Computational Intelligence in Music

Sound Art and Design Antonios Liapis

Visit to download the full and correct content document:
https://textbookfull.com/product/computational-intelligence-in-music-sound-art-and-de
sign-antonios-liapis/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Music Sound and Architecture in Islam Michael Frishkopf

https://textbookfull.com/product/music-sound-and-architecture-in-
islam-michael-frishkopf/

Computational Intelligence in Sports Iztok Fister

https://textbookfull.com/product/computational-intelligence-in-
sports-iztok-fister/

Presence Through Sound SOAS Studies in Music 1st

Edition Keith Howard

https://textbookfull.com/product/presence-through-sound-soas-
studies-in-music-1st-edition-keith-howard/

Code as Creative Medium: A Handbook for Computational

Art and Design Golan Levin

https://textbookfull.com/product/code-as-creative-medium-a-
handbook-for-computational-art-and-design-golan-levin/
Advances in Computational Intelligence Proceedings of
Second International Conference on Computational
Intelligence 2018 Sudip Kumar Sahana

https://textbookfull.com/product/advances-in-computational-
intelligence-proceedings-of-second-international-conference-on-
computational-intelligence-2018-sudip-kumar-sahana/

Computational Intelligence in Digital Pedagogy Arpan

Deyasi

https://textbookfull.com/product/computational-intelligence-in-
digital-pedagogy-arpan-deyasi/

Sound and Image Aesthetics and Practices Sound Design

1st Edition Andrew Knight-Hill (Editor)

https://textbookfull.com/product/sound-and-image-aesthetics-and-
practices-sound-design-1st-edition-andrew-knight-hill-editor/

Trends in Mathematics and Computational Intelligence

María Eugenia Cornejo

https://textbookfull.com/product/trends-in-mathematics-and-
computational-intelligence-maria-eugenia-cornejo/

Music and Sound in the Life and Literature of James

Joyce: Joyces Noyces Gerry Smyth

https://textbookfull.com/product/music-and-sound-in-the-life-and-
literature-of-james-joyce-joyces-noyces-gerry-smyth/
 Antonios Liapis
Juan Jesús Romero Cardalda
Anikó Ekárt (Eds.)

Computational
LNCS 10783

Intelligence in Music,
Sound, Art and Design
7th International Conference, EvoMUSART 2018
Parma, Italy, April 4–6, 2018
Proceedings

123
Lecture Notes in Computer Science 10783
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Antonios Liapis Juan Jesús Romero Cardalda
•

Anikó Ekárt (Eds.)

Computational
Intelligence in Music,
Sound, Art and Design
7th International Conference, EvoMUSART 2018
Parma, Italy, April 4–6, 2018
Proceedings

123
Editors
Antonios Liapis Anikó Ekárt
University of Malta Aston University
Msida Birmingham
Malta UK
Juan Jesús Romero Cardalda
University of A Coruña
A Coruña
Spain

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-77582-1 ISBN 978-3-319-77583-8 (eBook)
https://doi.org/10.1007/978-3-319-77583-8

Library of Congress Control Number: 2018934316

LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

© Springer International Publishing AG, part of Springer Nature 2018

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. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

EvoMUSART 2018—the 7th International Conference and the 14th European Event
on Biologically Inspired Music, Sound, Art and Design—took place during April 4–6,
2018, in Parma, Italy. It brought together researchers who use biologically inspired
computer techniques for artistic, aesthetic, and design purposes. Researchers presented
their latest work in the intersection of the fields of computer science, evolutionary
systems, art, music, and aesthetics. As always, the atmosphere was fun, friendly, and
constructive.
EvoMUSART has grown steadily since its first edition in 2003 in Essex, UK, when
it was one of the Applications of Evolutionary Computing workshops. Since 2012 it
has been a full conference as part of the Evo* co-located events.
EvoMUSART 2018 received 33 submissions. The peer-review process was rigorous
and double-blind. The international Program Committee, listed herein, was composed
of 60 members from 20 countries. EvoMUSART continued to provide useful feedback
to authors: For the papers sent for full review, there were on average 3.5 reviews per
paper. The number of accepted papers was 12 long talks (36% acceptance rate) and
eight posters accompanied by short talks, yielding an overall acceptance rate of 61%.
As always, the EvoMUSART proceedings cover a wide range of topics and
application areas, including: generative approaches to music and visual art; medical art
therapy; visualization in virtual reality; interactive evolutionary computation; and the
art theory of evolutionary computation. This volume of proceedings collects the
accepted papers.
As always, the standard of submissions was high, and good-quality papers had to be
rejected. We thank all authors for submitting their work, including those whose work
was not accepted for presentation on this occasion.
The work of reviewing is done voluntarily and generally with little official recog-
nition from the institutions where reviewers are employed. Nevertheless, professional
reviewing is essential to a healthy conference. Therefore we particularly thank the
members of the Program Committee for their hard work and professionalism in pro-
viding constructive and fair reviews.
EvoMUSART 2018 was part of the Evo* 2018 event, which included three addi-
tional conferences: EuroGP 2018, EvoCOP 2018 and EvoApplications 2018. Many
people helped to make this event a success.
We thank the invited keynote speakers, Una-May O’Reilly (MIT Computer Science
and Artificial Intelligence Laboratory, USA) and Penousal Machado (Computational
Design and Visualisation Lab at the University of Coimbra, Portugal) for their inspi-
rational talks.
We thank SPECIES, the Society for the Promotion of Evolutionary Computation in
Europe and Its Surroundings, for its sponsorship of the event and the local organizing
team lead by Stefano Cagnoni and Monica Mordonini, from the University of Parma,
Italy.
VI Preface

We thank Marc Schoenauer (Inria Saclay, France), for continued assistance in

providing the MyReview conference management system, and Pablo García Sánchez
(University of Cádiz, Spain) for Evo* publicity and website service.
Finally, and above all, we would like to express our most heartfelt thanks to
Jennifer Willies and Anna Esparcia-Alcázar (Universitat Politècnica de València,
Spain), for their dedicated work and coordination of the event. Without them Evo*
would not enjoy its current level of success as the leading European event on
bio-inspired computation.

April 2018 Antonios Liapis

Juan Romero
Anikó Ekárt
 Organization

EvoMUSART 2018 was part of Evo* 2018, Europe’s premier co-located events in the
field of evolutionary computing, which also included the conferences EuroGP 2018,
EvoCOP 2018, and EvoApplications 2018.

Organizing Committee
Conference Chairs
Antonios Liapis University of Malta, Malta
Juan Romero University of A Coruña, Spain

Publication Chair
Anikó Ekárt Aston University, UK

Program Committee
Mauro Annunziato ENEA, Italy
Dan Ashlock University of Guelph, Canada
Peter Bentley University College London, UK
Eleonora Bilotta University of Calabria, Italy
Tim Blackwell Goldsmiths College, University of London, UK
Andrew Brown Griffith University, Australia
Adrian Carballal University of A Coruña, Spain
Amilcar Cardoso University of Coimbra, Portugal
Peter Cariani University of Binghamton, USA
Vic Ciesielski RMIT, Australia
John Collomosse University of Surrey, UK
Kate Compton University of California Santa Cruz, USA
João Correia University of Coimbra, Portugal
Pedro Cruz Northeastern University, USA
Palle Dahlstedt Göteborg University, Sweden
Hans Dehlinger Independent Artist, Germany
Alan Dorin Monash University, Australia
Arne Eigenfeldt Simon Fraser University, Canada
José Fornari NICS/Unicamp, Brazil
Marcelo Freitas Caetano IRCAM, France
Philip Galanter Texas A&M College of Architecture, USA
Pablo Gervás Complutense University of Madrid, Spain
Andrew Gildfind Google, Inc., Australia
Gary Greenfield University of Richmond, USA
Scot Gresham Lancaster University of Texas, Dallas, USA
VIII Organization

Carlos Grilo Instituto Politécnico de Leiria, Portugal

Andrew Horner University of Science and Technology, Hong Kong,
SAR China
Takashi Ikegami The University of Tokyo, Japan
Christian Jacob University of Calgary, Canada
Patrick Janssen National University of Singapore, Singapore
Colin Johnson University of Kent, UK
Daniel Jones Goldsmiths College, University of London, UK
Anna Jordanous University of Kent, UK
Amy K. Hoover University of Central Florida, USA
Maximos Aristotle University of Thessaloniki, Greece
Kaliakatsos-Papakostas
Matthew Lewis Ohio State University, USA
Alain Lioret Paris 8 University, France
Louis Philippe Lopes University of Malta, Malta
Roisin Loughran University College Dublin, Ireland
Penousal Machado University of Coimbra, Portugal
Roger Malina University of Texas, Dallas, USA
Bill Manaris College of Charleston, USA
Tiago Martins University of Coimbra, Portugal
Jon McCormack Monash University, Australia
Eduardo Miranda University of Plymouth, UK
Nicolas Monmarché University of Tours, France
Marcos Nadal University of Vienna, Austria
Michael O’Neill University College Dublin, Ireland
Philippe Pasquier Simon Fraser University, Canada
Alejandro Pazos University of A Coruña, Spain
Somnuk Phon-Amnuaisuk Brunei Institute of Technology, Malaysia
Brian Ross Brock University, Canada
Jonathan E. Rowe University of Birmingham, UK
Antonino Santos University of A Coruña, Spain
Marco Scirea IT University of Copenhagen, Denmark
Daniel Silva University of Coimbra, Portugal
Benjamin Smith Indianapolis University, Purdue University,
Indianapolis, USA
Stephen Todd IBM, UK
Paulo Urbano Universidade de Lisboa, Portugal
Anna Ursyn University of Northern Colorado, USA
Dan Ventura Brigham Young University, USA
 Contents

Visual Art Inspired by the Collective Feeding Behavior

of Sand-Bubbler Crabs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Hendrik Richter

Dynamical Music with Musical Boolean Networks . . . . . . . . . . . . . . . . . . . 18

George Gabriel and Susan Stepney

Non-photorealistic Rendering with Cartesian Genetic Programming

Using Graphics Processing Units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Illya Bakurov and Brian J. Ross

Construction of a Repertoire of Analog Form-Finding Techniques

as a Basis for Computational Morphological Exploration in Design
and Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Ever Patiño and Jorge Maya

Medical Art Therapy of the Future: Building an Interactive Virtual

Underwater World in a Children’s Hospital . . . . . . . . . . . . . . . . . . . . . . . . 64
Ludivine Lechat, Lieven Menschaert, Tom De Smedt, Lucas Nijs,
Monica Dhar, Koen Norga, and Jaan Toelen

Expressive Piano Music Playing Using a Kalman Filter . . . . . . . . . . . . . . . . 78

Alexandra Bonnici, Maria Mifsud, and Kenneth P. Camilleri

Generative Solid Modelling Employing Natural Language

Understanding and 3D Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Marinos Koutsomichalis and Björn Gambäck

evoExplore: Multiscale Visualization of Evolutionary Histories

in Virtual Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Justin Kelly and Christian Jacob

Musical Organisms: A Generative Approach to Growing Musical Scores . . . . 128

Anna Lindemann and Eric Lindemann

Generating Drum Rhythms Through Data-Driven Conceptual Blending

of Features and Genetic Algorithms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Maximos Kaliakatsos-Papakostas

RoboJam: A Musical Mixture Density Network for Collaborative

Touchscreen Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Charles Patrick Martin and Jim Torresen
X Contents

Towards a General Framework for Artistic Style Transfer . . . . . . . . . . . . . . 177

Florian Uhde and Sanaz Mostaghim

Adaptive Interface for Mapping Body Movements to Sounds . . . . . . . . . . . . 194

Dimitrije Marković and Nebojša Malešević

On Collaborator Selection in Creative Agent Societies:

An Evolutionary Art Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Simo Linkola and Otto Hantula

Towards Partially Automatic Search of Edge Bundling Parameters . . . . . . . . 223

Evgheni Polisciuc, Filipe Assunção, and Penousal Machado

Co-evolving Melodies and Harmonization in Evolutionary

Music Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Olav Olseng and Björn Gambäck

Learning as Performance: Autoencoding and Generating Dance

Movements in Real Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Alexander Berman and Valencia James

Deep Interactive Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267

Philip Bontrager, Wending Lin, Julian Togelius, and Sebastian Risi

The Light Show: Flashing Fireflies Gathering and Flying

over Digital Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Paulo Urbano

Evotype: Towards the Evolution of Type Stencils . . . . . . . . . . . . . . . . . . . . 299

Tiago Martins, João Correia, Ernesto Costa,
and Penousal Machado

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

 Visual Art Inspired by the Collective
Feeding Behavior of Sand-Bubbler Crabs

Hendrik Richter(B)

Faculty of Electrical Engineering and Information Technology,

HTWK Leipzig University of Applied Sciences, Postfach 30 11 66,
04251 Leipzig, Germany
hendrik.richter@htwk-leipzig.de

Abstract. Sand-bubblers are crabs of the genera Dotilla and Scopimera

which are known to produce remarkable patterns and structures at trop-
ical beaches. From these pattern-making abilities, we may draw inspi-
ration for digital visual art. A simple mathematical model is proposed
and an algorithm is designed that may create such sand-bubbler pat-
terns artificially. In addition, design parameters to modify the patterns
are identified and analyzed by computational aesthetic measures. Finally,
an extension of the algorithm is discussed that may enable controlling
and guiding generative evolution of the art-making process.

1 Introduction
Among the various forms of inspiration from biology for digital visual art, swarms
and other types of collective animal behavior are considered to be particularly
promising of showing algorithmic ideas and templates that may have the poten-
tial to create non-trivial patterns of aesthetic value. Examples are ant- and ant-
colony-inspired visual art [12,25], but our imagination has also been stimulated
by the pattern-making of flies [1] and other swarms, flocks, or colonies [15,22].
In this paper, the collective behavior of another family of animals provides
a source of inspiration: crabs of the genera Dotilla and Scopimera, commonly
called sand-bubblers. Sand-bubbler crabs are known for their burrow-orientated
feeding behavior that produces radial patterns of regular shapes and designs in
the wet sand of tropical beaches [2,3,7]. The patterns consist of tiny balls that
are placed in curves or spirals, straight or bent lines, which finally form overall
structures, thus producing astonishing works of natural art [4]. The creation of
structures starts at receding tide. The patterns grow in complexity and scale
as time goes by until the returning tide cleans them off, in fact resetting the
art-making process. This pattern-making behavior, its visual results, and the
potentials of thus inspired algorithms producing digital visual art are our topics.
We approach these topics by discussing three interconnected main themes
in this paper. The first is to describe an aspect of animal behavior observed
in nature (the feeding pattern of a colony of sand-bubbler crabs) by economic
mathematical means in order to make it accessible for an algorithmic process
c Springer International Publishing AG, part of Springer Nature 2018
A. Liapis et al. (Eds.): EvoMUSART 2018, LNCS 10783, pp. 1–17, 2018.
https://doi.org/10.1007/978-3-319-77583-8_1
2 H. Richter

producing two-dimensional visual art. This goes along with identifying design
parameters that allow the patterns to vary within the restriction that they should
generally resemble structures as found in nature. The second theme is to study
the image-producing process by computational aesthetic measures. Here, the
focus is not on numerically evaluating an overall artistic value. There are some
strong arguments that such an evaluation is rather elusive and may not really
be achievable free from ambiguity, see also the discussion in Sect. 3. Instead, the
focus is put on how some aesthetic measures, namely Benford’s law measure,
Ross, Ralph and Zong’s bell-curve measure and a fractal dimension measure,
scale to the design parameters. A third theme is to use these measures to guide
and control the art-making process. This is done by updating the design para-
meters while the algorithmic process is running. By employing the relationships
between aesthetic measures and design parameters for updating the parameters,
the measures can be changed in an intended way. Hence, this theme studies
the possibilities of including generative development, which takes up some ideas
from evolving and evolutionary art.
The paper is structured as follows. Section 2 reviews the behavior of sand-
bubbler crabs, particularly the burrow-orientated feeding process. Inspired by
these behavioral patterns, Sect. 3 presents an algorithm to create visual art, stud-
ies computational aesthetic measures and generative development, and finally
shows examples of art pieces thus created. Concluding remarks and an outlook
end the paper.

2 Behavior of Sand-bubbler Crabs

Sand-bubbler crabs are tiny crustacean decapods of the genera Dotilla and
Scopimera in the family Dotillidae that dwell on tropical sandy beaches across
the Indo-Pacific region. They permanently inhabit intertidal environments as
they adapted to the interplay between high and low tide. Sand-bubblers are sur-
face deposit feeders. Thus, their habitat must be periodically covered by water
for their food supply to restore. They are also air-breathers. Thus, the habital
zone needs to be periodically uncovered from water for giving them time for
feeding. Observing these periodic changes, the crabs dig burrows in the sand
where they seek shelter and hide during high tide. In low tide (and daylight)
they swarm out in search for food. They pick up sand grains with their pincers
and eat the microscopic organic particles that are left over by the receding tide
and coat the grains. After they have cleaned the sand grains of eatable com-
ponents, they form them into small balls, also called pseudo-faecal pellets, and
throw them behind. Supposedly, they do so in order to avoid searching the same
sand twice. As high tide comes back in, they retreat to their burrows, seal them
to trap air, and wait for the next tidal cycle.
By this process of feeding, and particularly by molding sand into pellets and
distributing them, sand-bubbler crabs create intricate patterns in the sand that
remain there until the next high tide washes them away, see Fig. 1 that shows
such feeding patterns as observed by the author at Tanjung Beach, Langkawi,
 Visual Art Inspired by the Collective Feeding Behavior 3

Fig. 1. Feeding patterns of sand-bubbler crabs, observed by the author at Tanjung

Beach, Langkawi, Malaysia, in March 2017.

Malaysia in March 2017. The patterns in the sand start from a center point,
have a radial orientation and comprise of small sand balls along a tiny path.
The patterns vary in complexity depending on how many crabs there are con-
tributing, on how long the last high tide is away, and on several other factors
discussed below. The patterns in the sand are starting point and inspiration for
an algorithm producing two-dimensional visual art. Before presenting such an
algorithm in Sect. 3, it may be interesting to explore the feeding process and the
pattern-making a bit more deeply.
A main feature of the lifestyle of sand-bubblers is their burrow-orientated
feeding behavior [2,3,5,7]. This means that the entrance of the burrow is the
center of the feeding range and the feeding process is highly structured. Basically,
such a structured and stereotype behavior can be found generally among the 14
species of Dotilla and 17 species of Scopimera currently recognized, but there is
variety in the details of the feeding. This, in turn, offers to vary the art-making
algorithm in order to achieve different graphic effects.
The feeding starts from the burrow entrance, where the crab sets out to
progress sideways along a straight line radiating from the burrow. While moving
along, it sorts sand from organic-rich fine particles which are scraped off and
subsequently ingested. The crab only feeds on the upper millimeter or two of the
sand. The residual larger nutrient-deficient sand grains are molded into ball-like
globules, but not digested, which motivates calling them pseudo-faecal pellets.
When a pellet has reached a certain diameter, which scales to the size of the crab,
it is moved along the body and pushed away behind the animal. After moving
along the straight line radiating from the burrow and ejecting pellets for a certain
distance, the crab steps aside, turns around and moves back toward the burrow
entrance while continuing with feeding and pellet ejecting. When reaching the
entrance, its turns around again, and works the next line (see also Fig. 2 for a
schematic description). The pellets are placed in such a way that a straight path
to the burrow entrance always remains clear. This is essential for the crab to
retreat to the burrow in case of external disturbance or danger. Consequently,
the feeding progresses radially line by line while the lines (which are also called
trenches in the biological literature) rotate around the burrow entrance just like
the hands of a clock. In fieldwork it has been observed that the rotation may be
clockwise or anti-clockwise with approximately the same frequency [17,20]. By
4 H. Richter

Fig. 2. Algorithmically generating a pellet location with the burrow coordinates

(xi , yi )T , the trench angle θ1 of the (lower branch of the) first trench and the radial
coordinate rk ; see also the first term of the right-hand side of Eq. (1). The arrows indi-
cate the feeding direction of a sand-bubbler crab. The outbound trench j = 3 depicts
a radial string of six equally distanced pellets, while for j = 1 there is one pellet; for
the other trenches the pellets are not shown.

such a rotation of the feeding trench, an angular sector of a circle in the sand is
excavated, cleared of eatable components, and amassed with pellets. The angular
sector of uninterrupted feeding may vary; an average of 60◦ has been reported,
but also that occasionally a full circle has been completed [2,13,17], see also the
right-hand side of Fig. 1. Furthermore, it has been reported that there is a fairly
close relationship between the maximum length of a trench and the size of the
crab. Such a relationship is plausible as the maximum length together with the
crab size scales to the maximum time needed for the crab to seek shelter in the
burrow, if an external disturbance requires such an escape. For one species of
sand-bubblers, Dotilla wichmanni, even a measurement of the maximum length
has been recorded with the body length of eight to nine crabs [17]. Furthermore,
another stereotype of pattern has been identified as sometimes the distribution
of pellets along the trench follows a rhythmic pattern where the crab regularly
suspends pellet ejecting and “concentric rings” emerge, as shown for Dotilla
clepsydrodactylus [20].
In view of the structured burrow-orientated feeding behavior of sand-bubbler
crabs as described above, it seems not surprising that regular patterns in the sand
emerge. Ideally, a well-ordered collection of full circular areas with the burrow
entrance in the center should appear that are filled with radial strings of equally
distanced pellets. However, looking at patterns from nature (see, for instance,
Fig. 1) reveals them as irregular in several respects. There are reasons for these
irregularities. A first reason is that for a colony of sand-bubblers the coordinates
of the burrow entrances are randomly distributed across the beach where they
are dwelling. There is almost always a certain minimum distance between two
burrow entrances and the number of burrows for a given square of beach follows
a Poisson distribution, as shown for the species Dotilla fenestrata [11]. A second
 Visual Art Inspired by the Collective Feeding Behavior 5

reason is that the orientation of the first trench (and hence the overall orientation
of the angular sector) is not different from a random distribution [17,20]. There
is no indication that the orientation is related to the slope of the sand surface,
or any compass direction or the sun’s position. Another reason, for instance
reported for the species Scopimera inflata [7], is that a crab may move a short
distance away from the entrance of the burrow before starting to feed, which
results in a gap between burrow and trench. The width of the gap appears to
be randomly distributed around the crab’s size. A further reason is that for a
given burrow (and therefore a crab of given size) the trenches have not all the
same length, even if the feeding is undisturbed [11]. There is a weak correlation
between trench length and feeding time, with the trenches growing longer the
longer the crab feeds, but generally also trench length must be regarded as to
follow a random distribution. Finally, a crab may be externally disturbed at
any time, for instance by intruders such as other crabs or human observers, or
eating enemies such as shorebirds. As sand-bubblers are extremely perceptive
to movement, in such a case the crab rapidly escapes to the burrow. It may
reemerge a little later and resume eating. The feeding may or may not continue
along the trench last used before escape. It has been convincingly argued by field
experiments that sand-bubblers (for instance the species Dotilla wichmanni)
must possess remarkable orientation abilities that allow them to resume the
direction along which they were feeding, even in the absence of visual clues such
as a partly finished trench [17]. Whether or not a crab uses these abilities in a
given situation, however, also appears to follow a random distribution. Lastly,
a crab may interrupt its feeding even if there is no evident external trigger, for
instance to rest or to restore its water supply, which is done in the burrow.
To sum it up: The exact position of each pellet follows from the structure of
the burrow-oriented feeding behavior, but is subject to a substantial degree of
random. If we reframe the organized and sequential placement of a sufficiently
large number of pellets as a dynamic process, the pattern generation could be
interpreted as a random walk substantially biased by the feeding structure. This
interplay between determinism and random may be one of the reasons why the
patterns in the sand appear non-trivial and of some aesthetic value. Finally,
we may note that apart from feeding, there are at least two further aspects of
the sand-bubbler’s lifestyle that might be interesting as inspiration for visual
art, burrow-digging and interaction with other crabs. These aspects could be
addressed in future work.

3 Design of Art-Making Algorithms

3.1 Generating Patterns from Pellet Distribution

The feeding behavior of sand-bubbler crabs as described in Sect. 2 is now used to

design templates for algorithmically generating two-dimensional visual art. With
respect to the regular and irregular aspects of the sand patterns, the art-making
algorithm is also comprised of deterministic and random elements.
6 H. Richter

Algorithm 1. Generate patterns from pellet distribution

Set maximum number of burrows I
for i=1 to I do
Generate burrow coordinates (xi , yi )T
Set maximum number of trenches Ji
for j=1 to Ji do
Generate trench angle θj
Set maximum number of pellets Kij
for k=1 to Kij do
Generate pellet radial coordinate rk
2
Set mean μijk and variance σijk
Calculate pellet location (xijk , yijk )T by Eq. (1)
Return location and store for subsequent coloring and visualization
end for
end for
end for

Describing the algorithm starts with the notion that every pellet has a loca-
tion (xijk , yijk )T in a two-dimensional plane. The index (ijk) refers to the k-th
pellet (k = 1, 2, 3, . . . , Kij ) belonging to the j-th trench (j = 1, 2, 3, . . . , Ji )
of the i-th burrow (i = 1, 2, 3, . . . , I). The pellet location is algorithmically
specified by
     
xijk xi + rk · cos (θj ) N (μijk , σijk
2
)
= + , (1)
yijk yi + rk · sin (θj ) N (μijk , σijk
2
)

where (xi , yi )T are the coordinates of the i-th burrow, θj is the trench angle
belonging to the j-th trench, and rk is the radial coordinate of the k-th pellet.
For each burrow, the maximum number of pellets Kij remains to be specified
for a given i and j; the same applies to the maximum number of trenches Ji .
Figure 2 unpacks the description and shows an example with one burrow and six
trenches (three outbound and three returning).
The first term of the right-hand side of Eq. (1) can be seen as to represent
the deterministic aspect of the pellet ejecting that characterizes the burrow-
orientated feeding behavior of sand-bubblers. The second term represents the
random aspect with shifting the position by a realization of a random variable
2
normally distributed with mean μijk and variance σijk . Accordingly, every pel-
T
let location (xijk , yijk ) has three deterministic parameters: burrow coordinates
(xi , yi )T , trench angle θj and radial coordinate rk ; and two stochastic param-
2
eters: mean μijk and variance σijk of the normal distribution from which the
random shift in position is realized. Equation (1) not only specifies the loca-
tion of a pellet, it can also be used to generate patterns consisting of pellets,
trenches and burrows, see Algorithm 1. Therefore, for each burrow coordinates
(xi , yi )T , one sequence of trench angles (θ1 , θ2 , . . . , θJi ) and Ji sequences of pellet
radial coordinates (r1 , r2 , . . . , rKij ) are defined. A major guideline for defining
these parameter sequences is the intention to replicate characteristic features of
 Visual Art Inspired by the Collective Feeding Behavior 7

Table 1. Design parameters and default values for generating patterns from pellet
distribution, see Algorithm 1. The parameters of RTL are general and also apply to the
other templates.

Template Design parameter Symbol Default value

RTL (and general) Number of burrows I 3
Burrow coordinates (xi , yi )T N (0, 7)
Maximum number of trenches Jmax 50
Pellet distance dj 0.25
Mean random shifting μijk −1, 0, 1
2
Variance random shifting σijk 0.3, 0.8
Mean trench length μT 25
Variance trench length σT2 1
GTL Grow rate ΔμT 2
CCR Number of gaps λ 3
Gap width gλ 4
BTG Burrow gap width g0 8

the sand-bubbler pattern as discussed in Sect. 2. These patterns can be labeled

according to the graphic effects they convey and subsequently taken as tem-
plates to create two-dimensional visual art works. In the numerical experiments
reported in Sect. 3.4 the following templates are used: (i) random trench length
(RTL), (ii) growing trench length (GTL), (iii) concentric rings (CCR), and (iv)
burrow-to-trench gaps (BTG). Each template varies the art-making algorithm
and replicates a specific pattern found in nature. For random trench length
(RTL), the maximal number of trenches Ji is a random integer uniformly dis-
tributed on [1, Jmax ]. The first and the last trench angle (θ1 and θJi ) are chosen
as realizations of a uniform random variable distributed on [0, 2π]; the remaining
trench angles are calculated from the number Ji to be equidistant between θ1
and θJi . Each trench has a length j that is a realization of a random variable
normally distributed on [μT , σT2 ]. From the length j , the pellet radial coordi-
nates are calculated to be equidistant with distance dj . Growing trench length
(GTL) is similar to RTL (and uses the same general design parameters), but
the trench length j is calculated by a normal distribution where the mean μT
grows with trench j by a growth rate ΔμT . Thus, the first trench has mean μT ,
while the Ji -th trench has mean μT + (Ji − 1)ΔμT . The variance σT2 is constant.
Concentric rings (CCR) again have trenches with constant mean and variance
but do not always have equidistant pellet radial coordinates as there are λ gaps
of width gλ where no pellets are placed. Finally, burrow-to-trench gaps (BTG)
start with a gap of width g0 in pellet placement. Table 1 summarizes the design
parameters.
Algorithm 1 specifies the location of the pellets and hence the structure of the
patterns. In nature, the patterns are monochromatic as the pellets all have the
8 H. Richter

color of the sand they are built from, see Fig. 1. The artistic interpretation of the
patterns suggests using different colors for making them more visually appealing
and also conveying additional information. As the algorithm imposes a chrono-
logical order on pellet placement by a forward counting of burrows, trenches
and finally the pellets themselves, a possible design is to use a colorization that
reflects this order. Alternatively, each burrow can be associated with a specific
color and the order of the pellet placement can be characterized by shades or
brightness of this color. In addition, the burrows using the same template (see
Table 1) may take colors that are similar.

3.2 Analysis by Computational Aesthetic Measures

Next, the templates and their design parameters are analyzed by computational
aesthetic measures. This analysis aims at studying two questions. (i) To what
extend do the measures vary over the templates and parameters? This goes along
with assessing how sensitive these relationships are. (ii) Which templates and/or
parameters are most suitable to guide and control the art-making process toward
desired values of the aesthetic measures?
Three computational aesthetic measures are studied, Benford’s law measure
(BFL), Ross, Ralph and Zong’s bell curve measure (RRZ) and a fractal dimen-
sion measure (FRD), which were chosen because a recent study reported them
to be weakly correlated [6]. All measures are calculated for the images with
512 × 512 pixels. The Benford’s law measure BFL [6,21] is known for scaling to
the naturalness of an image. It is calculated by taking the distribution of the
luminosity of an image and comparing it to the Benford’s law distribution:

BFL = 1 − dtotal /dmax (2)

9
where dtotal = i=1 (HI (i) − HB (i)), dmax = 1.398 is the maximum possible
value of dtotal , HB = (0.301, 0.176, 0.125, 0.097, 0.079, 0.067, 0.058, 0.051, 0.046)
is the Benford’s law distribution and HI is the normalized sorted 9-bin histogram
of the luminosity of the image. The luminosity of the image is calculated for each
pixel of the image by taking the weighted sum of the red (R), green (G) and
blue (B) values: lum = 0.2126 · R + 0.7152 · G + 0.0722 · B.
Ross, Ralph and Zong’s bell-curve measure RRZ [6,23] is a measure of color
gradient normality. It compares the color gradients of the image to a normal
(bell-curve or Gaussian) distribution. Therefore, the color gradient of the colors
red, green and blue are calculated for each pixel, summarized and normalized
by a detection threshold, see [6,23] for details. Hence, we obtain a distribution
with mean μ and variance σ 2 . The distribution is discretized into 100 bins to
obtain a discrete color gradient distribution pi . For μ and σ 2 calculated from
the observed color gradient distribution, an expected discrete Gaussian distribu-
tion qi is computed. Finally, the Kullback-Leibler divergence between these two
distributions gives the bell-curve measures

RRZ = pi log (pi /qi ). (3)
i
 Visual Art Inspired by the Collective Feeding Behavior 9

Fig. 3. Aesthetic measures as a function of design parameters: number of burrows I,

maximum number of trenches Jmax and pellet distance dj . RTL (blue), GTL (magenta),
2 2
CCR (red), BTG (green); solid lines, σijk = 0.3, dotted lines, σijk = 0.8.

Note that in deviation from [6,23], the Kullback-Leibler divergence in Eq. (3) is
not amplified by the factor 1000.
The third aesthetic measure studied is the fractal dimension FRD [6,24] of
the image. Following empirical studies and an argument by Spehar et al. [24], a
fractal image of dimension 1.3 ≤ d ≤ 1.5 is most preferred by human evaluation,
compared to images that have a fractal dimension outside this range. Thus, a
fractal dimension d ≈ 1.35 is considered to be most desirable and the fractal
aesthetic measure
FRD = max (0, 1 − |1.35 − d|) (4)
can be defined, where d is the fractal dimension of the image calculated by
box-counting.
Figure 3 shows the measures for the templates RTL, GTL, CCR and BTG
for selected design parameters and two values of the variance of the random
2
shifting σijk . The results are means over 100 images generated by Algorithm 1
and the design parameters listed in Table 1. The design parameter number of
burrows I is varied over 2 ≤ I ≤ 10, the maximum number of trenches Jmax
over 20 ≤ Jmax ≤ 100 and the pellet distance dj over 0.05 ≤ dj ≤ 1. The default
values of these parameters all lie within the range studied. Furthermore, as the
aesthetic measures are calculated from the RGB values of the pixels, the colors
10 H. Richter

of the image have a profound effect on the values of the measures. As discussed
in Sect. 3.1, the colorization of the pellets is also a subject of algorithmic design.
Thus, to counter the effect of a given colorization and for not singularizing
particular colors, the results are calculated as means over 70 hue values randomly
selected.
From the results in Fig. 3, we see that the different templates (RTL, GTL,
CCR, BTG) generally produce characteristic curves over the design parame-
ters for the aesthetic measures BFL and RRZ. Also, the two values of random
shifting (which can be interpreted as noise levels) are clearly distinct. The gen-
eral trend is that BFL and RRZ decrease with increasing number of burrows I
and increasing maximum number of trenches Jmax , while there is an increase
with increasing pellet distance dj . An interesting exception is the template CCR
which for varying pellet distance dj behaves differently and crosses the curves
of the other templates. The results for the aesthetic measures FRD are slightly
different. For FRD, the templates RTL and BTG (blue and green lines) have
almost the same results and for all templates, different noise levels have only
negligible effect. A possible explanation is that the difference between RTL and
BTG is only that in the images there is a circle free from pellets around the
burrow entrance for BTG. The overall structure of the pellet placement is the
same, which is recognized by the calculation of the fractal dimension accounting
mainly for density of geometrical objects but less for their layout. The same
may apply for varying the noise level. Based on these results assessing how sen-
sitive the relationships between design parameters and aesthetic measures are,
some conclusions can be drawn about possible ways to guide and control the
art-making algorithm. Basically, there are two options. One is to switch to a
particular template but keep the values of design parameters, another is to keep
the template but change the values of design parameters. (In fact, it would also
be thinkable to change both template and parameter value at the same time,
but this may appear rather contrived.) For instance, to increase the aesthetic
measure BFL, employing the template BTG is suitable, while RRZ is largest for
RTL (but not very clearly) and FRD for GTL. Another possibility to raise BFL
is to decrease the maximum number of trenches Jmax . The same is also useful to
increase RRZ, particularly for the template CCR, but not for FRD. There are
more subtle schemes to manipulate the measures, but it is also clear that the
three aesthetic measures studied here react differently, yet even contradictory to
these manipulations. With this in mind, next section discusses possible designs
of such manipulations and their effects on the art-making process.

3.3 Guiding and Controlling Generative Evolution

Main features of generative and evolving digital art are that the creating process
develops over time, is guided and controlled by the algorithmic design, and thus
gains functional autonomy to produce complete art works [10]. A frequently used
design for guidance and control is to implement a feedback, for instance via eval-
uating the works by using a measure of their aesthetic value [18,22]. Such ideas
are, for instance, implemented by employing evolutionary algorithms for finding
 Visual Art Inspired by the Collective Feeding Behavior 11

“optimal” values of the aesthetic measures, which has shown to create interesting
works of art [1,6,14,22]. Algorithm 1 given in Sect. 3.1 does not really, in itself,
evoke such a perspective of autonomously guiding and controlling generative evo-
lution. The algorithm is suitable for producing visual art works. It also depends
on selecting templates and design parameters (see Table 1), which enables modi-
fying the works by changing the selection. Thus, promoting generative evolution
essentially requires measuring the aesthetic value of the algorithmically gener-
ated patterns and designing a feedback.
Of course, it is possible to evaluate the pattern by interaction with humans.
However, there are some issues with interactive evaluation [8,9,22]. The first
is that humans evaluating art are usually much slower in doing so than com-
puters algorithmically generating the works, which creates the proverbial “fit-
ness bottleneck”. A second reason is that human evaluation may change over
time, for instance caused by fatigue, but also by boredom, which may bias the
results towards superficial novelty rather than overall quality. A third reason
is that human evaluation of art is always conditioned by personal and cultural
“taste”. For this reason, it is no real help to distribute the evaluation to a
larger group of human evaluators. In such a case, either the tastes of different
subgroups become visible (as for instance shown for evaluating the beauty of
abstract paintings [19]), or generally the averaging effect of asking a large group
of people about their aesthetic choices is reproduced, but arguably “the unique
kind of vision expected from artists” [8] is not achieved.
An alternative are computational aesthetic measures [6,8,9,19,21], which
enable an automated rating without human interference or supervision.
Section 3.2 discussed such measures for evaluating sand-bubbler patterns. These
aesthetic measures of two-dimensional visual art are frequently calculated by
analyzing the (spatial) distribution of colored pixels (or groups of pixels). In
other words, computational aesthetic measures use properties such as pixel-wise
color information to deduce an overall rating of an image. However, there is a
significant conceptual gap between measuring computational image properties
and aesthetic beauty. Apart from the measures considered in Sect. 3.2, there is
a huge number of measures that try to bridge this gap, see e.g. [6,9,16,19] for
an overview. Some of them are strongly correlated, while others address differ-
ent aspects of the color distribution. Moreover, recent studies [6,19] suggested
that it seems rather difficult to establish stable correlations between beauty rat-
ings by humans and image properties, which casts some doubts on whether an
evaluation of the aesthetic value of visual art based on computational aesthetic
measures is really objective, meaningful and feasible. These studies have also
shown that computational aesthetic measures account more for certain visual
effects (or “visual styles” [6]) than for aesthetic value in general. For instance,
Benford’s law measure (BFL) assigns high values for images that have a grainy
texture, while Ross, Ralph and Zong’s bell curve measure (RRZ) likes distinct
color progression and the fractal dimension measure (FRD) values low colorful-
ness highly; see den Heijer and Eiben [6] for a discussion and comparison of 7
computational aesthetic measures.
12 H. Richter

Algorithm 2. Guide and control pattern-making

Set up look-up table with template and design parameters that increase computa-
tional aesthetic measures
Calculate expected aesthetic measures for the look-up table
Select random template and take default design parameter
while Termination criterion not met do
Generate burrow by Algorithm 1
Calculate aesthetic measures
if Measure < Expectation then
Select randomly increasing template or design parameter
else
Keep template and parameter
end if
end while

Fig. 4. Aesthetic measures for guiding and controlling pattern-making. The measures
for the templates (RTL, GLT, CCR, BTG) with the default values of the design param-
eters (see Table 1) are compared to the measures obtained by Algorithm 2. GCB means
using the measure BFL to guide and control, GCR uses RRZ, GCF uses FRD.

In the following, we study an alternative approach for guiding and controlling

generative evolution that tries to address the desire to promote certain visual
effects, see Algorithm 2. It utilizes the relationships between design parameters
and aesthetic measures discussed in Sect. 3.2. Therefore, a look-up-table is set
up that lists templates and design parameters that increase aesthetic measures.
After generating the pattern for each single burrow, an aesthetic measure is
calculated and compared to an “expected value” recorded before, see Fig. 3.
If the value is below the expectation, a measure-increasing template or design
parameter is activated. If not, the generation process continues unaltered. Which
template or parameter is selected is due to chance. These steps are repeated until
a termination criterion is met. In this implementation, the pattern generation
ends if a maximum number of burrows are placed in the image. (In the rare
case that no measure-increasing template or parameter is available, the setting
is also kept.) An interesting feature of such an algorithmic design is that it
immediately allows to control and intervene while the image is being created.
In other words, the control does not wait until the image is completed, but uses
feedback for intervening during the process creating the art works. Figure 4 shows
 Visual Art Inspired by the Collective Feeding Behavior 13

the aesthetic measures BFL, RRZ and FRD for Algorithm 2. The experimental
setup is the same as for the templates, see Fig. 3. Again, the results are averages
over 100 images and 70 random hue values. The aesthetic measures for the
templates in Table 1 are compared to using the measures BFL (GCB ), RRZ
(GCR ) and FRD (GCF ) for guiding and controlling the art-making process.
Apart from the results of the measures that are obtained by using the same
measures to guide and control, we also record the results of the other measures
to examine cross-effects (GCR and GCF for BFL, GCB and GCF for RRZ and
GCB and GCR for FRD). The results in Fig. 4 indicate that using templates
and design parameters is mostly suitable to manipulate the aesthetic measures
in an intended way. For BFL we obtain that the measure for GCB is increased as
compared to the templates RTL, GTL and CCR, while the images do contain a
mix of all templates. For FRD, we even find an increase in GCF compared to all
templates. However, for RRZ we get a rather strong decrease in GCR . A possible
explanation is that the measure RRZ accounts for color gradient normality, while
the effect of Algorithm 2 is basically in changing the selection of templates and
design parameters but only indirectly in colorization. Hence, the measure RRZ
is poorly sensitive to such a changed selection. These speculations are supported
by looking at cross-effects as we see that also the measures that are not used for
guidance and control show similar results. This additionally allows to conjecture
that for images such as the feeding patterns the aesthetic measures are more
correlated than initially assumed. Next, we look at the visual results produced
by both algorithms1 .

3.4 Examples of Art Works

Figure 5 shows a gallery of nine art works generated by Algorithm 1 with the
parameter values given in Table 1. Looking at these images, we can see that they
all display structures that have similarity to natural sand-bubbler patterns. How-
ever, colorization of the actual choices of the random-dependent design parame-
ters also offers diversity. In addition, another six images generated by Algorithm 2
are shown in Fig. 6. Although there is a substantial degree of similarity between
these images and the images generated by Algorithm 1, it is worth observing
some subtle differences. For instance, it is noticeable that typically the images
produced by Algorithm 2 are more widespread and grainier, see for instance
Fig. 6 upper left and lower middle, which may be attributed to the effect of pro-
moting the visual effects associated with the measure BFL. However, at least
in the opinion of the author, there was no real success in provoking a distinct
color progression or a particularly low colorfulness, as supposedly associated with
RRZ and FRD. More analytic work is needed to clarify the relationships between
aesthetic measures and visual effects for a given image-producing algorithm.

1
See https://feit-msr.htwk-leipzig.de/sandbubblerart/ for further images and videos.
14 H. Richter

Fig. 5. Gallery with examples of sand-bubbler inspired art works generated by

Algorithm 1.

Fig. 6. Gallery with examples of sand-bubbler inspired art works generated by

Algorithm 2.
 Visual Art Inspired by the Collective Feeding Behavior 15

4 Concluding Remarks and Outlook

An algorithm to create visual art has been presented that draws inspiration from
the collective feeding behavior of sand-bubbler crabs. Thus, the paper exempli-
fies another instance of an art-making process based on mathematical models
of animal behavior observed in nature. Apart from the algorithmic design and
its artistic output, we mainly studied the algorithmic process, particularly how
computational aesthetic measures scale to the design parameters and how these
measures can be used for controlling and guiding generative evolution. The inter-
est in analyzing the algorithmic process of digital art-making is from both a
computational and an artistic point of view. The rationale for the latter is that
art-theoretically and following an argument of Galanter [8,10], in generative art
the artistic work in itself is not seen as important as the artistic process. In a
reminiscence to the art movement of “truth to material”, generative art may
focus on “truth to process”. A possible interpretation states that it is far less
interesting to just ask if a human observer likes a particular piece of artificial art.
What is much more interesting is to analyze the creation procedure of generative
art and study the interplay between algorithmic design (and design parameters)
and the works thus created. Not only is the question of whether or not an artifi-
cially created image is more or less aesthetic from a human point of view merely
opening up an inexhaustible topic of debate and might be not answerable, this
question may actually not be very useful for advancing our understanding of
what constitutes artistic beauty and how it can be algorithmically (re-)created.
In this spirit the galleries of art works shown in Figs. 5 and 6 could be rather
seen as accompanying the analytic results, and not the other way around.
The analytic and visual results given here only use a small subset of the design
space opened up by Algorithms 1 and 2. Future work could further explore this
design space. For instance, the images only contain pellets of a given size and
color. It would be interesting to study the visual effects of pellets with different
sizes or pellets decorated in a polychromatic way and/or with their own visual
structure. Furthermore, the patterns presented are bound by the restriction that
they resemble (at least roughly) sand-bubbler patterns as can be observed in
nature. A possible extension is to soften this restriction. For instance, in nature
the scale of patterns may vary, but not dramatically. In the art works, there could
be (next to patterns of the size as seen in Figs. 5 and 6) very large structures
even exceeding the edges of the images. A possible effect would be to have a
background of pellets across the entire image, which may produce different levels
of pointillism, as similarly shown by Urbano’s sand painting artists [25].
16 H. Richter

References
1. Abbood, Z.A., Amlal, O., Vidal, F.P.: Evolutionary art using the fly algorithm. In:
Squillero, G., Sim, K. (eds.) EvoApplications 2017. LNCS, vol. 10199, pp. 455–470.
Springer, Cham (2017). https://doi.org/10.1007/978-3-319-55849-3 30
2. Ansell, A.D.: Migration or shelter? Behavioural options for deposit feeding crabs on
tropical sandy shores. In: Chelazzi, G., Vannini, M. (eds.) Behavioral Adaptation
to Intertidal Life. NATO ASI Series (Series A: Life Sciences), vol. 151, pp. 15–26.
Springer, Boston (1988). https://doi.org/10.1007/978-1-4899-3737-7 2
3. Bigalke, R.: On the habits of the crab Dotilla fenestrata, Hilgendorf, with special
reference to the mode of feeding. South Afr. J. Nat. Hist. 3, 205–209 (1921)
4. Burton, A.: Symbols in sand. Front. Ecol. Environ. 14, 456 (2016)
5. Chakrabarti, A., Chakrabarti, R., Hertweck, G.: Surface traces and bioturbate
textures from bubbler crabs: an indicator of subtropical to tropical tidal flat envi-
ronments. Senckenb. Marit 36, 19–27 (2006)
6. den Heijer, E., Eiben, A.E.: Investigating aesthetic measures for unsupervised evo-
lutionary art. Swarm Evol. Comput. 16, 52–68 (2014)
7. Fielder, D.R.: The feeding behaviour of the sand crab Scopimera inflata (Decapoda,
Ocypodidae). J. Zool. (Lond.) 160, 35–49 (1970)
8. Galanter, P.: The problem with evolutionary art is . . .. In: Di Chio, C., et al. (eds.)
EvoApplications 2010. LNCS, vol. 6025, pp. 321–330. Springer, Heidelberg (2010).
https://doi.org/10.1007/978-3-642-12242-2 33
9. Galanter, P.: Computational aesthetic evaluation: past and future. In: McCormack,
J., d’Inverno, M. (eds.) Computers and Creativity, pp. 255–293. Springer,
Heidelberg (2012). https://doi.org/10.1007/978-3-642-31727-9 10
10. Galanter, P.: Generative art theory. In: Paul, C. (ed.) A Companion to Digital Art,
pp. 146–180. John Wiley, Chichester (2016)
11. Gherardi, F., Russo, S.: Burrowing activity in the sand-bubbler crab, Dotilla fen-
estrata (Crustacea, Ocypodidae), inhabiting a mangrove swamp in Kenya. J. Zool.
(Lond.) 253, 211–223 (2001)
12. Greenfield, G., Machado, P.: Ant- and ant-colony-inspired alife visual art. Artif.
Life 21, 293–306 (2015)
13. Hartnoll, R.G.: Factors affecting the distribution and behaviour of the crab Dotilla
fenestrata on East African shores. Estuar. Coast. Mar. Sci. 1, 137–152 (1973)
14. House, A., Agah, A.: Autonomous evolution of digital art using genetic algorithms.
J. Intell. Syst. 25, 319–333 (2016)
15. Jacob, C.J., Hushlak, G., Boyd, J.E., Nuytten, P., Sayles, M., Pilat, M.: Swarmart:
interactive art from swarm intelligence. Leonardo 40, 248–254 (2007)
16. Johnson, C.G.: Fitness in evolutionary art and music: a taxonomy and future
prospects. Int. J. Arts Technol. 9, 4–25 (2016)
17. Luschi, P., Del Seppia, C., Crosio, E.: Orientation during short-range feeding in
the crab Dotilla wichmanni. J. Comp. Physiol. A. 181, 461–468 (1997)
18. Machado, P., Martins, T., Amaro, H., Abreu, P.H.: An interface for fitness func-
tion design. In: Romero, J., McDermott, J., Correia, J. (eds.) EvoMUSART 2014.
LNCS, vol. 8601, pp. 13–25. Springer, Heidelberg (2014). https://doi.org/10.1007/
978-3-662-44335-4 2
19. Mallon, B., Redies, C., Hayn-Leichsenring, G.U.: Beauty in abstract paintings:
perceptual contrast and statistical properties. Front. Hum. Neurosci. 8, 161 (2014)
20. Mohan, P.M., Pandey, P., Vijay, D., Dhivya, P.: Studies on morphology and ori-
entation of pseudofaecal pellets of sand bubbler crab. J. Coast. Env. 2, 129–142
(2011)
 Visual Art Inspired by the Collective Feeding Behavior 17

21. Neumann, A., Alexander, B., Neumann, F.: Evolutionary image transition using
random walks. In: Correia, J., Ciesielski, V., Liapis, A. (eds.) EvoMUSART 2017.
LNCS, vol. 10198, pp. 230–245. Springer, Cham (2017). https://doi.org/10.1007/
978-3-319-55750-2 16
22. Romero, J., Machado, P. (eds.): The Art of Artificial Evolution: A Handbook
on Evolutionary Art and Music. Natural Computing Series. Springer, Heidelberg
(2008). https://doi.org/10.1007/978-3-540-72877-1
23. Ross, B.J., Ralph, W., Zong, H.: Evolutionary image synthesis using a model of
aesthetics. In: Yen, G.G. (ed.) Proceedings of the IEEE Congress on Evolutionary
Computation, IEEE CEC 2006, pp. 1087–1094. IEEE Press, Piscataway (2006)
24. Spehar, B., Clifford, C.W., Newell, B.R., Taylor, R.P.: Universal aesthetic of frac-
tals. Comput. Graph. 27, 813–820 (2003)
25. Urbano, P.: The T. albipennis sand painting artists. In: Di Chio, C., et al. (eds.)
EvoApplications 2011. LNCS, vol. 6625, pp. 414–423. Springer, Heidelberg (2011).
https://doi.org/10.1007/978-3-642-20520-0 42
 Dynamical Music with Musical
Boolean Networks

George Gabriel(B) and Susan Stepney

Department of Computer Science, University of York, York, UK

{gg685,susan.stepney}@york.ac.uk

Abstract. An extended Boolean network model is investigated as a pos-

sible medium in which a human composer can write music. A Boolean
network is a simple discrete-time dynamical system whose state is char-
acterised by the states of its constituent Boolean-valued vertices. The
evolution of the system is predetermined by an initial state and the
properties of the activation functions associated with each vertex. By
associating musical events with the states of the system, its trajectory
from a particular start state can be interpreted as a piece of tonal music.
The primary source of interest in composing music using a deterministic
dynamical system is the dependence of the musical result on the initial
conditions. This paper explores the possibility of producing musically
interesting variations on a given melodic phrase by changing the initial
conditions from which the generating dynamical system is started.

Keywords: Music · Dynamical systems · Boolean networks

Computer-assisted composition

1 Introduction

Research on computer-assisted and algorithmic composition has produced many

ways of composing new music, such as evolutionary and unconventional com-
putational approaches [12,13]. These methods have been and continue to be
designed according to various purposes and intentions, with each design allow-
ing for differing levels of human involvement in the compositional process.
The theory of dynamical systems has been applied to the field of algorith-
mic composition for some time, using systems such as cellular automata [4] and
neural networks [2] to autonomously produce new music. Here we use a repre-
sentation of music as the trajectory of such a dynamical system; we investigate
a musical interpretation of the dependence of a dynamical system on its initial
conditions.
A dynamical system can be thought of as an abstract structure whose state
changes, either deterministically or non-deterministically, as time progresses. In
the non-deterministic case, transitions between states are decided upon proba-
bilistically. This notion has been applied in the algorithmic generation of music

c Springer International Publishing AG, part of Springer Nature 2018

A. Liapis et al. (Eds.): EvoMUSART 2018, LNCS 10783, pp. 18–33, 2018.
https://doi.org/10.1007/978-3-319-77583-8_2
Another random document with
no related content on Scribd:
 abandoned compasses, square, and drawing-
pen in despair, if Eugène had not been at hand to
set him right again.

Fig. 41.
 CHAPTER XVI.
THE CRITIC.

It was the end of November, and the weather had hitherto allowed
our builders to make use of every day. The autumn sun was
favourable to the enterprise, and at several points the house was
reaching the height of the window-heads of the ground floor.
Nevertheless it required all M. de Gandelau’s determination to
prevent the works from being suspended. By degrees the ground
was deserted by the able-bodied workmen, who were being called
away to the army. Those who remained had their attention
distracted, and were not making the best use of their time. It was
becoming difficult to get the hauling done, all the horses and carts
being pressed into the service of the country. The province was
furrowed in every direction by the tracks of regiments making for the
Loire. Many hours were spent in talk; and every one was anxiously
expecting news of the war, which assumed an increasingly gloomy
aspect. However, Orleans had been re-occupied by the French
troops, and all hope did not seem lost. Paris was resisting. In the
meantime an addition was made to the circle at M. de Gandelau’s
château in the person of a friend of the family, who, having had his
property occupied and injured by the Germans, had been obliged to
abandon it for fear of worse, and came to pay a visit to M. de
Gandelau on his way to the west of France, where he had relations.
He was a man of about fifty or sixty years of age, tall, and of frigid
aspect, though a perpetual smile seemed stereotyped on his face.
He might have been taken for a diplomatist of the old stamp.
The new-comer had read and travelled much, knew a little of
everything, was a member of several learned societies, and his
opinion carried a certain weight with it in his département. He had
been a candidate for the legislature; had embarked in manufacturing
speculations, in which he had lost a good deal of money; then in
agricultural enterprises, but as they threatened to ingulf the
remainder of his fortune, he ultimately rested content with the
theoretical side of things, and with publishing pamphlets on
questions of all kinds, printed at his own expense, and lavishly
circulated. Every one of these brochures professed to give a simple
solution of all the difficulties in question, whether in the domain of
politics, science, manufactures, commerce, and even art. Building
had been one of his hobbies; but as architects appeared to him
unpractical, extravagant, and imbued with prejudices, he had taken
the sole direction of his building operations, making his own
bargains, treating directly with the contractors, giving the plans, and
superintending the work. This whim had been a very costly one, and
one fine morning his building fell to pieces. As he had no more faith
in engineers than in architects, he had determined to lay out roads
on his estate, and have them made according to a system of his
own. His attempts in this line had not been more successful than
those in building. The roads persisted in being impracticable. But M.
Durosay (that was the gentleman’s name) was one of those persons
whom experience—even though acquired at their own expense—
teaches but little. In other respects he was a worthy man; he was
extremely polite and obliging—generous even—especially towards
those who had the art of flattering his whims, and who, through
interest or conviction, gave him credit for being an infallible judge in
matters of all kinds.
If any one had come to consult him on any subject at the moment
he was about to step into a railway carriage, he would have let the
train go rather than not give a formal judgment, with reasons in full. It
must be observed, however, that he judged everything by an à priori
system, and would listen with only partial attention to the particular
reasons that tended to modify its application to the case in question.
On the other hand, he would allow his positions to be discussed, and
did not manifest the least impatience if his opinion was not shared by
others. He was fond of repeating this aphorism: “Light emanates
from the shock of conflicting ideas;”—but with the understanding that
he always played the part of the producer, never that of the recipient.
A short time after his arrival, and when the gloomy subjects of
conversation which were the order of the day had been exhausted,
they began to talk about Paul’s house (as it was the custom at the
château to call it). M. Durosay asked to see the designs. “Building
and I are old acquaintances; I know something about it,” said he.
Eugène could not repress a smile; but the speaker took no notice
of it, his mishaps as a builder having left no painful recollections in
his mind.
“Capital!” said M. Durosay, when they had explained the plans to
him, and he had examined them. “I have seen houses in Belgium
something like this. There are very good ideas here; it will be a very
pleasant habitation if our friends the Prussians let you finish it.... Will
you allow me to make one or two remarks about it?”
“Certainly.”
“Not that I presume for a moment to suggest any change in these
plans, which appear to me admirable.... But I have had the
opportunity of a wide field of observation and comparison.... Well,
then, to tell you frankly my first impression, this seems to me to have
rather the character of a town-house, what we call a hôtel, than a
country-house.... You will excuse my saying so, will you not?... I do
not understand a country-house thus shut in: I should like to see a
portico round it, or at least a wide veranda;—windows opening out—
a more decided reflex of exterior life.”
“But, my dear friend,” said M. de Gandelau, “I expect that my
children will come and spend a good part of the year here; it is no
object with them to have one of those habitations in which people
reside only for the two or three summer months, and where they
entertain the idlers of the city; they want a good house, which will
perfectly exclude wind and wet, where they can live comfortably at
every season of the year.”
“Certainly—a very proper consideration; but what do you think of
those North Italian villas, where the climate is pretty severe in winter
and spring, but which are not the less charming with their porticos,
terraces, wide open entrance-halls, and their balconies looking out
over the country? All these habitations have a dignified aspect; they
ennoble life, we may say, and enlarge those narrow ideas to which
our age is only too prone.... And then, does it not seem to you that
there is a too manifest want of symmetry, at least in one of the
façades? Doesn’t this make the house look a little like those edifices
which have been built piecemeal, with a view to satisfy successive
requirements—in short, is there not a want of that unity which ought
to be found in every work of art?”
“But it is not a work of art that I wish to leave my daughter; it is a
good house—convenient and substantial.”
“Very good. But you will allow that if we can secure both kinds of
excellence, so much the better. For a person of such extreme
refinement and so charming in every respect as your daughter, it is
but proper that a habitation should be provided reflecting in its
exterior the charms and graces of its occupant. It would be a
pleasure to you, in visiting Madame Marie, to see in the distance her
little family grouped around her under a portico of delicate
architecture, or under a loggia.... But this seems to me more like the
house of some grave Flemish alderman. In these gables there is a
kind of severity which——”
“Come, come, my dear friend, gables are not severe; they are
gables—that’s all.”
“But indeed these gables with their high roofs have a severe
aspect, which by no means agrees with the idea one forms of a
house built for pleasure.”
“But it is not a house built for pleasure; it is a house built for people
who are going to live in it, not for summer loungers—especially as
we never have such people in our neighbourhood.”
“Still, however, I should have liked to give a warmth to these fronts
(which have a somewhat frigid aspect) by light and airy projections,
and a covered gallery, with a terrace over it.”
“Warmth? warmth? Why instead of that, you would give us the
rheumatism with your galleries. They may do very well at Nice or
Mentone, but they are not to be thought of in our part of the country.
We want the sun upon the walls of our habitations, while your
porticos are like mushroom-houses.”
 “I see, my dear friend,” resumed M. Durosay, after a pause, “that
you keep to your taste for what you call the practical side of things.
Yet see what a good opportunity you have of giving your daughter
one of those dwellings which, while satisfying the material
requirements of life, would possess that perfume of art which is too
rarely found in our country districts. A little exterior elegance is a
powerful charm which leaves an indelible trace in the mind. It is thus
that the Italians preserve the poetry of the brilliant eras of their
civilization. They are willing, at need, to sacrifice something of what
we call ‘comfort’—the material conveniences of life—to keep up
among them the noble traditions of high art.”
“I do not know what the traditions of high art are, or whether those
traditions preserve us from rain, wind, and sun; but I must confess
that your Italian villas in the environs of Verona and Venice appear
very dull and gloomy with their colonnades and closed shutters. I
have never had the wish to visit them, for I imagine one would be
very uncomfortable in them. If they build them so with a view to offer
tourists models of architecture, all well and good; but I make no
pretensions to amuse or interest tourists, and my daughter shares
my ideas in the matter.”
“Perhaps ... but just now your daughter is travelling in Italy; she is
going to sojourn on the shores of the Bosporus; who knows whether
on her return here she would not be charmed to meet a kind of
souvenir of the impressions she will not fail to have experienced
there, and whether the surprise you have in store for her would not
be still more delightful if you tried in some measure to revive those
impressions? What do you think of it, Mr. Architect?”
“As for myself,” said Eugène, “I am listening, and cannot but be
delighted to hear you discourse so ably on our art.”
“I may take it for granted, then, that you share my opinion, and that
you would be inclined to give this habitation, which you have so
skilfully arranged, some of those external charms in which perhaps it
is now deficient.”
“I cannot say that I should. M. de Gandelau, with his usual
courtesy, has left us quite at liberty, and has simply stated the limit of
expense to which he is prepared to go. As regards other
considerations, our programme having been agreed upon, we have
not been restricted to an excessive severity of style, nor forbidden
the adoption of what you consider the exterior charms of a dwelling-
house.”
“Well; although my friend with his practical mind does not appear
sensible to these charms, do not you, as an artist, think it desirable
to add something to these fronts, which are perhaps a little severe in
aspect, and which certainly with the help of your talent you could
render less cold? You know Italy; you have visited Pompeii: do you
not find in the architecture of those countries abundant suggestions
from which inspiration may be drawn—charming models, in fact?”
“Yes; I have visited Italy and France, but I must confess that I have
never been struck by the architectural works of those countries,
except so far as they preserved the imprint of the manners and
customs of those whose genius produced them. You mention
Pompeii. That which has vividly affected me in the remains of this
little provincial town of Italy is precisely this characteristic. Its small
dwellings exactly suited the habits of antiquity, the time when they
were erected, and the climate of the district. But from the study of
these habitations I infer that since we do not live on the shores of the
Gulf of Naples, and have customs very different from those which
suited the Pompeians, our dwellings ought not in any way to suggest
the peculiarities of theirs; that while, for example, it may have been
very agreeable to them to sup in an open triclinium, sheltered from
the wind by a velum, we cannot arrange dining-rooms after this
model in the Département de l’Indre; and that though it might have
been a luxury to them to sleep in a room whose area was only five or
six square yards and the door of which, left open, introduced you to
a court surrounded by a portico, this would be very inconvenient
here, as we should run great risk of catching cold if we left the door
open, or of being suffocated if we shut it.”
“But as you have mentioned ancient dwellings, allow me to remark
that those of Pompeii, even the most luxurious, do not exhibit
externally any of those magnificent features which you seem to
admire. The ancients reserved for the interior such luxury as they
affected, and it does not appear that they troubled themselves to
display anything of it to the passers-by. I have not a very clear idea
of what their villas,—their country-houses,—may have been; but
everything leads me to believe, as far as we can judge from the
remains preserved to us, that in them nothing was sacrificed to that
distinctively modern vanity which aims to make an external display of
architectural forms to strike the vulgar.
“I believe that those country palaces of Northern Italy with which
you have been so deeply smitten, are rather products of vanity than
abodes adapted to the habits of those who have erected them; in
fact, they have scarcely been inhabited, and the dilapidated
condition in which you see them does not date from yesterday.
Erected to satisfy vanity and the desire to make a show, they lasted
as habitations only as long as works due to vanity are accustomed to
last—that is, for a few years of the life of an individual; after which
they were abandoned.”
“You call vanity,” replied M. Durosay, “what I think to be love of art
—the desire to exhibit a work of art.”
“Probably we shall never agree upon that point,” answered
Eugène. “I think that art—in architecture at least—consists in being
truthful and simple. You see in it only a form that charms or repels
you: I look for something else; or rather I consider first whether this
form is really the expression of a requirement—whether a reason
can be given for its existence; and it charms me only so far as this
condition is fulfilled, according to my judgment.”
“You consider a barn, therefore, a work of art?”
“Certainly; if it is constructed so as to afford a suitable shelter for
what it is intended to hold, it is, in my view, more admirable than an
inconvenient palace, though decorated with colonnades and
pediments.”
“You ought to go to America.”
“Perhaps it would be wise to do so, if I knew that its people tried to
build simply in accordance with the tastes and requirements of the
inmates. But in America, as everywhere else now-a-days, they make
pretensions to style, and copy what they believe to be the beautiful
par excellence; that is, they follow, without discrimination, traditions
whose origin and principle they do not care to investigate.”
“Come,” said M. de Gandelau, who found the discussion rather
tedious, “we have travelled a good way from Paul’s house; but to
satisfy you, when you come and see my daughter in her new
dwelling, we will have a pasteboard portico put up in front of one of
the façades, and under the shade some Berri maidens dressed up
as Venetians, and some gentlemen in scarlet robes playing on the
guitar and the bassoon. It is getting late, and time to go to bed.”
 CHAPTER XVII.
PAUL INQUIRES WHAT ARCHITECTURE IS.

Eugène expected Paul to return to the discussion of the previous

evening, and in fact, when they were going early in the morning to
visit the works, Paul did not fail to throw out hints about it. But he did
not know how to give his curiosity a definite shape. His cousin would
not help him, but wished to give him full leisure to bring his ideas to a
focus.
“Is M. Durosay a judge of architecture?” said Paul, at last.
“Well, he talks about it like a person who has some acquaintance
with the art.”
“But yet you did not seem disposed to accede to what he asked.”
“What did he ask?”
“Why——you know very well what I mean——. He would have
liked Marie’s house to be——more——.”
“More what?”
“More——less severe; that it should have a portico and a loggia.
What is a loggia?”
“It is a wide covered balcony, most frequently closed on the two
sides, but opening in front—whether on the ground floor or the upper
stories—to the high road or the country.”
“And why should not a loggia be added to Marie’s house?”
“We might make one, or several.”
“Well, then?”
“Why then it must be placed in front of one of the apartments—the
drawing-room, for example, on the ground floor, in the middle of the
garden front, or if on the first floor, in front of the best bedroom.”
“And would not that have a good effect?”
 “Perhaps it might: but the apartment next to it, opening upon this
loggia, would be dark and gloomy, as the windows would be shaded
by its ceiling.”
“Ah! yes, that is true; but in fact we have loggias at the end of the
drawing-room, the billiard-room, and the dining-room.”
“Yes; only they are closed, instead of being open towards the
outside, and these apartments gain in area through them. These
loggias are therefore recesses—what they formerly called ‘bays.’ We
have thus all the advantages of a loggia without the inconveniences
which in our climate it would entail.”
“Why did you not say so to M. Durosay?”
“He could see it well enough; there was no need to mention it to
him.”
“He would have liked a portico, too.”
“For what purpose?”
“I do not know——. He said it would be pretty—that my sister and
her children would form a group under it, and that this would have a
pretty effect at a distance.”
“And would it be very agreeable to your sister to produce ‘a very
pretty effect’ at a distance?”
“Oh, I don’t think she would care about it.”
“But who are we building the house for?”
“Why, for my sister.”
“Not for strolling idlers, therefore. But the portico in question would
have the same inconveniences as the loggias; it would make the
apartments opening under the arcades or colonnades dark and
gloomy. Since then, in our country we spend more of our time in
rooms than under porticos, we should have to pay rather dearly for
the pleasure of forming groups for the gratification of passing
strangers.”
 “Doubtless we should. Besides, in front of the billiard-room we
have a conservatory, with steps down to the garden, which may
serve for a portico without darkening the room, as it will be glazed.”
“Certainly.”
“Perhaps M. Durosay did not observe this.”
“Oh! I daresay he did; but it has nothing imposing about it. He
would have liked a real covered portico, in the style of the Italian
porticos.”
“He seems to be very fond of Italian architecture.”
“Which?”
“Why, that he was talking about.”
“But there are many kinds of architecture in Italy, belonging to
different ages and latitudes, and varying with the habits of the
peoples who inhabit the peninsula.”
“You did not call his attention to that.”
“He must have known it.”
“I see that you don’t think M. Durosay earnest in his opinions.”
“M. Durosay is an excellent man; his opinions are sincere, and
therefore I regard them as serious; but he and I look at things from a
different point of view. He judges questions of art as a man of the
world, on a ground of sentiment, while I think we artists ought to
decide them by reasoning. Sentiment does not reason; it is like faith;
so it is impossible for us to understand one another, since we speak
a different language.”
Paul’s views on the subject were as yet far from clear. Hitherto he
had thought that architecture could be learned as we learn grammar
and spelling; and here was his cousin telling him that it found
expression in several languages, one of which might be known,
while the other remained quite unintelligible. He could not
understand what reasoning could have to do with a matter entirely
relating to form and appearance; yet he did not even know how to
put questions to his cousin on the subject with a view to gaining light
upon it. He was going along, therefore, with his head bent, striking
down with his stick the withered thistles that encumbered the side of
the path—Eugène, on his side, not seeming desirous of breaking the
silence. They arrived thus at the works; they were almost deserted.
“It froze last night,” said Branchu; “and it’s going to be a hard
frost.”
“Well, you must cover the stone-work with litter or straw, and we
shall have to stop. Put some scaffold planks on the walls, the straw
over them, and slabs with stones at intervals. Take care that the
planks project beyond the faces of the walls. If you have not straw
enough, put soil on the slabs, or turf clods. As to the cellar vaults,
spread a good layer of mould over them, and contrive some
openings in the haunches, so that the rain or melted snow may run
off. Come, set to work! Let all this be arranged for and finished to-
morrow evening; then we will stop till the end of the cold weather.”
“So much the better,” said Master Branchu, “for all the young
fellows are gone, and none but poor creatures are left at the works.”
“This suspension of the building,” said Eugène, when they were
returning to the château, “will permit us to work out the details of
construction without having to hurry over them.”
“Yes,” replied Paul; “but I should much like to know how you set
about it when you have to draw one of the details.”
“You must have learnt that in the two months we have been doing
this sort of work.”
“Not quite; I perceive that you say what you intend, and that what
you intend shows itself drawn on paper; now, I have tried to do the
same, but though I knew well what I intended, I could bring nothing
to paper; or if I did draw anything, it made me forget what I had in my
mind. Yet, surely, for everything one wants to do in architecture,
there must be a method, a process, a—what should I call it?—a
recipe.”
“Ah! now I see what you mean. But you must perceive, my young
cousin, that people often fancy they understand and intend, while
they really do not always know what they intend, and do not clearly
understand the question in hand. All this morning, for instance, you
have been revolving in your mind this question which you have only
just put to me; and I have wished to give you leisure to present it in a
definite form—to do which your brain has been obliged to work. Now,
thanks to the effort you have made, you will comprehend the answer
I am able to give you better. You remember those two lines of
Boileau’s—

“Ce que l’on conçoit bien s’énonce clairement,

Et les mots pour le dire arrivent aisément,”

and which are applicable to all the arts. The great thing is to
habituate one’s self to clear conceptions. Unfortunately we learn to
form phrases before we learn to reason, and try to express an idea
before it is completely elaborated in the brain. Then we fancy we can
supply what is incomplete in this idea by a happy combination of
words. In architecture we think of forms that have seemed attractive,
before knowing whether they will be exactly appropriate to what
reason and the rigorous observance of the necessities of
construction, or the requirements of the case, demand. In a speech,
the vulgar are readily seduced by a brilliant phrase, and perceive
only too late the intellectual void which this seductive form conceals.
Similarly, in architecture, the vulgar are seduced by a picturesque
aspect and an attractive form, and have to pay the penalty of their
error in the defects of the building. M. Durosay, in his admiration for
certain forms that had charmed him as a tourist, has not thought of
asking himself whether these externals were in harmony with the
requirements to be satisfied, and what the structure itself demanded;
the turn of the phrase has arrested his attention, and he has not
inquired whether there was a clearly developed idea behind it. We
might, therefore, have argued together in this way for days, without
hope of convincing one another—he being entirely occupied with the
form or fashion of the phrase, and not troubling himself as to whether
this form has a signification—whether this phrase expresses a clear
idea. All depends on this, my dear Paul, and, in my judgment, our
country, which is so near the verge of absolute ruin, will not recover
itself until it learns to reflect before it speaks. We build immense
edifices, costing fabulous sums, yet we have no clear idea as to
what they shall contain. Or rather, we think only of making the
casing, and leave it an open question whether we shall use it for
such or such a purpose. And I would have you observe that this
unfortunate habit prevails not with regard to public buildings only.
How many respectable men there are who, like M. Durosay, in
proposing to build a house for themselves, first determine to erect a
chalet, or an Italian villa, or an English cottage, according to the
fancy of the moment, and make it quite a secondary question
whether life will be comfortable in the case they are going to put up!
Consequently you will see Italian villas in the north of France and
Swiss chalets at Nice. Learn to reason, to observe before you
proceed to act, and you will be a good barrister, a good physician, a
good soldier, a good architect. If nature has endowed you with
genius, so much the better; it will supply a noble complement to your
faculties; but if you have not gained the habit of reasoning, genius
will be of no use to you, or rather, it cannot develop itself. Now, to
learn to reason, you must labour much and labour long, and not
allow yourself to be led astray by appearances, however attractive.
Unfortunately, our education and instruction in France lead us to
content ourselves with mere appearances, and to rely on traditions
which are regarded as articles of faith, and which consequently may
not be discussed. You will find M. Durosay’s portico confronting you
everywhere. The army, the government, literature, politics, and the
arts have their portico, which you must adopt, whatever has to be
done or wherever an entrance has to be provided; unless you have
sufficient energy, power for work, independence of character,
practical knowledge, persistent determination, and the authority
which that alone can secure, to say:—I will adopt your portico only
as far as I think it advantageous to make use of it. But to return to
your question as to whether any definite prescriptions or rules of
procedure can be given in architecture, I reply that there are practical
rules of procedure suitable to construction; but as the materials and
the means of execution vary continually, any such rules ought to be
modified by these variations. In architecture there is a method to be
followed in all cases that present themselves, but there are no
definite prescriptions or rules of procedure. This method is none
other than the application of your reasoning faculty to all particular
cases; for what is desirable in one set of circumstances is not so in
another. It is therefore on the observation of these circumstances—
of facts, customs, climate, and hygienic conditions—that your reason
must rely before forming the conception of your work. And when this
operation is complete, and all is properly arranged in your mind, then
you will be able to put on paper without hesitation the result of this
intellectual labour.”
“I think I apprehend your meaning; but how must I begin?”
“By acquiring the habit of observing everything, and reflecting on
everything you see, hear, or read. When you have a ditch before you
that you want to cross, do you not ask yourself whether your legs will
carry you to the other side? do you not know, as the result of
previous observation, whether you can jump the ditch or not, and do
you not decide accordingly? You do not ask yourself before jumping
whether Achilles or Roland was alleged by the poets to have leaped
much wider spaces. It is yourself, your own strength, that you consult
—not that of heroes—on pain of tumbling into the water. Exactly in
the same way, if you have to build a house for a person you know,
you first remind yourself that a house is made for people to live in;
then you represent to yourself the habits of the owner, you calculate
the number of apartments he requires, and what relations they will
have to each other. You know whether he lives alone, or entertains
much company; whether he will live in the house at such or such a
season; whether he affects luxury or lives quietly; whether he has
many servants, or employs only one, &c.: and when you have
thoroughly considered all these essential conditions, you will try to
put on paper the result of these observations. But if the first thing you
think of is putting this person and his family in a house like those of
Pompeii, or in a feudal château, it is a thousand to one that you will
build him an uncomfortable habitation—that you will be obliged to
sacrifice the convenience of its arrangements in order to assign them
a place in a building that belongs to a period and a civilization
differing from our own civilization and times.”
“I can quite understand that, but still we can learn how to make a
door, a window, or a staircase.”
 “That is to say, it is possible to explain how people in former times
set to work to make a door, a staircase, or a floor; but it is not
proposed, nor ought it to be proposed, in teaching you the methods
employed by our predecessors, to oblige you to do exactly what they
did, since you perhaps possess materials which they did not, and
your customs differ from theirs. The instructions given you run thus,
—at least they ought to run thus: ‘These are the results of the
experience acquired during past ages; make these your starting-
point; do as your predecessors have done; use your reasoning
faculty in applying the knowledge that has been acquired, but in
obedience to the requirements of the present. You ought not to be
ignorant of what has been done before you,—it is an accumulation
for the common good, a possession secured to mankind. You ought
to be acquainted with its existence and value; but, as a partner in its
advantages, add your store of intelligence; make a step in advance,
do not retrograde.’ But observe: there is only one means of
preventing retrogression in architecture, and that is making art the
faithful expression of the requirements of the time in which we live,—
making the building a casing suited to that which it is destined to
contain.”
“And is not this always done?”
“Not exactly. We are something like persons who have inherited
from their ancestors a costly stock of furniture,—a venerable and
venerated heirloom—who keep and make use of this furniture,
though it is inconvenient to them, and no longer suits the habits of
the times; who have even gone so far as to appoint a guardian for
this old lumber, who is enjoined not to allow it to be modified. If
therefore you, the master of the house, want to change the covering
of this furniture, or send some of the articles themselves, which are
more inconvenient than useful, to the lumber-room, the guardian you
pay and lodge assumes a dignified air, and declares that the function
with which you have invested him, and which he makes a point of
strictly discharging, forbids him to allow these modifications or
suppressions; that his honour is concerned in not allowing these
relics to suffer injury or change, since he is commissioned to
preserve them. For the sake of peace, you continue to make use of
this intolerable furniture, and you retain its guardian.”
“I do not quite understand you.”
“By and by you will. But observe, I have given you fair warning. If
you go into some old mansion crammed with antiquated furniture,
take care not to criticise it; for though the host and hostess may
content themselves with smiling, the guardian of those curiosities will
take good care that you never set foot in that house again.”
 CHAPTER XVIII.
THEORETICAL STUDIES.

The cold and the state of the times prevented the works from being
continued. The winter might be a long one. Eugène and Paul
prepared themselves, therefore, to employ this compulsory leisure to
advantage. It was decided between them that they should not merely
draw out the details necessary for finishing the works, but that
Eugène should take advantage of these winter days to enlighten
Paul on many points respecting which, as clerk-of-works, he was
deficient.
Paul took an increasing interest in this employment. Hitherto the
execution had immediately followed the labours of the study, and
example and practice came to ratify theory; but he was quite aware
that all his attention and desire to follow the lead of his chief were not
sufficient, and that at each step he found himself confronted by a
difficulty. The further the work advanced, the more utterly incapable
did he feel himself. He set to work, therefore, with a hearty desire to
learn; indeed, so much the more eagerly as all that surrounded him
assumed a more and more gloomy and desolate aspect. Paul had
never spent a winter in the country, although he used to come home
to the Christmas festivities; the few days spent at his father’s
château had passed away so quickly, that he had not time to
consider how things looked out of doors. Besides, the house was full
of guests at that time; the presence of his elder sister gave it
animation; everything had a holiday aspect. But the scene was quite
changed at the beginning of December, 1870; the neighbouring
villages were deserted, or occupied only for a few hours by troops ill-
clad, dying of hunger, generally going to fight without enthusiasm,
and leaving the exhausted and the sick in the cottages. Then would
come long lines of carriages that looked like so many funeral
processions.
The snow was beginning to cover the fields and to muffle distant
sounds. Seldom did any of the peasants come to the château. The
postman still paid his regular visits, but the letters and newspapers
he brought tended only to depress the spirits of the inmates.
Sometimes they gave shelter to members of the Garde Mobile, or to
soldiers of the line; but all were dumb: the officers themselves would
ask to be allowed to rest in their rooms under pretext of fatigue,
rather than go down to the drawing-room. M. de Gandelau, up early
in the morning, in spite of his gout, seemed to be omnipresent; he
was to be found everywhere, among the farms and at the
neighbouring town, facilitating the transport of munitions of war,
organizing hospitals, supplying provisions and lightening the
difficulties imposed by routine. “Set Paul to work, my friend,” he said
to Eugène every evening; “that is all the demand I make on your
friendship. I feel it is a considerable one, but grant it, I entreat you.”
In fact, the greater part of the day was passed in studying some
question relating to building; then the architect and his clerk-of-works
would go and take a walk before the evening, during which Eugène
did not fail to start some interesting topic. The country and natural
phenomena were the habitual subjects of these conversations; and
thus Paul was learning to observe and reflect, and it became every
day clearer to him how much knowledge must be acquired to
accomplish a task of even limited scope. His cousin did not fail to
reiterate the sentiment: “The more you know the more you will feel
your want of knowledge; and the highest acquisition in science is the
conviction that we know nothing.”
“What good is it to learn, then?” rejoined Paul, one day.
“That we may become modest; that we may occupy life with
something better than those things to which vanity prompts us; that
we may make ourselves of some little use to our fellows, without
exacting gratitude from them.”
Eugène made Paul draw a good deal, and always from nature, or
from drawings executed while he was present, for he had not
brought with him any specimens of architectural design. Besides
this, Paul made a fair copy of memoranda relating to the parts of the
house already erected. Thus he gained a complete acquaintance
with the structure of every part of the stone-work.