes a pretty smiling girl, with a curtsey, just as the savages are at
the car, and you are ushered into a very comfortable room; and the
men turn out to be guides. Well, thank Heaven it’s no worse! I had
fifteen pounds still left; and, when desperate, have no doubt should
fight like a lion.
 THE GIANT’S LOOM, GIANT’S CAUSEWAY.
The traveller no sooner issues from the inn by a back door, which he
is informed will lead him straight to the Causeway, than the guides
pounce upon him, with a dozen rough boatmen who are likewise
lying in wait; and a crew of shrill beggar-boys, with boxes of spars,
ready to tear him and each other to pieces seemingly, yell and bawl
incessantly round him. “I’m the guide Miss Henry recommends,”
shouts one. “I’m Mr. Macdonald’s guide,” pushes in another. “This
way,” roars a third, and drags his prey down a precipice; the rest of
them clambering and quarrelling after. I had no friends; I was
perfectly helpless. I wanted to walk down to the shore by myself, but
they would not let me, and I had nothing for it but to yield myself
into the hands of the guide who had seized me, who hurried me
down the steep to a little wild bay, flanked on each side by rugged
cliffs and rocks, against which the waters came tumbling, frothing,
and roaring furiously. Upon some of these black rocks two or three
boats were lying: four men seized a boat, pushed it shouting into the
water, and ravished me into it. We had slid between two rocks,
where the channel came gurgling in: we were up one swelling wave
that came in a huge advancing body ten feet above us, and were
plunging madly down another (the descent causes a sensation in the
lower regions of the stomach which it is not at all necessary here to
describe), before I had leisure to ask myself why the deuce I was in
that boat, with four rowers hurrooing and bounding madly from one
liquid mountain to another—four rowers whom I was bound to pay. I
say, the query came qualmishly across me why the devil I was there,
and why not walking calmly on the shore.
The guide began pouring his professional jargon into my ears.
“Every one of them bays,” says he, “has a name (take my place, and
the spray won’t come over you): that is Port Noffer, and the next,
Port na Gange; them rocks is the Stookawns (for every rock has its
name as well as every bay); and yonder—give way, my boys,—
hurray, we’re over it now: has it wet you much, sir?—that’s a little
cave: it goes five hundred feet under ground, and the boats goes
into it easy of a calm day.”
“Is it a fine day or a rough one now?” said I; the internal disturbance
going on with more severity than ever.
“It’s betwixt and between; or, I may say, neither one nor the other.
Sit up, sir. Look at the entrance of the cave. Don’t be afraid, sir;
never has an accident happened in any one of these boats, and the
most delicate ladies has rode in them on rougher days than this.
Now, boys, pull to the big cave. That, sir, is six hundred and sixty
yards in length, though some say it goes for miles inland, where the
people sleeping in their houses hear the waters roaring under them.”
The water was tossing and tumbling into the mouth of the little
cave. I looked,—for the guide would not let me alone till I did,—and
saw what might be expected: a black hole of some forty feet high,
into which it was no more possible to see than into a millstone. “For
Heaven’s sake, sir,” says I, “if you’ve no particular wish to see the
mouth of the big cave, put about and let us see the Causeway and
get ashore.” This was done, the guide meanwhile telling some story
of a ship of the Spanish Armada having fired her guns at two peaks
of rock, then visible, which the crew mistook for chimney-pots—what
benighted fools these Spanish Armadilloes must have been; it is
easier to see a rock than a chimney-pot; it is easy to know that
chimney-pots do not grow on rocks.—“But where, if you please, is
the Causeway?”
“That’s the Causeway before you,” says the guide.
“Which?”
“That pier which you see jutting out into the bay right ahead.”
“Mon dieu! and have I travelled a hundred and fifty miles to see
that?”
I declare, upon my conscience, the barge moored at Hungerford
Market is a more majestic object, and seems to occupy as much
space. As for telling a man that the Causeway is merely a part of the
sight; that he is there for the purpose of examining the surrounding
scenery; that if he looks to the westward he will see Portrush and
Donegal Head before him; that the cliffs immediately in his front are
green in some places, black in others, interspersed with blotches of
brown and streaks of verdure;—what is all this to a lonely individual
lying sick in a boat, between two immense waves that only give him
momentary glimpses of the land in question, to show that it is
frightfully near, and yet you are an hour from it? They won’t let you
go away—that cursed guide will tell out his stock of legends and
stories. The boatmen insist upon your looking at boxes of
“specimens,” which you must buy of them; they laugh as you grow
paler and paler; they offer you more and more “specimens”; even
the dirty lad who pulls number three, and is not allowed by his
comrades to speak, puts in his oar, and hands you over a piece of
Irish diamond (it looks like half-sucked alicompayne), and scorns
you. “Hurry, lads, now for it, give way!” how the oars do hurtle in
the rowlocks, as the boat goes up an aqueous mountain, and then
down into one of those cursed maritime valleys where there is no
rest as on shore!
At last, after they had pulled me enough about, and sold me all the
boxes of specimens, I was permitted to land at the spot whence we
set out, and whence, though we had been rowing for an hour, we
had never been above five hundred yards distant. Let all cockneys
take warning from this; let the solitary one caught issuing from the
back door of the hotel, shout at once to the boatmen to be gone—
that he will have none of them. Let him, at any rate, go first down to
the water to determine whether it be smooth enough to allow him to
take any decent pleasure by riding on its surface. For after all, it
must be remembered that it is pleasure we come for—that we are
not obliged to take those boats.—Well, well! I paid ten shillings for
mine, and ten minutes after would cheerfully have paid five pounds
to be allowed to quit it; it was no hard bargain after all. As for the
boxes of spar and specimens, I at once, being on terra firma, broke
my promise, and said I would see them all—first. It is wrong to
swear, I know; but sometimes it relieves one so much!

THE KEYSTONE, GIANT’S CAUSEWAY.

The first act on shore was to make a sacrifice to Sanctissima Tellus;
offering up to her a neat and becoming Taglioni coat, bought for a
guinea in Covent Garden only three months back. I sprawled on my
back on the smoothest of rocks that is, and tore the elbows to
pieces: the guide picked me up; the boatman did not stir, for they
had their will of me; the guide alone picked me up, I say, and bade
me follow him. We went across a boggy ground in one of the little
bays, round which rise the green walls of the cliff, terminated on
either side by a black crag, and the line of the shore washed by the
poluphloisboiotic, nay the poluphloisboiotatotic sea. Two beggars
stepped over the bog after us howling for money, and each holding
up a cursed box of specimens. No oaths, threats, entreaties, would
drive these vermin away; for some time the whole scene had been
spoiled by the incessant and abominable jargon of them, the
boatmen, and the guides. I was obliged to give them money to be
left in quiet, and if, as no doubt will be the case, the Giant’s
Causeway shall be a still greater resort of travellers than ever, the
county must put policemen on the rocks to keep the beggars away,
or fling them in the water when they appear.
And now, by force of money, having got rid of the sea and land
beggars, you are at liberty to examine at your leisure the wonders of
the place. There is not the least need for a guide to attend the
stranger, unless the latter have a mind to listen to a parcel of
legends, which may be well from the mouth of a wild simple peasant
who believes in his tales, but are odious from a dullard who narrates
them at the rate of sixpence a lie. Fee him and the other beggars,
and at last you are left tranquil to look at the strange scene with
your own eyes, and enjoy your own thoughts at leisure.
That is, if the thoughts awakened by such a scene may be called
enjoyment; but for me, I confess, they are too near akin to fear to
be pleasant; and I don’t know that I would desire to change that
sensation of awe and terror which the hour’s walk occasioned, for a
greater familiarity with this wild, sad, lonely place. The solitude is
awful. I can’t understand how those chattering guides dare to lift up
their voices here, and cry for money.
It looks like the beginning of the world, somehow: the sea looks
older than in other places, the hills and rocks strange, and formed
differently from other rocks and hills—as those vast dubious
monsters were formed who possessed the earth before man. The
hilltops are shattered into a thousand cragged fantastical shapes;
the water comes swelling into scores of little strange creeks, or goes
off with a leap, roaring into those mysterious caves yonder, which
penetrate who knows how far into our common world. The savage
rock-sides are painted of a hundred colours. Does the sun ever shine
here? When the world was moulded and fashioned out of formless
chaos, this must have been the bit over—a remnant of chaos! Think
of that!—it is a tailor’s simile. Well, I am a cockney: I wish I were in
Pall Mall! Yonder is a kelp-burner: a lurid smoke from his burning
kelp rises up to the leaden sky, and he looks as naked and fierce as
Cain. Bubbling up out of the rocks at the very brim of the sea rises a
little crystal spring: how comes it there? and there is an old grey hag
beside, who has been there for hundreds and hundreds of years,
and there sits and sells whisky at the extremity of creation! How do
you dare to sell whisky there, old woman? Did you serve old Saturn
with a glass when he lay along the Causeway here? In reply, she
says, she has no change for a shilling: she never has; but her whisky
is good.
This is not a description of the Giant’s Causeway (as some clever
critic will remark), but of a Londoner there, who is by no means so
interesting an object as the natural curiosity in question. That single
hint is sufficient; I have not a word more to say. “If,” says he, “you
cannot describe the scene lying before us—if you cannot state from
your personal observation that the number of basaltic pillars
composing the Causeway has been computed at about forty
thousand, which vary in diameter, their surface presenting the
appearance of a tesselated pavement of polygonal stones—that each
pillar is formed of several distinct joints, the convex end of the one
being accurately fitted in the concave of the next, and the length of
the joints varying from five feet to four inches—that although the
pillars are polygonal, there is but one of three sides in the whole
forty thousand (think of that!), but three of nine sides, and that it
may be safely computed that ninety-nine out of one hundred pillars
have either five, six, or seven sides; if you cannot state something
useful, you had much better, sir, retire and get your dinner.”
Never was summons more gladly obeyed. The dinner must be ready
by this time; so, remain you, and look on at the awful scene, and
copy it down in words if you can. If at the end of the trial you are
dissatisfied with your skill as a painter, and find that the biggest of
your words cannot render the hues and vastness of that tremendous
swelling sea—of those lean solitary crags standing rigid along the
shore, where they have been watching the ocean ever since it was
made—of those grey towers of Dunluce standing upon a leaden
rock, and looking as if some old old princess, of old old fairy times,
were dragon-guarded within—of yon flat stretches of sand where the
Scotch and Irish mermaids hold conference—come away, too, and
prate no more about the scene! There is that in nature, dear
Jenkins, which passes even our powers. We can feel the beauty of a
magnificent landscape, perhaps: but we can describe a leg of
mutton and turnips better. Come, then, this scene is for our betters
to depict. If Mr. Tennyson were to come hither for a month, and
brood over the place, he might, in some of those lofty heroic lines
which the author of the Morte d’Arthur knows how to pile up, convey
to the reader a sense of this gigantic desolate scene. What! you,
too, are a poet? Well, then Jenkins, stay! but believe me, you had
best take my advice, and come off.
The Irish Sketch-Book (London, 1843).
 THE GREAT GLACIER OF THE
SELKIRKS
(CANADA)
DOUGLAS SLADEN
If Banff represents the Rocky Mountains made easy, the Glacier
House represents the Selkirks made easy—a much more notable
performance, for these mountains had long been regarded as
impassable by engineering. The Glacier House is a few miles beyond
Rogers’ Pass, in the midst of the line’s greatest marvels of nature
and engineering. Just before comes the monarch of snow sheds; just
above the monarch of glaciers; just below the monarch of viaducts.
The Great Glacier of the Selkirks comes to a conclusion within a
couple of miles above it. The moraines and splintered forests at its
foot tell a frightful tale of destruction, and the glacier advances
every year; but only a few inches, so the hotel is safe for the
present.
The hotel is a pretty little châlet, mostly dining-room, with a trim,
level lawn in front containing a fine fountain. Eighteen miles broad is
the great Glacier of the Selkirks, one foot of which is planted so
threateningly above the hotel and the railway station, that it looks as
if it meant to stamp them out of existence with the stealth of a thief
in the night.
A marvellous and delightful walk it is from the hotel to the Glacier—
at first through dry woods of fir and spruce, and balsam and
tamarack, carpeted, wherever the sun breaks through, with purple
blueberries, wild raspberries, pigeon and salmon berries. Here you
might meet a grizzly bear any minute. You pause, if you are only a
man and a woman, on the lovers’ seat under the thousand-ton
boulder hurled down by the Glacier in the childhood of the earth.
Then you pass the fierce glacial torrent of grey-green water, so cold
or charged with impurities that fish refuse to live in it, swelling, as all
snow-fed rivers do, as the heat of a summer’s day waxes. Some of
its pools are huge and deep; some of its falls and rapids as fierce as
the cataract at Lorette, rounded boulders and splintered trunks
everywhere attesting its fury. The path crosses and recrosses the
river over bridges of tree-trunks, with smaller trunks loosely pinned
across them, like the little straw mats in which cream cheeses are
wrapped. As the path mounts, the scenery becomes more open, and
you are greeted, according to the season, with Canada’s gorgeous
lily or Canada’s prodigality of wild fruits; for you are in the track of
the glacier and the avalanche, and in the death of the forest is the
birth of blossoms and berries. All around you now is a scene of awful
grandeur—boulders as big as settlers’ huts, and giant tree trunks,
many of them blackened with fire, tossed together like the rubbish
on a dust-heap, and, brooding over all, the great Glacier like a
dragon crouching for the spring. One can hardly believe it is the
Glacier; the transitions are so abrupt. A turn of a path brings you
almost in contact with a piece of ice larger than any lake in the
British Islands. From under its skirts trickle tiny rills; a few feet
below, the rills league themselves into a river. Even a first-class
glacier is a disappointing affair if you go too close. Its blueness
disappears, also its luminosity, except in crevasses deep enough to
show you the pure heart of the ice. The surface is a dirty-looking
mixture of ice and snow. There were two lovely horizontal crevasses,
one so spacious and shining that it is called the Fairy Cavern. The
pleasure of standing in them is spoilt, because they look all the time
as if they were going to close on you. At another foot of the Glacier
there are immense moraines, looking like the earthworks of Dover
Castle. I examined them one October day when I went with a guide
to the top of the Glacier, eight thousand feet above sea-level, to see
the splendid Glacier-girdled head of Mount Fox on the other side of
the abyss.
I never intend to do any more mountain climbing through deep,
fresh snow. For the last hour or two of the ascent the snow was as
deep as one’s thighs at every step, and though the guide was towing
me by a rope tied round my waist, it was intolerably wearisome. To
begin with, he had to sound with his staff at every step and see that
we were on terra firma, and not on the soufflet of a crevasse; and
though there had been such a snowfall the night before, the sun was
as hot as summer overhead. The sight was worth doing once, with
the miles and miles of the sea of ice all round one, and the long
white slopes of virgin snow.
If it had not been for the aggressive visage of Mount Fox, it would
have answered to the description of the interior of Greenland given
me by Dr. Nansen, where the world consists of yourselves, the sun,
and the snow. We started at eight o’clock in the morning, but in
some way or other I was not quite as rapid as the guide had
calculated, for a couple of hours before nightfall he began to get
excited, if not alarmed. We were at the time clear of the deep snow,
and muddling about in a mixture of drifts and moraines; but after
dark he was not sure of his way until we struck the path at the foot
of the Glacier....
The Glacier House has not only its noble and easily accessible
glacier; it is in the very heart of the finest mountain scenery in the
Selkirks, which is so different to the scenery of the Rockies. The
Canadian Rockies are blunt-topped fisty mountains, with knuckles of
bare rock sticking out everywhere. The Selkirks are graceful
pyramids and sharp sierras, up to their shoulders in magnificent
forests of lofty pines. The trees on the Rockies are much smaller and
poorer. Right above the hotel, to the left of the overhanging Glacier,
is the bare steeple of Sir Donald, one of the monarchs of the range;
Ross Peak and Cheops frown on the descent of the line to the
Pacific; and the line of the Atlantic is guarded by the hundred
pinnacles of the rifted mountain, formerly known as the Hermit, and
now, with singular infelicity, re-christened, in an eponymous fit,
Mount Tupper.
 THE GREAT GLACIER OF THE SELKIRKS.
Sir Charles Tupper is one of Canada’s greatest men, but his name is
more suitable for a great man than a great mountain, especially
since there is a very perfect effect of a hermit and his dog formed by
boulders near the top of the mountain. The men in the railway camp
have got over this difficulty with the doggerel:

“That’s Sir Charles Tupper

Going home to his supper.”

We made two long stays at the Glacier House, and I never enjoyed
anything more in my life than the effect of the snug little châlet, with
its velvety lawn, in the stronghold of the giant mountains, brought
into touch with the great world twice a day by the trains east and
west, which echoed their approach and departure miles on miles
through the ranges.
On the Cars and Off (London, 1895).
 MAUNA LOA
(HAWAII)
LADY BRASSEY
At 6:30 a. m., we made the island of Hawaii, rather too much to
leeward, as we had been carried by the strong current at least
eighteen miles out of our course. We were therefore obliged to beat
up to windward, in the course of which operation we passed a large
bark running before the wind—the first ship we had seen since
leaving Tahiti—and also a fine whale, blowing close to us. We could
not see the high land in the centre of the island, owing to the mist in
which it was enveloped, and there was great excitement and much
speculation on board as to the principal points which were visible. At
noon the observations taken proved that Tom was right in his
opinion as to our exact position. The wind dropped as we
approached the coast, where we could see the heavy surf dashing
against the black lava cliffs, rushing up the little creeks, and
throwing its spray in huge fountain-like jets high above the tall
cocoanut-trees far inland.
We sailed along close to the shore, and by two o’clock were near the
entrance to the Bay of Hilo. In answer to our signal for a pilot, a
boat came off with a man who said he knew the entrance to the
harbour, but informed us that the proper pilot had gone to Honolulu
on a pleasure trip.
It was a clear afternoon. The mountains, Mauna Kea and Mauna
Loa, could be plainly seen from top to bottom, their giant crests
rising nearly 14,000 feet above our heads, their tree and fern clad
slopes seamed with deep gulches or ravines, down each of which a
fertilizing river ran into the sea. Inside the reef, the white coral
shore, on which the waves seemed too lazy to break, is fringed with