Automata Languages And Programming 41st

International Colloquium Icalp 2014 Copenhagen

Denmark July 811 2014 Proceedings Part Ii 1st
Edition Javier Esparza download
https://ebookbell.com/product/automata-languages-and-
programming-41st-international-colloquium-icalp-2014-copenhagen-
denmark-july-811-2014-proceedings-part-ii-1st-edition-javier-
esparza-4935126

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Automata Languages And Programming 41st International Colloquium Icalp

2014 Copenhagen Denmark July 811 2014 Proceedings Part I 1st Edition
Javier Esparza

https://ebookbell.com/product/automata-languages-and-programming-41st-
international-colloquium-icalp-2014-copenhagen-denmark-
july-811-2014-proceedings-part-i-1st-edition-javier-esparza-4935124

Automata Languages And Programming 38th International Colloquium Icalp

2011 Zurich Switzerland July 48 2011 Proceedings Part I 1st Edition
Piotr Berman

https://ebookbell.com/product/automata-languages-and-programming-38th-
international-colloquium-icalp-2011-zurich-switzerland-
july-48-2011-proceedings-part-i-1st-edition-piotr-berman-2449784

Automata Languages And Programming 38th International Colloquium Icalp

2011 Zurich Switzerland July 48 2011 Proceedings Part Ii 1st Edition
Rajeev Alur

https://ebookbell.com/product/automata-languages-and-programming-38th-
international-colloquium-icalp-2011-zurich-switzerland-
july-48-2011-proceedings-part-ii-1st-edition-rajeev-alur-2449944

Automata Languages And Programming 37th International Colloquium Icalp

2010 Bordeaux France July 610 2010 Proceedings Part Ii 1st Edition
Pierre Fraigniaud Auth

https://ebookbell.com/product/automata-languages-and-programming-37th-
international-colloquium-icalp-2010-bordeaux-france-
july-610-2010-proceedings-part-ii-1st-edition-pierre-fraigniaud-
auth-2533248
Automata Languages And Programming 38th International Colloquium Icalp
2011 Zurich Switzerland July 48 2011 Proceedings Part I 1st Edition
Piotr Berman

https://ebookbell.com/product/automata-languages-and-programming-38th-
international-colloquium-icalp-2011-zurich-switzerland-
july-48-2011-proceedings-part-i-1st-edition-piotr-berman-4140796

Automata Languages And Programming 37th International Colloquium Icalp

2010 Bordeaux France July 610 2010 Proceedings Part I 1st Edition
Burkhard Monien

https://ebookbell.com/product/automata-languages-and-programming-37th-
international-colloquium-icalp-2010-bordeaux-france-
july-610-2010-proceedings-part-i-1st-edition-burkhard-monien-4140798

Automata Languages And Programming 32nd International Colloquium Icalp

2005 Lisbon Portugal July 1115 2005 Proceedings 1st Edition Leslie G
Valiant Auth

https://ebookbell.com/product/automata-languages-and-programming-32nd-
international-colloquium-icalp-2005-lisbon-portugal-
july-1115-2005-proceedings-1st-edition-leslie-g-valiant-auth-4239120

Automata Languages And Programming 35th International Colloquium Icalp

2008 Reykjavik Iceland July 711 2008 Proceedings Part I 1st Edition
Bruno Courcelle Auth

https://ebookbell.com/product/automata-languages-and-programming-35th-
international-colloquium-icalp-2008-reykjavik-iceland-
july-711-2008-proceedings-part-i-1st-edition-bruno-courcelle-
auth-4240362

Automata Languages And Programming 35th International Colloquium Icalp

2008 Reykjavik Iceland July 711 2008 Proceedings Part Ii 1st Edition
Ran Canetti Auth

https://ebookbell.com/product/automata-languages-and-programming-35th-
international-colloquium-icalp-2008-reykjavik-iceland-
july-711-2008-proceedings-part-ii-1st-edition-ran-canetti-auth-4520930
 Javier Esparza
Pierre Fraigniaud
Thore Husfeldt
Elias Koutsoupias (Eds.)
ARCoSS
LNCS 8573

Automata, Languages,
and Programming
41st International Colloquium, ICALP 2014
Copenhagen, Denmark, July 8-11, 2014
Proceedings, Part II

123
Lecture Notes in Computer Science 8573
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison, UK Takeo Kanade, USA
Josef Kittler, UK Jon M. Kleinberg, USA
Alfred Kobsa, USA Friedemann Mattern, Switzerland
John C. Mitchell, USA Moni Naor, Israel
Oscar Nierstrasz, Switzerland C. Pandu Rangan, India
Bernhard Steffen, Germany Doug Tygar, USA
Demetri Terzopoulos, USA
Gerhard Weikum, Germany

Advanced Research in Computing and Software Science

Subline of Lectures Notes in Computer Science

Subline Series Editors

Giorgio Ausiello, University of Rome ‘La Sapienza’, Italy
Vladimiro Sassone, University of Southampton, UK

Subline Advisory Board

Susanne Albers, University of Freiburg, Germany
Benjamin C. Pierce, University of Pennsylvania, USA
Bernhard Steffen, University of Dortmund, Germany
Deng Xiaotie, City University of Hong Kong
Jeannette M. Wing, Microsoft Research, Redmond, WA, USA
Javier Esparza Pierre Fraigniaud
Thore Husfeldt Elias Koutsoupias (Eds.)

Automata, Languages,
and Programming

41st International Colloquium, ICALP 2014

Copenhagen, Denmark, July 8-11, 2014
Proceedings, Part II

13
Volume Editors
Javier Esparza
Technische Universität München, Germany
E-mail: [email protected]
Pierre Fraigniaud
LIAFA, Université Paris Diderot-Paris 7, France
E-mail: [email protected]
Thore Husfeldt
IT University of Copenhagen, Denmark
E-mail: [email protected]
Elias Koutsoupias
University of Oxford, UK
E-mail: [email protected]

ISSN 0302-9743 e-ISSN 1611-3349

ISBN 978-3-662-43950-0 e-ISBN 978-3-662-43951-7
DOI 10.1007/978-3-662-43951-7
Springer Heidelberg New York Dordrecht London

Library of Congress Control Number: 2014941781

LNCS Sublibrary: SL 1 – Theoretical Computer Science and General Issues

© Springer-Verlag Berlin Heidelberg 2014
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. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered and
executed on a computer system, for exclusive use by the purchaser of the work. Duplication of this publication
or parts thereof is permitted only under the provisions of the Copyright Law of the Publisher’s location,
in ist current version, and permission for use must always be obtained from Springer. Permissions for use
may be obtained through RightsLink at the Copyright Clearance Center. Violations are liable to prosecution
under the respective Copyright Law.
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.
While the advice and information in this book are believed to be true and accurate at the date of publication,
neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or
omissions that may be made. The publisher makes no warranty, express or implied, with respect to the
material contained herein.
Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
 Preface

This volume contains the papers presented at ICALP 2014: the 41st International
Colloquium on Automata, Languages and Programming, held during July 8–11,
2014, at IT University of Copenhagen. ICALP is the main conference and annual
meeting of the European Association for Theoretical Computer Science (EATCS)
and first took place in 1972. This year the ICALP program consisted of three
tracks:
– Track A: Algorithms, Complexity, and Games
– Track B: Logic, Semantics, Automata, and Theory of Programming
– Track C: Foundations of Networked Computation

In response to the call for papers, the three Program Committees received 484
submissions, a record number for ICALP. Track A received 319 submissions
(another record), track B received 106 submissions, and track C received 59
submissions. Each submission was reviewed by at least three Program Committee
members, aided by many subreviewers. The committee decided to accept 136
papers, which are collected in these proceedings. The selection was made by the
Program Committees based on originality, quality, and relevance to theoretical
computer science. The quality of the submissions was very high indeed, and
many deserving papers could not be selected.
The EATCS sponsored awards for both a best paper and a best student paper
for each of the three tracks, selected by the Program Committees.
The best paper awards were given to the following papers:
– Track A: Andreas Björklund and Thore Husfeldt, “Shortest Two Disjoint
Paths in Polynomial Time”
– Track B: Joel Ouaknine and James Worrell. “Ultimate Positivity Is Decidable
for Simple Linear Recurrence Sequences”
– Track C: Oliver Göbel, Martin Hoefer, Thomas Kesselheim, Thomas Schlei-
den, and Berthold Vöcking, “Online Independent Set Beyond the Worst-
Case: Secretaries, Prophets, and Periods”

The best student paper awards, for papers that are solely authored by stu-
dents, were given to the following papers:
– Track A: Sune K. Jakobsen, “Information Theoretical Cryptogenography”
– Track B: Michael Wehar, “Hardness Results for Intersection Non-Emptiness”
– Track C: Mohsen Ghaffari, “Near-Optimal Distributed Approximation of
Minimum-Weight Connected Dominating Set”

Apart from the contributed talks, the conference included invited presenta-
tions by Sanjeev Arora, Maurice Herlihy, Viktor Kuncak, and Claire Mathieu.
Abstracts of their talks are included in these proceedings as well.
VI Preface

The program of ICALP 2014 also included presentation of the Presburger

Award 2014 to David Woodruff, the EATCS Award 2014 to Gordon Plotkin,
and the Gödel Prize to Ronald Fagin, Amnon Lotem, and Moni Naor.
Two satellite events of ICALP were held on 7 July, 2014:

– Trends in Online Algorithms (TOLA 2014)

– Young Researcher Workshop on Automata, Languages and Programming
(YR-ICALP 2014)

We wish to thank all the authors who submitted extended abstracts for con-
sideration, the members of the three Program Committees for their scholarly
efforts, and all additional reviewers who assisted the Program Committees in
the evaluation process. We thank the sponsors Springer-Verlag, EATCS, CWI
Amsterdam, and Statens Kunstfond for their support, and the IT University of
Copenhagen for hosting ICALP 2014.
We are also grateful to all members of the Organizing Committee and to their
support staff.
The conference-management system EasyChair was used to handle the sub-
missions, to conduct the electronic Program Committee meetings, and to assist
with the assembly of the proceedings.

May 2014 Javier Esparza

Pierre Fraigniaud
Thore Husfeldt
Elias Koutsoupias
 Organization

Program Committee
Dimitris Achlioptas UC, Santa Cruz, USA
Pankaj Agrawal Duke University, USA
Paolo Baldan Università di Padova, Italy
Nikhil Bansal Eindhoven University of Technology,
The Netherlands
Michele Boreale Università di Firenze, Italy
Tomas Brazdil Masaryk University, Czech Republic
Gerth Stølting Brodal Aarhus University, Denmark
Véronique Bruyère University of Mons, Belgium
Jean Cardinal Université libre de Bruxelles, Belgium
Ning Chen Nanyang Technological University, Singapore
Giorgos Christodoulou University of Liverpool, UK
Andrea Clementi University of Rome Tor Vergata, Italy
Veronique Cortier CNRS, Loria, France
Anuj Dawar University of Cambridge, UK
Xiaotie Deng Shanghai Jiaotong University, China
Ilias Diakonikolas University of Edinburgh, UK
Benjamin Doerr MPI Saarbrücken, Germany
Chaled Elbassioni Masdar Institute, Abu Dhabi
Javier Esparza TU München, Germany
Kousha Etessami University of Edinburgh, UK
Panagiota Fatourou University of Crete, Greece
Michal Feldman Hebrew University, Israel
Maribel Fernandez Kings College London, UK
Antonio Fernández Anta Universidad Rey Juan Carlos, Spain
Amos Fiat Tel Aviv University, Israel
Pierre Fraigniaud CNRS and University of Paris Diderot, France
David Frutos Escrig Complutense University of Madrid, Spain
Pierre Ganty IMDEA Software Institute, Spain
Leszek Gasieniec University of Liverpool, UK
Phillip Gibbons Intel Labs, USA
Leslie Goldberg University of Oxford, UK
Vipul Goyal Microsoft, India
Peter Habermehl LIAFA, University of Paris 7, France
Magnus Halldorsson Reykjavik University, Iceland
Giuseppe Italiano University of Rome Tor Vergata, Italy
Marcin Kaminski University of Warsaw, Poland
VIII Organization

Haim Kaplan Tel Aviv University, Israel

Anna Karlin University of Washington, USA
Ioardanis Kerenidis University of Paris Diderot, France
Anne-Marie Kermarrec Inria Rennes, France
Robert Kleinberg Cornell University, USA
Michal Koucky Czech Academy of Sciences, Czech Republic
Elias Koutsoupias University of Oxford, UK
Robert Krauthgamer Weizmann Institute, Israel
Manfred Kufleitner University of Stuttgart, Germany
SΠlawomir Lasota Warsaw University, Poland
James Lee University of Washington, USA
Oded Maler CNRS-VERIMAG, France
Sebastian Maneth NICTA and UNSW, Australia
Madhavan Mukund Chennai Mathematical Institute, India
Ashwin Nayak University of Waterloo, Canada
Jens Palsberg UCLA, USA
Gopal Pandurangan Nanyang Technological University, Singapore
Boaz Patt-Shamir Tel Aviv University, Israel
Andrea Pietracaprina Università di Padova, Italy
Andrea Richa Arizona State University, USA
Luı́s Rodrigues Universidade Técnica de Lisboa, Portugal
Jared Saia University of New Mexico, USA
Piotr Sankowski University of Warsaw, Poland
Christian Scheideler Universität Paderborn, Germany
Thomas Schwentick TU Dortmund, Germany
Maria Serna UP Catalunya, Spain
Sonja Smets University of Amsterdam, The Netherlands
Christian Sohler TU Dortmund, Germany
Jiri Srba Aalborg University, Denmark
Jukka Suomela Aalto University, Finland
Ryan Williams Stanford University, USA
Philipp Woelfel University of Calgary, Canada
Steve Zdancewic University of Pennsylvania, USA

Additional Reviewers
Aaronson, Scott Agarwal, Rachit
Abe, Masayuki Aghazadeh, Zahra
Abraham, Ittai Agrawal, Shweta
Aceto, Luca Ajwani, Deepak
Adler, Isolde Akutsu, Tatsuya
Adsul, Bharat Al-Humaimeedy, Abeer
Afshani, Peyman Alamdari, Soroush
Agarwal, Alekh Alglave, Jade
 Organization IX

Allender, Eric Bogdanov, Andrej

Alon, Noga Bojanczyk, Mikolaj
Althaus, Ernst Boker, Udi
Alves, Sandra Bollig, Beate
An, Hyung-Chan Bollig, Benedikt
Anagnostopoulos, Aris Bonamy, Marthe
Ananth, Prabhanjan Bonchi, Filippo
Andoni, Alex Boneh, Dan
Andoni, Alexandr Bonifaci, Vincenzo
Ardenboim, Alon Bonnet, Edouard
Arkhipov, Alex Bonsangue, Marcello
Asarin, Eugene Bonsma, Paul
Aspnes, James Borgström, Johannes
Atig, Mohamed Faouzi Boutsidis, Christos
Atserias, Albert Boyar, Joan
Augustine, John Boyle, Elette
Avron, Haim Brakerski, Zvika
Babichenko, Yakov Brandstadt, Andreas
Bacci, Giorgio Braverman, Mark
Bacci, Giovanni Bremner, Michael
Bach, Eric Brettell, Nick
Balabonski, Thibaut Briet, Jop
Banerjee, Abhishek Brihaye, Thomas
Barrington, David Broadbent, Anne
Bartoletti, Massimo Brody, Joshua
Basset, Nicolas Bruni, Roberto
Bavarian, Mohammad Brzuska, Christina
Beame, Paul Buchbinder, Niv
Becchetti, Luca Buchin, Kevin
Bei, Xiaohui Buhrman, Harry
Belmonte, Rémy Byrka, Jaroslaw
Ben Avraham, Rinat Böhl, Florian
Ben-Amram, Amir Cai, Yang
Berger, Eli Caltais, Georgiana
Berry, Jonathan Canetti, Ran
Bertrand, Nathalie Canonne, Clément
Berwanger, Dietmar Cao, Yixin
Bhaskar, Umang Carraro, Alberto
Bitansky, Nir Cash, David
Blazy, Olivier Ceccarello, Matteo
Blesa, Maria J. Chakrabarti, Amit
Blömer, Johannes Chakraborty, Supratik
Bodirsky, Manuel Chalermsook, Parinya
Bodlaender, Hans L. Chan, Hubert
Bodlaender, Marijke Chan, Siu On
X Organization

Chan, Timothy Delahaye, Benoit

Chandran, Nishanth Delling, Daniel
Charatonik, Witold Delvenne, Jean-Charles
Chase, Melissa Delzanno, Giorgio
Chatterjee, Krishnendu Denysyuk, Oksana
Chechik, Shiri Dereniowski, Dariusz
Chekuri, Chandra Devanur, Nikhil
Chen, Jing Devroye, Luc
Chen, Xujin Diaz, Josep
Chen, Zhou Dietzfelbinger, Martin
Cheval, Vincent Diks, Krzysztof
Choudhury, Ashish Dima, Catalin
Chow, Sherman S.M. Diochnos, Dimitris
Chrobak, Marek Dobrev, Stefan
Chung, Kai-Min Doerr, Carola
Ciancia, Vincenzo Doyen, Laurent
Cicalese, Ferdinando Driemel, Anne
Clavier, Christophe Duflot, Marie
Clemente, Lorenzo Dumitrescu, Adrian
Codenotti, Paolo Dupuis, Frédéric
Cohen, Edith Durand, Arnaud
Cohen, Sarel Durand-Gasselin, Antoine
Cohn, Henry Durnoga, Konrad
Colcombet, Thomas Dvir, Zeev
Colini Baldeschi, Riccardo Dyer, Martin
Costello, Craig Edmonds, Jeff
Crescenzi, Pierluigi Efremenko, Klim
Cryan, Mary Efthymiou, Charilaos
Cygan, Marek Ehrgott, Matthias
Czerwiński, Wojciech Ehsanfar, Ebrahim
Dalmau, Victor Elbassioni, Khaled
Damaschke, Peter Elberfeld, Michael
Damgård, Ivan Elmasry, Amr
Dang, Thao Elsässer, Robert
Dani, Varsha Emmi, Michael
Dasgupta, Bhaskar Ene, Alina
Datta, Samir Enea, Constantin
David, Alexandre Enqvist, Sebastian
De Bonis, Annalisa Eppstein, David
de Caro, Angelo Epstein, Leah
De Caro, Angelo Erlebach, Thomas
De Liguoro, Ugo Escoffier, Bruno
de Wolf, Ronald Even, Guy
Decker, Normann Fahrenberg, Uli
Degorre, Aldric Fanelli, Angelo
 Organization XI

Farshim, Pooya Gille, Marc

Fefferman, Bill Giunti, Marco
Feige, Uriel Gkatzelis, Vasilis
Fekete, Sándor Glacet, Christian
Fernau, Henning Glen, Amy
Fijalkow, Nathanaël Gmyr, Robert
Filiot, Emmanuel Gogacz, Tomasz
Filmus, Yuval Goldberg, Paul
Fiorini, Samuel Gonzalez Vasco, Maria Isabel
Firmani, Donatella Gopalan, Parikshit
Fisman, Dana Gorbunov, Sergey
Flammini, Michele Gorecki, Pawel
Forbes, Michael A. Gorgunov, Sergey
Forejt, Vojtech Gorla, Daniele
Fortnow, Lance Grandoni, Fabrizio
Fotakis, Dimitris Greco, Gianluigi
Fountoulakis, Nikolaos Green, Oded
Franciosa, Paolo Grenet, Bruno
Frati, Fabrizio Grigorescu, Elena
Frieze, Alan Grigoryev, Dmitry
Fu, Hu Grossi, Roberto
Fu, Zhiguo GualÃ, Luciano
Fábregas, Ignacio Guo, Heng
Gaboardi, Marco Guo, Jiong
Gadducci, Fabio Guo, Siyao
Gaertner, Bernd Guruswami, Venkatesan
Galanis, Andreas Gutwenger, Carsten
Galesi, Nicola Göbel, Andreas
Gambs, Sebastien Haeupler, Bernhard
Garg, Ankit Haghpanah, Nima
Gaspers, Serge Haitner, Iftach
Gastin, Paul Hajiaghayi, Mohammadtaghi
Gavinsky, Dmitry Hansen, Kristoffer Arnsfelt
Gawrychowski, Pawel Hansen, Thomas Dueholm
Geck, Gaetano Hardt, Moritz
Geeraerts, Gilles Harju, Tero
Gelles, Ran Harrow, Aram
Genest, Blaise Harsha, Prahladh
Ghaffari, Mohsen Hatami, Hamed
Giakkoupis, George Haviv, Ishay
Giannakopoulos, Yiannis Hayes, Thomas
Giannopoulou, Archontia Hazay, Carmit
Giaquinta, Emanuele He, Meng
Gierasimczuk, Nina Heam, Pierre-Cyrille
Gilbert, Seth Heggernes, Pinar
XII Organization

Helmi, Maryam Kerber, Michael

Hirschkoff, Daniel Kesselheim, Thomas
Hlout, Loc Khandekar, Rohit
Hoefer, Martin Kiefer, Stefan
Hoffmann, Hella-Franziska King, Valerie
Hofheinz, Dennis Kiraly, Tamas
Hofman, Piotr Klauck, Hartmut
Huang, Chien-Chung Klein, Philip
Huang, Sangxia Klima, Ondrej
Huang, Xiangru Klin, Bartek
Huang, Zhiyi Klivans, Adam
Hunter, Paul Kniesburges, Sebastian
Husfeldt, Thore Kobayashi, Yusuke
Im, Hyeonseung Kobourov, Stephen
Indyk, Piotr Koebler, Johannes
Iovino, Vincenzo Koiran, Pascal
Irani, Sandy Kolay, Sudeshna
Isopi, Marco Kolliopoulos, Stavros
Ito, Takehiro Komjathy, Julia
Jacob, Riko Kontchakov, Roman
Jain, Rahul Kopczyński, Eryk
Jansen, Bart M.P. Kopelowitz, Tsvi
Jao, David Kopparty, Swastik
Jerrum, Mark Kortsarz, Guy
Jeż, Artur Kosowski, Adrian
Jeż, L
Π ukasz Kosub, Sven
Jiang, Minghui Kothari, Nishad
Jiang, Zhansheng Kothari, Pravesh
Joret, Gwenaël Koutis, Ioannis
Joux, Antoine Koutsopoulos, Andreas
Jurdzinski, Tomasz Kovacs, Annamaria
Jørgensen, Allan Grønlund Kratsch, Stefan
Kakimura, Naonori Krcal, Jan
Kantor, Erez Kretinsky, Jan
Kao, Ming-Yang Krishnaswamy, Ravishankar
Kapralov, Michael Krivosija, Amer
Kapur, Deepak Krug, Robert
Kara, Ahmet Krysta, Piotr
Karakostas, George Kucera, Antonin
Karhumäki, Juhani Kulikov, Alexander
Kausch, Jonathan Kulkarni, Janardhan
Kavitha, Telikepalli Kulkarni, Raghav
Kawamura, Akitoshi Kumar, Akash
Kayal, Neeraj Kumar, Amit
Keller, Orgad Kumar, K. Narayan
 Organization XIII

Kuperberg, Denis Lucier, Brendan

Kurz, Denis Löding, Christof
Kyropoulou, Maria M.S., Ramanujan
Labourel, Arnauld Ma, Minghui
Lachish, Oded Magniez, Frederic
Laekhanukit, Bundit Mahdian, Mohammad
Lagniez, Jean Marie Mahmoody, Mohammad
Lanik, Jan Makarychev, Konstantin
Laura, Luigi Makarychev, Yury
Lauria, Massimo Maletti, Andreas
Lauriere, Mathieu Malizia, Enrico
Laursen, Simon Mallmann-Trenn, Frederik
Lauser, Alexander Manea, Florin
Le Gall, Francois Maneva, Elitza
Le Scouarnec, Nicolas Mansour, Yishay
Lee, James Mardare, Radu
Lee, Troy Markey, Nicolas
Leonardos, Nikos Markou, Euripides
Lerays, Virginie Martens, Wim
Leroux, Jerome Martin, Barnaby
Levavi, Ariel Martin, Russell
Levin, Asaf Marx, Dániel
Levy, Jean-Jacques Marx, Maarten
Lewenstein, Moshe Masopust, Tomas
Li, Jian Mathieson, Luke
Li, Minming Matulef, Kevin
Li, Shi May, Alexander
Li, Yi Mayr, Richard
Li, Yingkai McColl, Robert
Libert, Benoit McGregor, Andrew
Libkin, Leonid McSherry, Frank
Lime, Didier Megow, Nicole
Lin, Anthony Widjaja Meier, Arne
Lin, Chengyu Meiklejohn, Sarah
Liu, Feng-Hao Meir, O.
Llana, Luis Mendel, Manor
Lodaya, Kamal Meng, Xianmeng
Lohrey, Markus Mens, Irini-Eleftheria
Lopez-Ortiz, Alejandro Mertzios, George
Loreti, Michele Meunier, Pierre-Etienne
Lotker, Zvi Miao, Peihan
Lovett, Shachar Michail, Dimitrios
Lozin, Vadim Michalak, Tomasz
Lu, Pinyan Mignot, Ludovic
Lu, Steve Milanic, Martin
XIV Organization

Milchtaich, Igal Niwinski, Damian

Miltersen, Peter Bro Nordstrom, Jakob
Misra, Pranabendu Novotný, Petr
Molinero, Xavier Nowotka, Dirk
Monemizadeh, Morteza Nutov, Zeev
Monmege, Benjamin Nuñez Chiroque, Luis
Montanaro, Ashley O’Donnell, Ryan
Montecchiani, Fabrizio O’Neill, Adam
Montenegro, Ravi Obdrzalek, Jan
Moore, Cristopher Ogierman, Adrian
Moran, Tal Olesen, Mads C.
Morere, Philippe Oliveira, Igor
Morris, Ben Onak, Krzysztof
Morsy, Ehab Ong, Luke
Moseley, Benjamin Ortega-Mallén, Yolanda
Movahedi, Mahnush Ortmann, Mark
Mucha, Marcin Ossona De Mendez, Patrice
Munagala, Kamesh Oualhadj, Youssouf
Munteanu, Alexander Paes Leme, Renato
Murawski, Andrzej Pagh, Rasmus
Murlak, Filip Palomino, Miguel
Muscholl, Anca Paluch, Katarzyna
Mvprao Pan, Jiangwei
Nagaj, Daniel Pan, Jiaxin
Nanongkai, Danupon Panagiotou, Konstantinos
Narayan Kumar, K. Panangaden, Prakash
Narodytska, Nina Pandey, Omkant
Natale, Emanuele Panigrahi, Debmalya
Nathan, Lemons Papadopoulos, Dimitrios
Navara, Mirko Papakonstantinou, Periklis
Navarra, Alfredo Paparas, Dimitris
Nederlof, Jesper Parys, Pawel
Neiman, Ofer Pasquale, Francesco
Nekrich, Yakov Pastro, Valerio
Newman, Alantha Patt-Shamir, Boaz
Nguyen, Hung Son Paulusma, Daniel
Nguyen, Huy Pauly, Arno
Nguyen, Trung Thanh Pavan, A.
Niehren, Joachim Peikert, Christopher
Nielsen, Jesper Sindahl Peng, Pan
Niewerth, Matthias Peressotti, Marco
Nikishkin, Vladimir Peretz, Ron
Nikoletseas, Sotiris Perez, Guillermo
Nikolov, Aleksandar Perifel, Sylvain
Nissim, Kobbi Perrin, Dominique
 Organization XV

Peserico, Enoch Roland, Jérémie

Pettie, Seth Romano, Paolo
Peña, Ricardo Ron, Dana
Picaronny, Claudine Rosa-Velardo, Fernando
Pieris, Andreas RosoΠlek, Robert
Pighizzini, Giovanni Rossi, Gianluca
Pilipczuk, Marcin Rossmanith, Peter
Pilipczuk, Michal Rosulek, Michael
Pin, Jean-Eric Rothvoss, Thomas
Plandowski, Wojciech Rubin, Natan
Polychroniadou, Antigoni Rubio, Fernando
Pottier, Franois Ruppert, Eric
Pottonen, Olli Saad, George
Pous, Damien Sablik, Mathieu
Pozzato, Gian Luca Sack, Joshua
Prabhakar, Pavithra Sadrzadeh, Mehrnoosh
Praveen, M. Saha, Chandan
Price, Eric Salvati, Sylvain
Pruhs, Kirk Sammartino, Matteo
Pucci, Geppino Sangnier, Arnaud
Pulina, Luca Sankur, Ocan
Pérez, Jorge A. Santaroni, Federico
Qiang, Ruixin Santhanam, Rahul
Qiao, Youming Santocanale, Luigi
Quyen, Vuong Anh Santos, Nuno
Rabani, Yuval Saptharishi, Ramprasad
Rabie, Mikael Sarkar, Susmit
Raecke, Harald Satti, Srinivasa Rao
Raghavendra, Prasad Sau, Ignasi
Raghunathan, Ananth Sauerwald, Thomas
Raghvendra, Sharathkumar Saurabh, Saket
Rahaman, Anisur Sawada, Joe
Rampersad, Narad Saxena, Nitin
Raskin, Jean-François Scarpa, Giannicola
Raz, Ran Scheder, Dominik
Regev, Oded Schmidt, Melanie
Rehak, Vojtech Schmidt-Schauss, Manfred
Reynier, Pierre-Alain Schmitz, Sylvain
Riba, Colin Schneider, Stefan
Richerby, David Schroder, Dominique
Riondato, Matteo Schröder, Lutz
Robinson, Peter Schuster, Martin
Roditty, Liam Schwartz, Roy
Rodriguez, Ismael Schweikardt, Nicole
Roetteler, Martin Schwiegelshohn, Chris
XVI Organization

Schwoon, Stefan Struth, Georg

Servais, Frédéric Su, Le
Servedio, Rocco Suchy, Ondrej
Seshadhri, C. Sun, Xiaoming
Setzer, Alexander Sun, Xiaorui
Shah, Rahul Suomela, Jukka
Shah, Simoni Suresh, S.P.
Shamir, Ohad Syrgkanis, Vasilis
Sharma, Vikram Sénizergues, Géraud
Shen, Alexandre Ta-Shma, Amnon
Shenoy R., Gautham Tamaki, Suguru
Shpilka, Amir Tamir, Tami
Shraibman, Adi Tan, Li-Yang
Sidiropoulos, Anastasios Tang, Bo
Siebertz, Sebastian Tao, Yufei
Sikdar, Somnath Tarjan, Robert
Silva, Alexandra Tavenas, Sébastien
Silvestri, Riccardo Telle, Jan Arne
Singh, Mohit Terhal, Barbara
Sitchinava, Nodari Terui, Kazushige
Sitters, Rene Terzi, Evimaria
Skowron, Piotr Thaler, Justin
Sokolova, Ana Thanh, Nguyen
Solomon, Shay Thapper, Johan
Sommer, Christian Thiagarajan, P.S.
Sousi, Perla Thilikos, Dimitrios
Spoerhase, Joachim Thorup, Mikkel
Sramek, Rastislav Thraves, Christopher
Srinivasan, Srikanth Toledo, Sivan
Srivastava, Piyush Toledoii, Sivan
Srivathsan, B. Tompits, Hans
Stachowiak, Grzegorz Torres Vieira, Hugo
Staiger, Ludwig Torunczyk, Szymon
Stainer, Julien Toruńczyk, Szymon
Starikovskaya, Tatiana Trevisan, Luca
Stefankovic, Daniel Trivedi, Ashutosh
Stehle, Damien Tschudi, Daniel
Stephan, Frank Tulsiani, Madhur
Stergiou, Christos Uehara, Ryuhei
Stoddard, Greg Ulus, Dogan
Strassburger, Lutz Umans, Chris
Straubing, Howard Umboh, Seeun
Strefler, Mario Uno, Yushi
Strejcek, Jan Upadhyay, Jalaj
Strothmann, Thim Valiant, Gregory
 Organization XVII

Valiente, Gabriel Wong, Prudence W.H.

Valiron, Benoı̂t Woodruff, David
van Breugel, Franck Wootters, Mary
van Melkebeek, Dieter Wright, John
Van Melkebeek, Dieter Wrochna, Marcin
van Stee, Rob Wu, Xiaodi
Varacca, Daniele Wulff-Nilsen, Christian
Vassilevska Williams, Virginia Wullschleger, Juerg
Vegh, Laszlo Xia, Ge
Velickovic, Boban Xiao, Tao
Venkitasubramaniam, Xie, Ning
Muthuramakrishnan Xing, Chaoping
Ventre, Carmine Xu, Xiaoming
Verschae, Jose Xue, Guoliang
Vidick, Thomas Yamada, Shota
Viet Tung, Hoang Yamakami, Tomoyuki
Viglietta, Giovanni Yamauchi, Yukiko
Vijayaraghavan, Aravindan Yang, Kaiyu
Vilaça, Xavier Yao, Penghui
Visconti, Ivan Yaroslavtsev, Grigory
Viswanathan, Mahesh Ye, Tao
Vogler, Walter Yekhanin, Sergey
Volkovich, Ilya Yi, Ke
Vrgoc, Domagoj Yiannakopoulos, Yiannis
Wachter-Zeh, Antonia Yin, Yitong
Wahlström, Magnus Yoshida, Yuichi
Walter, Tobias Young, Max
Walukiewicz, Igor Yu, Huacheng
Wang, Juntao Yuen, Tsz Hon
Wang, Kainan Zacharias, Thomas
Wanka, Rolf Zamani, Mahdi
Watson, Thomas Zang, Wenan
Wee, Hoeteck Zeh, Norbert
Weinstein, Omri Zhang, Bingsheng
Weiss, Armin Zhang, Chihao
Westermann, Matthias Zhang, Hongyang
Whistler, William Zhang, Jialin
Wieder, Udi Zhang, Jie
Wiese, Andreas Zhang, Jin
Wilkinson, Bryan T. Zhang, Shengyu
Wilson, David Zhang, Wuzhou
Winslow, Andrew Zhang, Yong
Witek, Maximilian Zhao, Zhiguang
Witkowski, Piotr Zhou, Hong-Sheng
Wollan, Paul Zhou, Yuan
XVIII Organization

Zhu, Zeyuan Allen Zivny, Stanislav

Ziegler, Martin Zuckerman, David
Zimand, Marius Zwick, Uri
Ziv-Ukelson, Michal Zych, Anna
 Table of Contents – Part II

Track B: Logic, Semantics, Automata, and Theory

of Programming
Symmetric Groups and Quotient Complexity of Boolean Operations . . . . 1
Jason Bell, Janusz Brzozowski, Nelma Moreira, and Rogério Reis

Handling Infinitely Branching WSTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Michael Blondin, Alain Finkel, and Pierre McKenzie

Transducers with Origin Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Mikolaj Bojańczyk

Weak MSO+U with Path Quantifiers over Infinite Trees . . . . . . . . . . . . . . 38

Mikolaj Bojańczyk

On the Decidability of MSO+U on Infinite Trees . . . . . . . . . . . . . . . . . . . . . 50

Mikolaj Bojańczyk, Tomasz Gogacz, Henryk Michalewski, and
Michal Skrzypczak

A Coalgebraic Foundation for Coinductive Union Types . . . . . . . . . . . . . . 62

Marcello Bonsangue, Jurriaan Rot, Davide Ancona,
Frank de Boer, and Jan Rutten

Turing Degrees of Limit Sets of Cellular Automata . . . . . . . . . . . . . . . . . . . 74

Alex Borello, Julien Cervelle, and Pascal Vanier

On the Complexity of Temporal-Logic Path Checking . . . . . . . . . . . . . . . . . 86

Daniel Bundala and Joël Ouaknine

Parameterised Linearisability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Andrea Cerone, Alexey Gotsman, and Hongseok Yang

Games with a Weak Adversary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Krishnendu Chatterjee and Laurent Doyen

The Complexity of Ergodic Mean-payoff Games . . . . . . . . . . . . . . . . . . . . . . 122

Krishnendu Chatterjee and Rasmus Ibsen-Jensen

Toward a Structure Theory of Regular Infinitary Trace Languages . . . . . . 134

Namit Chaturvedi

Unary Pushdown Automata and Straight-Line Programs . . . . . . . . . . . . . . 146

Dmitry Chistikov and Rupak Majumdar
XX Table of Contents – Part II

Robustness against Power is PSpace-complete . . . . . . . . . . . . . . . . . . . . . . . 158

Egor Derevenetc and Roland Meyer
A Nivat Theorem for Weighted Timed Automata and Weighted
Relative Distance Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Manfred Droste and Vitaly Perevoshchikov
Computability in Anonymous Networks: Revocable vs. Irrecovable
Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Yuval Emek, Jochen Seidel, and Roger Wattenhofer
Coalgebraic Weak Bisimulation from Recursive Equations
over Monads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Sergey Goncharov and Dirk Pattinson
Piecewise Boolean Algebras and Their Domains . . . . . . . . . . . . . . . . . . . . . . 208
Chris Heunen
Between Linearizability and Quiescent Consistency: Quantitative
Quiescent Consistency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Radha Jagadeesan and James Riely
Bisimulation Equivalence of First-Order Grammars . . . . . . . . . . . . . . . . . . . 232
Petr Jančar
Context Unification is in PSPACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
Artur Jeż
Monodic Fragments of Probabilistic First-Order Logic . . . . . . . . . . . . . . . . 256
Jean Christoph Jung, Carsten Lutz, Sergey Goncharov, and
Lutz Schröder
Stability and Complexity of Minimising Probabilistic Automata . . . . . . . . 268
Stefan Kiefer and Björn Wachter
Kleene Algebra with Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
Dexter Kozen and Konstantinos Mamouras
All–Instances Termination of Chase is Undecidable . . . . . . . . . . . . . . . . . . . 293
Tomasz Gogacz and Jerzy Marcinkowski
Non-uniform Polytime Computation in the Infinitary Affine
Lambda-Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Damiano Mazza
On the Positivity Problem for Simple Linear Recurrence Sequences . . . . . 318
Joël Ouaknine and James Worrell
Ultimate Positivity is Decidable for Simple Linear Recurrence
Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Joël Ouaknine and James Worrell
 Table of Contents – Part II XXI

Going Higher in the First-Order Quantifier Alternation Hierarchy on

Words . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
Thomas Place and Marc Zeitoun

Hardness Results for Intersection Non-Emptiness . . . . . . . . . . . . . . . . . . . . . 354

Michael Wehar

Branching Bisimilarity Checking for PRS . . . . . . . . . . . . . . . . . . . . . . . . . . . 363

Qiang Yin, Yuxi Fu, Chaodong He, Mingzhang Huang, and
Xiuting Tao

Track C: Foundations of Networked Computing

Labeling Schemes for Bounded Degree Graphs . . . . . . . . . . . . . . . . . . . . . . . 375
David Adjiashvili and Noy Rotbart

Bounded-Angle Spanning Tree: Modeling Networks with Angular

Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Rom Aschner and Matthew J. Katz

Distributed Computing on Core-Periphery Networks: Axiom-Based Design 399

Chen Avin, Michael Borokhovich, Zvi Lotker, and David Peleg

Fault-Tolerant Rendezvous in Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411

Jérémie Chalopin, Yoann Dieudonné, Arnaud Labourel, and
Andrzej Pelc

Data Delivery by Energy-Constrained Mobile Agents on a Line . . . . . . . . 423

Jérémie Chalopin, Riko Jacob, Matúš Mihalák, and Peter Widmayer

The Power of Two Choices in Distributed Voting . . . . . . . . . . . . . . . . . . . . . 435

Colin Cooper, Robert Elsässer, and Tomasz Radzik

Jamming-Resistant Learning in Wireless Networks . . . . . . . . . . . . . . . . . . . 447

Johannes Dams, Martin Hoefer, and Thomas Kesselheim

Facility Location in Evolving Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459

David Eisenstat, Claire Mathieu, and Nicolas Schabanel

Solving the ANTS Problem with Asynchronous Finite State

Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
Yuval Emek, Tobias Langner, Jara Uitto, and Roger Wattenhofer

Near-Optimal Distributed Approximation of Minimum-Weight

Connected Dominating Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Mohsen Ghaffari

Randomized Rumor Spreading in Dynamic Graphs . . . . . . . . . . . . . . . . . . . 495

George Giakkoupis, Thomas Sauerwald, and Alexandre Stauffer
XXII Table of Contents – Part II

Online Independent Set Beyond the Worst-Case: Secretaries, Prophets,

and Periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 508
Oliver Göbel, Martin Hoefer, Thomas Kesselheim,
Thomas Schleiden, and Berthold Vöcking

Optimal Competitiveness for Symmetric Rectilinear Steiner

Arborescence and Related Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
Erez Kantor and Shay Kutten

Orienting Fully Dynamic Graphs with Worst-Case Time Bounds . . . . . . . 532

Tsvi Kopelowitz, Robert Krauthgamer, Ely Porat, and Shay Solomon

Does Adding More Agents Make a Difference? A Case Study of Cover

Time for the Rotor-Router . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544
Adrian Kosowski and Dominik Pajak Σ

The Melbourne Shuffle: Improving Oblivious Storage in the Cloud . . . . . . 556

Olga Ohrimenko, Michael T. Goodrich, Roberto Tamassia, and
Eli Upfal

Sending Secrets Swiftly: Approximation Algorithms for Generalized

Multicast Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
Afshin Nikzad and R. Ravi

Bypassing Erdős’ Girth Conjecture: Hybrid Stretch and Sourcewise

Spanners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 608
Merav Parter

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621

 Table of Contents – Part I

Invited Talks
Sporadic Solutions to Zero-One Exclusion Tasks . . . . . . . . . . . . . . . . . . . . . 1
Eli Gafni and Maurice Herlihy
Verifying and Synthesizing Software with Recursive Functions
(Invited Contribution) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Viktor Kuncak

Track A: Algorithms, Complexity, and Games

Weak Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Scott Aaronson, Andris Ambainis, Kaspars Balodis, and
Mohammad Bavarian
Consequences of Faster Alignment of Sequences . . . . . . . . . . . . . . . . . . . . . . 39
Amir Abboud, Virginia Vassilevska Williams, and Oren Weimann
Distance Labels with Optimal Local Stretch . . . . . . . . . . . . . . . . . . . . . . . . . 52
Ittai Abraham and Shiri Chechik
Time-Expanded Packings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
David Adjiashvili, Sandro Bosio, Robert Weismantel, and
Rico Zenklusen
Deterministic Rectangle Enclosure and Offline Dominance Reporting
on the RAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Peyman Afshani, Timothy M. Chan, and Konstantinos Tsakalidis
The Tropical Shadow-Vertex Algorithm Solves Mean Payoff Games in
Polynomial Time on Average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Xavier Allamigeon, Pascal Benchimol, and Stéphane Gaubert
Tighter Relations between Sensitivity and Other Complexity
Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Andris Ambainis, Mohammad Bavarian, Yihan Gao, Jieming Mao,
Xiaoming Sun, and Song Zuo
On Hardness of Jumbled Indexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Amihood Amir, Timothy M. Chan, Moshe Lewenstein, and
Noa Lewenstein
Morphing Planar Graph Drawings Optimally . . . . . . . . . . . . . . . . . . . . . . . . 126
Patrizio Angelini, Giordano Da Lozzo, Giuseppe Di Battista,
Fabrizio Frati, Maurizio Patrignani, and Vincenzo Roselli
XXIV Table of Contents – Part I

Incremental Algorithm for Maintaining DFS Tree for Undirected

Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Surender Baswana and Shahbaz Khan

On the Role of Shared Randomness in Simultaneous Communication . . . 150

Mohammad Bavarian, Dmitry Gavinsky, and Tsuyoshi Ito

Short PCPs with Projection Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

Eli Ben-Sasson and Emanuele Viola

Star Partitions of Perfect Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

René van Bevern, Robert Bredereck, Laurent Bulteau, Jiehua Chen,
Vincent Froese, Rolf Niedermeier, and Gerhard J. Woeginger

Coordination Mechanisms for Selfish Routing over Time on a Tree . . . . . 186

Sayan Bhattacharya, Janardhan Kulkarni, and Vahab Mirrokni

On Area-Optimal Planar Graph Drawings . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

Therese Biedl

Shortest Two Disjoint Paths in Polynomial Time . . . . . . . . . . . . . . . . . . . . . 211

Andreas Björklund and Thore Husfeldt

Listing Triangles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Andreas Björklund, Rasmus Pagh,
Virginia Vassilevska Williams, and Uri Zwick

On DNF Approximators for Monotone Boolean Functions . . . . . . . . . . . . . 235

Eric Blais, Johan Håstad, Rocco A. Servedio, and Li-Yang Tan

Internal DLA: Efficient Simulation of a Physical Growth Model

(Extended Abstract) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Karl Bringmann, Fabian Kuhn, Konstantinos Panagiotou,
Ueli Peter, and Henning Thomas

Lower Bounds for Approximate LDCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

Jop Briët, Zeev Dvir, Guangda Hu, and Shubhangi Saraf

Holographic Algorithms Beyond Matchgates . . . . . . . . . . . . . . . . . . . . . . . . . 271

Jin-Yi Cai, Heng Guo, and Tyson Williams

Testing Probability Distributions Underlying Aggregated Data . . . . . . . . . 283

Clément Canonne and Ronitt Rubinfeld

Parallel Repetition of Entangled Games with Exponential Decay via

the Superposed Information Cost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
André Chailloux and Giannicola Scarpa

The Bose-Hubbard Model is QMA-complete . . . . . . . . . . . . . . . . . . . . . . . . . 308

Andrew M. Childs, David Gosset, and Zak Webb
 Table of Contents – Part I XXV

Characterization of Binary Constraint System Games . . . . . . . . . . . . . . . . . 320

Richard Cleve and Rajat Mittal
Fast Algorithms for Constructing Maximum Entropy Summary Trees . . . 332
Richard Cole and Howard Karloff
Thorp Shuffling, Butterflies, and Non-markovian Couplings . . . . . . . . . . . . 344
Artur Czumaj and Berthold Vöcking
Dynamic Complexity of Directed Reachability and Other Problems . . . . . 356
Samir Datta, William Hesse, and Raghav Kulkarni
One Tile to Rule Them All: Simulating Any Tile Assembly System
with a Single Universal Tile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
Erik D. Demaine, Martin L. Demaine, Sándor P. Fekete,
Matthew J. Patitz, Robert T. Schweller, Andrew Winslow, and
Damien Woods
Canadians Should Travel Randomly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
Erik D. Demaine, Yamming Huang, Chung-Shou Liao, and
Kunihiko Sadakane
Efficiency Guarantees in Auctions with Budgets . . . . . . . . . . . . . . . . . . . . . . 392
Shahar Dobzinski and Renato Paes Leme
Parameterized Complexity of Bandwidth on Trees . . . . . . . . . . . . . . . . . . . . 405
Markus Sortland Dregi and Daniel Lokshtanov
Testing Equivalence of Polynomials under Shifts . . . . . . . . . . . . . . . . . . . . . 417
Zeev Dvir, Rafael Mendes de Oliveira, and Amir Shpilka
Optimal Analysis of Best Fit Bin Packing . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
György Dósa and Jiřı́ Sgall
Light Spanners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Michael Elkin, Ofer Neiman, and Shay Solomon
Semi-Streaming Set Cover (Extended Abstract) . . . . . . . . . . . . . . . . . . . . . . 453
Yuval Emek and Adi Rosén
Online Stochastic Reordering Buffer Scheduling . . . . . . . . . . . . . . . . . . . . . . 465
Hossein Esfandiari, MohammadTaghi Hajiaghayi,
Mohammad Reza Khani, Vahid Liaghat,
Hamid Mahini, and Harald Räcke
Demand Queries with Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 477
Uriel Feige and Shlomo Jozeph
Algorithmic Aspects of Regular Graph Covers with Applications to
Planar Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
Jiřı́ Fiala, Pavel Klavı́k, Jan Kratochvı́l, and Roman Nedela
XXVI Table of Contents – Part I

Public vs Private Coin in Bounded-Round Information . . . . . . . . . . . . . . . 502

Mark Braverman and Ankit Garg

En Route to the Log-Rank Conjecture: New Reductions and Equivalent

Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514
Dmitry Gavinsky and Shachar Lovett

Improved Submatrix Maximum Queries in Monge Matrices . . . . . . . . . . . . 525

Pawel Gawrychowski, Shay Mozes, and Oren Weimann

For-All Sparse Recovery in Near-Optimal Time . . . . . . . . . . . . . . . . . . . . . . 538

Anna C. Gilbert, Yi Li, Ely Porat, and Martin J. Strauss

Families with Infants: A General Approach to Solve Hard Partition

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
Alexander Golovnev, Alexander S. Kulikov, and Ivan Mihajlin

Changing Bases: Multistage Optimization for Matroids and

Matchings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563
Anupam Gupta, Kunal Talwar, and Udi Wieder

Near-Optimal Online Algorithms for Prize-Collecting Steiner

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576
MohammadTaghi Hajiaghayi, Vahid Liaghat, and
Debmalya Panigrahi

Nearly Linear-Time Model-Based Compressive Sensing . . . . . . . . . . . . . . . 588

Chinmay Hegde, Piotr Indyk, and Ludwig Schmidt

Breaking the PPSZ Barrier for Unique 3-SAT . . . . . . . . . . . . . . . . . . . . . . . . 600

Timon Hertli

Privately Solving Linear Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612

Justin Hsu, Aaron Roth, Tim Roughgarden, and Jonathan Ullman

How Unsplittable-Flow-Covering Helps Scheduling with Job-Dependent

Cost Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625
Wiebke Höhn, Julián Mestre, and Andreas Wiese

Why Some Heaps Support Constant-Amortized-Time Decrease-Key

Operations, and Others Do Not . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
John Iacono and Özgür Özkan

Partial Garbling Schemes and Their Applications . . . . . . . . . . . . . . . . . . . . 650

Yuval Ishai and Hoeteck Wee

On the Complexity of Trial and Error for Constraint Satisfaction

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663
Gábor Ivanyos, Raghav Kulkarni, Youming Qiao,
Miklos Santha, and Aarthi Sundaram
 Table of Contents – Part I XXVII

Information Theoretical Cryptogenography . . . . . . . . . . . . . . . . . . . . . . . . . . 676

Sune K. Jakobsen

The Complexity of Somewhat Approximation Resistant Predicates . . . . . 689

Subhash Khot, Madhur Tulsiani, and Pratik Worah

Approximate Nonnegative Rank Is Equivalent to the Smooth Rectangle

Bound . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 701
Gillat Kol, Shay Moran, Amir Shpilka, and Amir Yehudayoff

Distance Oracles for Time-Dependent Networks . . . . . . . . . . . . . . . . . . . . . . 713

Spyros Kontogiannis and Christos Zaroliagis

Efficient Indexing of Necklaces and Irreducible Polynomials over Finite

Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 726
Swastik Kopparty, Mrinal Kumar, and Michael Saks

Coloring Relatives of Interval Overlap Graphs via On-line Games . . . . . . 738

Tomasz Krawczyk and Bartosz Walczak

Superpolynomial Lower Bounds for General Homogeneous Depth 4

Arithmetic Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 751
Mrinal Kumar and Shubhangi Saraf

Testing Forest-Isomorphism in the Adjacency List Model . . . . . . . . . . . . . . 763

Mitsuru Kusumoto and Yuichi Yoshida

Parameterized Approximation Schemes Using Graph Widths . . . . . . . . . . 775

Michael Lampis

FPTAS for Weighted Fibonacci Gates and Its Applications . . . . . . . . . . . . 787

Pinyan Lu, Menghui Wang, and Chihao Zhang

Parameterized Algorithms to Preserve Connectivity . . . . . . . . . . . . . . . . . . 800

Manu Basavaraju, Fedor V. Fomin, Petr Golovach,
Pranabendu Misra, M.S. Ramanujan, and Saket Saurabh

Nonuniform Graph Partitioning with Unrelated Weights . . . . . . . . . . . . . . 812

Konstantin Makarychev and Yury Makarychev

Precedence-Constrained Scheduling of Malleable Jobs with

Preemption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823
Konstantin Makarychev and Debmalya Panigrahi

Unbounded Entanglement Can Be Needed to Achieve the Optimal

Success Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 835
Laura Mančinska and Thomas Vidick

QCSP on Semicomplete Digraphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847

Petar Dapić, Petar Marković, and Barnaby Martin
XXVIII Table of Contents – Part I

Fast Pseudorandomness for Independence and Load Balancing

[Extended Abstract] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 859
Raghu Meka, Omer Reingold, Guy N. Rothblum, and
Ron D. Rothblum
Determining Majority in Networks with Local Interactions and Very
Small Local Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 871
George B. Mertzios, Sotiris E. Nikoletseas,
Christoforos L. Raptopoulos, and Paul G. Spirakis
Lower Bounds for Oblivious Subspace Embeddings . . . . . . . . . . . . . . . . . . . 883
˜
Jelani Nelson and Huy L. Nguyên
On Input Indistinguishable Proof Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 895
Rafail Ostrovsky, Giuseppe Persiano, and Ivan Visconti
Secure Computation Using Leaky Tokens . . . . . . . . . . . . . . . . . . . . . . . . . . . 907
Manoj Prabhakaran, Amit Sahai, and Akshay Wadia
An Improved Interactive Streaming Algorithm for the Distinct
Elements Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
Hartmut Klauck and Ved Prakash
A Faster Parameterized Algorithm for Treedepth . . . . . . . . . . . . . . . . . . . . . 931
Felix Reidl, Peter Rossmanith, Fernando Sánchez Villaamil, and
Somnath Sikdar
Pseudorandom Graphs in Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . 943
Omer Reingold, Ron D. Rothblum, and Udi Wieder
Sampling-Based Proofs of Almost-Periodicity Results and Algorithmic
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 955
Eli Ben-Sasson, Noga Ron-Zewi, Madhur Tulsiani, and Julia Wolf
The Mondshein Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 967
Jens M. Schmidt
Balanced Allocations: A Simple Proof for the Heavily Loaded Case . . . . . 979
Kunal Talwar and Udi Wieder
Close to Uniform Prime Number Generation with Fewer Random
Bits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 991
Pierre-Alain Fouque and Mehdi Tibouchi
Optimal Strong Parallel Repetition for Projection Games on Low
Threshold Rank Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1003
Madhur Tulsiani, John Wright, and Yuan Zhou
Sparser Random 3-SAT Refutation Algorithms and the Interpolation
Problem (Extended Abstract) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1015
Iddo Tzameret
 Table of Contents – Part I XXIX

On Learning, Lower Bounds and (un)Keeping Promises . . . . . . . . . . . . . . . 1027

Ilya Volkovich

Certificates in Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1039

Yaoyu Wang and Yitong Yin

Optimal Query Complexity for Estimating the Trace of a Matrix . . . . . . . 1051

Karl Wimmer, Yi Wu, and Peng Zhang

Faster Separators for Shallow Minor-Free Graphs via Dynamic

Approximate Distance Oracles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1063
Christian Wulff-Nilsen

Spatial Mixing of Coloring Random Graphs . . . . . . . . . . . . . . . . . . . . . . . . . 1075

Yitong Yin

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087

 Symmetric Groups and Quotient Complexity
of Boolean Operations

Jason Bell1 , Janusz Brzozowski2, Nelma Moreira3 , and Rogério Reis3

1
Department of Pure Mathematics, University of Waterloo,
Waterloo, ON, Canada N2L 3G1
[email protected]
2
David R. Cheriton School of Computer Science, University of Waterloo,
Waterloo, ON, Canada N2L 3G1
[email protected]
3
CMUP & DCC, Faculdade de Ciências da Universidade do Porto,
Rua do Campo Alegre, 4169–007 Porto, Portugal
{nam,rvr}@dcc.fc.up.pt

Abstract. The quotient complexity of a regular language L is the num-

ber of left quotients of L, which is the same as the state complexity of
L. Suppose that L and L are binary regular languages with quotient
complexities m and n, and that the subgroups of permutations in the
transition semigroups of the minimal deterministic automata accepting
L and L are the symmetric groups Sm and Sn of degrees m and n,
respectively. Denote by ◦ any binary boolean operation that is not a
constant and not a function of one argument only. For m, n ≥ 2 with
(m, n) ∨∈ {(2, 2), (3, 4), (4, 3), (4, 4)} we prove that the quotient complex-
ity of L ◦ L is mn if and only either (a) m ∨= n or (b) m = n and
the bases (ordered pairs of generators) of Sm and Sn are not conjugate.
For (m, n) ∈ {(2, 2), (3, 4), (4, 3), (4, 4)} we give examples to show that
this need not hold. In proving these results we generalize the notion of
uniform minimality to direct products of automata. We also establish
a non-trivial connection between complexity of boolean operations and
group theory.

Keywords: Boolean operation, quotient complexity, regular language,

state complexity, symmetric group, transition semigroup.

1 Motivation

The left quotient, or simply quotient, of a regular language L over an alphabet

Σ by a word w ∈ Σ ∗ is the regular language w−1 L = {x ∈ Σ ∗ : wx ∈ L}. It
is well known that a language is regular if and only if it has a finite number
of quotients. Consequently, the number of quotients of a regular language, its
quotient complexity [1], is a natural measure of complexity of the language. Quo-
tient complexity is also known as state complexity [15], which is the number of
ω
For a complete version of this work see http://arxiv.org/abs/1310.1841.

J. Esparza et al. (Eds.): ICALP 2014, Part II, LNCS 8573, pp. 1–12, 2014.
c Springer-Verlag Berlin Heidelberg 2014
2 J. Bell et al.

states in the complete minimal deterministic finite automaton (DFA) recognizing

the language. We prefer quotient complexity because it is a language-theoretic
concept, and we refer to it simply as complexity.
The problem of determining the complexity of an operation [1,8,15,16] on
regular languages has received much attention. It is defined as the maximal
complexity of the language resulting from the operation, taken as a function
of the complexities of the operands. When operations are performed on large
automata it is important to have some information about the size of the result
and the time it will take to compute it. The quotient complexity of an operation
gives an upper bound on its time and space complexity [15].
Languages that meet the upper bound on the complexity of an operation are
witnesses for this operation. Although witnesses for common operations on reg-
ular languages are well known, there are occasions when one has to look for new
witnesses:
1. One may be interested in a class of languages that have the same com-
plexity with respect to a given operation. For example, let Σ = {a, b} and
let |w|a be the number of times the letter a appears in the word w ∈ Σ ∗ .
Then the intersection of the languages L = {w ∈ Σ ∗ : |w|a ≡ m − 1 mod m}
and L≥ = {w ∈ Σ ∗ : |w|b ≡ n − 1 mod n} has complexity mn. The languages
K = (b∗ a)m−1 Σ ∗ and K ≥ = (a∗ b)n−1 Σ ∗ also meet this bound; hence (L, L≥ ) and
(K, K ≥ ) are in the same complexity class with respect to intersection.
2. Whenever one studies complexity within a proper subclass of regular lan-
guages, one usually needs to find new witnesses. For example, in the class of
regular right ideals – languages L ⊆ Σ ∗ satisfying L = LΣ ∗ – languages K and
K ≥ are appropriate, but L and L≥ are not. The main result of the present paper
has been applied to right ideals in [4], where the proof that the witnesses used
there meet the bounds for boolean operations was greatly simplified with the
aid of our theorem.
3. When one studies combined operations – operations that involve more than
one basic operation, for example, the intersection of reversed languages – one
again need new witnesses [7].
Before stating our result, we provide some additional background information.
The syntactic congruence ↔L of L is defined as follows: For all x, y ∈ Σ ∗ ,
x ↔L y if and only if uxv ∈ L ⇔ uyv ∈ L for all u, v ∈ Σ ∗ . The set Σ + / ↔L
of equivalence classes of the relation ↔L is a semigroup with concatenation as
the operation; it is called the syntactic semigroup of L, which we denote by SL .
It is well known that the syntactic semigroup is isomorphic to the semigroup
SD of transformations performed by non-empty words on the set of states in
the minimal DFA D recognizing L; this semigroup is known as the transition
semigroup of D. If D has n states, the cardinality of the transition semigroup is
bounded from above by nn , and this bound is reachable.
The atoms [5,6] of a regular language are non-empty intersections of all left
quotients of the language, some or all of which may be complemented. A regular
language has at most 2n atoms, and their quotient complexities are known [5].
 Symmetric Groups and Quotient Complexity of Boolean Operations 3

The reverse of a word is defined inductively: the reverse of the empty word
ε is εR = ε, and the reverse of wa with w ∈ Σ ∗ and a ∈ Σ is (wa)R = awR .
The reverse of a language L is LR = {wR : w ∈ L}. For L with complexity n the
maximal complexity of LR is 2n , and this bound is reachable.
Whenever new witnesses are used, it is necessary to prove that these witnesses
meet the required bound. It would be very useful to have results stating that if
the languages in question have some property P , then they meet the upper bound
for a given operation. Some results of this type are now briefly discussed.
Let MSC denote the class of languages with maximal syntactic complexity
(languages with largest syntactic semigroups), let STT denote the class of lan-
guages whose minimal DFAs have set-transitive transition semigroups (for any
two sets of states of the same cardinality there is a transformation that maps
one set to the other), let MAL denote the class of maximally atomic languages
(languages that have all 2n atoms, all of which have maximal possible quotient
complexity), let MNA denote the class of languages with the maximal number
(2n ) of atoms, and let MCR denote the class of languages with a maximally
complex reverse (reverse of complexity 2n ). The following relations hold [3]:

MSC ⊂ STT = MAL ⊂ MNA = MCR.

The fact that MSC ⊂ MCR is a result of A. Salomaa, Wood, and Yu [12], and
the observation that MNA = MCR was made by Brzozowski and Tamm [6].
Our main theorem relates the complexity of proper binary boolean operations
on regular languages to the nature of the syntactic semigroups of the languages.
A boolean operation is proper if it is not a constant and not a function of one
variable only.
Let Sn denote the symmetric group of degree n. A basis [9] of Sn is an ordered
pair (s, t) of distinct transformations of Qn = {0, . . . , n − 1} that generate Sn .
Two bases (s, t) and (s≥ , t≥ ) of Sn are conjugate if there exists a transformation
r ∈ Sn such that rsr−1 = s≥ , and rtr−1 = t≥ .
Assume that a DFA D (respectively, D≥ ) has state set Qm (Qn ), and transition
semigroup Sm (Sn ). Let L (L≥ ) be the language accepted by D (D≥ ). Our main
theorem is a generalization of a result of Brzozowski and Liu [2]:
Theorem 1. Let D and D≥ be binary DFAs with m and n states respectively,
where m, n ≥ 2 and (m, n) ⇒∈ {(2, 2), (3, 4), (4, 3), (4, 4)}. If the subgroups of per-
mutations in the transition semigroups of D and D≥ are Sm and Sn respectively,
and ◦ is a proper binary boolean operation, then the complexity of L ◦ L≥ is
mn, unless m = n and the bases of the transition semigroups of D and D≥ are
conjugate, in which case the quotient complexity of L ◦ L≥ is at most m = n.
The proof that the complexity of a binary boolean operation is maximal in-
volves two steps. First, one proves that the direct product of the minimal DFAs
of the languages is connected, meaning that all of its states are reachable from
the initial state. Second, one verifies that any two states in the direct product
are distinguishable by some word, that is, that they are not equivalent. Since
both reachability and distinguishability will be proved using only permutations,
4 J. Bell et al.

it is convenient to ignore other transformations and assume that the transition

semigroups of the DFAs we deal with are symmetric groups.
The remainder of the paper is structured as follows: Section 2 defines our
terminology and notation. Section 3 deals with the conditions under which the
direct product of two automata is connected. Section 4 studies uniformly mini-
mal semiautomata (automata without final states), that is, semiautomata which
become minimal DFAs if one adds an arbitrary set of final states, other than the
empty set and the set of all states. Section 5 contains our main result relating
symmetric groups to the complexity of boolean operations for all except a few
cases. Section 6 concludes the paper.

2 Preliminaries
Groups. Our results rely heavily on the theory of finite groups. We refer the
reader to [11,13], for example, for basic facts about groups.
Transformations. A transformation of a set Q is a mapping of Q into itself.
We deal only with finite non-empty sets and, without loss of generality, assume
that Q = Qn = {0, 1, . . . , n − 1}. If t is a transformation of Qn and i ∈ Qn , then
t(i) is the image of i under t. An arbitrary transformation is written in the form
 
0 1 ... n − 2 n − 1
t= ,
i0 i1 . . . in−2 in−1

where ik = t(k), 0 ≤ k ≤ n−1, and ik ∈ Qn . The composition of two transforma-

tions t1 and t2 of Qn is a transformation t1 ◦ t2 such that (t1 ◦ t2 )(i) = t1 (t2 (i))
for all i ∈ Qn . We usually omit the composition operator and write t1 t2 . The set
of all transformations of Qn is a monoid under composititon with the identity
transformation acting as the unit element 1.
A permutation is a mapping of Qn onto itself. A permutation t is a cycle of
of length k or a k-cycle , where k ≥ 2, if there exist pairwise different elements
i1 , . . . , ik such that t(i1 ) = i2 , t(i2 ) = i3 , . . . , t(ik−1 ) = ik , and t(ik ) = i1 ,
and t does not affect any other elements. A cycle is denoted by (i1 , i2 , . . . , ik ).
A transposition is a 2-cycle. Every permutation is a product (composition) of
transpositions, and the parity of the number of transpositions in the factorization
is an invariant. A permutation is odd (even) if its factorization has an odd
(even) number of factors. The symmetric group Sn of degree n is the set of all
permutations of Qn , with composition as the group operation, and the identity
as 1. The alternating group An is the set of all even permutations of Sn .
Given a subgroup H of Sn , we say that H acts transitively on Qn if for each
i, j ∈ Qn there is some t ∈ H such that t(i) = j. We say that H acts doubly
transitively on Qn if whenever i, j, k, Ω ∈ Qn with i ⇒= j and k ⇒= Ω there is some
t ∈ H such that t(i) = k, t(j) = Ω.
Semiautomata and Automata. A deterministic finite semiautomaton (DFS)
is a quadruple A = (Q, Σ, δ, q0 ), where Q is a finite set of states, Σ is a finite
non-empty alphabet, δ : Q×Σ → Q is the transition function, and q0 is the initial
state. We extend δ to Q × Σ ∗ in the usual way. A state q is reachable from the
 Symmetric Groups and Quotient Complexity of Boolean Operations 5

initial state if there is a word w such that q = δ(q0 , w). A DFS is connected if
every state q ∈ Q is reachable.
For a DFS A = (Q, Σ, δ, q0 ) and a word w ∈ Σ ∗ , the transition function
δ(·, w) is a transformation of Q, the transformation induced by w. The set of all
transformations induced by non-empty words is the transition semigroup SA of
A. For w ∈ Σ + , we denote by w : t the transformation t of Qn induced by w.
Given semiautomata A = (Q, Σ, δ, q0 ) and A≥ = (Q≥ , Σ, δ ≥ , q0≥ ), we define their
direct product to be the DFS A × A≥ = (Q × Q≥ , Σ, (δ, δ ≥ ), (q0 , q0≥ )).
A deterministic finite automaton (DFA) is a quintuple D = (Q, Σ, δ, q0 , F ),
where (Q, Σ, δ, q0 ) is a DFS and F ⊆ Q is the set of final states. The DFA D
accepts a word w ∈ Σ ∗ if δ(q0 , w) ∈ F . The set of all words accepted by D is
the language L(D) of D. The language accepted from a state q of a DFA is the
language Lq (D) accepted by the DFA (Q, Σ, δ, q, F ). Two states of a DFA are
distinguishable if there exists a word w which is accepted from one of the states
and rejected from the other. Otherwise, the two states are equivalent. A DFA is
minimal if all of its states are reachable from the initial state and no two states
are equivalent. Note that if |Q| ≥ 2 and D is minimal, then ∅  F  Q.

3 Connectedness
From now on we are interested in semiautomata A and A≥ whose transition
semigroups are symmetric groups generated by two-element bases. We assume
that permutations s and s≥ are induced by a in A and A≥ , and permutations t
and t≥ by b, that is, a : s, b : t in A and a : s≥ , b : t≥ in A≥ .
Example 1. Let Σ = {a, b}, A = (Q3 , Σ, δ, 0), and A≥ = (Q3 , Σ, δ ≥ , 0), where
a : s = (0, 1, 2), b : t = (0, 1) in A, and a : s≥ = (0, 1, 2), b : t≥ = (1, 2) in A≥ . Then
(s, t) and (s≥ , t≥ ) are conjugate, since rsr−1 = s≥ and rtr−1 = t≥ for r = (0, 1, 2).
If A≥≥ has s≥≥ = (0, 1) and t≥≥ = (0, 1, 2), then (s, t) and (s≥≥ , t≥≥ ) are not conjugate.
The transition semigroups of A, A≥ and A≥≥ all have 6 elements. Those of A
and A≥ , when viewed as semigroups generated by a and b, are identical, but
those of A and A≥≥ are not: for example, a3 = 1 in SA but a2 = 1 in SA⊥⊥ . 
Theorem 2. Let Σ = {a, b}, let A = (Qm , Σ, δ, 0) and A≥ = (Qn , Σ, δ ≥ , 0) be
semiautomata with transition semigroups that are symmetric groups of degrees
m and n respectively, and let the corresponding bases be B and B ≥ . For m, n ≥ 1,
the direct product A × A≥ is connected if and only if either (1) m ⇒= n or (2)
m = n and B and B ≥ are not conjugate.
Proof. Without loss of generality, assume that m ≤ n. Let H denote the tran-
sition semigroup of A × A≥ ; then H is a subgroup of Sm × Sn . Define homo-
morphisms π1 : H → Sm and π2 : H → Sn by π1 ((s, t)) = s and π2 ((s, t)) = t.
Observe that π1 and π2 are surjective, since the transition semigroups of A and
A≥ are Sm and Sn respectively. We let H0 denote the subgroup of H consisting
of all elements that map the set {0} × Qn to itself. Then H0 has index m in H
and thus π2 (H0 ) has index at most m in π2 (H) = Sn . Thus the order of π2 (H0 )
is at least n!/m ≥ (n − 1)!.
6 J. Bell et al.

Since a subgroup of Sn that does not act transitively on Qn is necessarily iso-

morphic to a subgroup of Si ×Sn−i for some i ∈ {1, . . . , n−1} [14, Section 2.5.1],
a subgroup of Sn whose order is strictly greater than (n − 1)! acts transitively on
Qn . Moreover, a subgroup of order (n − 1)! that does not act transitively on Qn
is isomorphic to S1 × Sn−1 ; that is, it is the stabilizer of a point. Thus π2 (H0 )
fails to act transitively on Qn if and only if m = n and π2 (H0 ) is the stabilizer
of a point.
Suppose that m < n or m = n and π2 (H0 ) is not the stabilizer of a point,
which is equivalent to assuming that π2 (H0 ) acts transitively on Qn . We claim
that the direct product A × A≥ is connected. To see this, notice that given (i, j)
and (i≥ , j ≥ ) in Qm × Qn , we can find t (respectively t≥ ) in H that sends (i, j)
to (0, k) (respectively (i≥ , j ≥ ) to (0, k ≥ )) for some k (respectively k ≥ ) in Qn , since
π1 (H) = Sm acts transitively on Qm . Since we have assumed that π2 (H0 ) acts
transitively on Qn , we can find t≥≥ ∈ H such that π2 (t≥≥ ) ∈ π2 (H0 ) sends (0, k)
to (0, k ≥ ). Hence (t≥ )−1 t≥≥ t sends (i, j) to (i≥ , j ≥ ), and so A × A≥ is connected.
Suppose next that m = n and π2 (H0 ) is the stabilizer of a point. By relabelling
if necessary, we may assume that π2 (H0 ) stabilizes 0. Then H cannot send (0, 0)
to (0, i) for i ⇒= 0 and so A × A≥ is not connected. We claim that the bases B
and B ≥ are conjugate.
To prove this claim, note that H has the property that if (s, t) ∈ H ⊆ Sn × Sn
and s(0) = 0, then t(0) = 0. We claim there is a permutation u ∈ Sn with
u(0) = 0 such that if (s, t) ∈ H sends (0, 0) to (j, k), then k = u(j). First
suppose that k1 , k2 ∈ Qn have the property that there is some j ∈ Qn such that
(j, k1 ) and (j, k2 ) are in the orbit of (0, 0) under the action of H. Then we can
pick h in H such that π1 (h)(j) = 0. Then (0, π2 (h)(k1 )) and (0, π2 (h)(k2 )) are
both in the orbit of (0, 0), which means that π2 (h)(k1 ) = π2 (h)(k2 ) = 0, giving
k1 = k2 . It follows that there is a map u : Qn → Qn with u(0) = 0 such that, if
(s, t) ∈ H sends (0, 0) to (j, k), then k = u(j). Since π2 (H) acts transitively on
Qn , the map u must be surjective and hence is a permutation, as claimed.
Let s1 , s2 ∈ Sn denote the elements in the transition semigroup corresponding
to a ∈ Σ, and let t1 , t2 ∈ Sn correspond to b ∈ Σ. Let H ≥ be the group generated
by (s1 , u−1 t1 u), (s2 , u−1 t2 u). Then H ≥ is conjugate to H (we conjugate H by
(1, u) to obtain H ≥ ); furthermore, H ≥ has the property that if (s, t) ∈ H ≥ sends
(0, 0) to (i, j), then i = j. Thus H ≥ acts transitively on the diagonal of Qn × Qn ;
if (s, t) ∈ H ≥ then s(i) = t(i) for all i ∈ Qn , which gives that s = t. Hence, if
(s, t≥ ) ∈ H, then u−1 t≥ u = s and so the bases B and B ≥ are conjugate. Thus if
A × A≥ is not connected, then m = n and the bases B and B ≥ are conjugate.
Now we show the converse: If m = n and the bases B = (s, t) and B ≥ =
(s , t ) are conjugate, then A × A≥ is not connected. If rsr−1 = s≥ , and rtr−1 =
≥ ≥

t≥ , let ψr : {s, t}+ → {s≥ , t≥ }+ be the mapping that assigns to x ∈ {s, t}+ the
element rxr−1 ∈ {s≥ , t≥ }+ . For any x, y ∈ {s, t}+ , if xy = z, then ψr (x)ψr (y) =
(rxr−1 )(ryr−1 ) = r(xy)r−1 = ψr (z). Hence the transition semigroups of A and
A≥ are isomorphic.
 Symmetric Groups and Quotient Complexity of Boolean Operations 7

The direct product A × A≥ is defined by (Qn × Qn , {a, b}, (δ, δ ≥ ), (0, 0)), where
(δ, δ ≥ )((i, j), a) = (s(i), rsr−1 (j)) and (δ, δ ≥ )((i, j), b) = (t(i), rtr−1 (j)) for any
i, j ∈ Qn .
If A × A≥ is connected, then for all (i, j) ∈ Qn × Qn there must exist a
word w ∈ Σ + such that (δ, δ ≥ )((0, 0), w) = (i, j) or, equivalently, there exists a
permutation p such that p(0) = i and rpr−1 (0) = j. There are now two cases:
1. If r−1 (0) ⇒= 0, we prove that state (i, r(i)) is unreachable for all i ∈ Qn .
If (i, r(i)) is reachable, then there exists a permutation p such that p(0) = i
and rpr−1 (0) = r(i). But then r−1 rpr−1 (0) = pr−1 (0) = i = p(0), and so
p−1 pr−1 (0) = r−1 (0) = 0, which is a contradiction.
2. If r−1 (0) = 0, we prove that state (i, i) is unreachable for some i ∈ Qn .
Since r cannot be the identity, there must exist an i such that r(i) ⇒= i. Suppose
(i, i) is reachable for that i. Then there exists a permutation p such that p(0) = i
and rpr−1 (0) = i. Thus i = rpr−1 (0) = rp(0) = p(0) and r(i) = i, which is a
contradiction.
Hence A × A≥ cannot be connected. 
⊇

Remark 1. If A × A≥ is connected, then it is strongly connected, since the tran-

sition semigroup of A × A≥ is a group.

4 Uniformly Minimal Semiautomata

Semiautomata that result in minimal DFAs under any non-trivial assignment of
final states were studied by Restivo and Vaglica [10]. We modify their definitions
slightly to suit our purposes. A strongly connected DFS A = (Q, Σ, δ, q0 ) with
|Q| ≥ 2 is uniformly minimal if the DFA D = (Q, Σ, δ, q0 , F ) is minimal for each
set F of final states, where ∅  F  Q.
Given a DFS A = (Q, Σ, δ, q0 ), we define the pair graph of A to be the
directed graph GA = (VA , EA ), where the set VA of vertices is the set of all two-
element subsets {p, q} of Q, and the set EA of edges consists of unordered pairs
({p, q}, {p≥, q ≥ }) such that {δ(p, a), δ(q, a)} = {p≥ , q ≥ }. The following is from [10]:

Proposition 1 (Restivo and Vaglica). Let A = (Q, Σ, δ, q0 ) be a strongly

connected DFS with at least two states. If the pair graph (VD , ED ) is strongly
connected, then A is uniformly minimal.

We prove a similar result for semiautomata with symmetric groups.

Proposition 2. Suppose that A = (Qn , Σ, δ, q0 ) is a DFS and the transition

semigroup SA of A is the symmetric group Sn . Then A is strongly connected
and uniformly minimal.

Proof. If SA = Sn , then SA contains all permutations of Qn , in particular, the

cycle (0, . . . , n−1); hence A is strongly connected. For any (i, j), (k, Ω) ∈ Qn ×Qn
with i ⇒= j, k ⇒= Ω, and {i, j} ⇒= {k, Ω}, any permutation that maps i to k and
j to Ω connects {i, j} to {k, Ω} in the pair graph of A. Hence the pair graph is
strongly connected, and A is uniformly minimal by Proposition 1. 
⊇
8 J. Bell et al.

Let the truth values of propositions be 1 (true) and 0 (false). Let ◦ : {0, 1} ×
∗
{0, 1} → {0, 1} be a binary boolean function. Extend ◦ to a function ◦ : 2Σ ×
∗ ∗
2Σ → 2Σ : If w ∈ Σ ∗ and L, L≥ ⊆ Σ ∗ , then w ∈ (L ◦ L≥ ) ⇔ (w ∈ L) ◦ (w ∈ L≥ ).
Also, extend ◦ to a function ◦ : 2Qm × 2Qn → 2Qm ×Qn : If q ∈ Qm , q ≥ ∈ Qn ,
F ⊆ Qm , and F ≥ ⊆ Qn , then (q, q ≥ ) ∈ (F ◦ F ≥ ) ⇔ (q ∈ F ) ◦ (q ≥ ∈ F ≥ ).
Suppose that A = (Q, Σ, δ, 0) and A≥ = (Q≥ , Σ, δ ≥ , 0) with |Q| = m and
≥
|Q | = n are uniformly minimal DFSs, and ◦ is any proper boolean function.
The pair (A, A≥ ) is uniformly minimal for ◦ if the direct product P = (Q ×
Q≥ , Σ, (δ, δ ≥ ), (0, 0), F ◦ F ≥ ) is minimal for all valid assignments of sets F and F ≥
of final states to A and A≥ , that is, sets such that ∅  F  Q and ∅  F ≥  Q≥ .
If n = 1, then A × A≥ is isomorphic to A and no boolean function ◦ is proper.
Hence this case, and also the case with m = 1, is of no interest. Henceforth we
assume that m, n ≥ 2.
We now consider pair graphs of DFSs with symmetric groups as their transi-
tion semigroups.

Example 2. Suppose that m = n = 2, and A and A≥ both have S2 as their

transition semigroup. There are two permutations in S2 : (0, 1) and 1, and there
are three bases: B1 = (a : (0, 1), b : (0, 1)), B2 = (a : (0, 1), b : 1), and B3 =
(a : 1, b : (0, 1)). Note that no two of these bases are conjugate.
For each basis, there are two possible final states, 0 or 1, and hence two DFAs;
thus there are six different DFAs. There are then twelve direct products Dji × Dψk
with non-conjugate bases, where Dji (Dψk ) uses basis Bi (Bk ) and has j (Ω) as
final state, for i, k = 1, 2, 3 and j, Ω = 1, 2.
For each pair of DFAs accepting languages L and L≥ respectively, we tested
the complexity of five boolean functions: L ∪ L≥ , L ∩ L≥ , L ⊕ L≥ , L \ L≥ and
L≥ \ L. Note that the complexity of each remaining proper boolean function is
the same as that of one of these five functions. For all twelve direct products of
DFAs with non-conjugate bases, all proper boolean functions reach the maximal
complexity 4, except for the functions L ⊕ L≥ and L ⊕ L≥ , which fail in all twelve
cases. Thus any two DFAs D = (Q2 , Σ, δi , 0, F ) and D≥ = (Q2 , Σ, δk , 0, F ≥ ),
where Q2 = {0, 1}, Σ = {a, b}, δi (δk ) is defined by basis Bi (Bk ), F = {j} and
F ≥ = {Ω}, are uniformly minimal for all proper boolean functions, except ⊕ and
its complement. So our main result applies only in some cases if m = n = 2. 

Proposition 3. Let A = (Qm , Σ, δ, 0) and A≥ = (Qn , Σ, δ ≥ , 0), with m, n ≥ 2

and max(m, n) ≥ 3, be DFSs with transition semigroups that are symmetric
groups, and let P be their direct product. Then the following hold:
1. The pair graph of P consists of strongly connected components – which we
will call simply components – of one of the following three types:
(a) T1 ⊆ C1 = {{(i, j), (k, Ω)} : i ⇒= k, j ⇒= Ω},
(b) T2 ⊆ C2 = {{(i, j), (i, Ω)} : j ⇒= Ω},
(c) T3 ⊆ C3 = {{(i, j), (k, j)} : i ⇒= k}.
2. Every state (i, j) of the direct product P appears in at least one pair in
each component.
3. Each component has at least mn/2 ≥ 3 pairs.
 Symmetric Groups and Quotient Complexity of Boolean Operations 9

Proof. The first claim follows since the transition semigroup of P is a group.
The second claim holds because the direct product is strongly connected, by
Remark 1. For the third claim, note that there are mn states in P, but they can
appear in pairs; hence the bound mn/2. Since we are assuming that mn ≥ 6,
the last claim follows. 
⊇
Now consider DFAs D = (Qm , Σ, δ, 0, F ) and D≥ = (Qn , Σ, δ ≥ , 0, F ≥ ), where
∅  F  Qm and ∅  F ≥  Qn . A state {(i, j), (k, Ω)} of the pair graph of the
direct product P of D and D≥ is distinguishing if and only if (i, j) is final and
(k, Ω) is not, or vice versa.
Example 3. Suppose m = 3, n = 4, δ is defined by the basis (a : (0, 1), b : (0, 1, 2))
of S3 , and δ ≥ by the basis (a : (0, 1), b : (1, 3, 2)) of S4 . One verifies that these bases
are not conjugate. The direct product P is connected and has twelve states.
If F = {2}, F ≥ = {0, 1} and intersection is the boolean function, then there
are no distinguishing pairs in the component of the pair graph T containing
{(0, 0), (0, 3)}. Hence any two states appearing in the same pair of T are equiv-
alent. Indeed, the minimal version of P has only six states. 
Example 4. Suppose m = n = 4, δ is defined by the basis (a : (0, 1, 2), b : (2, 3)),
and δ ≥ by the basis (a : (1, 3, 2), b : (0, 2, 1, 3)). If F = {0, 1} and F ≥ = {0, 1}, then
the complexity of L ⊕ L≥ is 4, but all the other complexities are 12. 
Lemma 1. Let D = (Q, Σ, δ, 0, F ) and D≥ = (Q≥ , Σ, δ ≥ , 0, F ≥ ), with |Q|, |Q≥ | ≥
2, be DFAs with transition semigroups that are groups, and let P = (Q ×
Q≥ , Σ, (δ, δ ≥ ), (0, 0), F ◦ F ≥ ) be their direct product. Then P is minimal if and
only if every component of the pair graph GP of P has a distinguishing pair.

5 Symmetric Groups and Boolean Operations

We begin with a well-known but apparently unpublished result.
Lemma 2. Let n be a positive integer, let G be either Sn or An , and let H be
a subgroup of G of index m ≤ n. Then the following hold:
(i) if n ⇒= 4 and m < n, then H is either An or Sn ;
(ii) if m = n and n ⇒= 6, then there is some i ∈ Qn such that H is the set of
permutations in G that fix i.
(iii) if m = n = 6, then there is an automorphism φ of S6 such that φ(H) is
the set of elements that fix 0.
The following lemma, like Theorem 2, deals with reachability. The conditions
in the lemma, however, are useful for determining reachability in the pair graph
of A × A≥ , rather than in A × A≥ itself.
Lemma 3. Let Σ = {a, b}, let A = (Qm , Σ, δ, 0) and A≥ = (Qn , Σ, δ ≥ , 0) be
semiautomata with transition semigroups that are symmetric groups of degrees
m and n respectively with m ≤ n, n ⇒= 4 and (m, n) ⇒= (6, 6). Let H be the
transition semigroup of A × A≥ , and let π1 and π2 be the natural projections
from H onto Sm and Sn respectively. If H0 = {h ∈ H : π1 (h)(0) = 0}, then
10 J. Bell et al.

1. π2 (H0 ) is either Sn or An , or is the stabilizer of a point in Qn .

2. π2 (H0 ) is the stabilizer of a point if and only if m = n, and in this case
the direct product A × A≥ is not connected.
Proof. For Part 1, since π1 (H) = Sm , for each i ∈ {0, . . . , m − 1} there is some
hi ∈ H such that π1 (hi ) takes 0 to i. For a given h ∈ H, π1 (h) takes 0 to j for
some j ∈ {0, 1, . . . , m − 1}, and thus h−1 j h ∈ H0 and so h ∈ hj H0 . However,
−1
since π1 (h) takes 0 to j, we have hi h ⇒∈ H0 and thus h ⇒∈ hi H0 for i ⇒= j.
Thus the cosets
 h0 H, . . . , hm−1 H are distinct, and H0 has index m in H. Since
π2 (H) ⊆ m−1 i=0 π2 (hi )π2 (H0 ), π2 (H0 ) has index at most m in π2 (H) = Sn . If
n ⇒= 4 and m < n, then π2 (H0 ) is either An or Sn by Lemma 2. If m = n and
n ⇒= 6, then π2 (H0 ) has index n in Sn and hence must be the stabilizer of a some
i ∈ Qn by Lemma 2.
For Part 2, suppose that m = n and π2 (H0 ) is the stabilizer of a point in Qn .
By relabelling if necessary, we may assume that π2 (H0 ) stabilizes 0. Hence, if
h ∈ H sends (0, 0) to (0, j) then j = 0. In particular, there is no h ∈ H that
sends (0, 0) to (0, 1) or that sends (0, 1) to (0, 0), and so A × A≥ is necessarily
not connected. 
⊇
Lemma 4. Let A = (Qm , Σ, δ, 0) and A≥ = (Qn , Σ, δ ≥ , 0) be semiautomata with
transition semigroups that are the symmetric groups of degrees m and n, re-
spectively with m ≤ n, m ≥ 2, n ≥ 5, and (m, n) ⇒= (6, 6). If A × A≥ is
connected, then the pair graph of A × A≥ has exactly three connected compo-
nents: C1 = {{(i, j), (k, Ω)} : i ⇒= k, j ⇒= Ω}, C2 = {{(i, j), (i, Ω)} : j ⇒= Ω}, and
C3 = {{(i, j), (k, j)} : i ⇒= k}.
Proof. We let H denote the transition semigroup of A × A≥ . We show that each
of C1 , C2 , C3 is strongly connected. Note that each of C1 , C2 , C3 is necessarily
a union of connected components.
We show that C1 is strongly connected. Suppose we have pairs {(i, j), (k, Ω)}
and {(i≥ , j ≥ ), (k ≥ , Ω≥ )} with i, k distinct, i≥ , k ≥ distinct, j, Ω distinct, and j ≥ , Ω≥ dis-
tinct. Since Sm acts doubly transitively on Qm when m ≥ 2, there is some s ∈ H
that sends (i, j) to (i≥ , j ≥≥ ) and (k, Ω) to (k ≥ , Ω≥≥ ) for some j ≥≥ , Ω≥≥ ∈ Qn .
Thus we may assume without loss of generality that i≥ = i and k ≥ = k. Let H0
be the subgroup of Sm × Sn consisting of all x ∈ H such that π1 (x) fixes i. By
Lemma 3, since we assume that A × A≥ is connected, π2 (H0 ) is not a stabilizer of
a point in Qn . Hence π2 (H0 ) is either Sn or An . Let H1 denote the subgroup of
Sm × Sn consisting of all x ∈ H such that π1 (x) fixes i and k. By the argument
used in Lemma 3 to show that {h ∈ H : π1 (h)(0) = 0} has index m in H, we see
that π2 (H1 ) has index at most m − 1 in π2 (H0 ). Thus π2 (H1 ) is a subgroup of
An or Sn of index at most n − 1, and hence must again be An or Sn by Lemma
2. Since An and Sn both act doubly transitively on Qn , there is some h ∈ H
that sends (i, j) to (i, j ≥ ) and (k, Ω) to (k, Ω≥ ) whenever Ω and Ω≥ are distinct. This
proves that C1 is indeed a strongly connected component.
Next, consider pairs {(i, j), (i, k)} with j, k distinct. For given {(i≥ , j ≥ ), (i≥ , k ≥ )}
with j ≥ , k ≥ distinct, there is some element s ∈ H such that π1 (s)(i) = i≥ and thus
s sends (i, j) to (i≥ , j ≥≥ ) and (i, k) to (i≥ , k ≥≥ ) for some j ≥≥ , k ≥≥ ∈ Qn with j ≥≥ ⇒= k ≥≥ .
 Symmetric Groups and Quotient Complexity of Boolean Operations 11

Now note that π2 ({x ∈ H : π1 (x)(i≥ ) = i≥ }) is either Sn or An by Lemma 3, and

thus acts doubly transitively on Qn . It follows that there is some s≥ ∈ H such
that s≥ sends (i≥ , j ≥≥ ) to (i≥ , j ≥ ) and (i≥ , k ≥≥ ) to (i≥ , k ≥ ). Then s≥ s sends {(i, j), (i, k)}
to {(i≥ , j ≥ ), (i≥ , k ≥ )} and thus C2 is strongly connected.
Finally, consider pairs {(i, j), (k, j)} and {(i≥ , j ≥ ), (k ≥ , j ≥ )} with i, k distinct and
≥ ≥
i , k distinct. From the argument used in proving C1 is strongly connected, we
see that we can find s ∈ H that sends {(i, j), (k, j)} to {(i≥ , j ≥≥ ), (k ≥ , j ≥≥ )} for some
j ≥≥ . As in the proof that C1 is strongly connected, we see that the image of the
set of h ∈ H for which π1 (h) stabilizes both i≥ and k ≥ under π2 acts transitively
on Qn ; hence we can find s≥ ∈ H that sends {(i≥ , j ≥≥ ), (k ≥ , j ≥≥ )} to {(i≥ , j ≥ ), (k ≥ , j ≥ )}.
Thus C3 is strongly connected. 
⊇

Corollary 1. Let m and n be positive integers with n ≥ m ≥ 2, n ≥ 5, and

(m, n) ⇒= (6, 6), and let A = (Qm , Σ, δ, 0) and A≥ = (Qn , Σ, δ ≥ , 0) be semiau-
tomata with transition semigroups that are the symmetric groups of degrees m
and n. Suppose that the direct product A × A≥ is connected and assume further
that sets of final states are added to A and A≥ and that ◦ is a proper binary
boolean function that defines the set of final states of the direct product P. Then
P is minimal for any such ◦.

Proof. By Lemma 4, the pair graph of A × A≥ has three strongly connected

components: C1 = {{(i, j), (k, Ω)} : i ⇒= k, j ⇒= Ω}, C2 = {{(i, j), (i, Ω)} : j ⇒= Ω},
and C3 = {{(i, j), (k, j)} : i ⇒= k}.
For (i, j) ∈ Qm × Qn , define f ((i, j)) to be 1 if (i, j) is a final state, and 0,
otherwise. We first claim that C1 has a distinguishing pair, that is, there are pairs
(i, j) and (k, Ω) in Qm × Qn with i ⇒= k and j ⇒= Ω such that f ((i, j)) ⇒= f ((k, Ω)).
Suppose no distinguishing pair exists in C1 . Assume without loss of generality
that f ((0, 0)) = 0. then f ((i, j)) = 0 whenever i ⇒= 0 and j ⇒= 0. Given k ∈ Qn ,
we pick Ω ∈ Qn \ {0, k}; this is always possible since n ≥ 3. Since {(0, k), (1, Ω)}
is in C1 and we have assumed that C1 has no distinguishing pairs, we must have
f ((0, k)) = f ((1, Ω)). But f (1, Ω) must be 0, for otherwise we would have the
distinguishing pair {(0, 0), (1, Ω)}. Hence f ((0, k)) = f ((1, Ω)) = 0. Thus we have
f ((i, j)) = 0 for every i ∈ Qm and every j ∈ Qn \ {0}. Similarly, we must have
f ((i, 0)) = f ((0, 1)) = 0 for i ∈ Qm \ {0}, and hence f is the zero function, a
contradiction.
The fact that C2 and C3 both have distinguishing pairs follows from the fact
that ◦ is a proper boolean function. By Lemma 1, we conclude that A × A≥ is
uniformly minimal. 
⊇

We have proved our main result in the case that m ≤ n and n ≥ 5 if

(m, n) ⇒= (6, 6). By symmetry we may always assume that m ≤ n. The case
(m, n) = (2, 2) was handled in Example 2, that of (m, n) = (3, 4), in Exam-
ple 3, and that of (m, n) = (4, 4), in Example 4. So the only cases to consider
are those with (m, n) ∈ {(2, 3), (2, 4), (3, 3), (6, 6)}; these cases are covered at
http://arxiv.org/abs/1310.1841.
12 J. Bell et al.

6 Conclusions
We have shown that if the inputs of two DFAs induce transformations that con-
stitute non-conjugate bases of symmetric groups, then the quotient complexity
of all non-trivial boolean operations on the languages accepted by the DFAs is
maximal, except for a few special cases when the sizes of the DFAs are small.
We believe that other similar results are possible and deserve further study.

Acknowledgments. This work was supported by the Natural Sciences and En-
gineering Research Council of Canada under grants No. 611456 and OGP0000871,
by the European Regional Development Fund through the programme COM-
PETE, and by the Portuguese Government through the FCT under projects
PEst-C/MAT/UI0144/2011 and CANTE-PTDC/EIA-CCO/101904/2008. We
thank Gareth Davies for his careful proofreading and constructive comments.

References
1. Brzozowski, J.: Quotient complexity of regular languages. J. Autom. Lang.
Comb. 15(1/2), 71–89 (2010)
2. Brzozowski, J.: In search of the most complex regular languages. Int. J. Found.
Comput. Sc. 24(6), 691–708 (2013)
3. Brzozowski, J., Davies, G.: Maximally atomic languages. In: Ésik, Z., Fülop, Z.
(eds.) 14th International Conference Automata and Formal Languages, AFL 2014,
Szeged, Hungary, May 27-29. EPTCS, vol. 151, pp. 151–161 (2014)
4. Brzozowski, J., Davies, G.: Most complex regular right-ideal languages. In: 16th
International Workshop on Descriptional Complexity of Formal Systems, DCFS
2014, Turku, Finland, August 5-8. LNCS 8614 (to appear, 2014)
5. Brzozowski, J., Tamm, H.: Complexity of atoms of regular languages. Int. J. Found.
Comput. Sc. 24(7), 1009–1027 (2013)
6. Brzozowski, J., Tamm, H.: Theory of átomata. Theoret. Comput. Sci. (article in
press, 2014)
7. Liu, G., Martin-Vide, C., Salomaa, A., Yu, S.: State complexity of basic language
operations combined with reversal. Inform. and Comput. 206, 1178–1186 (2008)
8. Maslov, A.N.: Estimates of the number of states of finite automata. Dokl. Akad.
Nauk SSSR 194, 1266–1268 (1970) (Russian); English Translation: Soviet Math.
Dokl. 11, 1373–1375 (1970)
9. Piccard, S.: Sur les bases du groupe symétrique. Časopis Pro Pěstovánı́ Matematiky
a Fysiky 68(1), 15–30 (1939)
10. Restivo, A., Vaglica, R.: A graph theoretic approach to automata minimality. The-
oret. Comput. Sc. 429, 282–291 (2012)
11. Rotman, J.: The Theory of Groups: An Introduction. Allyn and Bacon, Inc., Boston
(1965)
12. Salomaa, A., Wood, D., Yu, S.: On the state complexity of reversals of regular
languages. Theoret. Comput. Sci. 320, 315–329 (2004)
13. Suzuki, M.: Group Theory, vol. 1. Springer, Berlin (1982)
14. Wilson, R.: The Finite Simple Groups. Springer, Berlin (2009)
15. Yu, S.: State complexity of regular languages. J. Autom. Lang. Comb. 6, 221–234
(2001)
16. Yu, S., Zhuang, Q., Salomaa, K.: The state complexities of some basic operations
on regular languages. Theoret. Comput. Sci. 125(2), 315–328 (1994)
 Handling Infinitely Branching WSTS

Michael Blondin1 , Alain Finkel2 , and Pierre McKenzie1

1
Université de Montréal and ENS Cachan
{blondimi,mckenzie}@iro.umontreal.ca
2
ENS Cachan
[email protected]

Abstract. Most decidability results concerning well-structured transi-

tion systems apply to the finitely branching variant. Yet some models
(inserting automata, ω-Petri nets, ...) are naturally infinitely branching.
Here we develop tools to handle infinitely branching WSTS by exploiting
the crucial property that in the (ideal) completion of a well-quasi-ordered
set, downward-closed sets are finite unions of ideals. Then, using these
tools, we derive decidability results and we delineate the undecidability
frontier in the case of the termination, the control-state maintainability
and the coverability problems. Coverability and boundedness under new
effectivity conditions are shown decidable.

1 Introduction

Well-structured transition systems (WSTS) [12,11,2] as a general class of infinite-

state systems have spawned decidability results for important problems such as
termination, boundedness, control-state maintainability and coverability. WSTS
consist of a (usually infinite) well ordered set of states, together with a monotone
transition relation. WSTS have found multiple uses: in settling the decidability
status of reachability and coverability for graph transformation systems [4,22],
in the forward analysis of depth-bounded processes [26,27], in the verification of
parameterized protocols [10] and the verification of multi-threaded asynchronous
software [21]. WSTS remain under development and are actively being investi-
gated [13,14,18,25,5,24].
Most existing decidability results for WSTS apply to the finitely branching
variant. However, WSTS such as inserting FIFO automata [7], inserting au-
tomata [6] and ω-Petri nets [17], that can arbitrarily increase some values, are
intrinsically infinitely branching, and any finitely branching WSTS parameter-
ized with an infinite set of initial states (such as broadcast protocols [10]) also
inherits an infinitely branching state. For instance, Geeraerts, Heußner, Praveen
and Raskin argue in [17] that parametric concurrent systems with dynamic

Supported by the French Agence Nationale de la Recherche, REACHARD (grant
ANR-11-BS02-001), by the Fonds québécois de la recherche sur la nature et les
technologies, by the Natural Sciences and Engineering Research Council of Canada
and by the “Chaire DIGITEO, ENS Cachan - École Polytechnique”.

J. Esparza et al. (Eds.): ICALP 2014, Part II, LNCS 8573, pp. 13–25, 2014.
∗c Springer-Verlag Berlin Heidelberg 2014
14 M. Blondin, A. Finkel, and P. McKenzie

thread creation can naturally be modelled by some classes of infinitely branch-

ing systems, like ω-Petri nets, i.e. Petri net with arcs that can consume/create
arbitrarily many tokens.
An outcome of our work is that the finite tree construction technique can be
recovered, even in the infinitely branching case, for the purpose of deciding the
boundedness problem for example.
The primary motivation for this paper is to explore the decidability status
of the termination, boundedness, control-state maintainability and coverability
problems for infinitely branching (general) WSTS. For the coverability problem,
known to be decidable for WSTS fulfilling the so-called prebasis computability
hypothesis [2], we wish to draw from the recent algebra-theoretic characteriza-
tions of downward-closed sets [13] and conceive of a post-oriented computability
hypothesis suitable for the design of a forward algorithm. (Indeed, forward algo-
rithms are arguably more intuitive than backward algorithms and post-oriented
computability more easily verified than prebasis computability, where prebasis
computability means computability of a finite basis of the upward closure of the
set of immediate predecessors, the testing of which is provably undecidable in
some WSTS.) Our contributions are the following:

1. As technical tools, we simplify and extend the analysis of the completion of

a general WSTS and we relate the behavior of a WSTS to that of its com-
pletion. In particular, we provide a general presentation of the completion
that is much less daunting than the presentations currently available in the
literature. This sets the stage for exploiting the main property of the com-
pletion of a WSTS, namely, the expressibility of any downward-closed set as
a (unique, as shown here) finite union of ideals, in the design of algorithms.
2. We uncover a new termination property (called strong termination) that
is computationally equivalent to the usual termination property for finitely
branching WSTS but that subtly differs from it in the presence of infinite
branching. Indeed, we exhibit WSTS for which strong termination is decid-
able yet the usual termination is undecidable. A similar subtle issue arises
as well in our generalization of the maintainability problem to infinitely
branching.
3. We generalize most decidability results mentioned for finitely branching
WSTS earlier to the infinitely branching case. This requires carefully tracking
the effectiveness and the monotonicity conditions which support decidability.
When possible, we delineate the frontier between decidability for a problem
and the undecidabilty that results from dropping one of these conditions.
The new decidability results for (strong) termination and (strong) maintain-
ability exploit the completion. The new algorithm for coverability uses a
forward strategy coupled with a post-oriented computability hypothesis.

Our work further highlights the naturalness of the class of ω 2 -WSTS. Indeed
our decidability results apply in one blow to known classes of infinitely branching
WSTS like inserting FIFO automata [7], inserting automata [6], ω-Petri nets [17]
and broadcast protocols [10].
 Handling Infinitely Branching WSTS 15

Section 2 below introduces notation and preliminaries. Section 3 surveys

known decidabilities and exhibits some undecidabilities. Section 4 develops our
tools to handle infinite branching. Section 5 contains the bulk of our decidability
results for infinitely branching WSTS. Section 6 summarizes our contribution
and suggests future work.

2 WSTS
Let X be a set and ∈ a quasi-ordering on X (∈ reflexive and transitive), then ∈
is a well-quasi-ordering (wqo) if for every infinite sequence x0 , x1 , . . . of elements
xn ≡ X, there exist i < j such that xi ∈ xj . It is well-known that Nd is well-
quasi-ordered under (x1 , . . . , xd ) ∈Nd (x1 , . . . , xd ) where the latter means that
⊆i xi ∈ xi (Dickson’s Lemma). We extend N to Nω by adding an element ω
verifying ω ↔Nω x for all x ≡ Nω . The set Ndω is also well-quasi-ordered. We
simply write ∈ for ∈N and ∈Nω when there is no ambiguity.
Recall that a WSTS is an ordered transition system S = (X, − ⇔S , ∈) such that
∈ is a well-quasi-ordering on X, and the relation − ⇔S ⊂ X × X is monotone (or
compatible) with ∈ meaning that for all x, y, x such that x − ⇔S y and x ∈ x ,
≥
there exists a state y  such that x − ⇔S y  and y ∈ y  . WSTS thus satisfy a
general monotony by definition. There exist other variations of monotony:
strong: x−⇔S y ≥ x ↔x =⇒ x − S y  ↔ y,
⇔
stuttering: x −
⇔S y ≥ x ↔x =⇒ x = x0 − ⇔S xk −
⇔S . . . − ⇔S y  ↔ y, ⊆i xi ↔ x,
+
transitive: x −
⇔S y ≥ x ↔x =⇒ x ⇔S y  ↔ y,
−
≥
strict: x−⇔S y ≥ x >x =⇒ x ⇔S y  > y.
−
Strong monotony implies stuttering monotony which implies transitive monotony.
We denote, as usual, PreS (x) = {y : y − ⇔S x}, PostS (x) = {y : x − ⇔S y},
PreS (T ) = x∈T PreS (x) and PostS (T ) = x∈T PostS (x).
Throughout this paper, WSTS will be assumed effective in the following sense:
(1) the set of states X is r.e. (which suffices to compute PostS (x) when |PostS (x)|
is known and finite); (2) the transition relation is decidable, i.e., the WSTS comes
equipped with an algorithm that can decide, given x, y ≡ X, whether x − ⇔S y
or equivalently whether y ≡ PostS (x); (3) the quasi-ordering ∈ is decidable,
i.e., the WSTS also comes equipped with an algorithm that can decide, given
x, y ≡ X, whether x ∈ y. Forward analysis techniques for (finitely branching)
WSTS typically compute the finite set PostS (x), which is made possible by
assuming PostS computable. Because our new setting allows PostS (x) to be
infinite, we need to adapt this assumption. Our “post-effectivity” notion mildly
weakens the usual hypothesis of “being able to compute PostS ”:

Definition 2.1. A transition system S = (X, − ⇔S ) is post-effective if S is ef-

fective and f : X ⇔ N ◦ {“infinite”} given by f (x) = |PostS (x)| is computable.

Transition systems defined by a finite set of recursive functions are typical

examples of finitely branching systems and they will be called functional. Let Fd
16 M. Blondin, A. Finkel, and P. McKenzie

denote the set of WSTS whose transitions relation is prescribed by finitely many
increasing functions f from Nd to Nd (i.e. x ∈ y =⇒ f (x) ∈ f (y)) which are
also recursive (i.e., given by halting Turing machines); these WSTS are finitely
branching and post-effective. Inserting FIFO automata [7], inserting automata
[6] and ω-Petri nets [17] are post-effective infinitely branching WSTS.
Recall that an effective ordered transition system is said essentially finite
branching [2] if the subset maxpost(x) of maximal elements of PostS (x) is non
empty, finite and computable. Some WSTS, e.g. ω-Petri nets, are post-effective
but are not essentially finite branching and conversely, we can exhibit essentially
finite branching WSTS that are not post-effective.
Post-effectivity (Definition 2.1) is a weaker notion than “having a finite and
computable PostS ”. The weaker notion does imply “computable PostS ” for ef-
fective WSTS that are finitely branching. Hence it is natural to ask whether the
finitely branching property is decidable for post-effective WSTS. It is not:
Proposition 2.2. Testing, given a post-effective WSTS S and x0 ≡ X, whether
≥
there exists an execution x0 −
⇔S x such that PostS (x) is infinite is undecidable.
Let ≤ T and → T stand respectively for the set of states that are ↔ and ∈
some state in T . A set T is upward closed if T = ≤ T and downward closed if
T = → T . An upward basis of a set T is a set B such that T = ≤ B. An ideal I
is a downward closed set that is also directed, i.e., ⊆a, b ≡ I, ∅c ≡ I such that
a ∈ c and b ∈ c. We note Ideals(X) the set of ideals of an ordered set X. A
directed complete partial ordering (dcpo) is an ordered set (X, ∈) such that every
directed set D ⊂ X has a least upper bound (lub) in X: for instance, (N, ∈),
with the usual notations, is not a dcpo since the directed set N has no lub in
N; if we add the lub ω to N, then (Nω , ∈) is a dcpo. There is a way to add all
lubs to any ordered set (X, ∈), that is called the ideal completion, since each
element x ≡ X can be identified with → x ≡ Ideals(X) and since it is well-known
that (Ideals(X), ⊂) is a dcpo [3,13]. We will consider the following problems for
WSTS, where the input to each problem is an effective WSTS S = (X, − ⇔S , ∈)
and a state x0 ≡ X, together with an x ≡ X in the case of coverability, and a
set t1 , . . . , tn ≡ X in the case of the maintainability problem:
– Coverability: ∅ execution x0 − ⇔S x1 −⇔S . . . −⇔S xk ↔ x?
– Boundedness: Post≥S (x0 ) is infinite?
– Termination: ∅ infinite execution x0 − ⇔S x1 − ⇔S . . .?
– Strong termination: ∅k ≡ N s.t. x0 − ⇔S x1 − ⇔S . . . − ⇔S xm =⇒ m ∈ k?
– Control-state maintainability:∅ computation (i.e. an infinite execution x0 −⇔S
x1 −⇔S . . . or a finite execution x0 −⇔S x1 − ⇔S . . . −
⇔S xk that cannot be
further extended) such that ⊆i xi ≡ ≤ {t1 , . . . , tn }?
– Strong control-state maintainability: ⊆k ≡ N, ∅ execution x0 − ⇔S x1 − ⇔S
... −
⇔S xm such that m ↔ k and ⊆i xi ≡ ≤ {t1 , . . . , tn }?

3 Decidability for WSTS

Recall that a WSTS S = (X, −
⇔S , ∈) has a computable prebasis [11,2] if the
WSTS comes equipped with a computable function that maps each x ≡ X to
 Handling Infinitely Branching WSTS 17

some finite basis of the upward closed set ≤ PreS (≤ x). We summarize the four
main decidability results known about (essentially) finite branching WSTS:
Theorem 3.1 ([12,11,2]).
– Termination is decidable for post-effective finitely branching WSTS with
transitive monotony [12], and for essentially finite branching effective WSTS
with strong monotony [2].
– Boundedness is decidable for post-effective finitely branching WSTS with
strict transitive monotony and well partial ordering [11].
– Control-state maintainability is decidable for post-effective finitely branching
WSTS with stuttering monotony [11], and for essentially finite branching
effective WSTS with strong monotony [2].
– Coverability is decidable for effective WSTS with prebasis computability [11,2].
Theorem 3.1 states results exactly as they appear in the literature, but it
would not be difficult to unify some of the hypotheses made here. For instance,
termination can be shown decidable for essentially finite branching effective
WSTS with transitive monotony. We defer a systematic treatment of this unifi-
cation to a future version of the present paper.
Our goals in this paper are to extend the decidability of termination, bound-
edness and maintainability given by Theorem 3.1 to the more general case of
infinitely branching WSTS. Our goal for the coverability problem is to investi-
gate alternative effectivity hypotheses. We first note:
Theorem 3.2. Termination is undecidable for post-effective WSTS with tran-
sitive (and even strong and strict) monotony.
In Sect. 5, we prove boundedness decidable for post-effective infinitely branch-
ing WSTS with strict monotony and well partial ordering. By contrast, as exem-
plified by Petri nets with Reset [8], boundedness is well known to be undecidable
for post-effective finitely branching WSTS with non-strict yet transitive (even
strong) monotony and with well partial ordering. Concerning maintainability,
Theorem 3.3. Control-state maintainability is undecidable for post-effective
WSTS with stuttering (and even strong and strict) monotony.
We now turn to coverability. Existing proofs that coverability is decidable
need the prebasis hypothesis: Abdulla et al. use a backward algorithm [11,1]
that computes a finite basis of ≤ Pre≥ (≤ x) and Geeraerts et al. use a forward
algorithm [18] that requires further hypotheses (i.e. restriction to an adequate
domain of limits, a mathematical hypothesis subsequently shown superfluous
[16,13]). Note that coverability for post-effective (even finitely branching) WSTS
becomes undecidable without the prebasis hypothesis, as is the case for instance
for WSTS in F2 (recall definition from Sect 2, i.e., WSTS composed of recursive
increasing functions from N2 to N2 ) [15].
Prebasis computability is sufficient to ensure decidability of coverability. How-
ever, as we show in Prop. 3.4 below, prebasis computability is not necessary:
there is a class of WSTS, namely F1 , for which coverability is decidable yet no
prebasis function is computable.
18 M. Blondin, A. Finkel, and P. McKenzie

Proposition 3.4. Coverability for F1 is decidable, but no algorithm that takes

as input S ≡ F1 and x ≡ N can systematically output a finite basis of ≤ PreS (≤ x).

4 Handling Infinite Branching Finitely

In this section we prepare the ground for developping decision procedures ca-
pable of handling, under natural hypotheses, infinitely branching systems. First
we would like the ability to compute finite representations of each term in the
sequence → x, → PostS (→ x), → PostS (→ PostS (→ x)), . . .. This requires finitely repre-
senting downward closed sets, which is possible for wqo. This section describes
how this is done and presents effective tools for doing it.

4.1 Downward Closed Sets and Ideals

It has long been known that in a wqo, any upward closed set has a finite basis;
this is Dickson’s lemma in (Nk , ∈) and it is Higman’s lemma in (Σ ≥ , ∈) when ∈
is the subword relation. It has recently been discovered that a similar situation
occurs for downward closed sets in wqo.
Theorem 4.1. [13] Any downward closed subset in a wqo X is a finite union
of ideals.
The original proof of Theorem 4.1 needs a technical bridge between topological
completions and ordering completions of a set. A short and self-contained proof
of Theorem 4.1 was given by Goubault-Larrecq [19].
Theorem 4.3 below slightly refines Theorem 4.1. It shows that any downward
closed set uniquely decomposes as a certain finite union of ideals. This requires:
Proposition 4.2. Any ideal contained in a finite union of ideals is contained in
one of these ideals. In particular, testing the inclusion of an ideal I in a union
J1 ◦ J2 ◦ ... ◦ Jk of ideals is equivalent to testing whether I ⊂ Jj for some j such
that 1 ∈ j ∈ k.
m
A finite union D = i=1 Ii of ideals will be said to canonically decompose D if
the Ii ’s are pairwise incomparable under inclusion. This terminology is justified:
Theorem 4.3. Any downward closed subset in a wqo X admits a unique decom-
position as a finite union of pairwise incomparable ideals. Therefore, a downward
closed subset decomposes canonically as the union of its maximal ideals.
Ideals in a wqo cannot necessarily be manipulated effectively. For instance,
there exist some ordered countable sets X such that Ideals(X) is not countable.
Consider X = Σ ≥ , with the prefix ordering. Then Ideals(X) is isomorphic to
Σ ≥ ◦ Σ ω and is not countable when Σ contains at least two letters. However:
Proposition 4.4. A wqo X is countable iff Ideals(X) is countable.
Fortunately, inclusion between ideals is decidable for well-quasi-ordered sets
obtained by closing finite sets and closing naturals numbers under finite prod-
ucts, disjoint sums, multiset operator and Kleene star (respectively with their
natural associated orderings) [13]. Therefore inclusion of ideals of Nd and inclu-
sion of ideals of Σ ≥ are decidable.
 Handling Infinitely Branching WSTS 19

4.2 Completion of WSTS

F
Recall that for a functional WSTS S = (X, −⇔, ∈) where F is a finite set of
increasing recursive functions f : X ⇔ X, the functional completion [14] is
F
defined by S = (X, −
⇔, ⊂) where X = Ideals(X) and F is the set of functions
def
f : Ideals(X) ⇔ Ideals(X) defined by f (I) = → f (I) for every f ≡ F . We note
that f (I) is an ideal if I is an ideal. Here we extend the completion process to
any (infinitely) branching WSTS:

Definition 4.5. The completion S of a WSTS S = (X, − ⇔S , ∈) is the ordered

transition system S = (X, −
⇔S, ⊂) where X  = Ideals(X), and I ⇔ − S J if J
appears in the canonical decomposition of → Post(I).
F
Let S = (X, − ⇔, ∈) be a functional WSTS, then the following relation holds
between S, S and S for every ideal I ≡ Ideals(X):
  
PostS (I) = f (I) = → f (I) = J = → PostS (I).
f ∈F f ∈F J∈PostS (I)

Another good news is that:

Proposition 4.6. The completion S of any WSTS S is finitely branching.
Moreover the completion computes exactly the downward closure of the reach-
ability set of its original system.
Proposition 4.7. Let S = (X, − ⇔S , ∈) be a WSTS and Post≥S(→ x) = {J1 , . . . ,
≥
Jn }. We have → PostS (x) = J1 ◦ . . . ◦ Jn .
A natural question that arises is whether the completion of a WSTS is also a
WSTS. It does indeed have monotony:
⇔S , ∈) then S has strong monotony.
Proposition 4.8. Let S = (X, −
However, (Ideals(X), ⊂) is not always a wqo and therefore the completion is
not always a WSTS. In fact, it is known to be a wqo iff (X, ∈) is a so-called
ω 2 -wqo, a notion we will not define here. In general, a wqo is not necessarily a
ω 2 -wqo and the typical counter-example is the Rado ordering [20]. Now, a result
from Jancar [20] simplifies the characterization of ω 2 -wqos as follows: a wqo ∈
is a ω 2 -wqo iff ∈# is a wqo, where ∈# is the Hoare ordering defined by A ∈# B
iff ≤ B ⊂≤ A.
Extending the terminology to WSTS, we obtain the following result general-
izing the known result for functional WSTS [14]:

Theorem 4.9. Let S be a WSTS, then S is a WSTS iff S is a ω 2 -WSTS.

We end this section with the observations that a WSTS inherits the strict
monotony of its completion but not conversely, and that post-effectivity of a
WSTS is independent from the post-effectivity of its completion.
20 M. Blondin, A. Finkel, and P. McKenzie

Proposition 4.10. Let (X, − ⇔S , ∈) be a WSTS. If S has strict monotony, then

so does S. However, if S has strict monotony then S doesn’t necessarily have it.

Proposition 4.11. There exists a post-effective WSTS whose completion is not

post-effective. Conversely, there exists a non post-effective WSTS whose comple-
tion is post-effective.

4.3 Post-effectiveness of Completions in Concrete Examples

An affine net S is a WSTS in Fd in which the recursive functions are affine

and a Petri net can be seen as an affine net where all matrices are the identity.
An ω-Petri net [17] is an (extended) Petri net in which arcs can be labelled by
positive integers or by ω. The completions of affine nets, ω-Petri nets and Lossy
Channel Systems can be shown post-effective.

5 Decidability in Infinitely Branching Post-effective

WSTS

5.1 (Strong) Termination

We are able to strengthen the hypotheses of Theorem 3.2 and to obtain: ter-
mination is undecidable, even for post-effective ω 2 -WSTS with strong and strict
monotony, and with post-effective completion by reducing from structural ter-
mination for Transfer Petri nets [9].
When a WSTS is infinitely branching, its termination problem differs in a
subtle way from its strong termination problem. We show the latter decidable
under suitable hypotheses:

Theorem 5.1. Strong termination is decidable for ω 2 -WSTS with transitive

monotony and post-effective completion.

Proving Theorem 5.1 requires comparing executions in a system with execu-

tions in its completion:
k
 If I −
⇔S , ∈) be a WSTS, and I, J ≡ X.
Proposition 5.2. Let S = (X, − ⇔S J,
then for every xJ ≡ J there exists xI ≡ I, y ≡ ≤ xJ and k  ≡ N such that
k
xI −⇔S y. Moreover, if S has transitive monotony then k  ↔ k; if S has strong
monotony then k  = k.
k
Proposition 5.3. Let S = (X, − ⇔S , ∈) be a WSTS and x, y ≡ X. If x − ⇔S y,
k
then for every ideal I ⊇ → x there exists an ideal J ⊇ → y such that I −
⇔S J.

Proof sketch of Theorem 5.1. Consider a ω 2 -WSTS S = (X, − ⇔S , ∈) such that S

is post-effective. Finkel and Schnoebelen [11, Theorem 4.6] show that termina-
tion, and thus strong termination, is decidable for post-effective WSTS having
 Handling Infinitely Branching WSTS 21

transitive monotony. By hypothesis, S is a WSTS and S has strong (and tran-

sitive) monotony by Prop. 4.8. Therefore, strong termination for S is decidable.
From Prop. 5.2 and Prop. 5.3, no bound on the length of executions from x0
 Hence
exists in S iff no bound on the length of executions from → x0 exists in S.
decidability of strong termination from x0 in S follows from being able to decide
 Note that we have implicitly assumed that a
strong termination from → x0 in S.
representation of → x0 can be effectively computed. ∪


5.2 Boundedness
Drawing from [8], we know that boundedness is undecidable, even for finitely
branching post-effective ω 2 -WSTS with strong (but not strict) monotony and
post-effective completion. Petri net with reset arcs are such a class.
It is known that for finitely branching post-effective WSTS with strict tran-
sitive monotony and a well partial ordering (wpo), the boundedness problem is
decidable [11]. We generalize this result to (possibly) infinitely branching WSTS
and we note that the hypothesis of transitive monotony was not necessary in the
proof of [11]. The proof follows [11] by building a finite reachability tree, with
the extra step of testing whether PostS (x) is infinite for each new node.

Theorem 5.4. Boundedness is decidable for post-effective WSTS with strict

monotony and with well partial ordering.

5.3 (Strong) Control-State Maintainability

By a reduction from the termination problem, the hypotheses of Theorem 3.3
can be strengthened: control-state maintainability is undecidable, even for post-
effective ω 2 -WSTS with strong and strict monotony, and with post-effective com-
pletion. By contrast, the strong variant of the problem introduced in this paper
is decidable, under suitable hypotheses, for infinitely branching WSTS:

Theorem 5.5. Strong control-state maintainability is decidable for ω 2 -WSTS

with strong monotony and a post-effective completion.

Before proving Theorem 5.5, we need Prop. 5.6 and Prop. 5.7 to relate covering
executions in a WSTS to covering executions in its completion.
Proposition 5.6. Let S = (X, ⇔ − S , ∈) be a WSTS with strong monotony and
{t1 , . . . , tn } ⊂ X. Let I0 −
⇔S I1 −⇔S . . . −
⇔S Ik be an execution such that for all
0 ∈ j ∈ k we have Ij ≡ ≤X {→ t1 , . . . , → tn }. Then for every y ≡ Ik there exists
an execution x0 − ⇔S x1 − ⇔S . . . −
⇔S xk such that x0 ≡ I0 , xk ≡ ≤ y and for all
0 ∈ j ∈ k we have xj ≡ ≤ {t1 , . . . , tn }.

Proof. Let I0 be an execution of length 0 in S as described in the proposition,

and let y ≡ I0 . By hypothesis, there exists ti such that → ti ⊂ I0 and thus ti ≡ I0 .
Since I0 is an ideal, there exists x0 ≡ I0 such that x0 ↔ y and x0 ↔ ti . Therefore
the execution x0 of length 0 in S meets all requirements.
 Another Random Document on
Scribd Without Any Related Topics
nor his coat were so exaggerated as his friend’s, but all the world did not
turn to look at him; he had no fashionable mistress upon whom he
squandered money, but he strove to be on the best terms with the mistresses
of his friends; he did not waste his fortune, because he had none, and he had
not left his place, because he had never had any.
However, as one must needs try, in society, to have some talent, some
profession, or some rank, in default of fortune, Jéricourt had become an
author. He had not stopped to consider whether he had the necessary
vocation and intellect for that; he had said to himself: “I propose to be an
author;” and as one ordinarily effects his purpose by dint of perseverance
and unbounded self-assurance, Jéricourt, by persistently frequenting the
café where the young men who write for the stage ordinarily gather, had
insinuated himself among them, playing billiards with one, dominoes with
another; he had become one of their intimates, and then had begun to talk of
plays, of plots, of original ideas which he claimed to have had; and when
someone would say to him:
“That is old, that subject has already been treated fifty times!” he would
exclaim:
“I don’t see why it shouldn’t be treated fifty-one times! A thing that has
succeeded so often will succeed again. It is mere folly to try to do
something new; one risks failure; whereas, by following roads already
marked out, one is certain to arrive without hindrance.”
Jéricourt found people of his opinion; and thus it was that he became an
author by revamping what others had done before him. And he ended by
believing himself to be an inventor, a man of genius, and by making idiots
of the type of his friend Saint-Arthur believe it also. The number of fools is
infinite!
“Well, my pretty flower girl, I must have a wonderful, a stupendous
bouquet!” said Alfred, halting in front of Violette; “it’s for a lady who
knows what’s what, and who has already had the most beautiful bouquets
that are made in Paris,—isn’t that so, Jéricourt?—Sapristi! I haven’t a cigar;
Jéricourt, my dear fellow, make me a cigarette, will you?”
“You don’t like them.”
“Ah! it is true that I have become so accustomed to panatelas—I say!
look at that little woman yonder! She turned around to look at me. If I
weren’t in such a hurry, I’d follow her.”
 “Aha! would you be unfaithful to Zizi Dutaillis?”
“Oh! pardieu! a little amourette of a moment.—Make me a cigarette.—
Well, flower girl! you don’t show me anything.”
“Why, monsieur, you see what I have; choose for yourself.”
“Choose for myself! why, all this stuff is horrible! these bouquets are
good for nobody but circus riders! I don’t want any of these. I told you that
I wanted something wonderful, such a bouquet as never yet was seen.”
“I will make you up one!”
“All right! but hurry. Zizi is waiting for me, and she doesn’t like to wait;
her nerves are all upset when I am late.”
“Here’s your cigarette.”
“Thanks, my boy. Have you fire?”
“Always, when I am before this fascinating flower girl. Pray look at
those eyes! did you ever see anything more alluring?”
“True! for a flower girl’s eyes, they are very fair.”
“And that nose, that mouth, and that cruel air, which would be so
becoming to her if it were not genuine!”
“Ah! mademoiselle is cruel, is she?”
“Alas! yes, my dear Alfred.—Would you believe that for nearly a month
I have been sighing at her feet, and without making any progress?”
“The deuce, my boy, you don’t know how to go about it! You tempt me
to try my hand with the flower girl. If I should undertake it, I will wager
that the affair would go faster,—eh, my girl?”
As he spoke, Alfred tried to take Violette’s arm; but she struck him
across the fingers with a bunch of roses and lilacs that she held; and as there
were some thorns in the bunch, the young dandy made a wry face as he
withdrew his hand.
“Bigre! she has scratched my fingers! Is this flower girl a Lucretia?”
“I told you that she resisted me, and yet you choose to meddle!”
“She plays the prude; but if I had time! I am terribly afraid of being late.
Zizi will be angry; she plays to-night, and she is much more nervous when
she plays! You are coming to dine with us, Jéricourt, are you not?”
“Impossible.”
“Bah! why impossible?”
 “Because I dine with this lovely girl, with the flower girl.—Isn’t it true,
Violette, that you will dine with me to-day?”
“Monsieur, I thought that I answered you the other day in such a way
that you would not give me any more such invitations.”
“My dear love, you are too fascinating to remain virtuous long; why
shouldn’t you give me the preference? I will give you your own apartment,
pretty furniture, pretty dresses; the theatre every evening; that’s the life that
awaits you!”
“I prefer to sell my bouquets, monsieur.”
“That’s absurd! Unless you have some passion that closes your heart to
me, you ought to yield to me.”
“No, monsieur, I feel no sort of obligation to you.”
“Ha! ha! ha! poor Jéricourt! he fails in his suit to a flower girl! That will
make Zizi laugh! I’ll tell her at dinner.—I say, my pretty girl, don’t make
my bouquet all white, please. The other day I offered one like that to Zizi,
and she declared that it looked like a cauliflower.”
“There, monsieur, how is this? Do you like it?”
“Why, yes, it isn’t bad; it has some style! I think that it will produce an
effect.—Come,—Jéricourt, as Mademoiselle Violette refuses to dine with
you, it seems to me that you can accept my invitation. If I don’t bring you,
Zizi will be sulky; she is much livelier when you are there; that is easily
understood, for you make her laugh, you make puns, and she declares that
there is no such thing as a good dinner without puns.”
“I tell you again, Alfred, that Mademoiselle Violette will not be
inexorable; why, I propose to launch her in society, to make her the fashion,
for I have all the small newspapers at my disposal.”
“He is telling you the truth, my girl, and the small newspapers are the
only ones that are read nowadays, for they are much more amusing than the
large ones. For my part, I know nothing better than the Tintamarre! Dieu!
the Tintamarre; there’s a newspaper that always drives away the blues! I
learn puns from it and I repeat them to Zizi; but unfortunately I don’t
remember them very well, so that she doesn’t understand them.—Ah! what
a beautiful bouquet!—Well, my dear fellow, will you come?”
And the pretty young man with the light whiskers, holding his enormous
bouquet in one hand, tried with the other to lead away his friend, who, half
leaning over the flower girl’s counter, was gazing at her with his face close
to hers, although she did her utmost to move away from him.
It was at this moment that Chicotin Patatras, who had spied one of his
cronies a few steps from Violette’s booth, ran to him and tripped him up,—a
method of beginning a conversation decidedly fashionable among street
urchins. The friend, taken by surprise, fell upon the sidewalk, and as he
rose, saw Chicotin laughing and making fun of him, and apparently
challenging him to retaliate. He immediately started to run after him, which
was what young Patatras hoped that he would do. When he saw that his
comrade was about to overtake him, he jumped back in such a way as to
collide with the persons who stood in front of Violette’s booth.
Chicotin had hoped to fall on Jéricourt, but having failed to calculate the
distance accurately, he collided violently with the young dandy, Alfred de
Saint-Arthur. The shock was so sudden and so unexpected by Alfred, whose
back was turned to the passers-by, that he fell forward with his face against
the flower girl’s wares; and as the counter was not strong enough to hold
the weight of his body, it collapsed under the young man.
Violette uttered a loud exclamation when she saw her flowers scattered
over the concrete, and Monsieur de Saint-Arthur apparently trying to swim
among them.
Jéricourt, taken by surprise by this unexpected mishap, also received a
kick or two from Chicotin, for the latter, still pursued by his comrade, who
finally overtook him, began with him a struggle which was all in jest, but in
which, although they were merely fooling, the young men dealt each other
blows so lustily applied that everybody who was near them received some.
“Will you stop, or go somewhere else and fight, you clowns?” cried
Jéricourt. “Just look at what you have done! All of the girl’s flowers are on
the ground!”
“Ah! it’s that good-for-nothing Patatras again!” said Violette; “he is
always doing something of the kind; he must always be making trouble
somewhere! It’s outrageous; I am going to complain to the inspector.”
“Oh, I beg pardon, excuse me, Mamzelle Violette,” said Chicotin, rising,
“you must know that I didn’t fall against your stall on purpose. It’s
Chopard’s fault; why did he chase me when I tried to get out of the way
behind your customers?”
 “Why did you come and trip me up when I wasn’t saying anything to
you?”
“I’ll do it again when I choose, you long bobêche!”
“Oh, yes! just come and try! I’ll show you!”
“Sapristi! are you going to begin again, you scoundrels, instead of
picking up these bouquets which you knocked down?”
“But your friend has fallen too, monsieur, and he doesn’t get up!—Help
him! Perhaps he’s hurt himself!”
At these words from Violette, Jéricourt condescended at last to pay some
attention to his companion; with the assistance of Chicotin, he succeeded,
not without difficulty, in placing him on his feet; for Alfred was almost
suffocated; two rosebuds had been forced into each nostril, and had entered
far enough to close them hermetically; and as he had in addition a bunch of
gilly-flowers over his mouth, he could not breathe at all, and was beginning
to turn purple. Once upon his feet, he opened his mouth as if he proposed to
swallow everything in his neighborhood, and shook his head to try to rid
himself of the two rosebuds, whose thorny stems tickled the lower part of
his face unpleasantly. But he could not succeed; Jéricourt had to pull one of
the stems and Chicotin the other, to uncork his nose. This operation was not
performed without a number of shrieks from Monsieur de Saint-Arthur, but
his nose at last recovered its air current, and everybody’s mind was at ease.
When the young dandy recovered all his faculties, the thing that troubled
him most was that he had broken one of his suspenders, and that his
trousers on that side were not held in place.
“All sorts of misfortunes at once,” cried Alfred; “I have broken my left
suspender. But who was it, then, who came down on me like a bomb and
pushed me onto that counter?”
“Excuse me, master, my excellency, I did it by accident, and not on
purpose, for I was fooling with Chopard.”
“What, you scoundrel, was it you?—Ah! I recognize you; I have
employed you more than once.”
“Oh! I remember very well! You are one of those generous and
distinguished gentlemen that a man doesn’t forget. I have often opened your
carriage door, master, and you are always with such pretty ladies, ladies
from the theatre, and so well dressed, that everybody looks at you. Shall I
wait at Monsieur Bonvalet’s, master, to see if you want to send me to find
out how far they’ve got in the play?”
“All right, all right, we’ll see. After all, as he didn’t do it on purpose—
And my bouquet, what became of that in the scrimmage?”
“Here it is, monsieur,” said Violette; “luckily nothing happened to it.”
“It’s my broken suspender that worries me most; my trousers are all
creased on that side! I’d give thirty francs for a pair of suspenders.”
“Would you like mine, master?”
“No, thanks! That would look nice!”
“I’ll go and buy you a pair at the druggist’s on Rue du Temple.”
“What does the idiot say?” muttered Jéricourt; “suspenders at a
druggist’s! do you propose to buy them made of marshmallow paste?”
“At all events I can’t stop here any longer,” cried Alfred; “Zizi will make
a horrible row; she will be in an infernal humor; and if she sees that my
trousers are creased, it will be much worse! And she will see it, for she
always looks at them first when I join her; she is so particular about dress;
she said to me once: ‘A man who doesn’t have morocco straps to his boots
shall never step foot inside my door!’—Well, Jéricourt, are you coming?”
Tall Jéricourt decided at last to go away with his friend; for the flower
girl, busily engaged in picking up her flowers, did not seem disposed to
laugh, and he saw that he must needs abandon the idea of being listened to
for that day at least. So he walked away, arm-in-arm with Alfred de Saint-
Arthur, who, as he walked, did his utmost to hold his trousers up. When he
saw the two young men take their leave, Chicotin Patatras nodded his head
to Georget, who was not very far away, and who answered with a smile.
And Violette, as she tried to replace her flowers in order upon her counter,
did not fail to notice that pantomime.
 V

A CONCIERGE’S LODGE

In a house of respectable appearance on Rue d’Angoulême, about half-

past eleven one evening, the street bell was pulled so violently that it caused
Monsieur Baudoin, the concierge, to leap from his chair, upon which he was
beginning to doze, while his wife Hildegarde took advantage of his nodding
to open a small cupboard and take therefrom a bottle, the neck of which she
proceeded to introduce into her mouth, and took several swallows of a fluid
which she seemed to enjoy greatly.
Baudoin the concierge was a tall, thin man, with a pale face and light
hair, who had passed his fiftieth year, but was still very straight, and as
active as a young man. To his occupation of concierge, he added that of
clerk in a stage office, which kept him only until six o’clock. He was an
honest man, to whom one could fearlessly entrust his house and his
treasure; he did promptly whatever he was ordered to do, unless he did not
fully understand; but in that case it was not safe to reproach him, for
Baudoin lost his temper very readily, having an immeasurable self-esteem
and claiming that he never made a mistake. When he did lose his temper,
Baudoin swore like a trooper, and turned as red as a turkey-cock.
Hildegarde, the concierge’s wife, was two or three years older than her
husband; she had once been pretty and sentimental; she was not very well
preserved, and her inclination to sentiment having with age become
diverted to brandy, Madame Baudoin had neglected herself considerably;
there was a deplorable carelessness in her dress, which resulted in nothing
ever being in place. Baudoin, who was always neat and decently dressed,
often reproached his wife for her heedlessness in that respect, and as he had
also discovered her unlucky fondness for liquor, he sometimes added to his
reproaches lessons of an impressive sort, which made Hildegarde bellow
loudly, and promise never to give way again to her miserable failing; but
she never failed to forget that promise, whenever she thought that her
husband would know nothing about it.
 Moreover, Madame Baudoin was a genuine type of concierge: talkative,
inquisitive, gossiping, scandal-loving, incautious, not evil-minded at
bottom, but capable of setting the whole quarter at odds with remarks made
without ill-intent. Her husband often scolded her for it; but “what’s bred in
the bone will come out in the flesh!”
At the jingling of the bell, and the somersault performed by her husband,
Hildegarde, bewildered, and realizing that she had not time to replace the
bottle of brandy in the cupboard, hastily put it on the floor between her feet,
and then sat down, thus having her dear bottle in the place where the open
air tradeswomen put their foot-warmers.
“Didn’t someone ring?” said Baudoin, rubbing his eyes.
“Yes, my dear, yes, someone certainly did!” replied his wife, without
moving from her chair.
“Well, then, draw the cord, Hildegarde; you’re right near it.”
“You can draw it much easier yourself, my dear; you have only to put
out your arm and lean forward a bit.”
“Look here, why won’t you pull the cord, when you’re right beside it?
What sort of way is that to behave?”
“Why—why, you see I pulled it just now when the tenant of the first
floor came in, so it’s your turn.”
“Ah! so this is a new idea you’ve got into your head! Madame is afraid
that she will pull the cord oftener than I, who have just come in, all tired
and out of breath after running about Paris all day! What a lazy hussy!”
“Mon Dieu! is it possible for a man to be as ugly as this to his wife! to
be so unwilling to do anything for her! Ah! Monsieur Baudoin, how you
have changed!”
While this little dialogue was taking place between the couple employed
to take care of the house, the person who had rung remained at the door,
which is rarely pleasing when one returns home at night. A second peal at
the bell, much more violent than the first, announced that he was losing
patience.
Baudoin decided to pull the cord, but as he did so he said to his wife:
“Ah! bigre! you shall pay for this, Hildegarde! by all that’s good! I
promise you that.”
 Hildegarde made no reply, but continued to sit over her bottle. Someone
came in and closed the street door; and soon a man appeared at the
concierge’s lodge, and said curtly:
“Give me my light.”
“Oh! yes, Monsieur Malberg; this minute, Monsieur Malberg.—
Hildegarde, just take Monsieur Malberg’s candlestick from the cupboard by
you, light it at our lamp, or rather light it with a match, for the chimney of
our lamp is cracked and it might break in your hand.—You are just from the
theatre, I suppose, Monsieur Malberg? They say that they are giving a fine
play there just now; I don’t know which theatre, but no matter, it seems that
it’s fine, all the same! You have been to see it, of course?”
“I have been where I chose to go, and it is none of your business,”
replied the tenant, in a tone which did not invite further conversation.
“Well! what about my light? Are you going to give it to me to-night? or do
you intend to keep me waiting here as long as you did in the street?”
“What, Hildegarde, haven’t you lighted Monsieur Malberg’s light yet?
Look here, what are you about? God forgive me, Monsieur Malberg, but I
believe that my wife is getting deaf or idiotic; something’s the matter with
her to-night; it isn’t possible—yes, she may have been tippling. You know
her unlucky failing, which will lead her to perdition! and it isn’t for lack of
my trying to correct it by every means that I can think of.”
Whereupon Hildegarde, who had her reasons for not stirring from her
chair, made haste to reply:
“Oh, yes! the means you use are very nice! I advise you to boast of
them; you ought to be ashamed of them! a man with an education, who has
clerks under him, in an office, to raise his hand to his wife! Yes, Monsieur
Malberg, I don’t blush to confess that Monsieur Baudoin has the baseness
to strike me! that’s a nice thing to do, ain’t it?”
But the man to whom these questions were addressed, observing that no
one thought of giving him his light, pushed open the door of the lodge, took
his candlestick, lighted the candle with a piece of paper, and went upstairs
without another word to the concierge and his wife, who continued their
conversation.
“Well, Hildegarde, do you see what you’ve done? Here’s Monsieur
Malberg had to light his candle himself! what will he think of us?”
 “Oh! I don’t care what he thinks! he’s an agreeable man, that tenant! a
fellow who never talks, who hardly answers when you speak to him, and
always in a short, surly tone, as if he was always angry!”
“It is true that he seldom laughs; but still perhaps that’s his nature; there
are people who enjoy being dismal. However, he’s a man who occupies an
apartment at eleven hundred francs, and who pays on the dot, without
having to be reminded that it is rent day, and who has very handsome
furniture, and mirrors in every room, so that the proprietor has a very high
regard for him.”
“Oh! I don’t say that he’s a vagabond! but why doesn’t he keep a maid,
who’d come to our lodge in the evening and talk, as decent people always
do, instead of that miserable blackamoor, that yellow negro, who doesn’t
know how to do anything but wax his floor and polish his boots? as if you
could call that a servant! He ought to hire me to do his housework; that’s
my line!”
“You forget, Hildegarde, that the landlord doesn’t want you to do
housework. Of course, if you went away while I am at my office, there
wouldn’t be anybody but the cat to look after the lodge and answer
questions!”
“A fine job this is, where the concierge’s wife isn’t allowed to do
housework! That was my only ambition.”
“Oh, yes! the fact is that you were the cause of our being discharged
from the lodge we had before this, because you did housework for the men
on the fourth floor and drank all their liquor.”
“That isn’t true, it’s a slander!”
“Let’s not go back to that. I am mortified that Monsieur Malberg had to
light his candle himself; it’s a stain upon our good name.”
“Well then, you ought to have lighted it for him, if you have that on your
conscience!”
“Hold your tongue, Hildegarde; you’re very unreasonable to-night, you
have something bad to say about everybody. You find fault because
Monsieur Malberg has a yellow negro to work for him, and you don’t seem
to know that that is very distinguished. Swell people always have colored
servants in their employ.”
 “It’s a miserable fashion. But still, if that miserable Pingo or Ponceau—I
never know what his name is—was only agreeable.”
“Pongo!”
“Oh! what a dog of a name! Pongo! But he never talks, the blackamoor;
or else he talks to himself, and says things that I don’t understand; I believe
that he talks Morocco!”
“Come, Hildegarde, it’s almost twelve o’clock; go to bed, that’s the best
thing you can do.”
“Everybody hasn’t come in.”
“Yes they have, everybody except little Georget, who lives up under the
roof, with his mother.—By the way, how is the poor woman to-day?”
“Not very well; she’s had more fainting fits this afternoon, and I thought
she was going to put out her gas.”
“And her son hasn’t come home, at midnight! that’s what I call a ne’er-
do-well, a downright scamp! Hildegarde, heaven didn’t give us any
children, and I give thanks for it in my heart; because they aren’t always
honey for parents, and often absinthe rules the roost, as I see in the case of
Mère Georget!”
“Absinthe—absinthe—I don’t hate that! it helps the digestion!”
“Oh! bless my soul! you don’t hate any liquid; but I know that absinthe
is bad for the health; I’ve heard some of the clerks at the office talking
about two talented actors who played at the theatre and who put an end to
themselves with absinthe; without counting several others who are in a fair
way to do the same thing!”
“Bah! that’s all nonsense!”
“Come, Hildegarde, go to bed; I will come in a little while; and if little
Georget isn’t in at the quarter, I will leave him outside; I can’t waste my oil
for anyone who never makes it up to me. Well, you don’t move; are you
fastened to your chair to-night?”
“Go to bed first, Baudoin; I’ll sit up for the young man, and put the
lodge in order.”
“You know very well that I am not in the habit of going to bed before
you. I see your scheme: you will wait until I am asleep and then go to the
cupboard to say a word or two to the bottle!”
 “Oh! the idea of my going to the cupboard! It’s much more likely to be
you, for you like brandy too.”
“I like it reasonably, like a man with some self-respect, who doesn’t
choose to make a brute of himself.—Hildegarde, go to bed.”
“I don’t feel sleepy.”
“Hildegarde, we are going to have trouble! Will you go to bed at once?”
“You pester me——”
“Hildegarde, I shall be compelled to resort to severe means. Why, you
certainly are glued to your chair; this isn’t natural, I suspect some trick. Ah!
I see! I’ll bet that the bottle isn’t in the cupboard.”
And Baudoin rose to go to the cupboard, but as his wife was sitting in
front of it and did not move, he pushed her roughly aside, whereupon she
reeled, and almost instantly uttered a cry of distress so heartrending that her
husband feared that he had hurt her. But it was not Hildegarde who was
hurt, it was the bottle under her skirts, which she had involuntarily upset,
and which had broken, overflowing the lodge with all the liquor which it
contained.
“I say! what’s all this?” cried Baudoin, when he found a stream flowing
between his feet; but soon the odor which spread through the room left him
in no doubt as to the identity of the liquid.
“It is brandy; she had the bottle under her skirt; what a vile trick!”
“Yes, and you made me break it! that’s the worst of it, you brute! Such
splendid brandy!”
“Hildegarde, you persist in your debauchery; I am going to give you a
taste of the broomstick.”
“Touch me if you dare! I’ll call the watch! I’ll make a disturbance in the
house!”
Meanwhile Baudoin, who was in the habit of keeping his promises, had
gone to fetch the broomstick. At that moment, the bell at the street door
rang, and this time the woman made haste to open, hoping that it was
somebody who would protect her.
It was Georget, the young messenger of the flower market, who entered
the house, and in another instant the porter’s lodge, just as Baudoin raised
his broomstick over the head of his wife, who ran behind the young man,
crying:
 “Oh! monsieur, save an unfortunate woman, whose husband is trying to
murder her!”
“Sapristi! how strong it smells of brandy here!” said Georget, sniffing;
then, leaping upon the broomstick which the concierge held, he seized it
with both hands. But Baudoin held on, he would not let go, and a struggle
began between him and the young messenger, remarkably like the battles
around the flag, which we see in the war plays at the boulevard theatres;
only in this case the flag was a broomstick and the combatants were not in
uniform.
The struggle continued for some time, on nearly even terms; Baudoin
was stronger and little Georget more active. The concierge’s wife paid no
heed to the contestants; she had taken a small sponge, and was using it to
soak up the brandy from the floor; and when it was well saturated, she put it
to her lips.
Suddenly the broomstick broke, each of the contestants fell backward,
and the battle was at an end. Finding himself then on a level with Madame
Baudoin, who was kneeling on the floor with her body bent forward, still
soaking and sucking her sponge, Georget could not restrain a burst of
laughter; and the concierge, who was inclined at first to belabor his wife
with what remained of his broomstick, suddenly decided to lie down on the
floor, and to lap up the brandy with his tongue as thirsty dogs lap up the
water in the gutter.
 VI

THE GENTLEMAN OF THE THIRD FLOOR

Georget left Monsieur and Madame Baudoin fighting over the remains
of the brandy with sponge and tongue, and lighted one of the small, thin
candles which are rolled up like small rockets and which are sold for one or
two sous at the grocer’s. Then running quickly up six flights of stairs, he
reached a small door in which the key had been left; poor people are not
suspicious, especially as they have nothing which is worth the trouble of
stealing.
The young messenger walked through a small room, which received no
light except through a little round window, in which room was a cot bed
supplied with a very thin mattress and with an old window curtain which
served as bedclothes. This was Georget’s bedroom; but he did not stop
there. Opening the door at the end of the room, and trying to make no noise,
he entered another much larger one, where there was a little window. This
room, although the walls sloped, was large enough to contain a bed
surrounded by white curtains, an old mahogany bureau, a white wood table,
a small sideboard, several chairs, and on the mantel a tiny mirror
surrounded by a branch of consecrated boxwood. All this was more than
modest, but it was neat and clean; it indicated not destitution, but poverty.
Georget was walking very softly, concealing his light with his right
hand, when he heard a feeble voice from the bed:
“Is that you, Georget?”
“Yes, mother, it’s I. So you’re not asleep?”
“No, I haven’t been able to go to sleep, I don’t know why.”
“It must be because you are sicker; and you have not been well for
several days, although you didn’t admit it to me.”
“It’s nothing, just the lumbago, it will soon be gone. If you would just
give me something to drink, my dear, for I am very thirsty.”
“Yes, mother, in a minute. Wait until I light your candle and put out this
tallow thing of mine which smells worse than thirty-six lamps.”
 After lighting a bit of candle stuck in a bottle, Georget approached his
mother’s bed.
“Come, now you must tell me where your medicine is. But gracious
heaven, how red your face is, mother! and black circles round your eyes!
Are you worse?”
“Why, no, it is the heat of the bed that does that.”
“Give me your hand, and let me feel your pulse. Oh! how hot your hand
is! You are feverish, and very feverish too, I am sure.”
“Nonsense, as if you knew anything about it.”
“Oh! yes I do; you must have pain somewhere.”
“No, I am not in pain.”
“The first thing in the morning I’ll fetch a doctor.”
“I don’t want you to; what’s the use of a doctor, just because one has a
little fever; it will go away all right without him!”
“Where is your drink?”
“I thought I wouldn’t make any; I prefer water, I like it better.”
“Water, when one is feverish! Why, you didn’t do right! If I did such a
thing, you’d scold me and say that I acted like a child, and you would be
right. However, tell me where the sugar is; where do you keep the sugar,
mother?”
“Sugar! I don’t want any; it nauseates me; I prefer clear water.”
“Water without sugar, when one is as burning hot as you are! I never
heard of such a thing! You can’t mean it! Do you want to kill yourself? I am
going to give you some sweetened water, but I must heat it; it’s better so.”
“Oh, no! no!”
“I say yes; I will take care of you better than you do yourself. Come,
where is the sugar? where is the coal?”
And Georget ran all about the room, fumbling over all the furniture,
opening all the closets, looking in every corner, but he found nothing.
Thereupon a bright light broke upon his mind. He stopped in the middle of
the room, threw his cap on the floor, and cried in a tone of deep distress:
“Ah! I understand everything now! You didn’t make yourself any drink,
because you had no coal nor charcoal! You don’t want any sugar, because
you haven’t a single particle of it here! Yes, yes, that is it! You are out of
everything! You haven’t any money either, I am sure! and I, instead of
trying to earn some, so that you might have what you need, why, I do
nothing at all! I pass my day loafing, and at night I go to the theatre with
Patatras, who absolutely insisted on treating me. I go about enjoying myself
when my mother is sick, and I come home without a sou, without a single
piece of money; and I haven’t anything to buy her what might cure her! Ah!
I am a wicked son, a good-for-nothing! Forgive me, mother, forgive me; I
won’t do so any more! I will work, I swear to you that I will work now!”
And the young fellow fell on his knees beside his mother’s bed; and the
poor mother forgot her suffering, and tried only to comfort her son.
“What are you talking about, Georget? You, a good-for-nothing! Why,
you don’t mean that, my boy! Have I ever complained of your conduct?”
“Oh! I know very well that you never complain; you are too good!”
“You have been amusing yourself a little to-day; well, my boy, there’s no
harm in that; you must enjoy yourself while you’re young. Your friend
Chicotin took you to the theatre; the theatre is a decent amusement; it is
much better than going to the wine shop; you don’t make evil acquaintances
there, or destroy your health with unhealthy stuff that they give you for
wine. You haven’t earned anything to-day—that is too bad, but to-morrow
you will work, and you will be happier!”
“To-morrow! to-morrow! but you have had nothing to drink this
evening; you haven’t any sugar; and what will you take to quench your
thirst to-night? cold water, I suppose?”
“I am going to try to sleep; when one is asleep, one doesn’t need to
drink.”
“But when you wake in the morning, what shall I give you? for you have
no money here, nothing at all; isn’t that so, mother?”
“Dear me, yes, my dear; for unfortunately I haven’t been able to work
for a week; my sight has been all blurred!”
“Oh! you work too much, when you ought to rest.”
“Why so, Georget? I am not old enough yet to give up work, I am only
fifty-four! If a body was good for nothing at that age, it would be a great
pity!”
“I know that you’re not old, but still your health isn’t very good, and
then you didn’t use to need to work for your living.”
 “Oh! my dear, we must never say such things as that, and sigh over the
past! If one has been happy, so much the better; if one is so no longer, so
much the worse; regrets don’t help and only make our position worse!”
“All this doesn’t give you any nice, hot, sweet drink, and that is what
you must have!”
“Don’t despair, we are not altogether without resources. You know that I
have—your father’s watch; and if it is absolutely necessary, why——”
“What’s that you say! my father’s watch, which you think so much of!
the only thing of his that you have! part with that? No, I won’t have it. Wait
—suppose I should go and stand in front of some theatre?”
“What an idea! They are all over, all closed at this time of night.”
“Never mind; in front of a restaurant, I may still get something to do.”
“I don’t want you to go out; it’s too late.”
“Well then, in the house; pardi! sugar and coal—people lend each other
such things as that. Don’t be impatient, mother, I will come right back.”
“No, Georget, I don’t want you to ask the neighbors; don’t, I beg you!”
But Georget was not listening to his mother; he had already relighted his
tallow-dip and hurried from the room. When he reached the landing, the
young man stopped, for he was uncertain to whom he should apply for the
loan which he wished to obtain; but he did not hesitate long. He ran down
the stairs three or four at a time, and did not stop until he reached the
ground floor and knocked at the concierge’s lodge, saying to himself:
“Baudoin and his wife are not unkind; they dispute together, and fight
sometimes, but they haven’t bad hearts; they won’t refuse me. Besides, it is
only a loan, I will return it all.”
But Georget forgot that he had left the concierge and his wife engaged in
an occupation which was likely to plunge them into a profound slumber ere
long. In fact, after sucking and lapping brandy for some time, the husband
and wife had felt such an intense longing to sleep, that they had hardly
strength enough to reach their bed; and as the sleep caused by intoxication
is never light, Monsieur and Madame Baudoin did not hear the knocking at
their door; one might have fired a cannon under their noses, and they would
simply have said: “God bless you!”
Weary of knocking to no purpose, Georget walked away from the lodge,
murmuring:
 “Those concierges are regular beer kegs; I shall never succeed in waking
them, and I might as well give it up. Let me see, where can I apply? on the
first floor? The whole floor is occupied by a family of English people, who
hardly understand what you say to them and who don’t look very pleasant; I
should not be well received there, they would not be able to understand
what I said. On the second floor is a very pretty, very stylish lady, who
receives fine gentlemen, but who refuses to open her door when she has
company; her maid told me so the other day,—those are her orders. On the
third,—ah! that is the gentleman that they call the Bear in the house,
because he never talks with anybody, never receives any visitors, and hardly
answers when you speak to him. Much use there would be in trying to
borrow sugar and charcoal of that man! and still, if I thought that his black
man would open the door;—but no, Pongo sleeps so sound that he never
hears his master come home. So it would be the gentleman himself who
would open the door, and he would shut it in my face without answering
me. I don’t even dare to try!—On the fourth, on one side is an old woman,
so timid that she will never open her door after dusk; on the other, a student;
but he has gone into the country. And at the top, opposite us, no one; the
room is to let. Mon Dieu! whom shall I apply to then, if among all these
people I can see no hope of help for my poor mother, who has a high fever
and no cooling draft to lessen her suffering? Ah! Chicotin was quite right to
say that I am a fool to be in love, that I am too young. Mamzelle Violette
makes me forget my mother, my duty, my work. To think that I have done
nothing to-day! that I came home without a sou when I knew that my
mother was sick! Oh, I am a miserable, heartless villain! I shall never
forgive myself!”
As he said this, Georget went slowly upstairs, stopping frequently
because he was weeping; and he had stopped again, and rested his head
against the wall, to sob at his ease, when a door opened within two steps of
him.
He was then on the landing of the third floor, and it was Monsieur
Malberg who stood before him. When he saw the young messenger, who
still had the look of a mere boy, beating his head against the wall, and
giving free vent to his sobs, the gentleman who was called the Bear, and
who in fact had a rather stern expression and a rather rough voice, walked
toward Georget and asked:
“What are you doing here?”
 “Well! monsieur, you see, I am crying, I am unhappy.”
“And why are you crying?”
“Because my mother is sick, and because she has nothing that might help
her; because I didn’t work to-day, and came home to the house without a
sou; because I am a heartless wretch, and I deserve to be beaten!”
“Well, do you think that if you beat your head against the wall that that
will help your mother?”
“Oh! no, monsieur! but when a fellow doesn’t know which way to turn! I
went down and knocked at the concierge’s door; I wanted to borrow a little
sugar and some charcoal of them; but they didn’t answer; I suppose they
sleep too sound!”
“So you live in the house, do you?”
“Yes, monsieur, at the top, under the eaves; I live there with my mother,
who is the widow of my father, Pierre Brunoy, who was a soldier, a non-
commissioned officer, who left the service on account of a wound. Oh! he
was a fine man, was my father! He was a draughts-man, he had lots of
talent, and he used to make designs for ladies who embroider; we were
happy then; but he died. My mother undertook to keep a little smallwares
shop, to earn enough to educate me; but she didn’t succeed, for no one paid
her. Then, as she works very well on linen, she began to work for people,
and I, knowing that I ought to help mother, whose health isn’t very good,
and who has weak eyes, I said to myself: ‘I will be a messenger, for I could
never find a place, although I can read and write and figure; or else I should
have to work without pay for a long while and I must earn money right
away to help mother.’—So I started in as a messenger; for there isn’t any
foolish trade, so I was told;—and—that’s all, monsieur.”
The gentleman of the third floor listened attentively to Georget, and
when he had finished, said to him:
“Come with me.”
“Where, monsieur?”
“Into my room, of course.”
“What for, monsieur?”
“You will see; come.”
The youth placed his tallow-dip on the floor, and followed the
gentleman; his heart was still heavy, for he didn’t understand how the
person who occupied the handsome apartment on the third floor could need
his services so late. Monsieur Malberg passed through a reception room
very carefully polished, and into a beautiful dining-room. There he stopped,
opened a large sideboard, took out a loaf of sugar, which was hardly
touched, and placed it in Georget’s hand, saying:
“Take this!”
The poor boy looked at him with an almost dazed expression, and
murmured:
“What is this, monsieur?”
“Don’t you see that it is sugar?”
“Sugar, oh, yes! but this great loaf,—who’s all this for?”
“For your mother, of course! Didn’t you tell me that she hadn’t any and
that she was sick?”
“Oh! is it possible, monsieur, that you are so kind as—but this is too
much, monsieur, too much.”
“Hush, and come with me.”
This time Monsieur Malberg went into his kitchen, where Georget
followed him, holding the loaf of sugar in his arms. The gentleman pointed
to a large box without a cover, which stood under the stove, saying:
“Take that box; there’s charcoal in it.”
“Oh! how kind you are, monsieur! How can I thank you for——”
“It isn’t worth while, I don’t like thanks; take this box, I say.”
“Yes, monsieur, but be sure—I will return all this; pray believe me. Oh! I
will work to pay my debt.”
“Very well, very well! By the way, wait; I have some linden leaves here,
and some mallow; perhaps they will be good for your mother, and you
haven’t any in your room, I suppose?”
“No, I think not, monsieur.”
“Well, I’ll give you some then; come.”
Monsieur Malberg returned to his dining-room. Georget still followed
him, holding under one arm the loaf of sugar, and under the other the box
filled with charcoal. The gentleman opened the drawer of a small piece of
furniture, took out several paper bags, looked to see what they contained,
put two of them aside, and was about to give them to Georget, when he
stopped as if a sudden thought had struck him, and left the dining-room,
saying:
“I will return in a moment; what I want isn’t there; wait here.”
The young messenger was careful not to stir; he was so pleased that he
wondered whether he was not the play-thing of a dream; but he for whom
he was waiting soon returned, bringing several small packages of herbs,
saying:
“Here are some things which may be good for your mother,—linden,
orange leaves, mallow and violet; take them all, or rather let me put them in
your pocket, for you have no hand free.”
“Oh, monsieur! excuse me for the trouble I put you to. Mon Dieu! you
are too kind! I will pay you for this, monsieur; for we are not beggars, we
don’t ask alms, and I should be sorry for you to have that idea of us.”
“Very good! Your mother is sick and may need you; don’t leave her
alone any longer.”
“Yes, monsieur, you are right; my poor mother, she will be so happy, so
—so—Thanks, monsieur, oh, thanks a thousand times! Remember that I am
always here day and night, at your service.”
“I will remember; but go.”
And the gentleman pushed Georget before him, so that he soon found
himself on the landing once more. The door of the apartment closed, and he
reascended the staircase as quickly as he could, with his box of charcoal, his
loaf of sugar, and his tallow-dip still lighted.
At last he reached his room; this time he was not afraid of making a
noise when he went in; he was too happy not to wish to tell his mother
about it; but she was not asleep, and she gazed in amazement at her son
when he danced into the room, and placed the loaf of sugar on her bed,
crying:
“There, mother; you shall not drink plain cold water any more! Here is
sugar, here is charcoal, and in my pocket I have half a dozen herbs in
leaves. Ah! what luck! you will be cured right away! I can nurse you nicely
now.”
“What does this mean, my dear? where did you get all these things? You
hadn’t a sou just now. Explain yourself, Georget, I insist.”
 “Why, yes, yes, never fear, I am going to tell you the whole story; but let
me light the stove first, and then, while I blow my charcoal, I will tell you
how Providence came to our assistance. Where is the stove? Ah! there it is.
This will light very quickly, I know, although the bellows isn’t any too
good.”
“Did you get all these things in the house, my son?”
“Yes, mother; you see, first of all, I went down to borrow from the
concierge, Monsieur Baudoin; but it wasn’t any use for me to knock at their
door, I couldn’t wake them, they’re worse than deaf people. So then I was
coming up again in very low spirits, indeed, I believe I was crying, when
the door on the third landing opened, and the gentleman who lives there
came out to me. Oh! this thing proves, mother, that people very often say
foolish things, or that it’s very wrong to judge a person by his appearance.
For that gentleman that they call the Bear, that gentleman that never speaks
to anybody, and that everybody makes stupid jokes about, why, he took me
into his room, and gave me all these things for you, because I told him that
you were sick; and he didn’t even let me thank him!—Ah! you miserable
charcoal! you’ve got to burn! Now I am going to put some water over the
fire.”
“But, my dear, this is an enormous loaf of sugar, and it is almost whole;
you ought not to have borrowed so much as this.”
“As if that gentleman would listen to me! He says: ‘Take this!’ and if
you try to remonstrate, he shouts: ‘Hold your tongue!’ and it’s impossible to
prevent him from doing what he wants to.—Ah! my fire is going at last!”
“But this Monsieur Malberg—for the gentleman of the third floor is
named Malberg—I have never met him; what sort of looking man is he,
Georget? You must have had a good look at him, didn’t you?”
“Oh, yes! mother; why, he’s a man neither young nor old. At first sight, I
am sure that you would take him to be older than he really is, because when
a person never laughs, that makes him look older. He may be somewhat
over fifty years old; his face is not ugly, not by any means, but his features
have a sort of stern expression; his eyes are always gloomy and melancholy,
and there are great wrinkles on his brow; his eyebrows are heavy, and his
hair must have been black, but it’s a little gray now. When he fixes his great
brown eyes on you, it frightens you; and yet I got used to them, for his
expression is neither unkind nor contemptuous; it’s—I don’t know just how
to describe it—it’s sort of compassionate, or sorrowful; and his voice,
which sounds harsh at first, is much less so when he’s talked to you for
some time. You see, mother, that gentleman isn’t like most people; oh, no!
he makes you respect him, and it comes natural to obey him, and you don’t
dare to say anything.”
“Really, my dear, you make me long to know this gentleman; when I am
able to go out, I shall go to thank him. And did you tell him——”
“Just how we are placed, what we used to be, and what father did. Yes, I
told him everything. Did I do wrong?”
“No, my boy, we have done nothing which we need to be ashamed of or
to conceal.”
“Ah! my water is boiling; now I am going to make you some herb tea,
mother; which would you rather have?”
“Why, tell me first what you have in your pocket.”
“Wait and I’ll show you; I have a whole lot of bundles! Here, see what
this is.”
“Violet.”
“And this?”
“Linden leaves.”
“And—and—well! here’s something else now!”
“What’s the matter, Georget? Have you lost something?”
“Lost! oh, no! not by any means, mother! What I have just found in my
pocket certainly wasn’t there before! I am sure of that.”
“Why, what have you found in your pocket?”
“Here, look, mother!”
And the young messenger tossed upon the good woman’s bed four five-
franc pieces.
“Twenty francs, Georget! twenty francs! What does this mean? where
did you get all that money, my son?”
“I haven’t any idea, mother; and I am very sure that I didn’t have it when
I came home. I didn’t have a sou.”
“But this money didn’t get into your pocket of itself. Answer me,
Georget, and above all, don’t lie.”
 “Mon Dieu! how you say that, mother! Do you suppose that I am
capable of having stolen this money from someone, I should like to know?”
“No, my dear, I do not suppose that my son, that the child of my honest
Brunoy, would ever do a wicked action; but I have always carefully
preserved your father’s watch, and some time, without my knowledge, to
help me, you might have——”
“Pawned papa’s watch! Oh! never! I’d rather pawn myself! but wait,
mother; I remember now; yes, that must have been it.”
“What? tell me.”
“That gentleman on the third floor, when I had the loaf of sugar and the
box of charcoal in my arms, insisted on putting all these little bundles of
dried leaves in my pocket himself; and that’s the way he stuffed these five-
franc pieces into my pocket! Oh! I am sure of it now! for he went into his
bedroom alone, to get the money, no doubt. It was him, mother, it was him;
indeed, who else could have given me all this?”
“You are right, Georget, it can’t have been anybody else; people who
like to do good, think of everything, and it seems that he is very kind, this
Bear!”
“Yes, indeed, he is kind, but I shall not keep his money. I will work to-
morrow, and earn some; and he has put us under enough obligation by
lending us sugar and charcoal. Mother, we mustn’t keep these twenty francs
that he slipped into my pocket so slyly, so that I could not thank him, must
we? But still, it was very nice of him, all the same; he isn’t like other
people, that gentleman! I’ll bet that when he tosses a piece of money to a
poor man, he doesn’t try to make it ring on the sidewalk when it falls.”
“No, my dear, we mustn’t keep the twenty francs, for it is quite a large
sum, and it would be too hard to repay it.”
“I am going to take it back to the gentleman right away.”
“Oh! it must be quite late now; Monsieur Malberg has gone to bed, no
doubt, and is probably asleep; if you wake him up, he won’t like it. Wait till
morning, and when he’s up, you can take the money back to him, and thank
him again for both of us.”
“After all, you are right, mother; it will be better for me to let the
gentleman sleep, who has helped me to cure you. I will go to-morrow
morning, when his negro is up.—But the water is still boiling; give me what
I need for your tea.”
The invalid chose one of the herbs. Georget soon made the tea and
carried his mother a cup smoking hot and well sweetened; and when she
had drunk it, he filled the cup again and placed it on the table by the bed.
“If you are thirsty again in the night,” he said, “you must drink this; it
will be all ready; now try to go to sleep.”
“Yes, my dear, but it seems to me that I feel better already.”
“Well! mother; it is always like that; when a person has all that he needs
to get well, then the disease must go.”
“Oh! not always, my boy, for in that case rich people would never be
sick; but the thing that relieves one is contentment, happiness. It requires so
little to make poor people happy! and what has happened to us this evening
is real good fortune.”
“Oh, yes! it is a kind of good fortune that the rich do not know, but that
they can confer on others; and that must be a great pleasure too.—Good-
night, mother; if you need anything, call me.”
 VII

A DIFFICULT ERRAND

The next day, before six o’clock, Georget was up and dressed; he went
first to inquire concerning his mother’s health; the invalid had slept, and felt
better, although she was still too weak to rise. She smiled as she said:
“Up already, my dear?”
“I must earn a lot of money to-day, mother, in order to bring you all that
you need.”
“But I need nothing, as I have the material for making herb tea.”
“Oh! nobody knows! if you get better, perhaps a little beef soup won’t be
a bad thing for you. When a fellow is out on the boulevard early, he is more
apt to find work. There are maids who have bundles to send, people who
have to go into the country and are looking for a cab——”
“Poor Georget! what a miserable trade yours is! Knowing how to write
and figure as you do, you ought to have found a place in some office, or a
clerkship in some shop.”
“Oh, yes! and wait a year or two before earning any kind of a salary!
Don’t think about that any more, mother; I am very happy as I am! A clerk!
shut up all day in an office! oh! how sick I should get of that! then I should
never see her!”
“Who is it that you’d never see, my child?”
Georget blushed, but made haste to reply:
“I mean that I shouldn’t see you during the day, whenever I wanted to.
By the way, mother, I must go to see the gentleman on the third floor, the
gentleman who is so kind, although he doesn’t show it. I am going to return
his twenty francs.”
“Isn’t it a little too early? He isn’t up yet, probably.”
“Oh! I am very sure that he gets up early; he isn’t one of the kind to
coddle himself. Anyway, I’ll ask his valet, that mulatto who’s such a strange
creature, they say.”
 “Go, my dear, and thank the gentleman from me, until I can do it
myself.”
Georget cast a glance at the mirror to make sure that nothing was lacking
in his costume. When a man is in love, he becomes particular about his
looks, and Georget would have been very glad to please the pretty flower
girl of the Château d’Eau, who seemed to look upon him as a child; that
distressed the poor boy, he was sorry that he was not at least twenty years
old, because he thought that then she would pay more attention to him. For
we are never content with the passage of time; when we are young, we
think that it doesn’t move fast enough; later, we complain because it moves
too fast. And yet we know that the wisest course is to take it as it comes;
probably we are not often wise, as we are always growling about it.
Georget went down to the third floor, and rang softly at Monsieur
Malberg’s door; a very dark mulatto, whose hair age had not yet turned
white, and who spoke French very well for a colored man, and very ill for a
Parisian, opened the door and recognized the young messenger whom he
had met sometimes on the stairs.
“Hullo! it’s Monsieur Georget. Morning, Monsieur Georget! What you
come here for so early?”
“Monsieur Pongo, I would like to speak to your master, Monsieur
Malberg.”
“Oh! master not up yet, he still sleep; I get up sooner, to tidy the room,
rub floor here in the morning without waking master.”
“If he is still asleep, I will wait.”
“Yes, you sit down on a nice little chair, like this.”
“Thanks, Monsieur Pongo; I hope I am not in your way; go on with your
work.”
“Yes, yes, then I go very soft and see if monsieur still sleep.”
The mulatto went into another room. Georget sat down and waited. After
a few moments he heard voices in the next room and supposed that
Monsieur Malberg was awake. But still he was left alone, nobody came,
and Georget, beginning to be impatient, coughed, walked about the room
and stole softly to the door, which was ajar. He was surprised to find that
the mulatto was alone, but that as he did his work, he kept up a steady
conversation with all the furniture and other objects in the study, which to
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com