[Ebooks PDF] download (Ebook) Statistical Mechanics of Classical and Disordered Systems : Luminy, France, August 2018 by Véronique Gayrard, Louis-Pierre Arguin, Nicola Kistler, Irina Kourkova ISBN 9783030290764, 9783030290771, 303029076X, 3030290778 full chapters

Download Full Version ebook - Visit ebooknice.
com
(Ebook) Statistical Mechanics of Classical and

Disordered Systems : Luminy, France, August 2018
by Véronique Gayrard, Louis-Pierre Arguin, Nicola
Kistler, Irina Kourkova ISBN 9783030290764,
9783030290771, 303029076X, 3030290778
https://ebooknice.com/product/statistical-mechanics-of-
classical-and-disordered-systems-luminy-france-
august-2018-10799812
Click the button below to download
DOWLOAD EBOOK
Discover More Ebook - Explore Now at ebooknice.com

Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...
Start reading on any device today!
(Ebook) Biota Grow 2C gather 2C cook by Loucas, Jason;

Viles, James ISBN 9781459699816, 9781743365571,
9781925268492, 1459699815, 1743365578, 1925268497
https://ebooknice.com/product/biota-grow-2c-gather-2c-cook-6661374
ebooknice.com
(Ebook) Matematik 5000+ Kurs 2c Lärobok by Lena

Alfredsson, Hans Heikne, Sanna Bodemyr ISBN 9789127456600,
9127456609
https://ebooknice.com/product/matematik-5000-kurs-2c-larobok-23848312
ebooknice.com
(Ebook) SAT II Success MATH 1C and 2C 2002 (Peterson's SAT

II Success) by Peterson's ISBN 9780768906677, 0768906679
https://ebooknice.com/product/sat-ii-success-
math-1c-and-2c-2002-peterson-s-sat-ii-success-1722018
ebooknice.com
(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master

the SAT Subject Test: Math Levels 1 & 2) by Arco ISBN
9780768923049, 0768923042
https://ebooknice.com/product/master-sat-ii-math-1c-and-2c-4th-ed-
arco-master-the-sat-subject-test-math-levels-1-2-2326094
ebooknice.com
(Ebook) Cambridge IGCSE and O Level History Workbook 2C -
Depth Study: the United States, 1919-41 2nd Edition by
Benjamin Harrison ISBN 9781398375147, 9781398375048,
1398375144, 1398375047
https://ebooknice.com/product/cambridge-igcse-and-o-level-history-
workbook-2c-depth-study-the-united-states-1919-41-2nd-edition-53538044
ebooknice.com
(Ebook) Statistical Mechanics of Disordered Systems: A

Mathematical Perspective by Anton Bovier ISBN
9780511169113, 9780521849913, 0511169116, 0521849918
https://ebooknice.com/product/statistical-mechanics-of-disordered-
systems-a-mathematical-perspective-986864
ebooknice.com
(Ebook) The Statistical Mechanics of Irreversible

Phenomena by Pierre Gaspard ISBN 9781108473729, 1108473725
https://ebooknice.com/product/the-statistical-mechanics-of-
irreversible-phenomena-44533148
ebooknice.com
(Ebook) Internet of Things, Smart Spaces, and Next

Generation Networks and Systems: 18th International
Conference, NEW2AN 2018, and 11th Conference, ruSMART
2018, St. Petersburg, Russia, August 27–29, 2018,
https://ebooknice.com/product/internet-of-things-smart-spaces-and-
Proceedings by Olga Galinina, Sergey Andreev, Sergey
next-generation-networks-and-systems-18th-international-conference-
new2an-2018-and-11th-conference-rusmart-2018-st-petersburg-russia-
Balandin, Yevgeni Koucheryavy ISBN 9783030011673,
august-2729-2018-proceedings-7320374
ebooknice.com
9783030011680, 3030011674, 3030011682
(Ebook) Classical Hopf Algebras and Their Applications

(Algebra and Applications, 29) by Pierre Cartier, Frédéric
Patras ISBN 9783030778446, 3030778444
https://ebooknice.com/product/classical-hopf-algebras-and-their-
applications-algebra-and-applications-29-34794516
ebooknice.com
Springer Proceedings in Mathematics & Statistics
Véronique Gayrard
Louis-Pierre Arguin
Nicola Kistler
Irina Kourkova Editors
Statistical
Mechanics of
Classical and
Disordered Systems
Luminy, France, August 2018
Springer Proceedings in Mathematics &
Statistics
Volume 293
This book series features volumes composed of selected contributions from

workshops and conferences in all areas of current research in mathematics and
statistics, including operation research and optimization. In addition to an overall
evaluation of the interest, scientific quality, and timeliness of each proposal at the
hands of the publisher, individual contributions are all refereed to the high quality
standards of leading journals in the field. Thus, this series provides the research
community with well-edited, authoritative reports on developments in the most
exciting areas of mathematical and statistical research today.
More information about this series at http://www.springer.com/series/10533

Véronique Gayrard Louis-Pierre Arguin
• •
Nicola Kistler Irina Kourkova

•
Editors
Statistical Mechanics
of Classical and Disordered
Systems
Luminy, France, August 2018
123
Editors
Véronique Gayrard Louis-Pierre Arguin
Aix-Marseille Universite CNRS Baruch College, CUNY
Institut de Mathematiques de Marseille New York, NY, USA
Marseille, France
Irina Kourkova
Nicola Kistler Laboratoire de Probabilités
Goethe-Universität Frankfurt Pierre and Marie Curie University VI
Frankfurt am Main, Germany Paris, France
ISSN 2194-1009 ISSN 2194-1017 (electronic)

ISBN 978-3-030-29076-4 ISBN 978-3-030-29077-1 (eBook)
https://doi.org/10.1007/978-3-030-29077-1
Mathematics Subject Classification (2010): 82-01, 82-02, 82-06, 82B03, 82B05, 82B26, 82B31, 82B41,
82B43, 82B44, 82B80, 82C44, 60F10, 60G07, 60G15, 60G70
© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
Statistical mechanics aims at understanding the behavior of systems involving very

many interacting components through so-called statistical ensembles, namely,
probability measures over all possible states of the system. Initiated in the late
1800s by Boltzmann and Maxwell, who were motivated by deriving the principles
of thermodynamics of gases from the statistical properties of molecules, the
framework of statistical mechanics was mathematically formalized by Gibbs a few
decades later. With time, statistical mechanics has grown into an extensive and
ever-evolving body of knowledge that encompasses several areas of physics and
probability theory, with deep and far-reaching links with areas as diverse as biology
and computer science. Virtually any area of research dealing with a large number of
agents displaying collective behavior, whether these be molecules, living organisms
or nodes of a network, can benefit from its formalism. The contributions to this
volume, originating from the international conference Advances in Statistical
Mechanics which took place at the CIRM in Luminy in August 2018, nicely reflect,
we believe, the current state of affairs of such a success story.
In fact, all research articles appearing in this volume deal, in one way or another,
with the many facets of modern statistical mechanics, and can be (loosely) grouped
in two major sub-fields: those pertaining to the somewhat classical statistical
mechanics of ordered systems, and those pertaining to the more recent field of
disordered systems.
The paper by Külske et al. and the one by van Enter fall in the first category. The
former addresses, under the lenses of the Widom-Rowlinson model, the “Gibbs vs
non-Gibbs transition”; the latter shows how models from statistical mechanics can
shed light on the delicate issue whether (and to which extent) Markov chains and
Markov fields lead to conceptually equivalent objects. Both issues are currently the
subject of intensive research activities.
The majority of contributions fall into the second category, that of disordered
systems. A further categorization into sub-groups is here possible, as one may
distinguish between “statics vs. dynamics”, i.e., between properties of systems at
equilibrium, as opposed to systems which are yet to relax. Arguin-Persechino study
via large deviations techniques at the level of entropy and free energy the extremes
v
vi Preface
of a GREM in random magnetic field. The paper by Bolthausen provides a new

proof of the replica symmetric solution of the SK-model which relies on a
Morita-type argument, and the TAP equations. Guerra addresses the highly efficient
but mathematically puzzling replica trick, recasting it into the framework of
interpolations. Kersting et al. show how the (R)SB-Parisi solution of the SK-model
emerges via high- temperature expansions from ad hoc Gibbs potentials in finite
volume. The paper by Cerny improves our understanding of the Metropolis
dynamics of Derrida’s REM, insofar it shows that a deterministic normalization
suffices for the convergence of the clock process. Gayrard-Hartung also address
dynamical properties of the REM, and provide a comprehensive analysis of the
phase diagram in the case of random hopping dynamics. The contribution by Wang
et al. addresses the predictability of the zero-temperature Glauber dynamics of Ising
ferromagnets in case of light- or heavy-tailed distributions.
The remaining two contributions are somewhat challenging to characterize in
terms of the aforementioned dichotomy “classical vs. disordered”, but they are
nevertheless deeply rooted in statistical mechanics. The paper by Faggionato pro-
vides a concise review of results establishing large deviation principles and
Gallavotti-Cohen dualities for Markov chains, and which thus provide insights into
the thermodynamics of (bio)molecular motors/pumps. Finally, the paper by Caputo
et al. deals with line ensembles for Brownian polymers: this is a field of probability
theory, which however bears strong connections with the statistical mechanics of
SOS-models.
The international conference, which eventually led to the publication of these
proceedings, would have never been possible without the (financial, logistic and
organizational) support of the CIRM in Luminy, the city of Marseille, the Aix
Marseille Université, the Institut de Mathématiques-Labex Archimède-CARMIN-
FRUMAM (Marseille), the LYSM-LPSM-Groupe de Mathématiques de
l'Aléatoire-Groupe de Modélisation Mathématique (Paris), the CNRS, the NSF, the
DFG, the University of Frankfurt, and the Bonn International Graduate School of
Mathematics. We express our heartfelt gratitude to these institutions, and to the
affiliated people who helped us all along this journey.
The meeting at the CIRM has also been a wonderful opportunity to celebrate the
60th birthday of our friend and colleague Anton Bovier. The works appearing in
these proceedings touch upon a tiny fraction only of his broad scientific interests,
but it seems fair to say that statistical mechanics at large has played a major role in
his distinguished career: we thus dedicate this volume to him, with the best wishes
for many happy returns.
New York, USA Louis-Pierre Arguin

Marseille, France Véronique Gayrard
Frankfurt am Main, Germany Nicola Kistler
Paris, France Irina Kourkova
Contents
Ordered Systems
Gibbs-Non Gibbs Transitions in Different Geometries:
The Widom-Rowlinson Model Under Stochastic Spin-Flip
Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Christof Külske
One-Sided Versus Two-Sided Stochastic Descriptions . . . . . . . . . . . . . . 21
Aernout C. D. van Enter
Disordered Systems
The Free Energy of the GREM with Random Magnetic Field . . . . . . . . 37
Louis-Pierre Arguin and Roberto Persechino
A Morita Type Proof of the Replica-Symmetric Formula for SK . . . . . . 63
Erwin Bolthausen
Concentration of the Clock Process Normalisation
for the Metropolis Dynamics of the REM . . . . . . . . . . . . . . . . . . . . . . . 95
Jiří Černý
Dynamic Phase Diagram of the REM . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Véronique Gayrard and Lisa Hartung
The Replica Trick in the Frame of Replica Interpolation . . . . . . . . . . . . 171
Francesco Guerra
From Parisi to Boltzmann . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Goetz Kersting, Nicola Kistler, Adrien Schertzer and Marius A. Schmidt
vii
viii Contents
Nature Versus Nurture: Dynamical Evolution in Disordered

Ising Ferromagnets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Lily Z. Wang, Reza Gheissari, Charles M. Newman and Daniel L. Stein
Miscellaneous
Tightness and Line Ensembles for Brownian Polymers
Under Geometric Area Tilts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Pietro Caputo, Dmitry Ioffe and Vitali Wachtel
Large Deviations and Uncertainty Relations in Periodically Driven
Markov Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Alessandra Faggionato
Contributors
Louis-Pierre Arguin Department of Mathematics, Baruch College and Graduate

Center, City University of New York, New York, NY, USA
Erwin Bolthausen University of Zurich, Zurich, Switzerland
Pietro Caputo Dipartimento di Matematica e Fisica, Roma Tre University, Rome,
Italy
Jiří Černý Department of Mathematics and Computer Science, University of
Basel, Basel, Switzerland
Alessandra Faggionato Dipartimento di Matematica, Università di Roma ‘La
Sapienza’, Roma, Italy
Véronique Gayrard I2M, CNRS Centrale Marseille, Aix Marseille University,
Marseille, France
Reza Gheissari Courant Institute, New York University, New York, NY, USA
Francesco Guerra Dipartimento di Fisica, Istituto Nazionale di Fisica Nucleare,
Sezione di Roma, Università di Roma “La Sapienza”, Roma, Italy
Lisa Hartung Institut für Mathematik, Johannes Gutenberg-Universität Mainz,
Mainz, Germany
Dmitry Ioffe The Faculty of Industrial Engineering and Management, Technion,
Haifa, Israel
Goetz Kersting J.W. Goethe-Universität, Frankfurt, Germany
Nicola Kistler J.W. Goethe-Universität, Frankfurt, Germany
Christof Külske Fakultät für Mathematik, Ruhr-University of Bochum, Bochum,
Germany
ix
Visit https://ebooknice.com to
discover a wide range of
eBooks across various genres.
Enjoy exclusive deals and
discounts to enhance your
reading experience. Start your
digital reading journey today!
x Contributors
Charles M. Newman Courant Institute, New York University, New York, NY,
USA;
NYU-ECNU Institute of Mathematical Sciences at NYU Shanghai, Shanghai,
China
Roberto Persechino Service des Enseignements Généraux, École de Technologie
Supérieure, Montréal, QC, Canada
Adrien Schertzer J.W. Goethe-Universität, Frankfurt, Germany
Marius A. Schmidt University of Basel, Basel, Switzerland
Daniel L. Stein Department of Physics and Courant Institute, New York
University, New York, NY, USA;
NYU-ECNU Institutes of Physics and Mathematical Sciences at NYU Shanghai,
Shanghai, China;
Santa Fe Institute, Santa Fe, NM, USA
Aernout C. D. van Enter Bernoulli Institute, University of Groningen,
Groningen, The Netherlands
Vitali Wachtel Institut für Mathematik, Universität Augsburg, Augsburg,
Germany
Lily Z. Wang Center for Applied Mathematics, Cornell University, Ithaca, NY,
USA
Ordered Systems
Gibbs-Non Gibbs Transitions in Different
Geometries: The Widom-Rowlinson
Model Under Stochastic Spin-Flip
Dynamics
Christof Külske
Abstract The Widom-Rowlinson model is an equilibrium model for point particles

in Euclidean space. It has a repulsive interaction between particles of different colors,
and shows a phase transition at high intensity. Natural versions of the model can
moreover be formulated in different geometries: in particular as a lattice system or
a mean-field system. We will discuss recent results on dynamical Gibbs-non Gibbs
transitions in this context. Main issues will be the possibility or impossibility of
an immediate loss of the Gibbs property, and of full-measure discontinuities of the
time-evolved models.
Keywords Gibbs measures · Stochastic time-evolution · Gibbs-non Gibbs

transitions · Widom-Rowlinson model
AMS 2000 subject classification 60K57 · 82B24 · 82B44
1 Introduction
Recent years have seen a variety of studies of Gibbs-non Gibbs transitions of mea-
sures which appear as image measures of Gibbs measures, under certain local trans-
formation rules. What is a Gibbs measure? There is a well-defined theory to define
Gibbs measures on lattices, where the probability space is given by the set of all func-
tions from lattice sites to a finite alphabet. The central object is that of a specification
[2, 15]. For other geometries, other but related approaches are adequate, see below.
The unifying idea is that Gibbs measures are measures whose conditional probabil-
ities to see a single symbol at a given site, are nice (continuous) functions of their
conditioning, see below. It has been discovered that Gibbs measures under natural
Collaborations with Benedikt Jahnel, Sascha Kissel, Utkir Rozikov.
C. Külske (B)
Fakultät für Mathematik, Ruhr-University of Bochum, Postfach 102148, 44721 Bochum,
Germany
e-mail: [email protected]
© Springer Nature Switzerland AG 2019 3
V. Gayrard et al. (eds.), Statistical Mechanics of Classical and Disordered Systems,
Springer Proceedings in Mathematics & Statistics 293,
https://doi.org/10.1007/978-3-030-29077-1_1
4 C. Külske
deterministic or stochastic transformations may lead to non-Gibbsian measures [10].

More specifically the study of stochastic time evolutions, even very simple ones,
applied to Gibbs measures, has shown very interesting transition phenomena, the
most prototypical example for this is the Glauber-evolved Ising model in [8]. Indeed,
stochastic time evolutions may destroy the Gibbs property of the image measure at
certain transition times, a phenomenon we call dynamical Gibbs-non Gibbs transi-
tions, see below. The purpose of this note is to take the Widom-Rowlinson model
[39] and variations thereof as a guiding example, apply an independent symmetric
stochastic spin flip dynamics to it, and describe our findings of what may and what
may not happen along the time-evolved trajectory of measures. We treat and com-
pare a hard-core version and a soft-core version of the model in various geometries,
namely in Euclidean space, on the lattice, as a mean-field model, and on a regular
tree. Our aim here is to provide an overview; for detailed statements and proofs we
refer to the original articles.
2 Gibbs on Lattice, Sequentially Gibbs, Marked Gibbs

Point Processes, and the Widom-Rowlinson Model
We start by recalling the notion of an infinite-volume Gibbs measure for lattice

systems. For the purpose of the discussion of the Widom-Rowlinson model and all
measures appearing under time-evolution defined below from it, it is sufficient to
restrict to the local state-space {−1, 0, 1} for particles carrying spins plus or minus,
and holes. Our site space is the lattice Zd . The space of infinite-volume configurations
is = {−1, 0, 1}Z .
d
2.1 Specifications and Gibbs Measures on the Lattice
The central object in Gibbsian theory on a countable site space which defines the
model is a specification. This covers both cases of infinite lattices and trees. It is a
candidate system for conditional probabilities of an infinite-volume Gibbs measure
μ (probability measure on ) to be defined by DLR equations μ(γ ( f |·)) = μ( f ).
A specification γ is by definition a family of probability kernels γ = (γ )Zd ,
indexed by finite subvolumes , where γ (dω|η) is a probability measure on ,
for each fixed configuration η. It must have the following properties. The first is the
consistency which means that
γ (γ (dω|·)|ω̃) = γ (dω|ω̃) (1)
for all finite volumes ⊂ Zd . It is suggested by the tower property of condi-

tional expectations.
Gibbs-Non Gibbs Transitions in Different Geometries … 5
The second is the Fc -measurability of γ ( f |·), for any bounded measurable
observable f . Here the sigma-algebra Fc is generated by the spin-variables outside
of the finite volume .
The last property is the properness γ (1 A |·) = 1 A for A ∈ Fc . It means that
the randomization of the kernel takes place only inside of , and an event which
is determined by what is outside of will indeed be determined by looking at the
boundary condition alone.
An important additional regularity requirement is quasilocality of the specification
which means that the function ω → γ ( f |ω) should be quasilocal for f quasilocal,
and this has to hold for all finite volumes . A quasilocal function is a uniform limit of
local functions, that is of functions which depend only on finitely many coordinates.
More specifically a Gibbsian specification on the infinite-volume state space =
{−1, 0, 1}Z for an interaction potential = ( A ) AZd and a priori measure α ∈
d
M1 ({−1, 0, 1}) by definition has probability kernels
1
γ,,α (ω |ωc ) := e− A∩=∅ A (ω) α(ωi ) (2)
Z (ωc ) i∈
where Z (ωc ) is the normalizing partition function. If is a finite-range potential

(meaning that A is only nonzero for finitely many A’s), obviously all sums are finite,
when we insist that A takes finite only values. Finiteness of the sums also holds, if
is uniformly absolutely convergent. For hard-core models the specification kernels
acquire an indicator, see the example (4) below.
The first statistical mechanics task in this setup is the following. Given a specifi-
cation γ = (γ )Zd in the above sense, find the corresponding Gibbs measures
G(γ) := {μ ∈ M1 (), μγ = μ, for all Zd } (3)
In general G(γ) may be empty, contain precisely one measure, or more than one
measures. If |G(γ)| > 1 we say that the specification γ has a phase transition. The
Gibbs measures G(γ) form a simplex, meaning that each measure has a unique
decomposition over the extremal elements, called pure states. Pure states can be
recovered as finite-volume limits with fixed boundary conditions.
Existence and extremal decomposition of proper infinite-volume measures becomes
even more involved for systems with random potentials. In general, for systems like
spin-glasses, the construction of infinite-volume states by non-random sequences of
volumes which exhaust the whole lattice is problematic, and for such systems the
higher-level notion of a metastate (a measure on infinite-volume Gibbs measures) is
useful [1, 2, 5, 27, 34].
6 C. Külske
2.2 Hard-Core and Soft-Core Widom-Rowlinson Model on

Lattice and in Mean-Field
We will consider here the version of the hard-core Widom-Rowlinson model on Zd

as in [17]. It has the a priori measure α ∈ M1 ({−1, 0, 1}) as its only parameter. Its
specification kernels are given by
1
γ,α
hc
(ω |ωc ) := I hc
(ω ω c) α(ωi ), (4)
hc
Z (ωc ) i∈

where the hard-core indicator Ihc (ω) = i∈ I(ωi ω j =−1, ∀ j∼i) forbids +− neighbors
to occur with positive probability. Related hard-core models have been studied on
lattices and trees, see for example [14, 32, 35].
The soft-core Widom-Rowlinson model on Zd has an additional repulsion param-
eter β > 0. In the specification kernels, which are by definition given by
I(ωi ω j =−1)

1 −β
γ,β,α
sc
(ω |ωc ) := α(ωi ),
b
{i, j}∈E
e (5)
sc
Z (ωc ) i∈
configurations with +− neighbors are suppressed, but not forbidden.

These definitions of a specification immediately extend to any graph with count-
ably infinite vertex set, where each vertex has a finite number of nearest neighbors.
In particular we may study this model on a regular tree with k + 1 neighbors, see
[23].
The mean-field formulation is different, as the model is defined as a whole
sequence of finite-volume Gibbs measures, indexed by the system size N ∈ N. The
elements in the sequence for the Mean-Field soft-core Widom-Rowlinson model
with repulsion parameter β > 0 are the measures
1 β
N
μ N ,β,α (ω[1,N ] ) := e− 2N 1≤i, j≤N I(ωi ω j =−1)
α(ω j ) (6)
Z N ,β,α j=1
for ω[1,N ] = (ωi )i=1,...,N ∈ {−1, 0, 1} N . For more details see [22].
2.3 Sequential Gibbsianness for Mean-Field (and

Kac-Models on Torus)
There is an intrinsic formulation of the Gibbs property which is suitable also in

situations where a finite-volume Hamiltonian can not be read off directly from the
explicit definition of the measures. It focusses on conditional probabilities instead,
suggested by analogy to the lattice situation [18, 28].
Take (μ N ) N ∈N a sequence of exchangeable probability measures μ N on the finite-

volume state space {−1, 0, 1} N . The large N -behavior of such a sequence defines our
model. The model is called sequentially Gibbs iff the volume-limit of the single-spin
probabilities in the finite-volume measures
lim μ N (dω1 |ω[2,N ] ) = γ(dω1 |ν) (7)

N ↑∞
exists whenever the empirical distributions of a configuration (ωi )i=2,3,4,... converge,
1
N
δω → ν. (8)
N − 1 i=2 i
This has to hold for all limiting empirical distributions of conditionings ν ∈

M1 ({−1, 0, 1}).
If there is some ν for which it is possible to obtain different limits for different
boundary conditions (ωi )i=2,3,... , and (ω̄i )i=2,3,... we call this ν a bad empirical mea-
sure. A model fails to be sequentially Gibbs, if there is at least one bad empirical
measure.
As a general consequence of this definition, the sequential Gibbs property implies
that the limiting kernel ν → γ(dω1 |ν) is continuous. In our Widom-Rowlinson case
where ν takes values in a finite-dimensional simplex, all topologies are the same, and
equal to the Euclidean topology. Clearly the mean-field Widom-Rowlinson model
defined above, is sequentially Gibbs.
A similar notion of the sequential Gibbs property can be extended to cover Kac-
models on the torus, and transformed Kac-models which have the same index set.
These models are again described by sequences whose asymptotics one wants to
capture, but they have a spatial structure. As in the mean-field models, there is
again a single-site limiting kernel, however the limiting empirical distribution ν
which appeared as a conditioning in the mean-field model is replaced by a whole
profile of spin densities on the unit torus. For details of these definitions and results,
see [16, 18].
2.4 Marked Gibbs Point Processes in Euclidean Space
Here the good definition of Gibbs measure is in some analogy to the lattice situation
[6, 20, 37]. We restrict again for the sake of our exposition to the specific simple
mark space which covers the Widom-Rowlinson model of point particles in Euclidean
space, and the time-evolved version we will discuss below. In this case the mark space
is {−1, 1}. It does not contain zero. The spatial degrees of freedom are described by
the set of locally finite subsets of Rd . A marked particle configuration is a pair ω =
(ω − , ω + ) describing the positions of minus-particles (and plus-particles respectively)
where each ω − , ω + ∈ . The configuration space of such marked configurations is
8 C. Külske
. For the measurable structure we need the σ-algebras F, F . These are the σ-
algebras for marked particles generated by the counting variables. They count the
number of plus- and minus particles in Borel sets A in the whole Euclidean space
(or all such sets A ∈ respectively, where may be any measurable subset of
Euclidean space).
A specification shall become, as on countable graphs, a candidate system for
conditional probabilities of Gibbs measure μ to be defined by DLR equations μγ =
μ for all measurable bounded subsets in Euclidean space. Hence, one defines a
specification to be a family of proper probability kernels γ = (γ )Rd with the
properties of consistency, that is γ γ = γ for all measurable volumes ⊂
Rd . One also needs Fc -measurability of γ ( f |·), for any bounded test observable
f . Properness means here that γ (1 A |·) = 1 A for A ∈ Fc .
We will further assume quasilocality of the specification. This means the compat-
ibility of the kernels γ with the local topology on the space of marked point clouds.
In this topology convergence for a sequence of marked particle clouds means that
the clouds must become constant in each bounded volume.
2.5 Widom-Rowlinson Model in Euclidean Space
We assume spatial dimension d ≥ 2, and fix the two-color local spin space (mark
space) of {−, +}. The model will be obtained as a modification of the base measure
P by which we denote a two-color homogenous Poisson Point Process in the infinite
volume, with intensities λ+ for plus colors and λ− for minus colors.
The (hard-core) Widom-Rowlinson specification is the Poisson-modification with
the specification kernels
1
γ (dω |ω c ) := χ(ω ω c )P (dω ) (9)
Z (ω c )
where the indicator χ is one iff the interspecies distance (the distance between points
of different sign) is bigger or equal than 2a, and P (dω ) denotes the two-color
Poisson process in the bounded volume . The picture shows a typical configuration
at large λ+ = λ− , in a finite volume.
This picture survives the thermodynamic limit: By results of [3, 4, 36] it is known
that in d ≥ 2, λ+ = λ− large, the continuum WiRo has a phase transition, and how
this is related to percolation of large clusters of overlapping disks.
In more general models of marked point particles, specifications which are Poisson
modifications may be obtained in terms of exponential factors with finite-volume
Hamiltonians which are formed with potentials. For such potentials one may consider
multibody potentials which are known from statistical physics, but we may also allow
for hyperedge potentials and define
1
γ (dω |ω c ) := e− ηω: η∩=∅ (η,ω) P (dω ) (10)
Z (ω c )
A hyperedge potential (η, ω) is by definition allowed to depend on the marked point

cloud ω not only via the particle positions and marks on hyperedge η (which is just a
finite subset of points in the cloud), but also on a whole neighborhood of η, up to some
horizon. This generalization is useful in models of stochastic geometry, involving
e.g. energies depending on the cells of a Voronoi tessellation. Hyperedge potentials
were successfully used in [6] where a general existence theory of infinite-volume
Gibbs measures is developed. We also refer to [21] for representation theorems.
There is it shown how one can go from a measure μ under continuity assumptions of
finite-volume conditional probabilities, to a hyperedge potential . These theorems
are a generalization of Kozlov-Sullivan theorems [25, 38] known on the lattice to
the continuum, and make use in a constructive way of the weak nonlocality allowed
by the hyperedge potential concept.
3 Dynamical Gibbs-Non Gibbs Transitions
We now come to time-evolutions. Consider again the Euclidean space Widom-

Rowlinson model, and fix some cloud of particles carrying the marks plus or minus.
We define a continuous-time stochastic dynamics by the following rule. Particle
locations stay fixed, holes stay fixed. The signs of the particles however change
stochastically, independently of each other, according to a temporal Poisson process
for each particle with rate one. In this way, at every particle location, the probability
to go from + to − in time t is given by
1
pt (+, −) = pt (−, +) = (1 − e−2t ). (11)
2
Starting with the same signed particle configuration shown above, after a small time
t, a fraction of the particles has kept their signs, as shown in the picture (or flipped
back). Of course, there is loss of memory in each fixed bounded volume, which is
exponentially fast in the time t. Interesting things however happen if we consider the
infinite volume, as we will discuss.
10 C. Külske
We will apply the same stochastic dynamics also on the spatially discrete model on
the lattice, and also to the mean-field model, at each finite system size N .
We want to understand better the structural Gibbsian properties of the measures
along a trajectory given by the time-evolution. For the purpose of concreteness we
focus on the Euclidean model. We say the model shows a dynamical Gibbs-non Gibbs
transition if the initial measure μ is Gibbs for a quasilocal specification, and for some
time t the time-evolved measure μt = μPt is not compatible with any quasilocal
specification. Here Pt is the semigroup giving the distribution to find an infinite-
volume configuration after time t when starting with a given initial configuration,
which is integrated over with respect to the starting measure μ. In our example
above Pt is the symmetric independent spin-flip dynamics, and does not involve a
randomization of the spatial degrees of freedom. However it is clear that one would
like to study more generally also dependent dynamics, and also possibly irreversible
dynamics, compare [19, 31]. Such studies have been performed at first for the Ising
model on the lattice, for work on this and related work see [8, 9, 11–13, 26, 28, 30].
3.1 Relation to Disordered Systems
To fix ideas, let us go to the lattice setup. That a time-evolved lattice measure μt (in
our the lattice Widom-Rowlinson model under symmetric spin-flip) is non-Gibbs is
indicated by very long-range dependencies in its conditional probabilities, that is
η → μt (ηi |ηZd \i ) (12)
behaves discontinuously w.r.t. the local topology. More precisely, as the r.h.s. is
only defined up to measure-zero sets, this means that there is no version which is
continuous.
A useful strategy (at least for independent dynamics) is the following. Consider
the two-layer measure, that is the joint distribution of spins and time zero and at in
the future at time t > 0 given by
μ̄t (dω, dη) = μ(dω)Pt (ω, dη). (13)
Analyze hidden phase transitions in first-layer measure constrained on the future

configuration η. By this we mean the measure
Visit https://ebooknice.com to
discover a wide range of
eBooks across various genres.
Enjoy exclusive deals and
discounts to enhance your
reading experience. Start your
digital reading journey today!
μ̄t (dω|η). (14)
A relation to disordered systems with quenched order appears when we view

the configuration η = (ηi )i∈Zd in the role of a quenched disorder configuration. By
playing with (suitable finite-volume approximations of) conditional probabilities a
picture emerges in which absence of phase transitions in the first-layer model implies
Gibbsian behavior of the time-evolved model. The opposite implication, that the
presence of a phase transition in the first-layer model implies the absence of the
Gibbs property of the time-evolved model is true in many examples, and proved for
a specific class of mean-field systems.
Methods which are different from the two-layer picture are used for depen-
dent dynamics. For mean-field systems and Kac-systems there are also path-large-
deviation principles available which lead to fixed-end-point variational problems for
trajectories of empirical measures. While in an abstract sense this is a solution, the
analytical understanding of the structure of minimizers of such problems can be quite
hard (see however [16]). It is an open challenge to fully develop the analogous theory
on the lattice, with ideas as suggested in [9].
3.2 Results on Dynamical GnG Transitions for Euclidean

Model
A marked infinite-volume configuration ω ∈ (that is a signed point cloud) is called

good for a specification γ iff for any Euclidean ball B we have

γ B ( f |ω c ) − γ B ( f |ω c ) → 0
B B
as ω ⇒ ω in the sense of local convergence.

We denote by (γ) the set of good configurations for the specification γ. We say
that a specification γ is called quasilocal iff (γ) = , that is if all point clouds are
good.
A measure on signed point clouds μ is called ql (quasilocally Gibbs) iff there exists
a specification γ such that (γ) = . μ is called asql (almost surely quasilocally
Gibbs) iff there exists γ for which at least μ((γ)) = 1, that is the good points are
a full-measure set.
Let us now describe the results on Gibbsian transitions in time and intensity
for μ+ obtained in [20]. The measure μ+ is the measure in the Euclidean Widom-
Rowlinson model obtained as an infinite-volume limit of finite-volume measures
with the maximal boundary condition of overlapping plus-discs. By FKG (stochastic
monotonicity) arguments this measure exists for all choices of parameters. We define
1 λ+ + λ−
tG := log (15)
2 λ+ − λ−
12 C. Külske
for λ+ > λ− . It will serve as a reentrance time into the Gibbsian region. We say
that the model with intensity parameters λ+ , λ− is in the high-intensity (percolating)
regime iff μ+ (B ↔ ∞) > 0 for some ball B (there is positive probability that there
is an infinite cluster of overlapping discs containing B). Then the behavior of the
time-evolved measure μ+ t is summarized in the following table.
λ+ > λ− λ+ = λ−
time 0 < t < tG t = tG tG < t ≤ ∞ 0<t≤∞
high intensity non-asql asql, non-ql ql non-asql
low intensity asq, non-ql asql, non-ql ql asql, non-ql
Main striking features are the immediate loss of the Gibbs property, and the
appearance of full-measure sets of bad configurations (discontinuity points of any
specification). More precise statements and detailed proofs can be found in [20]. The
proofs use cluster representations for conditional probabilities of the time-evolved
measure.
The appearance of typical bad configurations can be heuristically understood: Infi-
nite clusters in the time-zero model, together with the requirement that overlapping
disks have the same sign, gives a strong rigidity in the first-layer model constrained
on the future configuration η at time t. Indeed, conditional on fixed locations in
the percolating cluster, this cluster can only carry uniform plus signs, or unifom
minus signs, at time zero. Keeping locations in a conditioning η fixed and varying
the signs arbitrarily far away provides then a very effective mechanism to induce a
phase transition in the first-layer model. One shows that this implies jumps in certain
conditional probabilities at time t. Hence every percolating point cloud can be a bad
configuration, in the appropriate parameter regimes. As percolation is typical, this
implies full-measure sets of bad configurations.
3.3 Results on Dynamical GnG Transitions for the

Mean-Field Widom-Rowlinson Model
Before we come to dynamics, we need to describe the equilibrium behavior of the

mean-field model. The pressure of the mean-field model can be computed using large-
deviation techniques (Varadhan’s lemma) in terms of a variational formula where
extremal points need to be found in the space of single-site probability distributions.
Denoting by L 1N the fraction of spins with spin 1 at system size N , and using similar
notations for the other spin values, we have

N
1 −1
e−N β L N (ω)L N (ω)
1
p(β, α) = lim log α(dω j )
N →∞ N j=1
= sup (−βν(1)ν(−1) − I (ν|α))

ν∈M1 ({−1,0,1})
where I (ν|α) is the relative entropy of a single-site distribution ν w.r.t. the a priori
measure α. Correspondingly, the empirical distribution satisfies a large deviation
principle with speed N , and rate function given by the negative of the expression
under the sup plus a suitable constant.
A discussion of the variational problem (see [22]) shows: The symmetric model
at any α(1) = α(−1) > 0 has a second-order phase transition driven by repulsion
strength β > 0 at critical repulsion strength given by βc = 2 + e α(0)
α(1)
.
An explicit solution for the mean-field Widom-Rowlinson model at time t = 0
can be obtained as follows. We parametrize the empirical spin distribution ν via
coordinates (x, m) where x plays the role of occupation density, and m plays the role
of the magnetization on the occupied sites, writing
x x
ν(−1), ν(0), ν(1) = ( (1 − m), 1 − x, (1 + m)) (16)
2 2
Next we parametrize the a priori measure α via coordinates (h, l) where h =
α(1)
1
2
log α(−1) plays the role of an external magnetic field, and l := log 1−α(0)
α(0)
describes a bias on the occupation variables. Using these coordinates, the pressure
can be written as
βx 2
p(β, α) = log α(0) + sup − + x(l − log(2 cosh(h)) − J (x)
0≤x,|m|≤1 4
part for occupation density
βxm 2
+x + hm − I (m) − log 2
4
Ising part at occupation-dependent temperature
with an entropy for spins I (m) = 1−m 2

log( 1−m
2
) + 1+m
2
log( 1+m
2
) and an entropy
for occupation variables J (x) = (1 − x) log(3(1 − x)) + x log( 3x2 ). It turns out that
the symmetric antiferromagnetic model (β < 0) has a first-order transition when
crossing a line in β, α(0)-space, where jumps occur in occupation density x, at fixed
zero magnetization m = 0.
We are mostly interested in the ferromagnetic model. In this situation we obtain
that repulsion parameter β > 0, a priori measure α = α(h, l), and typical values
(m, x) of the empirical distribution, are related via the parametrization
14 C. Külske
2
(I (m) − h)(1 + e−l+log(2 cosh(h))+ m (I (m)−h)−m I (m)+I (m) )
1
β = β(m; α) =
m

x = x(m; α) = (1 + e−l+log(2 cosh(h))+ m (I (m)−h)−m I (m)+I (m) )−1
1
We remark as a corollary that the model has mean-field critical exponents: Fix any
α(0) ∈ (0, 1). Let βc be the corresponding critical value for the symmetric model.
Then there are positive constants such that
m(β, h = 0) m(βc , h)
lim 1 = c, lim 1 = c
β↓βc (β − βc ) 2 h↓0 h3
The main point is the study of the dynamical Gibbs-non Gibbs transitions under rate-
one symmetric independent spin-flip, keeping holes fixed, according to transition
probabilities (11).
Recall the notion of sequentially Gibbs and the notion of bad empirical measure,
see (7), (8) and the text below. For the sake of this review let us just present the time
evolution of the set of bad empirical measures in the regime of an inverse temperature
of the time-zero model in the region of β > 3 in a plot (compare [22] for the full
statement of the theorem describing all dynamical transitions). In the plot the inverse
temperature of the time zero model is β = 5 and we are starting from a symmetric
model with α(+) = α(−).
Here time increases from the top left to the top right, then from bottom left
to bottom right. The main features are the following: There is a short-time Gibbs
regime for all β, α. Small repulsion strength β ≤ 2 implies the Gibbs property of the
time-evolved model for all times t. The set of bad empirical measures at given β has
dimension one as a subset of the simplex, in the interior of its existence time-interval.
It can be a union of disconnected curves, a branching curve (which has a Y-shape,
see picture), or a line which is growing with time (growing antenna).
These features can be understood by a bifurcation analysis of a rate function

describing the first-layer model constrained on an empirical distribution. For our
model, this analysis is closely related to that of a time-evolved Ising model in the
following way. The bad measures α f in the time-evolved mean-field WiRo model
after time t satisfy for any symmetric a priori measure α

α f (1) − α f (−1) βα f ({1, −1})
B W i Ro (β, t) = α f ∈ M1 ({−1, 0, 1}), ∈ B I sing ,t .
α f ({1, −1}) 2
where B I sing (β I , t) denotes the set of bad magnetizations for the time-evolved Curie-
Weiss Ising model with initial inverse temperature β I . It is known from [28] that
B I sing (β I , t) turns out to be either empty, contain the magnetization value zero, or
to be given by a symmetric pair.
What can we say about the typicality of bad points in the time-evolved mean-
field Widom-Rowlinson model? Typicality means in the mean-field context that
the minimizers of the large-deviation rate function of the time-evolved model are
contained in the set of bad magnetizations. It is an analytical principle for time-
evolved mean-field Ising models that there is an atypicality of bad configurations.
This follows from the principle of preservation of semiconcavity for time-evolved
rate functions which are defined via integrals over Lagrange densities [26]. In simple
words this regularity statement means that kinks in a rate function can never appear
at local minima.
Our model does not fall in the Ising class, but the corresponding statement can
be proved explicitly. It is very nicely illustrated in our model by the following plot.
The repulsion strength of the model at time zero is β = 4 > 3 is the low-temperature
region. The red Y-shaped set denotes the set of bad empirical measures at a fixed
intermediate time. Its form is independent of the initial a priori distribution α, as
long as this is symmetric. By comparison the typical configurations for any α, after
time t are solid blue. They arise as time-evolution of the dotted blue lines describing
typical measures at time zero.
3.4 Lattice Widom-Rowlinson Model Under Time-Evolution
We describe mainly the lattice soft-core Widom-Rowlinson model. To prove the Gibbs
property of the time-evolved model in appropriate regions, Dobrushin-methods are
useful, as we outline now. Let γ := (γ )Zd be a quasilocal specification on the
lattice. The Dobrushin interdependence matrix, is defined by
16 C. Külske
Ci j (γ) = sup γ{i} (·|ω) − γ{i} (·|η)T V,i (17)

ωZd \{ j} =ηZd \{ j}
for sites i = j. The main theorem, dueto Dobrushin, states: If the Dobrushin condi-
tion holds, namely if c(γ) := supi∈Zd j∈Zd Ci j (γ) < 1, then |G(γ)| = 1. The theory
also allows to control the unique Gibbs measure under perturbations of the specifi-
cation, understand correlation decay in the measure, and derive more useful conse-
quences [15]. We show the Dobrushin region (the region in parameter space for which
c(γ) < 1) for the spatially homogeneous soft-core model on Z2 . The plot shows, for
different values of the repulsion strength β, the Dobrushin region (dark shaded) in the
space of a priori measures α ∈ M({−1, 0, 1}), projected to the α(1), α(−1)-plane.
1.0 1.0 1.0 1.0
0.8 0.8 0.8 0.8
0.6 0.6 0.6 0.6
0.4 0.4 0.4 0.4
0.2 0.2 0.2 0.2
0.0 0.0 0.0 0.0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
1
β< 2 β = 0.75 β = 1.05 β=2
It turns out that the boundary of the Dobrushin region consists of finitely many
pieces of ellipses. We have the following theorem for the homogeneous model:
(1) Let 0 ≤ βd < 1. Then for all α ∈ M({−1, 0, 1}) the soft-core Widom-
Rowlinson model satisfies the Dobrushin condition.
(2) For every β > 0 there exists an := (β) > 0 such that the soft-core model
satisfies the Dobrushin condition if dT V (α, δ1 ) < or dT V (α, δ−1 ) < .
With Dobrushin techniques one controls not only the translation-invariant model, but
also the first-layer model constrained on future configurations. With this one may
prove also the Gibbs property for the time-evolved model for small times: Let α ∈
M({−1, 0, 1}), β ≥ 0, and let μ ∈ G(γβ,α ) by any Gibbs measure. Then there exists
a time tc > 0 such that for all t < tc the time-evolved measure μt is a Gibbs measure
for some quasilocal specification γt . For the proof see [24], it extends methods of [29]
to a situation of degenerate time evolutions (where not all transitions are allowed)
to control all first-layer models for possible to control all first-layer models (14) for
possible end-conditionings η. We remark that the method does not make use of the
lattice structure, but gives the same result of short-time Gibbsianness for any graph
with bounded degree, for instance a regular tree.
For the opposite direction we prove: In the soft-core model on the lattice, at
sufficiently large repulsion strength, the maximal measure μ+ t is non-Gibbs, for times
t which are sufficiently large. For the proof it suffices to exhibit one non-removable
bad configuration for the single-site probability of the time-evolved measure. We
may choose in our case a fully occupied checkerboard configuration of alternating
plus- and minus-spins, and show that this configuration is bad, noting that we are
reduced basically to an Ising situation for this conditioning.
For the hard-core lattice model under time-evolution, Dobrushin techniques can
not be applied, as some entries of the Dobrushin matrix will necessarily become equal
to one. This is not just a shortcoming of the proof. Indeed, we find an immediate loss
of the Gibbs property, as in the Euclidean model, for the proof see [24].
3.5 The Widom-Rowlinson Model on a Cayley Tree
Let us now for our graph consider a Cayley tree, which is the infinite graph which
has no loops, and where each vertex has precisely k + 1 nearest neighbors. The
Widom-Rowlinson model in the hard-core version, and in the soft-core version, is
again defined by the specification kernels of (4) and (5).
We need to start with a good understanding of the Widom-Rowlinson model in
equilibrium. The tree-automorphism-invariant Gibbs measures which are also tree-
indexed Markov chains (also known as tree-invariant splitting Gibbs-measures) are
uniquely described via boundary laws, which appear as solutions of a parameter-
dependent two-dimensional fixed point equation (appearing as a tree recursion).
As a general abstract fact, extremal Gibbs measures for tree models are always
splitting Gibbs measures, the opposite is in general not true [15]. For certain classes
of hard-core models on trees the characterizations of solutions can be found in
[35], at least for low enough degree of the tree. For the equilibrium states of the
ferromagnetic soft-core model on the Cayley tree we find the following [23]. In
the antiferromagnetic model with symmetric intensities there is a transition in the
hole-density, somewhat similar to that in the mean-field model briefly described
above. It can be very explicitly analyzed for any order k, with explicit transition lines
in the interaction-intensity diagram. For the ferromagnetic model with symmetric
intensities, for the trees with 3 and 4 nearest neighbors, the critical lines for the
ferromagnetic phase transition are again explicit, with complete description of all
tree-invariant splitting Gibbs measures. For higher k, there are only bounds on critical
curves, which we conjecture to be sharp, see [23].
What about spin-flip time evolution of these measures on the tree? The Gibbsian
behavior of a time-evolved model can be very different from the behavior in other
geometries. For the Ising model on a Cayley-tree under independent stochastic spin-
flip in [7] the following was proved: The set of bad measures may depend on the
choice of the initial Gibbs measure of the time-evolved state. There can be multiple
transition times in the model with zero external magnetic field, and full-measure sets
of bad measures. For the time-evolved Widom-Rowlinson model on the Cayley tree,
this is an open problem.
Acknowledgements I am very grateful I met Anton, for all the discussions I had with him, and for
all inspiration he gave, during my Ph.D., in later years, until today. I wish him many many more
years, I am looking forward to many more of his contributions, to mathematics and beyond!
18 C. Külske
References
1. Arguin, L.-P., Damron, M., Newman, C., Stein, D.: Uniqueness of ground states for short-range
spin glasses in the half-plane. Commun. Math. Phys. 300, 641–657 (2010)
2. Bovier, A.: Statistical Mechanics of Disordered Systems: A Mathematical Perspective, vol. 18.
Cambridge University Press (2006)
3. Bricmont, J., Kuroda, K., Lebowitz, J.L.: The structure of Gibbs states and phase coexistence for
nonsymmetric continuum Widom–Rowlinson models. Z. Wahrsch. Verw. Gebiete 67, 121–138
(1984)
4. Chayes, J.T., Chayes, L., Kotecký, R.: The analysis of the Widom-Rowlinson model by stochas-
tic geometric methods. Commun. Math. Phys. 172, 551–569 (1995)
5. Cotar, C., Jahnel, B., Külske, C.: Extremal decomposition for random Gibbs measures: from
general metastates to metastates on extremal random Gibbs measures. Electron. Commun.
Probab. 23, 1–12 (2018)
6. Dereudre, D., Drouilet, R., Georgii, H.-O.: Existence of Gibbsian point processes with
geometry-dependent interactions. Probab. Theory Relat. Fields 153, 643–670 (2012)
7. van Enter, A.C.D., Ermolaev, V.N., Iacobelli, G., Külske, C.: Gibbs-non-Gibbs properties for
evolving Ising models on trees. Ann. Inst. Henri Poincare Probab. Stat. 48, 774–791 (2012)
8. van Enter, A.C.D., Fernández, R., den Hollander, F., Redig, F.: Possible loss and recovery of
Gibbsianness during the stochastic evolution of Gibbs measures. Commun. Math. Phys. 226,
101–130 (2002)
9. van Enter, A.C.D., Fernández, R., den Hollander, F., Redig, F.: A large-deviation view on
dynamical Gibbs-non-Gibbs transitions. Moscow Math. J. 10, 687–711 (2010)
10. van Enter, A.C.D., Fernández, R., Sokal, A.D.: Regularity properties and pathologies of
position-space renormalization-group transformations: scope and limitations of Gibbsian the-
ory. J. Stat. Phys. 72, 879–1167 (1993)
11. Enter, A.C.D., Külske, C., Opoku, A.A., Ruszel, W.M.: Gibbs-non-Gibbs properties for n-
vector lattice and mean-field models. Braz. J. Probab. Stat. 24, 226–255 (2010)
12. Ermolaev, V.N., Külske, C.: Low-temperature dynamics of the Curie-Weiss model: periodic
orbits, multiple histories and loss of Gibbsianness. J. Stat. Phys. 141, 727–756 (2010)
13. Fernández, R., den Hollander, F., Martínez, J.: Variational description of Gibbs-non-Gibbs
dynamical transitions for spin-flip systems with a Kac-type interaction. J. Stat. Phys. 147,
1094–1112 (2014)
14. Gallavotti, G., Lebowitz, J.: Phase transitions in binary lattice gases. J. Math. Phys. 12, 1129–
1133 (1971)
15. Georgii, H.-O.: Gibbs Measures and Phase Transitions. De Gruyter, New York (2011)
16. Henning, F., Kraaij, R., Külske, C.: Gibbs-non-Gibbs transition in the fuzzy Potts models with
a Kac-type interaction: closing the Ising gap. Bernoulli J. 25, 2051–2074 (2019)
17. Higuchi, Y., Takei, M.: Some results on the phase structure of the two-dimensional Widom-
Rowlinson model. Osaka J. Math. 41, 237–255 (2004)
18. den Hollander, F., Redig, F., van Zuijlen, W.: Gibbs-non-Gibbs dynamical transitions for mean-
field interacting Brownian motions. Stoch. Process. Appl. 125, 371–400 (2015)
19. Jahnel, B., Külske, C.: Attractor properties for irreversible and reversible interacting particle
systems. Commun. Math. Phys. 366, 139–172 (2015)
20. Jahnel, B., Külske, C.: The Widom-Rowlinson model under spin flip: immediate loss and sharp
recovery of quasilocality. Ann. Appl. Probab. 27, 3845–3892 (2017)
21. Jahnel, B., Külske, C.: Gibbsian representation for point processes via hyperedge potentials
(2017). arXiv:1707.05991
22. Kissel, S., Külske, C.: Dynamical Gibbs-non Gibbs transitions in Curie-Weiss Widom-
Rowlinson models. Markov Process. Relat. Fields 25, 379–413 (2019)
23. Kissel, S., Külske, C., Rozikov, U.: Hard-core and soft-core Widom-Rowlinson models on
Cayley trees. J. Stat. Mech. 4(043204), 22 (2019)
24. Kissel, S., Külske, C.: Dynamical Gibbs-non Gibbs transitions for hard-core and soft-core
Widom-Rowlinson models on the lattice (in preparation)
25. Kozlov, O.K.: A Gibbs description of a system of random variables. Problemy Peredaci Infor-
macii 10, 94–103 (1974)
26. Kraaij, R., Redig, F., van Zuijlen, W.: A Hamilton-Jacobi point of view on mean-field Gibbs-non
Gibbs transitions (2017). arXiv:1711.03489
27. Külske, C.: Metastates in disordered mean-field models: random field and Hopfield models. J.
Stat. Phys. 88, 1257–1293 (1997)
28. Külske, C., Le Ny, A.: Spin-flip dynamics of the Curie-Weiss model: loss of Gibbsianness with
possibly broken symmetry. Commun. Math. Phys. 271, 431–454 (2007)
29. Külske, C., Opoku, A.A.: The posterior metric and the goodness of Gibbsianness for transforms
of Gibbs measures. Electron. J. Probab. 13, 1307–1344 (2008)
30. Külske, C., Redig, F.: Loss without recovery of Gibbsianness during diffusion of continuous
spins. Probab. Theory Relat. Fields 135, 428–456 (2006)
31. Liggett, T.: Interacting Particle Systems. Springer-Verlag, New York (1985)
32. Mazel, A., Stuhl, I., Suhov, Y.: Hard-core configurations on a triangular lattice and Eisenstein
primes (2018). arXiv:1803.04041
33. Mazel, A., Stuhl, I., Suhov, Y.: A classical WR model with q particle types. J. Stat. Phys. 159,
1040–1086 (2015)
34. Newman, C.M., Stein, D.: Spin Glasses and Complexity. Princeton University Press (2013)
35. Rozikov, U.A.: Gibbs Measures on Cayley Trees. World Sci. Publ, Singapore (2013)
36. Ruelle, D.: Existence of a phase transition in a continuous classical system. Phys. Rev. Lett.
27, 1040–1041 (1971)
37. Ruelle, D.: Statistical Mechanics: Rigorous Results. World Scientific, River Edge, NJ (1999)
38. Sullivan, W.G.: Potentials for almost Markovian random fields. Commun. Math. Phys. 33,
61–74 (1973)
39. Widom, B., Rowlinson, J.S.: New model for the study of liquid-vapor phase transition. J. Chem.
Phys. 52, 1670–1684 (1970)
Exploring the Variety of Random
Documents with Different Content
been suggested, was doubtless inaugurated at a time when descent
was being changed from the female to the male line.
It was perhaps in the latter part of the Middle Status of barbarism
that descent and the rights of succession began to be traced through
males. When, through causes which will be noticed later in this
work, property began to accumulate in the hands of men, children
became the recognized heirs of their fathers and the foundation for
the present form of the family was laid. However, long after descent
began to be reckoned through males, absolute paternity was not
necessary to fatherhood. During the earlier ages of male supremacy,
fatherhood, like brotherhood, was a loose term and signified simply
the head of a house, or the “lord” or owner of the mother. It
mattered little whether a man had previously lent his wife to a
friend, or whether he had shared her favours with several brothers,
all the children “born on his bed” belonged to him and were of his
family.
Later in these pages will be observed the fact that the change in
reckoning descent, which occurred at a comparatively late period in
the history of the human race, is directly connected with the means
of subsistence. So long as land was held in common by the members
of the gens, and so long as women were able to manage the means
of support, their independence was secure, and they were able to
exercise absolute control over their own persons, their homes, and
their offspring. Under these conditions men were obliged to please
the women if they would win their favours.
From facts which have been demonstrated by various writers on
the subject of the early conditions of the human race, it is more than
probable that women were the original tillers of the soil, and that,
during the first period of barbarism, while the hunters and warriors
were engaged in war and the chase, occupations best suited to their
taste, women first discovered the art of producing farinaceous food
through cultivation, and through this discovery a hitherto exclusive
diet of fish and game was changed for a subsistence in part
vegetable.
It is conjectured also that the first domestication of animals was
brought about through a probable “freak of fancy.” That individuals
among these animals were first caught by hunters, conveyed by
them to their homes, and there tamed through the tenderness and
sympathy of women, is considered more than likely. There are,
however, so far as I know, no actual facts upon which to base such a
conclusion.
The increase of subsistence through horticulture and the
domestication of animals marks an important era in the history of
mankind. By this means the human race was enabled to spread itself
over distant areas, and through the improved condition of nutrition
alone, by which the physical conditions were improved and the
mental energies strengthened, the arts of life were multiplied and
the course of human activities directed into higher and more
important channels. Indeed, through the numerous benefits derived
from the one source of increased and improved subsistence, the
entire mode of life was changed or materially modified.
The religious idea, which subsequently comprehended a
complicated system of mythology based on phallic worship, at this
early age, consisted simply of a recognition of the bounties of earth.
The principal office connected with the religious ceremonies of the
Iroquois tribe of Indians, at the stage of development in which it
was first known to Europeans, seems to have been “Keeper of the
Faith,” a position occupied alike by both sexes. The Keepers of the
Faith were chosen by the wise members of the group; they were
censors of the people, with power to report the evil deeds of
persons to the council. “With no official head, and none of the marks
of a priesthood, their functions were equal.”105 For the most part,
their religious services consisted of festivals held at stated seasons
to celebrate the return of the bounties of Nature. A notable fact in
connection with this subject is, that during the earlier ages of
barbarism the religious idea was thoroughly monotheistic, and
idolatry was unknown, religious worship, for the most part,
consisting of a ceremony of thanksgiving, with invocations to the
Great Mother-Nature to continue to them the blessings of life. As
altruism waned and egoism advanced, however, supernaturalism, or
a belief in unseen forces, became more and more pronounced, until,
in the Latter Status of barbarism, when the supremacy of man had
become complete, the gens became merely the “centre of religious
influence and the source of religious development.”
The earlier governmental functions were administered through a
council of chiefs elected by the gentes. The thoroughly democratic
character of the gens may be observed in the fact that any member,
female or male, who desired to communicate with the council on
matters of public interest, might express her or his opinion either in
person or through an orator of her or his own selection.106 Hence,
we observe that government originated in the gens, which was a
pure democracy.
Regarding the council of the gens, Mr. Morgan remarks:
It was a democratic assembly because every adult male and

female member had a voice upon all questions brought before
it. It elected and deposed its sachem and chiefs, it elected
Keepers of the Faith, it condoned or avenged the murder of a
gentilis, and it adopted persons into the gens. It was the germ
of the higher council of the tribe, and of that still higher of the
confederacy, each of which was composed exclusively of chiefs
as representatives of the gentes....
All the members of an Iroquois gens were personally free,
and they were bound to defend each other’s freedom; they
were equal in privileges and in personal rights, the sachem and
chiefs claiming no superiority; and they were a brotherhood
bound together by the ties of kin. Liberty, equality, and
fraternity, though never formulated, were cardinal principles of
the gens. These facts are material because the gens was the
unit of a social and governmental system, the foundation upon
which Indian society was organized.... At the epoch of European
discovery the American Indian tribes generally were organized
in gentes with descent in the female line. The gens was the
basis of the phratry, of the tribe, and of the confederacy of
tribes.107
From the foregoing it would seem that the gens—the earliest

organization of society of which we have any accurate knowledge—
was founded on the “mother-right” or on the supremacy of women.
We are assured that the gentile organization is not confined to the
Latin, Grecian, and Sanskrit-speaking tribes, but that it has been
found “in other branches of the Aryan family of nations, in the
Semitic, Uralian, and Turanian families, among the tribes of Africa
and Australia, and of the American aborigines.”108
A tribe was composed of several gentes, the chiefs of which
formed the council. This council was invested with the power to
declare war and to regulate terms of peace, to receive embassies
and make alliances; it was in fact authorized to perform all the
governmental functions of the tribe. The duties performed by the
council of chiefs may be regarded as the first attempt at
representative government. In process of time, as the affairs of the
tribe became more complicated, a need arose for a recognized head,
one who when the council was not in session could lead in the
adjustment of matters pertaining to the general interest of the
group. In response to this demand, one of the sachems was
invested with a slight degree of authority over the other chiefs.
Hence arose the military chieftain of the Latter Status of barbarism.
That the powers delegated to the incumbent of this office differed
widely from those of a modern monarch, is shown in the fact that as
he had been elected by the members of the group he could by them
be deposed. We have seen that the powers exercised by sachem
and chief were alike transmitted through women. The mother is the
natural guardian of the family; so soon therefore as the actions of
the leaders of the group were not in accord with those principles of
equality and justice which had characterized society since its
organization, they were deposed, or, as in the case of the Senecas
described by Ashur Wright, they had their “horns knocked off”
through the influence of women.
At the head of the family, or gens, producing and controlling the
principal means of subsistence, and forming the line of descent and
inheritance, women, until the closing ages of the Middle Status of
barbarism, were without doubt the leading spirits, and thus far the
progress of mankind had been in strict accord with those principles
which since the separation of the sexes had governed development.
In process of time, however, the simple form of government which
has been described was found inadequate to meet the demands
arising from the more complicated requirements of increasing
numbers and the general growth of society; therefore, during the
opening ages of the Latter Status of barbarism, a form of
government was evolved which was better suited to their changed
conditions. When the idea of a coalescence of tribes, or of a
combination of forces for common defence had taken root, and
when under such confederation the council of chiefs had become co-
ordinated with a military leader for the general management and
defence of the community, it was thought that an important step
had been taken in progressive governmental functions. Yet, along
with the higher development of the governmental idea is to be
observed also a growing tendency toward the usurpation of power.
Scarcely was the office of military chieftain created, than we find the
people inaugurating measures with which to protect themselves
against encroachments upon their liberties, and devising means
whereby they might be enabled to check the personal ambition of
their leaders.
The extreme egoism developed within the male constitution was
already manifesting itself in the excessive greed for gain, and in the
inordinate thirst for military glory; hence, as a safeguard against
usurpation, in the earliest stages of the Latter Status of barbarism,
we find the tribe electing two military chieftains instead of one, two
leaders invested with equal powers and responsibilities and
subjected to the same restrictions and limitations in the exercise of
authority. The Spartan government upon its first appearance in
history is characterized by the existence of two war-chieftains, who,
by historians of later ages, have been designated as kings; a closer
investigation, however, of the functions performed by them shows
that they were lacking in nearly all the prerogatives which
characterize a modern sovereign.
So jealously had the rights of the people been guarded that the
basileus or war-chief of the Latter Status of barbarism, who is said to
represent the germ of our present king, emperor, and president, had
not succeeded in drawing to himself the powers exercised by a
monarch of modern times. The selection of a military leader, during
the Latter Status of barbarism, doubtless represents the first
differentiation of the civil from the military functions of government,
and indicates a virtual acknowledgment of the fact that society had
outgrown the primary and more simple form of government
administered by the council of chiefs.
The third stage in the development of the idea of government was
represented by a council of chiefs, a military commander, and an
assembly of the people. In this further growth of the administrative
functions may be discovered the same solicitude for individual liberty
and the rights of the community which had characterized the former
stage of development, and also the fact that still greater precautions
were deemed necessary to insure the people against tyranny and
the usurpation of their established rights. The council of chiefs,
although representing a pure democracy, and co-ordinated with two
military chieftains, between whom was an equal division of power
and responsibility, was found to be an insufficient safeguard against
despotism; hence the measures devised for the management of the
confederacy must henceforth be subjected to an Assembly of the
People, which, although of itself unable to originate or propound any
plan of government, was invested with the power to accept or reject
any measures offered for adoption by the council.
The gens was able to carry mankind through to the opening ages
of civilization, at which time the council of chiefs was transformed
into a senate, and the Assembly of the People assumed the form of
the popular assembly, from which have been derived our present
Congress and the two houses of the English Parliament.
By a careful study of the growth of government, it is discerned
that liberty, fraternity, and equality were the original and natural
inheritance of the human family, and that tyranny, injustice, and
oppression are excrescences which subsequently fastened
themselves upon human institutions through the gradual rise of the
egoistic principle developed in human nature. We have seen that
until the beginning of the Latter Status of barbarism, the gens
constituted a sovereign power in the tribe; women controlled the
gens, and sachem and chief were alike invested with the authority
necessary for leadership because they could trace their descent to
some female ancestor who was the acknowledged head of the
people, and whose influence and patronage must have extended
over all the individuals included within the recognized bond of
kinship.
With the deposing power in the hands of women, and with the
precautions which were taken by them against injustice or
usurpation of rights, it is plain that unless some unusual or
unprecedented circumstances had come into play, they never could
have lost that supremacy which, as the natural result of their
development, had been maintained by females since the separation
of the sexes.
CHAPTER IV
THE ORIGIN OF MARRIAGE
I will be master of what is mine own;
She is my goods, my chattels; she’s my house,
My household stuff, my field, my barn,
My horse, my ox, my ass, my everything.
The Taming of the Shrew.

It is an obvious fact that so far as her sex relations are concerned
the position of civilized woman is lower than that of the female
animal.
The question which presents itself at this stage of our inquiry is:
What were the causes which led to the overthrow of female
supremacy or what were the processes by which man gained the
undisputed right to the control of woman’s person? By contrasting
the industrial position of women under gentile institutions with that
of later times, after they had become the sexual slaves of men, it
will be seen that the question of economics is deeply involved in this
change. Although the early independence of women is now
recognized, the fact of their industrial supremacy is for the most part
ignored. Indeed the part performed by woman in originating and
developing human industries is seldom referred to by those dealing
with this subject.
As the activities best suited to the tastes of primitive man were
confined to war and the chase, those occupations and pursuits which
were necessary for the preservation of the group were carried on by
women. The reason for this is obvious. Fathers were not regarded as
being related to their offspring. The mother was the only recognized
parent. As the land was held in common, women were economically
free. They were absolutely independent of men for their support.
Under these conditions the importance of women’s position may be
easily perceived.
Not only did women establish the first industries, but they
invented and constructed the tools and implements by which these
industries were carried on. Women were the first tillers of the soil. It
was they who conceived the idea of preserving seeds whereby
farinaceous food might be produced. Corn was not only raised by
them but by them it was ground and further prepared for use. They
built clay granaries in which to store their food products and tamed
the cat to protect them. Implements for tilling the soil, and devices
for grinding the grain were invented by women. They were the first
architects and the first builders. They first conceived the idea of
making cloth with which to protect the body. They were the first
spinners and the first weavers. They invented the first spindles and
the first looms. Their attempts at decoration were the beginning of
art.
As these pioneers in industry were without means of
transportation other than their backs, some of the difficulties which
they encountered may be readily perceived. It must be borne in
mind that for primitive women there was no accumulated store of
knowledge and no previous race-experiences; neither were there
any established rules or precedents to guide them. All methods and
utilities had to be worked out by woman’s unaided brain. When the
conditions under which these pioneers in industry laboured are
considered, and when one reflects on the obstacles which must have
presented themselves at every step along their untried pathway, it
would almost seem that their early achievements were quite as
remarkable as are those which have since been accomplished by
men.
The fact is observed that woman assumed the rôle of protector
and provider, not as is commonly asserted because she was
compelled by man to become a beast of burden, but because she
was the recognized guardian not only of infant life but of the public
welfare. Later, after the primitive groups began to coalesce to form
the tribe, after wife-capture became prevalent and men thereby
secured the right to the control and ownership of individual women,
a right which they still claim, then and not till then did women
become beasts of burden. Then and not till then did man gain the
right to the control of woman’s person.
It is now known that wife-capture is the origin of our present form
of marriage, and that the establishment of the family with man at its
head rests on the same basis. It is also known that through forcible
marriage and the economic conditions which it entailed, woman
became a dependent, a mere appendage to her male mate. The
dominion of man and the assumed inferiority of woman are the
direct results of the authority which he was able to exercise over her
in the marital relation.
We have seen that prior to the decline of female influence women
taken prisoners in war were not regarded as the legitimate property
of their captors. On the contrary, female captives were adopted into
the gens and invested with the same status of personal
independence enjoyed by the original members of the group. Later,
however, female prisoners began to be regarded as the special booty
of their captors, and as belonging exclusively to them; and although
in primitive times marriage outside the limits of related groups was
prohibited, owing to the esteem in which military chieftains came to
be held, this claim was at length allowed them. Any courageous
young warrior, conscious of his popularity, might gather about him a
band of his clansmen and march against a neighbouring tribe, the
women taken prisoners during such expeditions being the special
prizes of their captors.
These prisoners were entitled to none of the privileges of the
community into which they were taken; and as the hostility felt
toward unrelated tribes had become so strong as to be shared by
women, the captive woman could no longer look for pity even from
her own sex.
From this time in the history of the race may be traced the decline
of woman’s power and the subjection of the natural female impulses.
As, at this stage, within the limits of their own tribe, women held the
balance of power in their own hands, and as they still exercised
unqualified control over their own persons, the acknowledged
ownership of one woman, who, being a “stranger,” was without
power or influence, would be an object much to be desired, and one
for which a warrior would not hesitate to brave the dangers of a
hostile camp. Hence, female captives were in demand, and the
women of warring tribes were sought after singly and in groups. In
process of time wars for wives became general and under the new
regime women had the fear of captivity constantly before their view
as a condition more to be dreaded than death.
In the Mahabharata of India it is stated that formerly “women
were unconfined and roved about at their pleasure, independent.”
Finally, marriage was instituted and a woman was bound to a man
for life. One of the eight forms of legalized marriage in the code of
Manu was that of capture de facto and was called Racshasa. This
particular form of conjugal union was practised exclusively by the
military classes, among which, the women taken in battle were the
acknowledged booty of their captors. A definition of this kind of
marriage is as follows: “The seizure of a maiden by force from her
house while she weeps and calls for assistance, after her kinsmen
and friends have been slain in battle or wounded, and their houses
broken open, is the marriage called Racshasa.”
Capture as the prescribed form of marriage for warriors may be
traced through thousands of years and among various peoples. Of
the three legalized forms of marital union in Rome, that by capture
was the one in use among the plebeians, the patricians at the same
time practising Confarreatio and Usus. In Arabia, as late as
Mohammed’s time, the carrying off of women was recognized as a
legal form of marriage.109
That capture constituted a legal form of marriage among the
Israelites, or that women taken captives in war were appropriated as
sexual slaves, is shown by their religious history, in which the
instructions given to the Lord’s chosen people after they had taken a
city was to “smite every male thereof with the edge of the sword:
But the women, and the little ones, and the cattle, and all that is in
the city” they were to take unto themselves. This, it will be noticed,
is to be done “unto the cities which are very far off,” and which “are
not of the cities of these nations.”110
When the Israelites 12,000 strong marched against the Midianites,
they were commanded by Moses to slay all the males, adults and
children, and all the women except the virgins. These virgins of
whom there were 32,000 were to be spared and utilized as wives by
the victorious Israelites. The fact will be noted that these women
had been taken from their own people, hence they were wholly
without influence or power. They were dependents and therefore
subject to the will of their masters. They were sexual slaves or
wives.
In Australia, among the North American Indians, the tribes of the
Amazon and the Orinoco, in Hindustan and Afghanistan, marriage by
actual capture is still practised, and many of the details connected
with the modus operandi have been given by various writers. The
following from Sir George Gray, relative to this form of marriage as it
exists at the present time among some of the native Australian
tribes, is quoted by Mr. J. F. McLennan.
Although a woman give no encouragement to her admirers,
many plots are laid to carry her off, and in the encounters
which result from these, she is almost certain to receive some
violent injury, for each of the combatants orders her to follow
him, and in the event of her refusing, throws a spear at her. The
early life of a young woman at all celebrated for beauty is
generally one continued series of captivity to different masters,
of ghastly wounds, of wanderings in strange families, of rapid
flights, of bad treatment from other females, amongst whom
she is brought a stranger by her captor; and rarely do you see a
form of unusual grace and elegance, but it is marked and
scarred by the furrows of old wounds; and many a female thus
wanders several hundred miles from the home of her infancy,
being carried off successively to distant and more distant
points.111
In an account describing the search for wives by the natives of

Sydney, Collins says:
The poor wretch is stolen upon in the absence of her

protectors. Being first stupefied with blows, inflicted with clubs
or wooden swords, on the head, back, and shoulders, every one
of which is followed by a stream of blood, she is then dragged
through the woods by one arm, with a perseverance and
violence that it might be supposed would displace it from its
socket. This outrage is not resented by the relations of the
female, who only retaliate by a similar outrage when they find
an opportunity. This is so constantly the practice among them
that even the children make it a play game, or exercise.112
By various travellers and explorers, the fact has been observed

that certain symbols representing force in their marriage ceremonies
are in use among nearly if not all extant tribes which have reached a
certain stage of growth. To such an extent, in an earlier age, has the
forcible carrying-off of women prevailed, that among most of these
tribes a valid marriage may not be consummated without the
appearance of force in the nuptial ceremonies. In reference to these
symbols, we have the following passage from Mr. McLennan:
Meantime, we observe that, whenever we discover symbolical

forms, we are justified in inferring that in the past life of the
people employing them, there were corresponding realities; and
if, among the primitive races which we examine, we find such
realities as might naturally pass into such forms on an advance
taking place in civility, then we may safely conclude (keeping
within the conditions of a sound inference) that what these now
are, those employing the symbols once were.113
Among primitive tribes, the area controlled by each was small,

therefore vigilance in maintaining their possessions was one of their
chief duties, and hostility to surrounding tribes a natural condition.
Subsequently, however, when friendly relations began to be
established with hitherto hostile tribes, they are found entering into
negotiations to furnish each other with wives. It was at this time that
marriage by sale or contract was instituted, an arrangement by
which the elder men in the tribe could be accommodated with
foreign wives, at the same time that their own daughters and sisters
became to them a source of revenue.
In Uganda many men obtain wives by exchanging daughters and
sisters with each other. Of this practice C. Staniland Wake says:
This is not an unusual mode of proceeding in different parts

of the world. The perpetuation of the monopoly of women
enjoyed to a great extent by the older men of the tribe among
the Australians is, according to Mr. Howitt, encouraged by those
having sisters or daughters to exchange with each other for
wives.114
Not unfrequently actual capture is practised side by side with

fiction—violent seizure being in active operation among the same
tribes at the same time with the symbol, the frequency of actual
violence depending partly on the extent to which hostility prevails
between the tribes, and partly on the degree of “uniformity
established by usage in the prices paid for wives.” Among certain
tribes, when a dispute arises concerning the price to be paid for a
bride, if the man is able to seize the woman and carry her off to his
tent, the law recognizes her as his wife and nothing is left for the
relations to do in the matter but to accept his terms as to the price.
The peoples among which actual capture is at present practised,
and those among which wives are procured by sale or contract,
represent two different stages in the development of the institution
of marriage, and it is owing to this fact that the symbols used
among the latter may be traced to the realities in which they
originated.
Of the Bedouins of Mt. Sinai, Burckhardt says that marriage is a
matter of sale and purchase, in which the inclination of the girl is
disregarded.
The young maid comes home in the evening with the cattle.
At a short distance from the camp she is met by the future
spouse and a couple of his young friends, and carried off by
force to her father’s tent. If she entertains any suspicion of their
designs she defends herself with stones, and often inflicts
wounds on the young men, even though she does not dislike
the lover, for, according to custom, the more she struggles,
bites, kicks, cries, and shrieks, the more she is applauded ever
after by her own companions.115
In reference to the Mezeyne Arabs the same writer observes that

a similar custom prevailed within the limits of the Sinai Peninsula,
but not among the other tribes of that province.
A girl having been wrapped in the Abba at night, is permitted

to escape from her tent, and fly into the neighbouring
mountains. The bridegroom goes in search of her next day, and
remains often many days before he can find her out, while her
female friends are apprised of her hiding-place, and furnish her
with provisions. If the husband finds her at last (which is sooner
or later, according to the impression that he has made upon the
girl’s heart), he is bound to consummate the marriage in the
open country, and to pass the night with her in the mountains.
The next morning the bride goes home to her tent, that she
may have some food; but again runs away in the evening and
repeats these flights several times, till she finally returns to her
tent. She does not go to live in her husband’s tent until she is
far advanced in pregnancy; if she does not become pregnant,
she may not join her husband till a full year from the wedding-
day.116
Cranz says that in Greenland “some females, when a husband is

proposed to them will fall into a swoon, elope to a desert place, or
cut off their hair.... In the latter case they are seldom troubled with
further addresses.” The refractory bride is dragged
forcibly into her suitor’s house, where she sits for several days
disconsolate, with dishevelled hair, and refuses nourishment.
When friendly exhortations are unavailing, she is compelled by
force and even with blows to receive her husband. Should she
elope, she is brought back and treated more harshly than
before.117
Wherever friendly relations have been established between the

tribe of the wife and that of the husband, he pays a price to her
relatives for the privilege of removing her to his camp. This purchase
price, together with the simulated hatred of the woman’s friends,
signifies a sacrifice on the part of the wife and her family. In Nubia
when a man marries he presents his wife with a wedding-dress, and
gives her also a pledge for three or four hundred piastres, half of
which sum is paid her in case of a divorce. Divorces, however, are
very rare.118
Among the Circassians, after the preliminaries have been settled
by the parents, the lover meets his bride-elect by night in some
secluded spot, and with the assistance of two or three of his best
friends seizes her and carries her away. Sometimes the pretended
capture takes place in the midst of a noisy feast. The woman is
usually conducted into the presence of a mutual friend, where, on
the following day, her friends, simulating anger, seek her and
demand a reason for her abduction. Although the affair is usually
settled at once by the bridegroom paying the accustomed price for
his bride, custom requires that there shall be still further
manifestations of anger on the part of her friends; so, on the
following day, all the relatives of the bride, armed with sticks,
proceed to the place where the bride is in waiting, there to meet the
bridegroom and his friends who have come to carry off the bride. A
sham fight ensues, in which the bridegroom and his party are always
victorious. Among certain of the Arabian tribes the bridegroom must
force his bride to enter his tent, and in France, as late as the
seventeenth century, a similar custom prevailed.
In describing a wedding dance in Abyssinia, Parkyns observes:
This dance is performed by men armed with shields and

lances, who with bounds, feints, and springs attack others
armed with guns, so as to approach them, and at the same time
avoid their fire, while the gunners make similar demonstrations,
and at last fire off their guns either in the air or into the earth,
and then, drawing their swords, flourish them about as a finish.
Finally the bridegroom fires off a gun and immediately rushes

across to where the bride and her female relations are stationed.119
Tylor informs us that a Scandinavian warrior generally sought to
gain his bride by force, that he conceived it beneath his dignity to
win her by pacific means. That the affair might appear more heroic,
he waited until the object of his choice was about to wed another,
and was actually on her way to the nuptial ceremony, when with his
friends he would surprise the wedding cortege, seize the bride, and
carry her off. It has been said of Scandinavian marriages that they
were matters of deep anxiety to the friends both of the bride and
groom, who, until the wedding was over, remained at home in
suspense fearing an attack of the kind already mentioned. It was
customary for a party of young men to station themselves at the
church door, and, as soon as the ceremony was completed, to carry
the news to the homes of the wedded pair. “Within a few
generations the same old practice was kept up in Wales, where the
bridegroom and his friends, mounted and armed as for war, carried
off the bride,” and in Ireland they used even to hurl spears at the
bride’s people, though at such a distance that no one was hurt.120
In the Amazon valley the bride is always carried away by violence.
Among the Zulus, although a purchase price is paid for a woman,
custom requires that a wife, after having been captured, shall make
three attempts to return to her own home.
Of the marriage customs in ancient Sparta, Plutarch says: “In their
marriages the bridegroom carried off the bride by violence.”121 In
Rome we have the familiar example of the Sabine women, who were
captured or carried off by force.
A notable fact in connection with the subject of capture is, that
the mother of the bride, or, in case the mother is dead, the nearest
female relative, is the individual who assumes the part of the
principal defender in this ceremony. She it is who attempts to rescue
the bride, and who more than any other mourns the fate of the
captured wife. Among primitive peoples, with the exception of the
symbol of wife-capture in marriage ceremonies, there is perhaps
none more significant than that typifying the hatred of the mother
for the captor of her daughter. Customs indicating estrangement or,
actual aversion to sons-in-law, usually, if not always, accompany
marriage by capture.
The fact that the change in the relative positions of the sexes, as
indicated by the sadica and ba’al forms of marriage in Arabia, was
not easily or speedily accomplished, is apparent not only in the
symbols of wife-capture everywhere practised among peoples in a
certain stage of development, but is strongly suggested also by the
aversion found to exist among these same peoples between
mothers-in-law and sons-in-law, whether appearing as a reality or as
a symbol.
Among the Arawaks of South America, it is unlawful for the
son-in-law to look upon the face of his mother-in-law. If they
live in the same house a partition separates them, and if by
chance they must enter the same boat, she must precede him
so as to keep her back toward him.
Among the Caribs, all the women talk with whom they will, but
the husband dare not converse with his wife’s relations except on
extraordinary occasions.122 Mr. Tylor refers to the fact that
In the account of the Floridian expedition of Alvar Nuñez,

commonly known as Cabeca de Vaca, or Cow’s Head, it is
mentioned that the parents-in-law did not enter the son-in-law’s
house, nor he theirs, nor his brother-in-law’s, and if they met by
chance, they went a buckshot out of their way, with their heads
down and eyes fixed on the ground, for they held it a bad thing
to see or speak to one another.
It is observed by Richardson, an author quoted by Tylor, that

among the Crees, while an Indian lives with his wife’s family, his
mother-in-law must not speak to or look at him. In some portions of
Australia, “the mother-in-law does not allow the son-in-law to see
her, but hides herself if he is near, and if she has to pass him makes
a circuit, keeping carefully concealed within her cloak.”
Among some of the tribes in Central Africa, from the moment a
marriage is contracted, the lover may not behold the parents of his
future bride. When a young man wishes to marry a girl, he
dispatches a messenger to negotiate with her parents regarding the
presents required and the number of oxen demanded. This being
arranged, he may not again look upon the father and mother of his
intended wife; “he takes the greatest care to avoid them, and if by
chance they perceive him they cover their faces as if all ties of
friendship were broken.” We are told, however, that this indifference
is only feigned, that they feel the same friendship as before, and in
conversation extol one another’s merit. Mr. Caillie says that this
custom extends beyond the relations; if the lover is of a different
camp, he must avoid all the inhabitants of the lady’s camp, except a
few intimate friends who are permitted to assist him in his love-
making. A little tent is set up for him in the neighbourhood, under
which he is to remain during the day. If he has occasion to cross the
camp he must cover his face. He may not see the face of his
intended throughout the day, but at nightfall he may creep silently to
her tent and remain with her until the dawn. These clandestine visits
are continued for a month or two when the marriage is solemnized.
At the wedding festival the women collect round the bride singing
her praises and extolling her virtues.123
Gubernatis is authority for the statement that, in many parts of
Italy the bride is compelled to go through the process of weeping on
her wedding-day, also for the fact that one of the marriage customs
prevalent in Sardinia is identical with that which appeared among
the plebeians at Rome, namely, the pretence of tearing the bride
from the arms of her mother.124
From the facts which have been obtained relative to the practice
of wife-capture, it is only natural to suppose that the mother of the
captured wife would be her chief ally and defender; that such has
been the case seems to be clearly shown by the symbols of distrust
and aversion everywhere manifested between mothers-in-law and
sons-in-law among the various existing uncivilized races. The
practice of wife-capture exists either as a reality or as a symbol
entering into the marriage ceremonies among the tribes of Central
Africa, the Indians of North and South America, in Australia, in New
Zealand, in Arabia, in the hill tracts of India, among the Fuegians,
and in the islands of the Pacific Ocean, and wherever this system is
found the symbol of hatred between mother-in-law and son-in-law
also prevails.
The simulated anger and sham violence connected with marriage
ceremonies among friendly peoples, which are so far removed from
a time when actual capture was practised as to be ignorant of the
true significance of these symbols, show the extent to which
marriage is based on the idea of force on the one side and unwilling
submission on the other.
As the numerous Arabian clans in the time of Mohammed
represented the varying stages of advancement from the second
period of barbarism to civilization, the constitution of Arab society at
that time affords an excellent opportunity for observing the growth
of the institution of marriage, and the various processes by which
the former supremacy of women was overthrown.
One of the principal objects of war at the time of the Prophet is
said to have been the capture of women for wives, a practice which
was recognized as lawful. Under Islam the custom of forcibly
carrying off women for wives was universal and was carried on side
by side with the system of marriage by contract or sale. The position
of the captured woman, however, differed somewhat from that of
the purchased wife. The former, having been forcibly carried away
from her home, lost the protection of her friends, while the
purchased wife, although she relinquished the authority which had
formerly been exercised by women within the gens, and although
she surrendered her person to her “lord,” did not forfeit her right to
the protection of her own family in case of abuse.
Although in Arabia, under the form of marriage by sale or
contract, the wife lost the right to the control of property belonging
to her own gens, she did not, as in Rome, forfeit her claim to the
protection of her kindred. If she received ill treatment within the
home of her husband, her relatives, who were still her natural
protectors, were bound to redress her wrongs. In Rome, on the
contrary, under a system representing a later stage in the
development of marriage, the wife was adopted into the stock of her
husband whose rights over her person were supreme, at the same
time that her kindred renounced the right to interfere in her behalf.
It is to the fact, that in early Arabia the wife never relinquished
her hold upon her own relations, that we are to look for an
explanation of the high social position of Arabian women. We are
assured that it is “an old Arab sentiment, and not Moslem,” that
women are entitled to the highest respect, and that as mothers of
the tribe they “are its most sacred trust.”
According to Professor W. R. Smith in Mohammed’s time, in
addition to the two forms of marriage mentioned, namely, that by
capture and that by sale or contract, there existed also a more
ancient form known as the sadica—a form of conjugal union which
was a remnant of the matriarchal system. By observing the facts
connected with this last-named institution, we shall be enabled to
understand something of the position occupied by women during the
earlier ages of human existence before wife-capture became
prevalent.
Among certain tribes just prior to Islam, upon the event of
marriage, the man presented the woman with a sum of money,
which offering was simply an acknowledgment of the favour which
she was conferring upon him. The husband went to live with the
wife in her tent, and as the contract was for no specified length of
time, he was at liberty to go whenever he tired of the conditions
imposed on him by his wife and her relations. Any children, however,
that were born as a result of this union belonged to the mother and
became members of her hayy. If she desired him to go, she simply
turned the tent around, “so that if the door had faced east it now
faced west, and when the man saw this he knew that he was
dismissed and did not enter.” In relation to these marriage customs
Professor Smith says: “Here, therefore, we have the proof of a well-
established custom of that kind of marriage which naturally goes
with female kinship in the generation immediately before Islam.” Of
this kind of marriage the same writer observes:
The motă marriage was a purely personal contract, founded

on consent between a man and a woman, without any
intervention on the part of the woman’s kin.... Now the fact that
there was no contract with the woman’s kin—such as was
necessary when the wife left her own people and came under
the authority of her husband—and that, indeed, her kin might
know nothing about it, can have only one explanation: in motă
marriage the woman did not leave her home, her people gave
up no rights which they had over her, and the children of the
marriage did not belong to the husband. Motă marriage, in
short, is simply the last remains of that type of marriage which
corresponds to a law of mother-kinship, and Islam condemns it,
and makes it “the sister of harlotry,” because it does not give
the husband a legitimate offspring, i. e., an offspring that is
reckoned to his own tribe and has rights of inheritance within
it.125
Before the separation of the Hebrews and Aramæans, the wife

remained within her own tent where she received her husband, the
children of such unions taking her name and becoming her heirs.
This kind of conjugal union is known to have been in existence in
many portions of the world. In Ceylon it is designated as the beena
marriage.
In ancient Arabia, not only did women control their own homes,
but they owned flocks and herds, and were absolutely independent
of male relations. As late as the fourteenth century of our era,
although the women of certain Arabian tribes were willing to marry
strangers, they never followed them to their homes.
Among the Bedouins it is a rare thing for a woman at marriage to
leave her home and kindred. When a woman marries a man he
settles among her kinsmen, and, as she presents him with a spear
and a tent by way of dowry, it would seem that he is expected to
join her relations and assist in the common defence. The marks of
authority under gentile rule are the possession of a tent and a lance;
yet we find that these are the objects which, under matriarchal
usages, the wife tenders her husband when he enters her family;
the first doubtless as a symbol of her protection, the second as
indicating her authority and the services which he is expected to
render her and her people. Until a late period in Rome it was the
custom, during the solemnities of marriage, to pass a lance over the
head of the wife in token of the power which the husband was about
to gain over her.126
Under the two types of marriage—namely, motă and ba’al—the
positions of women were so diametrically opposed that both could
not continue, hence when under the pressure brought to bear upon
them, women began to accept the ba’al form of marriage within
their own hayy, motă unions were doomed. Of the more ancient
form of marriage in Arabia, under which the woman chooses her
mate, evidences of which are still extant in that country, and that by
capture under which she becomes the slave of her lord, Professor
Smith says:
There is then abundant evidence that the ancient Arabs

practised marriage by capture. And we see that the type of
marriage so constituted is altogether different from those unions
of which the motă is a survival, and kinship through women the
necessary accompaniment. In the one case the woman chooses
and dismisses her husband at will, in the other she has lost the
right to dispose of her person and so the right of divorce lies
only with the husband; in the one case the woman receives the
husband in her own tent, among her own people, in the other
she is brought home to his tent and people; in the one case the
children are brought up under the protection of the mother’s kin
and are of her blood, in the other they remain with the father’s
kindred and are of his blood.
All later Arabic marriages under the system of male kinship,
whether constituted by capture or by contract, belong to the
same type; in all cases, as we shall presently see in detail, the
wife who follows her husband and bears children who are of his
blood has lost the right freely to dispose of her person; the
husband has authority over her and he alone has the right of
divorce. Accordingly the husband, in this kind of marriage, is
called not in Arabia only, but also among the Hebrews and
Aramæans, the woman’s “lord” or “owner,” and wherever this
name for husband is found we may be sure that marriage is of
the second type, with male kinship, and the wife bound to the
husband and following him to his home.127
Notwithstanding the humane enactments of Mohammed in the

interest of women, their position steadily declined, such enactments
having been overbalanced by the establishment of marriages of
dominion, by the growing idea that sadica or motă marriages were
not respectable, and that women could not depend upon their
relations to take their part against their husbands. The history of
religion shows that its growth has always followed the same course
as have the ideas concerning the relative importance of the sexes.
The god-idea and the fundamental doctrines of religion are always
found to be in harmony with the established principles and ideas
relative to sex domination and superiority. The religion of
Mohammed was essentially masculine, all its principles being in strict
accord with male supremacy; it is not, therefore, singular that when
the weight of religion was added to the already growing tendency
toward ba’al marriages that sadica marriages were doomed.
In Arabia, as elsewhere, the duties of the purchased wife were
specific. The present which under the older form of marriage had
been given to the bride as a love-token, or as an acknowledgment of
the husband’s devotion to her, subsequently took the form of a
purchase price, and was claimed by her father and brothers as a
compensation for the loss sustained by the group through the
removal of her offspring, whose services belonged to their mother’s
people. In other words, the husband paid a price to the wife’s
relations for the right to raise children by her which should belong
exclusively to his kin—children which should she remain within her
own home would belong to her kindred. The wife was therefore
removed to the husband’s hayy, where, so far as the sexual relation
was concerned, his rights over her were supreme.
We have observed that wherever the possessions of the gens
continued to be the property of all its members, and were controlled
by women, the man at marriage went to live with the woman; so
soon, however, as men began to claim the soil, and property began
to accumulate in their hands, the wife went to reside with her
husband and his family as a dependent. Among various tribes, the
form of marriage in use depends on the means of the contracting
parties; if the man is able to pay to the woman’s father or brothers
the full price charged for her, she goes to him as his slave—she is his
property as much as is his dog or his gun; if, however, he is unable
to pay the amount charged, he goes to live with her and her family,
and becomes their slave.
In Japan, among the higher classes, upon the marriage of the
eldest son, his bride accompanies him to his paternal home; but, on
the other hand, when the eldest daughter marries, her husband
takes up his abode with her parents. Eldest daughters always retain
their own names, which their husbands are obliged to assume. As
the wife of an eldest son becomes a member of her husband’s
family, and the husband of an eldest daughter joins the family of his
wife and assumes her name, the eldest son of a family may not
marry the eldest daughter of another family. Regarding the younger
members of the household, if the husband’s family provides the
house, the wife takes his name, while if the bride’s family furnishes
the home the bridegroom assumes the name of the wife.128
In the marriage customs of various nations, and in their ideas
relative to the ownership and control of the home, may be observed
something more than a hint of the principal causes underlying the
decline of female power. Wherever women remain within their own
homes, or with their own relations, they are mistresses of the
situation; but when they follow the fathers of their children to their
homes, they become dependents and wholly subject to the will and
pleasure of their husbands.
It is plain, however, that under a system of marriage by sale or
contract, although a woman might exercise little influence in the
home of her husband, so long as her relations stood ready to defend
her she would enjoy an immunity from abuse. The fact that a
woman can count upon her relations for protection against her
husband, shows plainly that in a certain stage of marriage by
contract or sale, women are not the abject slaves which they have
been represented to be. Although in the Fiji Islands a man may seize
a woman and take her to his home, she does not remain with him
unless agreeable to her inclinations.129
Amongst the Abipones, a man, on choosing a wife bargains

with her parents about the price. But it frequently happens that
the girl rescinds what has been agreed upon between the
parents and the bridegroom, obstinately rejecting the very
mention of marriage.130
Among the Charruas of South America, divorce is quite optional.

In Sumatra, if a man carries off a virgin against her will, he incurs a
heavy fine, or if a man carries off a woman under pretence of
marriage, “he must lodge her immediately with some reputable
family.”131
Although in the earlier ages of marriage by sale or contract,
daughters were regarded as the property of their fathers, still that
stage had not been reached at which women were accounted simply
as sexual slaves. The Arabs practised marriage by sale or contract,
yet they jealously watched over their women,—they “defended them
with their lives and eagerly redeemed them when they were taken
captive.” They thought it better to bury their daughters than to give
them in marriage to unworthy husbands.132 According to the
testimony of J. G. Wood, Kaffir women are very tenacious about their
relations, probably, it is thought, for the reason that husbands are
more respectful toward wives who have friends near them, than they
are to those who have no relations at hand to take their part.133
Usually among the Kaffirs, according to Mr. Shooter, although a man
pays a price to the parents of the woman whom he wishes to marry,
the affair is by no means settled; on the contrary, he must undergo
the closest scrutiny by her before she will consent to accept him.
Bidding him stand, she surveys first one side of him, then the other,
the relations in the meantime standing about awaiting her decision.
Upon this subject Mr. Wood remarks: “This amusing ceremony has
two meanings: the first that the contract of marriage is a voluntary
act on both sides; and the second, that the intending bridegroom
has as yet no authority over her.”134
Although under the system of marriage by sale or contract a
woman has a voice in the selection of her husband, and although
she can count on her kinsmen to protect her against abuse, still,
practically, the contract brings the wife under the same condition as
a captured wife; she follows her husband to his home, where, as a
dependent, he exercises control over her person and her children. In
Arabia prior to the time of the Prophet the wife could claim the
protection of her kindred against her husband, yet the principle
underlying marriage by contract and that by capture was the same,
except that under the former the husband paid a price for the
woman’s sexual subjection, while under the latter, not only in sexual
matters, but in all others as well, he was her “lord” and master.
The Prophet says: “I charge you with your women, for they are
with you as captives (awânî).” Professor Smith informs us that
according to the lexicons awânî is actually used in the same sense as
married women generally.135 For long ages after ba’al marriages had
been established, so degrading was the office of wife that women of
rank were considered too great to marry.
After relating some interesting accounts of certain practices in
common with the custom of capture among the Brazilian tribes, Sir
John Lubbock says:
This view also throws some light on the remarkable
subordination of the wife to the husband, which is so
characteristic of marriage, and so curiously inconsistent with all
our avowed ideas; moreover it tends to explain those curious
cases in which Hetairæ were held in greater estimation than
those women who were, as we should consider, properly and
respectably married to a single husband. The former were
originally fellow-countrywomen and relations; the latter captives
and slaves.136
With the development of the egoistic principle, or when

selfishness and the love of gain became the rule of action, the
protection of her kindred, which in an earlier age a woman could
count on against her husband, was withdrawn, and daughters came
to be looked upon as a legitimate source of gain to their families. On
this subject C. Staniland Wake remarks:
Women by marriage became slaves, and it was the universal

practice for a man who parted with his daughter to be a slave to
require a valuable consideration for her. Moreover, as a man can
purchase as many slaves as he likes, so he can take as many
wives as he pleases.137
Thus arose polygamy.

In Rome, in the Latter Status of barbarism and the opening ages
of civilization, woman, at marriage, forfeited all the privileges
belonging to her as a member of her own family, while within that of
her husband no compensatory advantages were granted her. Even a
proprietary right in her own children was denied her, and from a
legal point of view the wife became the daughter of her husband,
and not unfrequently the ward of her own son.
After the power gained by man over woman during the latter ages
of barbarism had reached its height, the family was based not on the
marriage of a woman and a man, but upon the power of a man over
a woman and her offspring, or upon the absolute authority of the
male parent. In Rome a man’s wife and children were members of
his family not because they were related to him but because they
were subject to his control. At this stage in the development of the
family, the father had the power of “uncontrolled corporal
chastisement” and of life and death over his children.138 If it was his
will to do so, he could even sell them. Indeed, a son’s freedom from
paternal tyranny could be gained only by the actual sale of his
person by his father. Relating to the control exercised by the father
over his children, it is observed that he had the right “during their
whole life to imprison, scourge, keep to rustic labour in chains, to
sell or slay, even though they may be in the enjoyment of high state
offices.”139 If a father granted freedom to his son, that son was no
longer a member of his family.
That, with the exception of force, there is no quality in the male
constitution capable of binding together the various individuals born
of the same father, is apparent from the past history of the human
race. Mr. Parkyns, referring to the character of the Abyssinians,
observes that the worst point in the constitution of their society is
the want of affection among relations, “even though they be children
of one father.” He says that the animosities which keep the tribes in
a constant state of warfare do not exist among the offspring of the
same mother and father, but, as the children of polygamous fathers
are more numerous than own brethren, fraternal affection is a rare
thing.140 A comparison between the family group under archaic
usages at a time when woman’s influence was in the ascendency,
and the Roman family under the older Roman law, will serve to show
the wide difference existing between the altruistic and egoistic
principles as controlling agencies in the home and in society.
A significant fact in connection with this subject is here suggested,
that, although for untold ages women were leaders of the gens, so
long as their will was supreme, no human right was ever invaded,
and no legitimate manly prerogative usurped; but, on the contrary,
all were equal, and the principles of a pure democracy were firmly
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade
Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.
Let us accompany you on the journey of exploring knowledge and

personal growth!
ebooknice.com

[Ebooks PDF] download (Ebook) Statistical Mechanics of Classical and Disordered Systems : Luminy, France, August 2018 by Véronique Gayrard, Louis-Pierre Arguin, Nicola Kistler, Irina Kourkova ISBN 9783030290764, 9783030290771, 303029076X, 3030290778 full chapters

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

[Ebooks PDF] download (Ebook) Statistical Mechanics of Classical and Disordered Systems : Luminy, France, August 2018 by Véronique Gayrard, Louis-Pierre Arguin, Nicola Kistler, Irina Kourkova ISBN 9783030290764, 9783030290771, 303029076X, 3030290778 full chapters

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

[Ebooks PDF] download (Ebook) Statistical Mechanics of Classical and Disordered Systems : Luminy, France, August 2018 by Véronique Gayrard, Louis-Pierre Arguin, Nicola Kistler, Irina Kourkova ISBN 9783030290764, 9783030290771, 303029076X, 3030290778 full chapters

Uploaded by

Copyright:

Available Formats

Download Full Version ebook - Visit ebooknice.

(Ebook) Statistical Mechanics of Classical and

Click the button below to download

Discover More Ebook - Explore Now at ebooknice.com

Start reading on any device today!

(Ebook) Biota Grow 2C gather 2C cook by Loucas, Jason;

(Ebook) Matematik 5000+ Kurs 2c Lärobok by Lena

(Ebook) SAT II Success MATH 1C and 2C 2002 (Peterson's SAT

(Ebook) Master SAT II Math 1c and 2c 4th ed (Arco Master

(Ebook) Statistical Mechanics of Disordered Systems: A

(Ebook) The Statistical Mechanics of Irreversible

(Ebook) Internet of Things, Smart Spaces, and Next

(Ebook) Classical Hopf Algebras and Their Applications

This book series features volumes composed of selected contributions from

More information about this series at http://www.springer.com/series/10533

Nicola Kistler Irina Kourkova

ISSN 2194-1009 ISSN 2194-1017 (electronic)

© Springer Nature Switzerland AG 2019

Statistical mechanics aims at understanding the behavior of systems involving very

of a GREM in random magnetic ﬁeld. The paper by Bolthausen provides a new

New York, USA Louis-Pierre Arguin

Nature Versus Nurture: Dynamical Evolution in Disordered

Louis-Pierre Arguin Department of Mathematics, Baruch College and Graduate

Abstract The Widom-Rowlinson model is an equilibrium model for point particles

Keywords Gibbs measures · Stochastic time-evolution · Gibbs-non Gibbs

AMS 2000 subject classification 60K57 · 82B24 · 82B44

Collaborations with Benedikt Jahnel, Sascha Kissel, Utkir Rozikov.

deterministic or stochastic transformations may lead to non-Gibbsian measures [10].

2 Gibbs on Lattice, Sequentially Gibbs, Marked Gibbs

We start by recalling the notion of an infinite-volume Gibbs measure for lattice

2.1 Specifications and Gibbs Measures on the Lattice

γ (γ (dω|·)|ω̃) = γ (dω|ω̃) (1)

for all finite volumes ⊂ Zd . It is suggested by the tower property of condi-

M1 ({−1, 0, 1}) by definition has probability kernels

where Z (ωc ) is the normalizing partition function. If  is a finite-range potential

G(γ) := {μ ∈ M1 (), μγ = μ, for all Zd } (3)

2.2 Hard-Core and Soft-Core Widom-Rowlinson Model on

We will consider here the version of the hard-core Widom-Rowlinson model on Zd

configurations with +− neighbors are suppressed, but not forbidden.

2.3 Sequential Gibbsianness for Mean-Field (and

There is an intrinsic formulation of the Gibbs property which is suitable also in

Take (μ N ) N ∈N a sequence of exchangeable probability measures μ N on the finite-

lim μ N (dω1 |ω[2,N ] ) = γ(dω1 |ν) (7)

exists whenever the empirical distributions of a configuration (ωi )i=2,3,4,... converge,

This has to hold for all limiting empirical distributions of conditionings ν ∈

2.4 Marked Gibbs Point Processes in Euclidean Space

2.5 Widom-Rowlinson Model in Euclidean Space

A hyperedge potential (η, ω) is by definition allowed to depend on the marked point

3 Dynamical Gibbs-Non Gibbs Transitions

We now come to time-evolutions. Consider again the Euclidean space Widom-

3.1 Relation to Disordered Systems

η → μt (ηi |ηZd \i ) (12)

μ̄t (dω, dη) = μ(dω)Pt (ω, dη). (13)

Analyze hidden phase transitions in first-layer measure constrained on the future

μ̄t (dω|η). (14)

A relation to disordered systems with quenched order appears when we view

3.2 Results on Dynamical GnG Transitions for Euclidean

A marked infinite-volume configuration ω ∈ (that is a signed point cloud) is called

where Z (ωc ) is the normalizing partition function. If is a finite-range potential

A hyperedge potential (η, ω) is by definition allowed to depend on the marked point

as ω ⇒ ω in the sense of local convergence.

Ci j (γ) = sup γ{i} (·|ω) − γ{i} (·|η)T V,i (17)