The Origins of Unfairness Social Categories and Cultural Evolution 1St Edition Cailin Oconnor Full Chapter

The Origins of Unfairness: Social
Categories and Cultural Evolution 1st

Edition Cailin O'Connor
Visit to download the full and correct content document:
https://ebookmass.com/product/the-origins-of-unfairness-social-categories-and-cultur
al-evolution-1st-edition-cailin-oconnor/
The Origins of Unfairness
The Origins of
Unfairness
Social Categories and Cultural
Evolution
Cailin O’Connor
1
3
Great Clarendon Street, Oxford, ox2 6dp,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Cailin O’Connor 2019
The moral rights of the author have been asserted
First Edition published in 2019
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2018967875
ISBN 978–0–19–878997–0
Printed and bound by
CPI Group (UK) Ltd, Croydon, cr0 4yy
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
To Jim
Acknowledgments
So many people have contributed to the production of this book. Many

thanks first to my colleagues Jeffrey Barrett, Jean-Paul Carvalho, John
Duffy, Simon Huttegger, Louis Narens, Mike McBride, Brian Skyrms,
Kyle Stanford, and James Weatherall for feedback at various stages of
this project. Thanks also to the many graduate student researchers who
worked on the project including Calvin Cochran, Emma Cushman, Travis
LaCroix, Aydin Mohseni, Sarita Rosenstock, Hannah Rubin, and Mike
Schneider. And thanks to other students at UC Irvine for feedback along
the way, including Nikhil Addleman, Gerard Rothfus, Gregor Greslehner,
and Ben Conover. Thanks to the Experimental Social Science Laboratory
at UC Irvine, and those involved in running it, for facilitating experimen-
tal work on the cultural Red King effect.
Several visiting fellowships proved crucial in creating the space and
time to write this book. Special thanks to the Pittsburgh Center for the
Philosophy of Science, and to John Norton. Thanks to my fellow fellows
and postdocs, especially Agnes Bolinska, for ideas and feedback. Thanks
to the MCMP at LMU Munich, the London School of Economics, and
Australia National University for visiting fellowship positions during the
time I was writing this book. And thanks so much to the many people at
each of these universities who listened to talks on the project, and gave
comments and feedback.
Many thanks to both Ellen Clarke and Kevin Zollman for extensive
comments on the manuscript, and thanks to Liam K. Bright and Remco
Heesen for comments along the way. To all the other people who gave
comments and feedback at talks but who I have forgotten about – thank
you!
There were several collaborators whose work and insights were invalu-
able. These include Liam K. Bright, Justin Bruner, Calvin Cochran, Travis
LaCroix, Aydin Mohseni, Hannah Rubin, and Mike Schneider. Special
thanks to Justin Bruner, whose paper on the cultural Red King was the
inspiration that eventually led to this book, and who collaborated with me
x acknowledgments
on so many of the papers described here. And thanks to Nicole Bourbaki

for ongoing support.
This material is based on work supported by the National Science
Foundation under STS grant 1535139 “Social Dynamics and Diversity in
Epistemic Communities”. Many thanks to NSF for their support, and to
Fred Kronz in particular.
So many thanks to my family and friends for putting up with me during
the painful process of writing a first book. Maureen and James Weatherall,
all my thanks for the extensive childcare and emotional support! Jim
Weatherall, you are my rock. Eve and Vera, you are ridiculous.
List of figures
1.1. Payoff table for a simple, correlative coordination game 28

1.2. Payoff tables for two simple, correlative coordination
games. (a) shows one where outcome B vs. B is preferred
to A vs. A by both actors. (b) shows one where actors have
different preferences over the two outcomes 30
1.3. Payoff table for the dancing game 31
1.4. Payoff tables for two simple, complementary coordination
games. In (a) actors both prefer one coordination outcome
to the other. In (b) actors have conflicting preferences over
the two coordination outcomes 32
1.5. Payoff table for hawk–dove 33
1.6. A general payoff table for a two–person, two-strategy game 34
1.7. A division-of-labor game with an egalitarian, but less
efficient, option 35
2.1. A general payoff table for a symmetric two-person,
two-strategy game 50
3.1. A dancing game 68
3.2. The phase diagram for a single population playing the
dancing game 69
3.3. A phase diagram for the dancing game evolved with the
two-population replicator dynamics 71
3.4. A phase diagram for the leader–follower game evolved
with the two-population replicator dynamics 73
3.5. Correlative coordination game with a preferable outcome
for both players 80
3.6. A stag hunt 81
4.1. Payoff table for an MFEO game where α and β determine
the benefits to the two actors for reaching the preferable
equilibrium 89
xii list of figures
4.2. Proportions of simulations that go to the jointly less

preferable equilibrium for a perfectly divided population
playing the MFEO game where α = β 90
4.3. Phase diagrams for perfectly divided populations playing
MFEO games where α = β varies. As α = β increases, the
size of the basin of attraction for the better equilibrium
also increases 91
4.4. An asymmetric leader–follower game 93
5.1. Payoff tables for two simplified Nash demand games 107
5.2. Basins of attraction for fair and unfair outcomes between
types in the Nash demand game with three strategies 112
5.3. Basins of attraction for fair and unfair outcomes between
types in the Nash demand game with various numbers of
strategies 114
5.4. Payoff table for a three-strategy Nash demand game with
threatpoints 119
5.5. Basins of attraction for a model where one type has a
higher disagreement point in the Nash demand game 120
5.6. Payoff table for a Nash demand game where one player
carriers out a threat in response to a High demand 121
5.7. Nash demand game with different outside options for each
player 127
5.8. Proportions of outcomes where an initially powerful
group maintains power for a Nash demand game where
disagreement points are determined by past success 130
6.1. A general game where we can potentially observe a Red
King/Queen effect 136
6.2. A Nash demand game with demands 4 and 6 136
6.3. Phase diagram for two populations playing the
two-strategy Nash demand game 137
6.4. Phase diagrams for two populations playing a Nash
demand game where one population evolves m times as
quickly as the other. We see a Red King effect 137
list of figures xiii

quickly as the other. As m increases, the Red King grows
stronger, but is bounded 139
quickly as the other. We see a Red Queen effect 140
6.7. Basins of attraction for two types playing the Nash
demand game with a minority group 142
demand game. With restricted starting points, a general
Red Queen can translate to a Red King 147
demand game. If actors tend to display in-group
preference, the cultural Red King is strengthened 148
6.14. Payoff table for a Nash demand game with background
payoffs for one actor 156
6.15. A population with two dimensions of demographic
category, gender and race 158
6.16. Basins of attraction for actors with four intersectional
types playing a Nash demand game 160
6.17. Basins of attraction for actors with four intersectional
types playing a Nash demand game 161
7.1. Basins of attraction for collaboration for two types playing
a Nash demand game with an outside option 166
xiv list of figures
7.2. A network with two types represented by black and white

nodes. Edges represent interactive links between agents 167
7.3. An analog of the cultural Red King effect in a network
model of agents playing the Nash demand game. Results
are averaged over parameter values with H = 6 170
7.4. An evolving network where eventually agents break all
out-group links, leading to total homophily 172
7.5. Increasing discrimination corresponds to increasing
homophily in network bargaining models 173
8.1. A three-strategy complementary coordination game 183
8.2. A three-strategy complementary coordination game with
better equilibria for both players 184
8.3. A three-strategy complementary coordination game
where players have some conflict of interest 184
8.4. Basins of attraction for the three equilibria in the
household coordination game as a function of α, β = .6 186
8.5. Basins of attraction for the three equilibria in the
household coordination game as a function of β, α = .1 186
9.1. Phase diagrams for two populations. In the first, the top
right equilibrium is harder to escape 202
9.2. A Nash demand game where actors come to feel badly
when making High demands of another player 204
9.3. A Nash demand game where actors come to feel badly
when making High demands of another player, but not
enough to disrupt the pure strategy equilibria 204
Introduction
Imagine you are in a group of ten people. In a minute, you will all be
randomly paired with a partner. At the count of three, without a chance
to talk or communicate in any way, you must dance the tango. If you both
step forward, you’ll collide. If you both step back, you’ll look stupid. If one
of you steps forward, and the other back, you’ll do the dance successfully.
This is an example of a coordination problem—a situation where actors
have similar interests but nonetheless face difficulties in coordinating
their action. Presumably neither you nor your partner really cares which
one of you steps forward and which back, at least not as much as you care
about executing complementary actions. In other words, what you really
care about is coordination.
Now imagine a slightly different scenario. You are in a group of five men
and five women who will each be paired with a partner of the opposite
gender. And again, at the count of three, you dance the tango. This is
another example of a coordination problem, related to the first, but with
an extra element, which is that the group is divided into two observably
different types.
One thing that is immediately obvious about these coordination prob-
lems is that one is easier to solve than the other. In the second case,
just a small amount of information (something like one person shouting
“women step back”!) would be enough to get the entire group coordinat-
ing effectively. If a group can be easily divided into types, and can agree
ahead of time that certain types take certain actions, this eliminates the
need for extensive planning later.
Likewise, one could imagine a scenario where instead of a group of
men and women one was in a group with people of two different races,
or observably different religions, or redheads and brunettes, or tall and
short people, elderly and young people, goths and band geeks. In any of
The Origins of Unfairness: Social Categories and Cultural Evolution. Cailin O'Connor,
Oxford University Press (2019). © Oxford University Press.
DOI: 10.1093/oso/9780198789970.003.0001
2 introduction
these cases the coordination problem is easier to solve because there are
visible traits that the actors can take advantage of when coordinating their
action.
Now imagine a slightly different scenario. You are in a room with ten
really hungry people and five pizzas. Everyone is going to split a pizza
with a random partner. There are many ways to divide the pizza—you
could get one slice and your partner five, or you could each eat three
slices, or you could even decide that one partner will get all the pizza—
and in order to have a peaceful, happy lunch each pair is going to have to
choose a division. This sort of situation is usually referred to as a bargain-
ing problem, but notice that it too demands a sort of social coordination.
No one wants to leave pizza behind, or to argue over the pizza.
One way to avoid squabbles is to decide ahead of time on some division
that everyone will follow. Probably the one that sounds most natural to
you is the 50–50 split. This sounds attractive, of course, because it’s fair.
Everyone gets the same amount of pizza. But there is something else
attractive about this split. Suppose the group instead agrees that everyone
should divide their pizza 80–20. Once random pairing happens, there
then must be further deliberation over who gets 20% and who 80%. This
is because the 80–20 split, unlike the 50–50 one, is asymmetric. In fact,
50–50 is the only symmetric division of pizza available to the group
(assuming they want to eat the whole pizza), and thus the only one that
completely solves the lunch problem ahead of time.
Now once again imagine the same set-up but with five women and five
men, each pair of whom will go on a date. In this case, the group can solve
their problem by agreeing on the 50–50 split, but they can also solve their
problem by agreeing that each woman gets 80% and each man 20% of
the pizza once pairings are made. The addition of gender here means that
one single decision is enough to coordinate on an inequitable division in
a way that wasn’t possible for a uniform group.
• • •
Humans continually face real coordination problems. Consider, for
example, division of labor. Dividing labor in an organized way is crucial
to the success of human groups. Households where everyone cleans
and no one cooks are unsuccessful, as are societies where everyone is
a soldier and no one is a farmer. Just like the silly dancing problem
introduction 3
described above, in these real problems members of a group have to take

complementary social roles to be successful.
Dividing resources is also a ubiquitous part of human interaction. In
the workforce, coworkers must decide who will obtain the benefits, in
terms of salary, bonuses, time off, etc., of their joint action and who will do
the work to produce these goods. In the household, partners must decide
who does how much of the work (dividing the resource of free time) and
who receives which economic goods produced.
The central aim of this book will be to explore the ways in which
social categories—especially gender, but also categories like race and
religion—interact with and contribute to social solutions to problems of
coordination and resource division. In particular, this book uses formal
frameworks—game theory and evolutionary game theory—to explore
the evolution of norms and conventions that piggyback on seemingly
irrelevant factors like gender and race to solve these problems. As we
will see, these frameworks elucidate a variety of topics from the innate-
ness of gender differences, to collaboration in academia, to household
bargaining, to minority disadvantage, to homophily. In particular, these
frameworks help show how inequity can emerge from simple processes of
cultural change. In groups with gender and racial categories, the process
of learning conventions of coordination and resource division is such
that in a wide array of situations some groups will tend to get more and
others less. Clark and Blake (1994) wrote that: “. . . explanation of the
origins of institutionalized social inequality and political privilege must
resolve the central paradox of political life—why people cooperate with
their own subordination and exploitation in non-coercive circumstances”
(17). The answer to this paradox on the framework here will be that
even if everyone learns to do what is best for themselves at every stage of
cultural evolution, the group will still tend to end up in situations where
people of one sort take advantage of people of another sort. No one is
behaving irrationally in this story, and at every point everyone is making
the most of their social environment. Once we look through a cultural
evolutionary lens, there is no paradox—inequity is the expected outcome
of basic cultural evolutionary processes.
According to Smith and Choi (2007), “In the history of social thought,
accounts of the rise of inequality tend to sort into two categories: those
that emphasize the benefits that hierarchy brings to all . . . and those that
4 introduction
emphasize exploitation or coercion by one segment of society” (118).

This book is divided into two parts. Part I can be loosely understood
as modeling the first sort of inequity, and Part II the second. Another
way to think about this is as follows: Part I shows why social categories
can play a useful social function, despite leading to some inequity. Part II
shows why social categories facilitate processes that lead to inequity with
no beneficial function.
In the first half of the book I focus on the role social categories can play
as symmetry breakers in certain types of coordination problems. I start
by introducing coordination games—simple models that can represent
and inform real coordination problems. I use these games to show how
groups that use categories like gender to coordinate behavior can be
more efficient and more successful, in many cases, than those that do
not. The result, however, is a differentiation of roles that, while mutually
beneficial, are often inegalitarian. As I argue, the success of these social
patterns means that we should expect groups, via social learning or
cultural evolution, to adopt social categories for just this purpose. In the
case of gender, groups take advantage of existing biological sex differences
to create gender roles and conventions.
In the second half of the book, I shift focus to analyze the emer-
gence of more pernicious inequity between social groups. I introduce
bargaining games, intended to represent situations of resource division.
Once categories have been adopted, the cultural dynamics that lead to
bargaining norms are radically changed. New norms that are inequitable,
but not especially efficient, arise. And once they do, they can be self-
perpetuating. In other words, the development of types sets the stage for
serious inequity to spontaneously emerge and to persist between social
groups. I analyze, in particular, the conditions under which one social
group will tend to gain an advantage in terms of bargaining and resource
division.
One theme that runs through both halves of the book is that surpris-
ingly minimal conditions are needed to robustly produce phenomena
related to inequity that we usually think of as psychologically complex.
It takes very little to generate a situation in which social categories
(such as gender) are almost guaranteed to emerge. The preconditions
under which models move toward outcomes that look like discrimina-
tion, inequity, and distributional injustice are, again, very minimal. Once
overview 5
inequity emerges in these models, it takes very little for it to persist

indefinitely.
Of course, in the real world, gender and inequity are psychologically
complex—in-group preference, stereotyping, and biases are clearly part
of the story. Demonstrating that you don’t need these factors to generate
inequity, though, is important for a number of reasons. The models we
will consider show us that even if we eliminate psychological biases, this
will not solve the problem of inequity. They indicate that we need to think
of inequity as part of an ever-evolving process. It is not something we can
expect to fix and be done with.
Along these lines, at the end of the book, I use concepts from evo-
lutionary game theory to address social change. As I argue, a cultural
evolutionary framework can provide sometimes surprising insights into
shifting inequitable norms of resource division. Even when moral edu-
cation seems to have little effect on existing inequitable norms, it can
nonetheless change the underlying strategic situation in ways that erode
the stability of these norms, allowing for change later. Alternatively,
existing fair norms can be eroded in ways that make them unstable, even
though observable behavior has not changed. In general, as just described,
the ultimate picture I present is one where those concerned with social
justice must remain vigilant against the dynamic forces that push toward
inequity.
This book also demonstrates how useful in principle tools from game
theory and evolutionary game theory can be in elucidating causes of and
potential solutions to inequity. And, in particular, it develops a general
framework that can be used to further explore questions related to social
categories like gender, race, caste, and class, and the inequities that emerge
between these categories. I hope the tools developed here prove fruitful
to others in philosophy and the social sciences who are interested in the
emergence of inequity.
0.1 Overview
The body of the book, as mentioned, is divided into two parts.
Part I, which includes Chapters 1 through 4, focuses on coordination,
and, in particular, the case of gendered division of labor. Part II, including
Chapters 5 through 7, shifts gears to focus on inequitable divisions of
6 introduction
resources between social groups more generally. In Chapters 8 and 9, I

draw together insights from both halves of the book.
In Chapter 1, I start by introducing a case that will be used to illus-
trate the role social categories can play in coordination—that of gen-
dered division of labor. I then begin to build the evolutionary modeling
framework used throughout the book by introducing coordination prob-
lems and coordination games. As I point out, once one builds models of
coordination problems it is easy to see that different ones provide different
sorts of challenges to groups. Some problems can be solved when an entire
group arrives at a convention for a single, universal behavior. This will
not work for problems where actors must engage in different behaviors
to succeed. This distinction sets the stage for the rest of Part I, where I
show how social categories can facilitate coordination in the latter kinds
of problem.
In Chapter 2, I introduce the notion of type-conditioning, or differ-
ential treatment of interactive partners based on irrelevant social types
(often referred to as “tags” in the literature). What does this sort of
behavior entail in the real world? How does it function in a model? As
I will show, once type-conditioning is possible in coordination games, it
allows for populations to reach outcomes that can provide a benefit in
terms of group efficiency, and often in terms of individual payoff, but that
may be inegalitarian in character. In the case of gendered division of labor,
groups use gender to divide roles efficiently, but in a way that sometimes
advantages one side. Chapter 3 extends this discussion by showing how
the presence of types and type-conditioning in a model radically alters
cultural evolutionary processes. In particular, I demonstrate that groups
engaging in coordination problems with types reach the beneficial, but
inegalitarian, outcomes just described.
Chapter 4 fleshes out the upshot of these evolutionary models to the
case of gender. First, I show how we can understand conventionality as
coming in degrees, and how this should inform our understanding of the
innateness of gender roles. The second main argument from this chapter
is that from a completely undifferentiated society, the behaviors associ-
ated with gender, and perhaps what we would even want to call gender
itself, can emerge endogenously as part of a solution to coordination
problems.
Part II begins with Chapter 5 where I present an analysis of the role
of power in the emergence of bargaining conventions. As I show, small
overview 7
power asymmetries, like those that emerge in Part I of the book, can trans-
late to advantages for a powerful group. Furthermore, these advantages
can persist in conditions where winning one bargaining contest impacts
the power of a social group and improves their chances of winning future
bargaining contests. In other words, they can compound. In Chapter 6,
I focus on asymmetries in learning environments (rather than power)
and the role they play in the emergence of inequity. I show how minority
status in particular can lead to disadvantage as a result of the different
learning environments that minority and majority members tend to
inhabit. In addition, at the end of this chapter, I consider the role of power
and learning asymmetries in intersectional populations. Throughout, I
highlight how little is needed to generate inequity of a pernicious sort
between social groups.
Chapter 7 extends this analysis to ask: what happens once inequitable
bargaining conventions arise in a social group? In particular, does dis-
crimination lead the oppressed to avoid their oppressors? To address this
question, I look at network models, which explicitly represent interactive
structures between individuals. As will become clear, discrimination can
lead to segregation. Those who suffer discrimination from out-group
members tend to choose in-group members to interact with instead.
Though, as I will illustrate, when one group has advantages with respect
to resources and power, a disadvantaged group will sometimes tolerate
discrimination to gain access to those resources.
Chapter 8 models in greater depth a particular case of interest to
economists and sociologists—the emergence of household division of
labor and household bargaining. The models address how inequity can
emerge in the household, and also why certain patterns of coordination
are likely to arise. In doing this, I draw on both parts of the book, and
show how the different sorts of inequity addressed in Part I and Part II
can interrelate. This exploration leads into the final chapter where I
focus on changing inequitable social patterns of bargaining. In particular,
I use the cultural evolutionary framework developed in the book to
ground a discussion of the conditions that facilitate or hinder norm
change. As I point out, social dynamical patterns may mean that we
are thinking about such change in the wrong way. Instead of concep-
tualizing inequity as a social ill to solve, a more fruitful approach will
treat inequity as a continuing process, requiring continuing effort to
counteract it.
8 introduction
0.2 Explanation and Models of Cultural

Evolution
Before continuing to Part I of the book, I’d like to say a word about the
methodology used.
This book attempts to understand deep mathematical regularities in
some of the social dynamical patterns that arise around gender and
inequity. For the most part, this work is highly abstract and highly
idealized. In almost every case, I provide the models discussed with
interpretations—matching up elements of the models to elements of real-
world situations and, where appropriate, arguing that this match is a
good one. I also use empirical work, when possible, to assess whether
the models discussed are, in fact, providing insight into the phenomena
under discussion. Because the book addresses many related phenomena
these discussions necessarily vary in levels of carefulness. Some of the
models presented are tied to the relevant phenomena quite tightly. Others
are suggestive of the phenomena, but the details are not filled in as
meticulously. This means that the explanatory role of the models will
differ from case to case.
In some cases, the models discussed can be thought of as providing
“how-possibly” information. If something can evolve in an evolutionary
model under basic conditions, we come to believe that these conditions
are enough to possibly support the evolution of that behavior in the
real world. In other cases, I take the models to have deeper explanatory
power, giving us insight into how some patterns of behavior may have
potentially emerged. The difference here is not in the models, but in
the epistemic role they play. In the “how-potentially” cases, the models
are intended to increase our confidence in the potential of a process
to have really occurred.1 In particular, many of the examples of how-
potentially modeling in the book will involve what Weisberg (2007, 2012)
describes as minimalist idealization, where the model pares away causally
irrelevant factors to reveal candidates for the underlying causal variables
responsible for a phenomenon.2 In still other cases, I will argue that
1 These models can be especially important in directing us toward future empirical
research (Rosenstock et al., 2017).

2 Along these lines, Potochnik (2007) presents a picture of why evolutionary models that
appeal to payoffs while abstracting away from the details of the mechanistic interactions that
explanation and models of cultural evolution 9
the models discussed play an important epistemic role by outlining the

minimal conditions for certain social patterns—especially inequitable
ones—to arise, regardless of how these patterns were actually generated
in the real world (O’Connor, 2017b). This kind of “how-minimally”
modeling is especially useful in thinking about intervention. For instance,
suppose we intervene on real groups via implicit bias training. If inequity
emerges under minimal conditions that do not include biases, we should
not expect this intervention to fully solve our problem.
Importantly, sometimes the same model will play multiple epistemic
roles. For example, I provide models of the emergence of gender roles
that I think potentially illuminate how these patterns emerged in the real
world. At the same time, they demonstrate how such roles can possibly
emerge from minimal preconditions. Altogether, the explanatory picture
that emerges echoes Downes (2011), who emphasizes the wide set of
explanatory roles that models can play.3
When it comes to cultural evolution, there are many modeling choices
to employ. Sometimes the debates about the efficacy of these modeling
frameworks (or the efficacy of modeling cultural evolution at all) get
quite hot.4 Part of the problem is that cultural evolution is itself a varied
and sometimes disunified phenomenon. There is no reason to think that
humans, with our big brains and our cornucopia of cultural practices,
should undergo cultural change of the same sort in every case. In fact,
there is room for many different modeling practices to successfully rep-
resent cultural evolution, with the appropriateness of the practice varying
from case to case. (This said, for interested readers Mesoudi et al. (2006)
gives an influential unifying account of cultural evolution, and Henrich
(2015) gives an extensive overview of the varied processes of cultural
evolution and gene-culture co-evolution.)
Of course, any one modeler will have to choose a framework to
elucidate the phenomena they hope to represent. The sorts of behaviors
actually drive evolution are still useful. These models capture what we might describe as core
causes of evolutionary progress.
3 See also, for example, Nersessian (1999); O’Connor and Weatherall (2016).
4 Cultural attraction theorists accuse those using population biology-type models of
trying to fit the square peg of cultural evolution into the round hole of biological evolution.
In response, cultural attraction theory is accused of circular reasoning. Evolutionary game
theorists are criticized for over-simplification, while accusing others of building models that
lack causal transparency.
10 introduction
this book focuses on include things like gendered division of labor, racial
bias, and norm emergence in the workforce. These behaviors are the result
of many processes. They are shaped by (at least) rational (and not so
rational) decision-making, individual learning as a result of past events,
social learning from successful or prominent social models, parent-to-
offspring cultural transmission, and peer-to-peer transmission. In other
words, the real processes shaping these behaviors are massively complex,
and essentially unmodelable in their full detail. Rather than trying to
pull these processes apart, I will focus on a simple change process that
captures some of what happens—especially adaptive changes—in many
of these individual processes. It will not be a perfect representation, but
it can provide understanding while doing well enough. This method
reflects a choice to elevate causal transparency, simplicity, and tractable
explanation over complexity and accuracy. Philosophers of modeling
have argued that models must always trade off desiderata, and this work
is no different (Weisberg, 2012).
One might ask: why use models at all? Why not stick to empirical
data in exploring these issues? Stewart (2010), in a paper modeling the
emergence of racial inequity, compellingly justifies the use of models in
this sort of case. Gender norms and norms of inequity emerge in the
context of dynamical, human interaction. Empirical results gathered at
a single time will fail to capture these interactions. Even if we wanted to
gather dynamical data on the emergence of broad social conventions and
norms, this data is often removed from us in time. Also, it often involves
countless interactions across many, many social actors. In short, it is not
practical to gain a full understanding of the dynamics of the emergence
of conventions and norms in human society via empirical means. Models
can fill the gap. There is something more to say, applying specifically to
cases where social interventions are called for (as in the topics studied
here). Social interventions are costly in terms of time and effort. They also
pose a risk when they impact the lives of those involved. Models present
a way to study counterfactual dependencies in the social realm with
minimal risk, and relatively little cost. They can then be used to direct
further empirical study that is well grounded in theoretical prediction.5
5
Thanks to Liam K. Bright for pulling out this role for the models in this book.
1
Gender, Coordination
Problems, and Coordination
Games
Women in the Ashante tribe of West Africa make pottery to be used day to
day for cooking and storing food. Men, on the other hand, are responsible
for woodworking. In the Hadza tribe, men tend to hunt meat, while
women focus on the acquisition of vegetables. In the United States during
the 1960s, women were primarily responsible for preparing breakfast,
while men did the lawn care.
These patterns are part of what is referred to in humans as the gendered
division of labor. Across all observed societies, it is the case that men and
women have, at least to some degree, divided labor between them. This
creates an explanandum for social scientists—why do we see such pat-
terns? It isn’t as if human groups had to arrange themselves in such a way.
Labor could have been divided by individual preferences or strengths. Or
labor could be undivided, so that each individual does a bit of whatever
job needs doing. This has led to questions like: do men and women have
different innate preferences that cause them to naturally choose different
jobs? Is there a cultural function fulfilled by this division of labor?
Part I of this book will illustrate (among other things) how social cat-
egories, like gender, can break symmetry in certain sorts of coordination
situations, and so allow groups with categories to coordinate better than
groups without them. Because of this functionality, as I will argue, cultural
evolution has taken advantage of social categories, shaping many of our
conventions around them. In order to tell this story, I’m going to make
use of gendered division of labor as a key case. This is, in part, because
The Origins of Unfairness: Social Categories and Cultural Evolution. Cailin O'Connor,
Oxford University Press (2019). © Oxford University Press.
DOI: 10.1093/oso/9780198789970.003.0002
14 gender, coordination problems and games
it is so well studied in the social sciences, and in part because it is a

paradigm example of how irrelevant differences between individuals can
nonetheless become completely central to social coordination. Crucially,
previous authors have argued that gendered division of labor is the
starting point for gender inequality (Okin, 1989; Ridgeway, 2011), and
the cultural evolutionary framework I develop will help inform how
natural processes of learning and cultural transmission might lead groups
to inequitable norms and conventions of this sort. As we will see in
Chapter 4, a cultural evolutionary framework can also shed light on some
puzzling features of the gendered division of labor.
All this is not to say that gendered division of labor is the only
interesting case where the framework developed here might apply. Both
caste and class are social categories that seem to be part of solutions to
coordination problems. Since these cases are very different in their details,
though, it will be beyond the scope of this book to carefully illustrate how
and where this framework applies more broadly. In Part II of the book,
we will consider models that apply straightforwardly to a broader set
of cases.
In the Introduction, I described two simple coordination scenarios—
one where people want to coordinate their dance steps, and another where
they need to decide on a division of pizza. Coordination in the broadest
sense of the word is central to this book. In particular, coordination prob-
lems define the set of strategic, social scenarios where social categories
end up mattering deeply to the evolution of conventions and norms. I’ll
start the chapter with a brief discussion of gender and gendered division
of labor, drawing out the features most relevant to the framework devel-
oped here. Then I’ll discuss generally what coordination problems are and
introduce the models used to represent them—coordination games. I’ll
draw on previous work in economics to explain why division of labor
is, itself, a coordination problem in the standard sense. (Along the way,
we’ll discuss the notions of convention and norm, and the use of game
theoretic models to represent them.) As we’ll see, not all coordination
games are equal. While some can be solved by conventions and norms
that are identical for everyone in a society, others, those that require
people to take different, complementary actions, pose a special problem.
Coordinating behavior in these sorts of games requires extra information
to break symmetry between those who are interacting—who is the one
gender and gendered division of labor 15
who steps forward, and who back? Division of labor, as it turns out, is
just such a problem. Actors have to decide who will do which of several
complementary actions. This will set the stage for the next two chapters,
which will explore how social categories like gender can provide a means
of symmetry breaking in these sorts of cases.
1.1 Gender and Gendered Division of Labor

1.1.1 What is gender?
The answer to this question is not straightforward. Gender is not a simple
concept, nor is it a unified one. Across academic disciplines different
definitions of gender are employed for different reasons.1 Since this book
is not a work of gender theory, but a work of evolutionary game theory
that addresses social categories, we will want to understand the aspects
of gender that have the most to do with strategic social behavior. In
particular, I will focus more here on the role gender plays in such strategic
behavior, rather than, say, how gender shapes personal experience.
Sociologists Candace West and Don Zimmerman, in their seminal
paper, “Doing Gender,” point out that “[i]n Western societies, the
accepted cultural perspective on gender views women and men as
naturally and unequivocally defined categories of being . . . Competent
adult members of these societies see differences between the two as fun-
damental and enduring” (West and Zimmerman, 1987, 128). The claim
is that we tend to think of men and women as inherently different, and as
belonging to clear, distinct categories. These categories seem so natural
that until relatively recently, there was little push to examine them. Money
(Money et al., 1955) was the first academic to use the term “gender role” to
refer to something that associates with biological sex, but is not identical
to it. Theorists have subsequently endorsed the distinction between
biological sex and gender, where, roughly, the former tracks inborn bio-
logical differences, and the latter constructed social categories. (Things
are not actually so simple because, for example, cultural differences
1 In philosophy, Haslanger (2000) provides a discussion of different types of definitions
of gender, and introduces what she calls a pragmatic definition—one that is useful for
positive social change.
shape the development of bodies (Butler, 2011a), but such consider-

ations are beyond the scope of this exploration.2 See also Haslanger
(2015b) for an analysis of the ways culture influences our concept
of “sex”.)
A useful notion of gender for our purposes goes something like this.
Based on innate biological sex differences, we can determine sex cate-
gories, or categories that neatly divide the human population into types
based on sex (West and Zimmerman, 1987; Ridgeway and Smith-Lovin,
1999). These categories piggyback on biological sex differences, but are
separate from them because, for example, intersex people tend to be
assigned to one category or the other and trans people may switch sex
categories (Money and Ehrhardt, 1972).3 Using these sex categories,
societies develop patterns of social behavior that constitute gender. These
patterns are governed by normative expectations for a person’s behavior
across many behavioral arenas.
So what are these normatively governed patterns of behavior? There
are several broad classes of patterned behavior related to gender that I
will focus on here as especially useful for understanding the evolutionary
models that will be developed, though these will not capture all gendered
patterns of behavior. First, cross-culturally, gender, as a rule, is used to
divide labor, but this division is often conventional. In the next section,
this will be elaborated. Second, sex category and gender identity is usually
signaled, often elaborately, through both appearance and behavior. In
Chapter 2, I will further discuss this sort of signaling in service of a
more general discussion of social categories and their instantiation in
models. Third, gendered behaviors are reproduced in human populations
via learning and punishment. In Chapter 3, we will come back to the
relevance of this to evolutionary models of gender.
2 Butler (2011a) argues that sex should be understood “no longer as a bodily given
on which the construction of gender is artificially imposed, but as a cultural norm which
governs the materialization of bodies” (xi).
3 Additionally, in some societies, there are more than two sex categories. For example,
some societies have a third category consisting of biological males who are socially like
women, such as hijras in India and some Native American two-spirit people. Other African
and Native American societies have third categories for biological females who behave like
men by taking the social responsibilities of fathers and husbands (Martin and Voorhies,
1975; Blackwood, 1984; Williams, 1992; Thomas et al., 1997).
gender and gendered division of labor 17
1.1.2 Gendered division of labor

As mentioned, every culture divides labor by gender.4 This is true even
though human societies have taken on radically different modes of
organization cross-culturally and over the course of history ranging,
for example, from traditional foraging societies, to agricultural societies,
to industrial, and to post industrial, societies (Basow, 1992). Within each
of these categories are countless structural differences related to political
organization, social structures, marriage rules, etc.
There are two sorts of ways in which labor is divided, the first relating
to who does which tasks and the second to the overall amount of work
done (Blood and Wolfe, 1960). Under the title “division of labor,” I will be
concerned with division of complementary jobs. The division of overall
amount of labor will fall under “division of resources” and “household
bargaining” in particular, and will be discussed at length later in the book,
especially in Chapter 8.
Although every society divides labor by gender, this division takes
on very different forms (Murdock and Provost, 1973; Dahlberg, 1981;
Costin, 2001; Marlowe, 2010). Murdock and Provost (1973), in a classic
article look at fifty “technological” activities in 185 societies including
things like food collection, production, and preparation, material abstrac-
tion/processing, and manufacturing of articles. They coded these based
on whether the activities were performed exclusively by men, predom-
inantly by men, by both sexes equally, predominantly by women, or
exclusively by women. They found that some activities, like hunting large
game, metal and woodworking, mining, and (puzzlingly) making musical
instruments, were performed almost exclusively by men across cultures.5
Other activities—such as spinning, laundering, cooking (especially veg-
etable food), and dairying—tended to be performed by women. A larger
range of activities, to varying degrees, were performed by one gender
in some societies and the other in others. These included activities like
making rope, planting crops, carrying burdens, caring for small animals,
house-building, etc. Some activities were performed by both genders in
4 Evidence discussed in Gibbons (2011) also suggests that this division of labor by gender
in humans is ancient.
5 Even activities like big game hunting, though, are sometimes performed by women
cross-culturally (Bird and Codding, 2015).

some societies, though it has been subsequently pointed out that many
of these activities, when more carefully parsed, consist of sub-activities
that are divided by gender.6 Subsequent research reports similar findings.
Costin (2001), for example, looks at the manufacture of crafts for trade
and sale and finds a substantial division of labor by gender though great
variety cross-culturally in who makes what. The key observation here
is that for a wide swath of activities, while each group divides them
by gender, there is variability across groups as to which gender does
the job.
In modern societies, like more traditional ones, household labor also
tends to be divided by gender (Blood and Wolfe, 1960; Thrall, 1978;
Pinch and Storey, 1992; Bott and Spillius, 2014). In a classic study Blood
and Wolfe (1960) surveyed families and asked who performed which of
eight household tasks. They found a significant division of household
labor with some tasks usually performed by the husband (repairs, lawn
care, shoveling snow) and some usually performed by the wife (cooking
breakfast, cleaning the living room, and doing dishes), and a few that were
performed by either gender (paying bills and buying groceries). Thrall
(1978), in looking at household division of labor that included children
as well as adults, found that only 20% of the tasks they considered did not
fall largely to one gender or another.
Some have attempted to explain these divisions via appeal to innate
sex differences in humans (as we will discuss in Chapter 4), though the
massive cross-cultural variation in patterns of division of labor presses
against such an explanation. A more promising line of explanation in
game theory appeals to strategic aspects of coordination (Becker, 1981).
The framework I will now begin to develop is in line with this second
sort of explanation, though it takes a bigger-picture approach to the role
of social categories and coordination, and emphasizes the importance of
cultural evolution to understanding these phenomena. Let’s get started
by delving into the sort of strategic situation where social categories like
gender can improve outcomes—coordination problems.
6 Costin (2001) gives a few examples. “Among the Ashante in west Africa, women
produced utilitarian domestic pottery for sale in the marketplace, while men produced ritual
ceramic vessels on order for elite patrons” (2). Here pottery-making would count as one
technological activity, but we can see that classifying this culture as one where men and
women both make pottery misses an important division of labor.
coordination problems 19
1.2 Coordination Problems

What are the basic features of a coordination problem? Informally, we can
boil them down to two. First, actors in a coordination problem usually
have some level of common interest in that they want to coordinate. For
example, in the tango problem described in the Introduction, actors want
to end up at the same sorts of outcomes (where one steps forward and
the other back). Second, coordination problems are problems because
despite the common interest of the actors, it is nontrivial for them to
coordinate their actions to meet these interests. This is usually the case
because there are multiple ways for coordination to happen, and suc-
cessfully coordinating involves settling on one of them. Schelling (1960)
oriented game theorists’ attention to this sort of problem. When actors
must choose between many equally, or nearly equally, good possible joint
coordination outcomes, how do they pick?
We can distinguish two classes of coordination problems. The first I
will call a correlative coordination problem. In this sort of problem, actors
in a social sphere need to coordinate action by making the same choice,
or correlating what they do. A classic example is choosing what side of
the road to drive on. In any society, everyone would like all members to
drive on one side of the road, but it doesn’t really matter which. To give
another example, most societies have standard working hours and this
standard allows for all sorts of further coordination (when restaurants
are open, when trains run a heavier schedule, when people schedule
meetings, when child care is available, etc.) In the US those hours are 8–5
and in Spain they are 9–2 and 4–8. It matters more that these happen
at the same time than that they happen at any particular time. Another
example relates to language. In each society, members would like to agree
on which word means what, but the actual word itself doesn’t matter
(again, within reason).
There are also correlative coordination problems that do not need to
involve broad societal choices, but more interpersonal ones. For example,
a family might wish to do something together, and must somehow all
decide what that will be. Perhaps the group could go to the movies, or
to the beach, or to the fair, and while members may have preferences
about which, their main preference is that they all do the same thing. The
same sort of problem can arise for a couple on a date, or friends trying to
spend time together, or work colleagues who schedule regular meetings
but must decide where and when.
The second class of coordination problems I will call complementary

coordination problems. In these cases, actors need to coordinate action,
but they have to do so by using different strategies, or complement-
ing each other, rather than by all doing the same thing. The tango
problem introduced at the beginning of the book is a good example
of a complementary coordination problem. People coordinate by taking
complementary roles, stepping forward and back, rather than by both
stepping forward or both stepping back. There are many coordination
problems that have this complementary character. Consider things like
getting on and off an elevator or stopped subway car. The people on the
inside and the outside need to coordinate by doing different things, one
group moving while the other waits. Or suppose two people want to order
at McDonald’s. Who goes first? What if two people arrive at a doorway at
the same time—who enters and who waits (or holds it open)?
Division of labor falls squarely into this type of coordination problem
(which will be further elaborated later in the chapter). Consider division
of labor within a modern household. A single household needs a person
to clean the bathroom, take out the trash, do the grocery shopping, etc.7
However, there are many jobs to keep a household running, and it makes
sense to divide them among various members. In this case, everyone
wants to coordinate who does what, but often by not doing the same thing.
(As I will elaborate later, this sort of division of labor is most important for
jobs that require some skill. For entirely skill-less jobs everyone can more
easily do a bit of everything.) In the same way, a village needs a police
officer, a bank clerk, a grocer, etc., and ideally just one person (or a small
number) doing each of these.
Another sort of complementary coordination problem relates to lead-
ership. Human groups often need to coordinate flexibly, by responding to
changes around them. Coordinated group action is facilitated by having
one individual, or a small number, making decisions. For instance, a
military without a strict hierarchy of leaders would be utterly useless.
Even in romantic relationships, friend pairings, or small tribes, it is often
easier to have a decision-maker than having to always make decisions
by committee. In all these cases, actors can coordinate more effectively
by taking different roles—leader and follower—rather than by taking the
7 A similar point can be made about traditional households and groups if we substitute
these jobs for gathering, hunting, toolmaking, etc.

coordination problems 21
same role. Now, though, these roles specify not a particular behavior, but
a pattern of behavior (i.e. making the decisions, rather than driving on
the left side of the road).8,9
One last category of complementary coordination problem we’ll
address, which will become particularly important in Part II of the book,
involves the division of resources. Whenever humans jointly produce
resources, or whenever they otherwise obtain sharable resources, they
must decide who will get what. These problems have a complementary
coordination character because in order to successfully divide resources,
actors must have compatible expectations or demands for what they
receive. We cannot both take home 60% of a carrot harvest. If our
company earns fifty thousand dollars in surplus this year, we cannot
each take home 30k. When two actors form a household, if they jointly
take too much free time (arguably a precious resource) the outcome will
be credit card debt and a pile of dirty dishes. Traditionally, these sorts
of resource division cases have fallen under the heading of “bargaining”
rather than that of “coordination,” and they are usually represented by
a bargaining game. I will wait until Chapter 5 to introduce models that
specifically represent this sort of complementary coordination.
Lewis (1969), in his famous work on convention, calls the differences
between correlative and complementary coordination problems “spuri-
ous” (10). As he points out, by redefining behaviors in a coordination
game, one can go from one type of problem to the other. For instance,
Lewis describes a real-world coordination problem he encountered while
living in Oberlin, Ohio. All phone calls in the town were automati-
cally cut off after three minutes. The coordination problem here was
to determine which party would call back. This, on first glance, looks
8 Notice that in the tango, the two partners must both know who will perform which
of the two basic versions of the steps (the one that starts by going forward or the one that
starts by going back). They also must select one partner to lead. Since there is no preset
choreography for the entire dance, if both dancers try to lead, they will fail to coordinate,
and ditto if they both try to follow. Both sorts of divisions are crucial for coordination in
human groups.
9 Millikan (2005) distinguishes what she calls “leader–follower” conventions from other
sorts of conventions. These are conventions where one actor observably takes the leadership
part of an established convention, and the other actor is then able to take up the follower role
and coordinate. These are distinct from the types of problems just described, because they
involve rigid, or semi-rigid behaviors that actors simply figure out how to divide via a leader–
follower distinction. I refer here to flexible behaviors where the problem is to determine who
will be the leader in general.
like a complementary problem—one actor must call and the other must
not in order to be successful. As Lewis argues, though, one can instead
think of the possible actions here not as “call back” and “wait,” but as
“call back if one is the original caller” and “call back if one is not the
original caller.” On this rebranding, a solution to the problem entails
everyone taking the same action. As will become clear, this redescription
of the problem is only available if there is some way to break symmetry
between the two actors—some extra information available to determine
who does what. (In his case, that asymmetry is provided by the fact that
one person must always be the caller and the other receive the call.) For
this reason, Lewis is overlooking a key difference between these sorts of
problems. The restatement is only possible in some cases, and how to
make a complementary problem a correlative one is a thorny topic in its
own right.
1.3 Coordination, Convention, and Norm

In human groups coordination problems tend to be solved by conven-
tions. When it comes to driving, the convention in the US is to go right,
and the convention in India is to go left. As discussed, conventional
working hours differ in Spain and America. In my childhood family,
the convention is to go for a long walk after dinner, instead of playing
Scrabble. My husband does the cooking, and I do the laundry, whereas
we might have gone the other way.
Lewis (1969) was one of the first philosophers to bring game theory
to bear on social conventions.10 Indeed, he starts with coordination
games in order to define what conventions themselves are. On Lewis’s
account conventions are behavioral regularities in groups of actors faced
with repeated coordination problems. His definition is quite detailed,
but, approximately, for such regularities to constitute conventions it is
necessary that members of the group mostly conform to them, they
expect others to conform to the same patterns, and they have generally
similar preferences over outcomes in the problem (78). In addition, Lewis
10 Though discussion of social convention by philosophers goes much farther back.
Hume, in The Treatise on Human Nature, describes a convention with a coordination

character. “Two men who pull at the oars of a boat, do it by an agreement or convention,
tho’ they have never given promises to each other”(Hume, 1888, Bk III, Pt II, Sec II).
coordination, convention, and norm 23
stipulates a requirement of common knowledge, which is approximately

that each actor involved knows that the things just listed hold true, and
that the others know this, etc. If these conditions obtain, actors should
be expected to continue to conform to their conventional solution to
whatever coordination problem it is they face. Each expects that changing
behavior will detriment them, and so continues to make the same choices.
And, jointly, the actors continue to successfully solve their problem.11
On this definition, only groups of actors with high levels of
rationality—really only groups of humans—can have conventions. Since
Lewis’s Convention, though, a body of work has emerged in philosophy
showing that solutions to coordination problems can emerge endoge-
nously through processes of learning or biological evolution, both for
simple and for more complex actors. Skyrms (2010), for example, inves-
tigates how signaling conventions can emerge in evolutionary models
as solutions to signaling games, which are a branch of coordination
problem (and which Lewis uses to explain linguistic convention). The
striking thing about these sorts of models is that they have served as
successful representations of incredibly diverse sorts of populations—
bacteria, vervet monkeys, humans, businesses looking to hire, etc. This
diversity of explanatory success raises a question: are the solutions
to these problems in human groups importantly different from the
solutions elsewhere? In other words, is there an important difference
between Lewisean conventions in human societies, and, say, alarm calls
in animals? What about sex roles in plants?12
In this book, I employ evolutionary models with low to medium
rationality requirements. I think it is useful to conceive of the solutions
that arise in these as conventional, even when they do not involve expec-
tations or common knowledge, or any of the human-level rationality
requirements of Lewis. Conventions in this sense should be thought of
as behavioral regularities across groups that solve coordination problems
11 Schelling (1960), in his seminal work on coordination, has something similar in mind
as the typical solution to a coordination problem—that actors have mutual expectations,

and expectations about each others’ expectations, driving them to a solution.
12 Cao (2012), in a discussion of whether signaling games can represent neuronal sig-
naling, makes a point relevant to distinguishing between these sorts of conventions. As she
observes, if we look at different cases where signals “might have been otherwise” (the usual
bare requirement for conventionality), some of these are easy to change now (like language),
and others very hard (like bacterial signals). The sense of “could have been otherwise” in the
latter cases appeals to deep evolutionary counterfactuals.
broadly defined. This is not a careful definition, but the goal of this book
is not to provide an analysis of convention. The claim here, note, is not
that every convention is a solution to a coordination problem. We will
simply focus on a particular set of conventions that are.13
I would like to draw a few distinctions that will be useful later. The first
distinction is between solutions to coordination problems that consist
in behavioral regularities, on the one hand, and such solutions that
have obtained normative force. Conventions, on the definition here,
need not carry normative force. If bacteria have evolved a chemical
signal that solves a coordination problem, this constitutes a convention.
It should be obvious that if some bacterium fails to send this signal,
the other bacteria will not shun or punish the dissenter, and there will
be no expectations that something different should have happened. In
human groups, on the other hand, conventional solutions to coordination
problems often also constitute norms in the sense that members of
the population feel that they themselves and others ought to act in a
particular way.14
Arguably, most human conventions acquire some sort of normative
force, and previous authors have argued that this is always the case.
Gilbert (1992), for example, argues that conventions are norms because
they consist in joint acceptance that a group ought to behave in a certain
way. On the definition from Lewis (1969, 97) conventions are norms
because one will be going against one’s best interest by switching behavior,
and so others will believe one ought to conform to the conventional
behavior. Furthermore, others will respond badly to a failure to conform
since it matters to their payoffs.15 Weber (2009) defines convention as a
13 The notion that every convention solves a coordination problem has been successfully
challenged. Gilbert (1992) offers a thorough critique of Lewis’s account of convention where
she gives examples of social conventions that cannot be represented by Lewis’s proper
coordination equilibria. Millikan (2005), likewise, points out that conventions like saying
“Damn” when you stub your toe are not well represented as solutions to coordination
problems. Binmore (2008) argues against Lewis’s common knowledge requirement for
conventions, using an evolutionary game theoretic perspective.
14 Although it is slightly orthogonal to this discussion, readers might be interested
in Anderson (2000), who discusses several approaches to understanding social norms,

including rational-choice and evolutionary-based approaches.
15 Sugden (2000) outlines in detail why Lewisian conventions obtain normative force via
an analysis of mutual expectations. And Guala (2013) uses an experiment to show that it is
very easy for conventions in the Lewis sense to gain normative force, though he argues that
they gain an intrinsic normativity rather than a “should” based on a consideration of others’
payoffs as described by Lewis.
coordination, convention, and norm 25
“binding” custom, where failure to meet it will lead to “sanctions of dis-

approval” (127), implying that normative force is attached to conventions
by definition.
This running together of conventions and norms in the human case
obscures the fact that there is a continuum along which conventions
hold normative force. For example, it is a convention that people wear
formal attire to a wedding. Failing to do so will annoy and clearly is in
violation of a social norm, but not to an extreme degree. Failing to drive
on the correct side of the road, however, is an egregious norm violation
and will tend to create quite a lot of consternation. Furthermore, there
exist human conventions which meet the definition here, and even which
meet Lewis’s requirements, but almost entirely lack normative force.16
Millikan (2005), who defines conventions as patterns of behavior that
are reproduced, where part of this reproduction depends on the force of
precedent, agrees that a distinction should be made between conventions
and norms because many conventions are not typically followed (and thus
are missing the “should”).17 An analysis from Bicchieri (2005) is perhaps
most useful here. She distinguishes between conventions—behavioral
regularities which actors wish to follow if they expect others to follow
them because of the strategic structure of the interaction—and social
norms, which actors wish to follow if others expect them to or will punish
them for deviance. As she points out, stable conventions can become this
sort of social norm over time, though this does not necessarily happen.
She points out that this is especially likely when breaking a convention
will lead to “negative externalities.”18
Throughout the book, I will distinguish between conventions and
norms, though, as mentioned, I think these are best understood as
existing on a continuum. Because the work of the book consists in
16 My husband and I have a convention of watching “The Office” together at the end of
the day, but there is no sense in which either of us feels that we ought to do this. (Probably we
ought to get to work on the laundry and dishes.) Failure to abide by this convention would
lead to absolutely zero disapproval or censure by the other party.
17 She gives an example of handing out cigars after having a child. This is not a convention
of the sort I am concerned with here because it does not solve a coordination problem.
18 Arguably, by definition, breaking a convention that solves a coordination problem will
lead to a negative externality since it will lower a partner’s payoff from what is expected.
There are degrees to which this can happen, though. I might disappoint my husband by
deciding not to watch The Office, but if I decide not to follow driving rules I might kill
someone.
using evolutionary models to understand the emergence of behavioral

regularities, I will focus on conventions to a much greater degree.
A second distinction has to do with the type of underlying coordi-
nation problem. Lewis (1969) claims that conventions are arbitrary, in
the sense that they could have been otherwise. This arbitrariness is a
key aspect of all accounts of convention. In her extensive critique of
Lewis’s account, though, Gilbert (1992) points out that there are coordi-
nation problems, in his sense, where “one of the two proper coordination
equilibria gives each player a payoff vastly superior to the other, while
the other gives each player a payoff little better than zero” (342). In
other words, one way of coordinating will be strongly preferred by both
players. For such problems, she claims, solutions will not be arbitrary in
the right sense for them to be conventions. The tension in this critique
can be resolved by pointing out that there is a continuum along which
conventions are more or less arbitrary (as well as a continuum along
which they are more or less normative). For some problems, there are
multiple possible outcomes that might be equally good solutions. The
left and the right sides of the road fall under this heading. In other sorts
of problems, there are multiple solutions, but they vary with respect to
goodness. For example, I cited working hours as a solution to a social
coordination problem, but it is not the case that any hours will work. If
working hours were from midnight until 8 am, people would be unhappy
and unwell.19 For still other problems, there might be one solution that is
very clearly better than the others—in a pair of friends learning to rock
climb, having the one with experience do the lead climb is clearly better
than having the inexperienced climber do it, though the complementary
roles could potentially be filled by either member. Coordination problems
can be thought of as existing on a spectrum from those where there
are many equally good solutions, and those where some solutions are
more attractive (Simons and Zollman, 2018). I will call solutions on the
former end of the spectrum more conventional and those on the latter end
less conventional (and sometimes more functional). In Chapter 4, I will
return to this theme to give a simple formal measure intended to capture
where on this spectrum a coordination problem lies. As we will see, this
understanding of conventions as having varying degrees of arbitrariness
will help elucidate the conventionality of patterns of gendered division
19
See, for example, Davis et al. (2001).
coordination games 27
of labor. Let’s now turn to game theory to start building models of

coordination problems.
1.4 Coordination Games

Game theory was developed as a framework for modeling strategic
interactions among humans. By “strategic,” I mean any interaction that
involves multiple actors who choose how to behave, where these actors
care about what their partners do.20 Coordination problems obviously
fall under this heading—each person involved wants to make a choice
based on what their partner does. If you step forward, I want to step
back, and vice versa. Game theory attempts to explain and understand
behavior by simplifying such interactions, modeling them, and then
using a relatively bare set of assumptions about human choice to predict
or explain strategic outcomes.
In game theoretic models of coordination problems—coordination
games—actors have to coordinate strategies to be successful, and there
are multiple ways to do so. In the last section, I sometimes described
coordination problems with multiple actors. (In the military, for example,
the coordination problem is solved when many actors align in a proper
hierarchy that facilitates flexible action.) In this section, and throughout
the book, I will focus on problems with only two actors. This is not
because some of the interactions I will address cannot be fruitfully
represented by more complex models, but because the goal here is to
provide explanatory clarity, sometimes at the expense of more fine-
grained representation. Furthermore, small games have been found, in
many cases, to provide deep insight into behavioral interactions despite
their simplicity (Sigmund et al., 2001).
1.4.1 Correlative coordination games
A game involves three things: actors, strategies, and payoffs.21 Actors in
a game are those involved in the strategic interaction. Strategies define
what each actor can do (step forward or back in the tango, for example).
20 Note that this definition does not require that these interactions involve conflict (as is
sometimes assumed about game theory).

21 Usually games also define information for the actors, or what each actor knows about
the interaction. In this book, this element will be downplayed, since it is less relevant to
emerging or evolving behaviors than it is to rationality-based analyses.
Player 2
A B
A 1, 1 0, 0
Player 1
B 0, 0 1, 1
Figure 1.1 Payoff table for a simple, correlative coordination game
Payoffs determine outcomes for each actor given the set of strategies they
have chosen.
Figure 1.1 shows what is called a payoff table for the simplest type
of correlative coordination game—one with two players, where both
coordination outcomes are equally preferred. The actors are player 1 and
player 2, who each have two strategies—A or B. A could be “drive on
the right side of the road” and B could be “drive on the left,” and for
this reason I will sometimes call this the driving game. Rows in the table
correspond to possible strategies for player 1 and columns to possible
strategies for player 2. Entries to the table represent payoffs to the two
players for any combination of strategies, with player 1’s payoff listed first.
So if both players choose A, they each get 1. Ditto if they both choose B. If
they choose A and B, they get nothing. They succeed only by correlating
action.
What, in this figure, do the payoff numbers correspond to? The answer
given by game theorists is utility, an abstract representation of whatever
it is a player prefers or likes. Most game theoretic analysis proceeds by
assuming that actors try to maximize their utility by choosing the strategy
that is expected to provide the best payoff as determined by a calculation
involving beliefs about the strategic situation. Sometimes it is easy to
say what the best strategy will be. Other times, there may be multiple
reasonable strategies to choose from. In many cases, expected behavior
in games accords to what is called a Nash equilibrium (Nash, 1951). This
is a set of strategies where no actor can deviate and improve her payoff.
For this reason, these sets of strategies are thought of as stable and likely
to arise in the real world.22
22 Evidence from experimental economics indicates that, indeed, humans often learn to
play Nash equilibria in the lab, though not always. (See Smith (1994) for examples of cases
where experimental play does and does not conform to Nash equilibrium predictions.)
Besides the Nash equilibrium concept, there are a host of other solution concepts developed
by game theorists to predict and explain strategic behavior. A discussion of these concepts
is beyond the scope of this book.
In any particular game, the absolute numbers are in some ways less
important than the comparisons between the numbers for each player
(though they still matter for plenty of things). Here it matters that player
1 prefers 1 to 0 and player 2 likewise, but this strategic scenario could also
be represented by a game where the entries had 100 and 0, or 2 and -50 for
the coordination and non-coordination outcomes. If these changes were
made, the ordering would still capture the idea that each player prefers
the coordination outcomes, and does not prefer one of these over the
other. Once we use these games in evolutionary models, the significance
of these numbers will shift, though, as will the method of analyzing the
model. Instead of representing utility the numbers will instead determine
how evolutionary change happens, and the details of payoffs will often be
very significant. More on this in Chapter 3.
The game in Figure 1.1 has two Nash equilibria.23 In fact, this is a gen-
eral property of coordination games—that there be at least two plausible
outcomes actors might end up at. The Nash equilibria here are the strategy
pairings where both actors choose A or both choose B. In either of these
pairings, neither actor can switch strategies and improve her payoff. (If
either switches, she goes from getting a 1 to getting a 0.) Also note that,
in this case, neither player prefers one Nash equilibrium over the other
because both players get the same payoff (1) for either equilibrium. This
means that they are both happy to coordinate in whatever way.
Correlative coordination games do not always have this character.
Consider the games presented in Figure 1.2. In both of these games,
there are two coordination equilibria, and they are the same as for the
game in Figure 1.1 (A vs. A, and B vs. B). In both of these cases, though,
there is a difference between the two equilibria. In (a), one coordination
outcome is better than the other for both actors. While both players
prefer to coordinate over not coordinating, they also both prefer B vs.
B to A vs. A. This game represents scenarios where, for example, the less
preferred outcome could represent working hours from midnight to 8
instead of 8 to 5. Note that this game moves slightly away from a pure
23 To be more precise, the game has two pure strategy Nash equilibria. These are Nash
equilibria where both actors play pure strategies, or choose the same action all the time.
One can also consider mixed strategies, where actors make two choices probabilistically, but
for now I will ignore these. For the most part, mixed strategies will only be discussed in
this book when they are significant from an evolutionary point of view, and this will not be
often.
Player 2
(a) A B
A 1, 1 0, 0
Player 1
B 0, 0 2, 2
Player 2
(b) A B
A 2, 1 0, 0
Player 1
B 0, 0 1, 2
Figure 1.2 Payoff tables for two simple, correlative coordination games. (a) shows
one where outcome B vs. B is preferred to A vs. A by both actors. (b) shows one
where actors have different preferences over the two outcomes
convention character and toward a functional character, as discussed in

the last section, for this reason.
In the game presented in (b), the actors now no longer have interests
that perfectly line up. In each previous example, they preferred the same
outcomes. Now, each actor prefers to coordinate, but player 1 prefers that
both play A and player 2 prefers B. This game has traditionally been
referred to (suggestively) as the battle of the sexes. The story is that a man
and a woman would like to go out together, but she prefers the opera and
he prefers the baseball game. Osborne and Rubinstein (1994) uses a non-
gender normative story where two friends want to go to the opera, but one
prefers Bach and the other Stravinsky. I’ll refer to it as the Bach–Stravinsky
game. Note that this game has a strongly conventional character, despite
the modification to create conflict of interest. Here, as in the driving game,
there is not an outcome that is obviously better. (At least from a general
point of view. Players 1 and 2, of course, have views about which outcome
is better.)
The Bach–Stravinsky game may be called a conflictual coordination
game, indicating that while the coordination character holds, there is
now some conflict of interest between the actors. Games are sometimes
referred to as either conflict of interest or common interest games, but,
in fact, many of the most interesting games in game theory are both. For
zero-sum games, one player’s gain is another’s loss, for complete common
interest games one player’s gain is the other’s gain. Bach–Stravinsky is a
perfect example of something in the middle. Schelling (1960) refers to this
sort of game as a “mixed-motive” game because it represents a “mixture of

mutual dependence and conflict, of partnership and competition” (89).24
1.4.2 Complementary coordination games
In this section, I’ll introduce the games that we’ll spend the most attention
on throughout the book, and, as we will see, that best represent gendered
division of labor—complementary coordination games.25
Figure 1.3 shows the simplest example of a complementary coordina-
tion game. This two-person game is identical to that in Figure 1.1, but
now the actors only receive payoffs when they choose complementary
actions. The two Nash equilibria of this game are A vs. B and B vs. A.
In these strategy pairings, if either party switches actions she goes from
getting 1 to getting 0. For simplicity’s sake, throughout the book, I’ll call
this game the dancing game, because in this game A could represent “step
forward” and B “step back.” A and B could likewise be “cook” and “clean,”
in a household. Or “lead” and “follow” in a management situation. Or
“make pottery” and “make wood crafts.”
As with the simplest correlative coordination game, actors playing the
dancing game do not care which equilibrium they arrive at. They simply
care about coordination. In other words, there is no conflict of interest in
this game. Again, though, as with correlative coordination games, there
are variations on this complementary coordination game that change the
Player 2
A B
A 0, 0 1, 1
Player 1
B 1, 1 0, 0
Figure 1.3 Payoff table for the dancing game
24 He describes the following way of determining whether actors in a game have common
interests, conflicts of interest, or something in between. Make a chart where the x-axis
represents player 1’s payoffs and the y-axis represents player 2’s payoffs. Mark down a point
representing each possible outcome. If the slope of lines between these points is always
negative, the game is a conflict-of-interest game, and vice versa for common interest. A
game in the middle will have both positive and negative slopes between outcomes.
25 It is typical to call these “anti-coordination” games. There are a few reasons I do not
use this label here. First, it seems to imply that actors do not wish to coordinate, which is the
opposite of the truth. Second, they are sometimes referred to as “discoordination” games,
and I have seen both of these labels applied to games where one actors attempts coordination
and the other prefers to avoid it, like matching pennies. So I start with fresh terminology.
(a) Player 2
A B
A 0, 0 2, 2
Player 1
B 1, 1 0, 0
(b) Player 2
A B
A 0, 0 1, 2
Player 1
B 2, 1 0, 0
Figure 1.4 Payoff tables for two simple, complementary coordination games. In
(a) actors both prefer one coordination outcome to the other. In (b) actors have
conflicting preferences over the two coordination outcomes
strategic character of the situation. Consider the games in Figure 1.4.

They are complementary versions of those introduced in Figure 1.2.
In the game in (a), the two actors coordinate by taking complementary
actions, but the overall payoff is better for both if they choose one partic-
ular pairing. This game could represent a scenario where, for example,
one of the partners is better suited to one of the two complementary
social roles. Perhaps one spouse really likes doing laundry and the other
likes doing dishes. Obviously both will prefer the division of household
labor that accords with these preferences. Or suppose that one person
likes sitting down and thinking all day while another likes being outside
walking around and talking to people. It is obvious which of these two
should be a bank clerk and which a police officer. Throughout the book,
I will call this the MFEO (made for each other) game since it represents
scenarios where individuals fit well into complementary social roles. This
game, like its corollary in the previous section, has a more functional
character and less of a conventional one, since there is a preferable
outcome for both players.
The game in (b) represents scenarios where actors need to take com-
plementary actions, but one is preferable to both. For example, perhaps all
members of a company would like to be the CEO, rather than the lowly
clerk. When getting drinks at Peet’s, everyone would like to order first,
rather than second. In dividing labor, perhaps everyone prefers a job that
gives them direct control over food resources. I will call this game the
leader–follower game for scenarios like the first one mentioned here—
where both actors would like to play a social role that has higher status
Another random document with
no related content on Scribd:
Jo aikoi hän lähteä noutamaan jotakin työtä. Vaan yht'äkkiä kiintyi
hänen katseensa isoon muurahaiseen, joka laukkasi hänen
polvellansa.
"Kas sinua veitikkaa!" Luunapilla hän singautti sen oikeaan

olinpaikkaansa. "Saavatkohan makaajat olla niiltä rauhassa?" Hän
astui hiljaa ensin toisen, sitte toisen riippuverkon luo. Niissä ei
näkynyt mitään eikä puissakaan, joihin ne oli sidottu.
Lehtori tarkasteli katseillaan myöskin saalia, jolla rouva Streng

makasi. "Ahah, tuossapa on muutamia!" Hän kiireimmiten karkoitti
rauhanhäiritsijät. Mutta mitä enemmän hän niitä hätyytti, sitä
äkäisemmin yhä uusia ryhtyi rynnäkköön. Viimein oli lehtorilla kylliksi
työtä pitäessään suuren saalin reunoja puhtaina kutsumattomista
vieraista.
Rouva Lajunen raotti silmiään. Hän ei ensin käsittänyt, missä oli.

Olo vain tuntui niin suloisen raittiilta: hikinen kuumuus ei vaivannut,
niin kuin muulloin kammarin sohvalla herätessä, eikä pääkään ollut
raskas. Hän ummisti jälleen silmänsä, huolimatta ottaa asiasta
tarkempaa selkoa, aikoen vain nauttia oikein kyllälti tätä
odottamatonta hyvinvointiansa.
Ei hän kuitenkaan enää uudelleen nukkunut. Tuulen henki hiukan

humahutti puiden latvoja. Lintu pyrähti lentoon. Niistä hän viimein
muisti olopaikkansa ja että muitakin oli siellä metsikössä.
Kuului vähän oksain raksetta. Rouva avasi silmänsä ja näki

miehensä puuhailevan rouva Strengin tienoilla, koetti nousta, mutta
ei osannut.
Lehtori huomasi liikettä vaimonsa riippuverkosta ja läksi tutkimaan,
joko hän heräsi, sekä auttoi hänet alas.
"Mitä sinä siellä puuhasit?"
"Karkoittelin muurahaisia", selitti lehtori viattomasti.
"Onko täällä niitä?" — Rouva kauhistui omasta puolestaan.
"Näkyy olevan maassa, vaan ei puissa."
Samassa tuli Mari tuomaan kahvia, kun oli nähnyt herrasväen

lähdön mäelle päin eikä Katri antanut hänen jättää kahvia
seisomaan.
Siihen keskeytyivät rouva Lajusen arvelut ja utelemiset. Mieli

myöskin oli nyt niin virkeänä ja iloisena, että hän ei huolinut
pahastua semmoisista mitättömyyksistä kuin tuosta, jotka oli nähnyt.
Tähän puheen hälinään ja kahvikompeiden kilinään toisetkin

heräsivät kokonaan virkistyneinä.
Nytpä kahvilla olikin maku, jollaista rouva Lajunen harvoin muisti

tunteneensa.
"En minä oikein käsitä, miten tämä olo nyt tuntuu niin virkeältä ja
hauskalta", virkkoi lehtori.
"Kyllä minä aina voin toimittaa yhtä hauskaa, jos vain suostutte
minun ohjeihini. Noustaan aikaisin, ollaan liikkeessä ja puuhataan
jotakin", neuvoi lehtori.
"Mitäpä täällä on puuhaamista", vastusti häntä rouvansa.

"Jos sinua ei mikään miellytä, niin vaikea sitä toisen on tehdä
sinulle miellyttäväksi. Itsensä pitää kunkin osaaman hankkia itselleen
huvia."
"Ehk'et ole ihan väärässä", myönsi viimein rouva. Mitäpä siinä

vastusteleminen auttoi, kun oma kokemus nyt todisti toista.
Vielä selvemmäksi varttui tämä kokemus parin kolmen viikon

kuluessa, joll'aikaa rehtori ja rouva Lajunen lakkasivat ikävöimästä
sateen tuloa, korttipeliä ja kaupungin seuroja sekä vähitellen
huomata maalla elämisen edes jotakuinkin siedettäväksi, ell'eivät
sitä vielä tunnustaneetkaan erittäin kaihottavan hupaiseksi.
Omituinen oli todellakin rouvain luonteiden vastakkaisuus. Yhä

useammin alkoi lehtori varsinkin nyt, kuin enemmän oleskeltiin
yksissä, tavata itseään vertailemasta heitä toisiinsa ja
ihmettelemästä, miksi hän tuota eroa ei ennen ollut huomannut,
kunnes hän viimein arvasi syyksi, että hänellä ei ollut vielä ennen
ollutkaan tällaista vertailemisen tilaisuutta.
Mikäli tämä asia hänelle selvisi, mitä suloisemmaksi, etevämmäksi

ja käytöllisemmäksi tosi ihmiseksi hän huomasi rouva Strengin, sitä
aremmaksi hän muuttui käytökseltään häntä kohtaan, varsinkin
oman vaimonsa läsnä ollessa. Hänellä oli salaisuus, vaarallinen
heille kaikille, jos se vain tuli syrjäisen tietoon, joka sitä osaisi käyttää
väärin, sovun sotkemiseksi.
Lehtorin mieli oli alkuaan ollut avoin kuin virheetön peili, mutta sen
pinta oli vähitellen himmennyt sikäli, kuin hän oli huomannut
rouvansa kykenemättömäksi käsittämään sellaista syvällistä
selvyyttä.
Nytkin hänestä tuntui oikein raskaalta tuon salaisuuden pitäminen.
Mutta jos hän olisi sanonut vaimolleen, että rouva Streng oli häntä
ymmärtävämpi tai muuta sen tapaista, niin olisi vaimo sen käsittänyt
moitteeksi ja ollut liian ylpeä sitä kärsimään, samalla olisi vaimo
myöskin saanut mustasukkaisuuden aihetta, joka sitte olisi tuottanut
suurta kiusaa sekä hänelle itselleen että myöskin lehtorille.
Jonkun ajan jaksoivat rehtori ja rouva Lajunen kestää uutta

elinjärjestystä, aikaista nousemista ja liikkeellä oloa, niin kauan kuin
se kaikki uutuudellaan viehätti. Mutta eräillä ihmisillä on omituinen
taito saada käsissänsä kaikki pian vanhenemaan. Niinpä nytkin
rehtori kohta kokonaan kyllästyi "itsensä vaivaamiseen" ja palasi
entisen ikävyyden seuraan. Rouva Lajunen jaksoi vähän kauemmin
nauttia luonnon virkistystä, mutta uupui hänkin. Hänellä olikin
muutos ollut liian suuri, niin että se, äkisti toimeen pantuna,
enemmän uuvutti, kuin kukaan osasi edeltä päin arvata.
Mikäli entiselleen palaaminen toisilla tapahtui, jäivät lehtori ja

rouva Streng kahden sekä kroketti-peliin että soutelu- ja
kävelyretkille. Lehtori tosin koetti kartella yhdessä oloa, mutta kun
rouva Streng, aina hilpeä ja iloinen, sattui kumppaniksi, ei hän voinut
häntä kieltää, kun näet tahtoi sentään olla laimin lyömättä tavallista
kohteliaisuutta.
Vähitellen kasvoi lehtorin mielessä omituinen kaipaus, ystävän

puutteen tunne, sellaisen ystävän, jolle aina voisi puhua suoraan
kaikki ajatuksensa ja joka ne oikein käsittäisi. Ja tämä kaiho kasvoi
sitä suuremmaksi, mitä selvemmin hän käsitti, että hänen vierellänsä
usein oli sellainen ystävä, mutta hänellehän hän ei saattanut puhua.
Tuo tunne vahvistui viimein niin voimakkaaksi, että voimakas mies oli
kokonaan musertua sen taakan alle.
Rehtori turhamaisuudessaan uskoi itseänsä sekä muodon
muhkeudella että muilla ominaisuuksilta kaikista miehistä
etevimmäksi, joten hänelle ei edes johtunut mieleenkään, että
lehtorin seura ehkä saattaisi tulla vaaralliseksi hänen vaimollensa.
Mitäpä tuosta, jos lehtori souteli häntä, kun kerran näkyi olevankin
kuin luotu raskaan työn tekoon.
Jos rouva Lajunen viihtyi rehtorin seurassa, niinkuin näytti

viihtyvän, niin mitäpä outoa siinäkään olisi rehtorin mielestä ollut.
Päin vastoin ei rehtoria olisi ollenkaan kummastuttanut, jos rouva
Lajunen olisi pitänyt häntä miestänsä etevämpänä, vaikkapa
mielistynytkin häneen, koska lehtori ei kuitenkaan näyttänyt pitävän
suurta lukua vaimonsa seurasta.
Näin rehtori usein mietiskeli istuessaan kiikkulaudalla lehtorin

pihassa rouvan vastapäätä, joka kiikkutuolissaan lueskeli romaaneja.
Harvoin he mitään puhuivat, ja rouva taisi välistä unhottaa koko
rehtorin läsnä olonkin, koskapa hän joskus jotenkin suoravaisesti
huuteli ja komenteli piikojansa sieltä ulkoa käsin, kuitenkin nyt jo
paljon sävyisemmin kuin ensi päivinä, ollessansa tyytymätön maalla
oloon.
Rouva Lajusen mielessä, kun hänen miehensä yhä enemmin

ikään kuin häipyi naapurirouvan seuraan, rupesi vähitellen
kasvamaan jonkinlaista outoa levottomuutta, ja se suurensi tuon
muurahaissodan hänen silmissään aika isoksi tapaukseksi. Tosin
hän ei vielä ollut nähnyt muita samaan suuntaan viittaavia
todistuksia, mutta saattoihan niitä olla tapahtunut hänen
näkemättänsäkin. Kuitenkaan ei hänelle johtunut mieleen tehdä
samoin miehellensä kuin jo alkoi uskoa miehensä tekevän hänelle,
mutta päätti salaa tarkastella hänen käytöstänsä.
IV.
Joutui mansikka-aika ja hyvä marjakesä tulikin. Päivärinteet olivat

kirjavinaan punaisia, viljelemättömiä Pohjolan hedelmiä, makeampia
kuin mitkään viljellyt marjat.
Eräänä aamuna lehtori Lajunen jo aikaisin läksi liikkeelle,

oikeastaan kävelylle, vaan otti kuitenkin korisen käteensä eteen
ehkä sattuvan marjikon varalle.
Hän kiersi järven rantaa lähimmälle mäelle peltojen taa ja astuskeli

mitään erittäin ajattelematta, mutta muistamatta mansikoitakaan. Ei
hän myöskään ollut vaipunut luonnon ihailemiseen, kuten tavallisesti
ennen.
Viimein hänen eteensä ilmestyi parin pensaan takaa joku, joka

kyyrysillään todellakin kokoeli Luojan ja luonnon antimia, eikä se
ollut kukaan muu kuin rouva Streng, joka oli sinne ehtinyt ennen
häntä.
Lehtori yritti peräytymään, mutta liian myöhään; rouva jo huomasi

hänet. Hänen siis täytyi tervehtiä.
"Hyvää huomenta!" vastasi iloinen ääni. "Kyllä näitä marjoja on

täällä teillekin, ei teidän sen tähden tarvitse paeta."
"En minä sen tähden yrittänytkään." — Hän ei koettanut salata

paon aiettansa.
"Minkä tähden sitte? Sanokaapas, miksi te olette viime aikoina

ruvenneet pelkäämään minua. Olenko minä muuttunut julmemmaksi
vai miten?"
"Ette suinkaan. Päin vastoin olette liiankin hyvä." — Lehtori
tarkoitti yleensä kaikkia niitä rouvan ominaisuuksia, jotka häntä
miellyttivät, eikä erittäin niin sanottua hyvyyden avua.
"Ei kukaan voi olla liian hyvä, eikähän hyvyyden pitäisi ketään
pakoon ajaa. Ettekö enää pidä minua ystävänänne?"
"Pidän liiaksikin."
"Nyt minä en teitä ollenkaan ymmärrä."
"En minä itsekään. Sen vain tunnen, että minun tarvitsisi puhua,
mutta en voi kellekään."
"Ettekö minullekaan?"
"Kaikkein vähimmin teille."
"No, kyllä ainakin yhdelle voitte puhua, tuonne!" Hän osoitti kohti
taivasta.
Lehtori punastui. Oliko hän unhottanut Jumalansa! Todellakin

viime aikoina, silloin kuin oli hautonut tuota salaisuuttansa, joka nyt
oli ollut vähällä pujahtaa ennen aikojaan hänen suustansa, kenties,
kuten hän tähän asti luuli, sotkemaan toisen, ehkäpä useammankin
henkilön rauhaa.
Hän astui reippaasti rouvan luo, antoi kättä ja lausui vain yhden
sanan: "kiitos!"
"Mitäpä siinä on erittäin kiittämistä. Tietystihän Jumalalle voi kuka

hyvänsä puhua kaikki asiansa."
"Niin tosiaankin, mutta hän yritti minulta unohtumaan. Nyt olen
selvillä, ja tuo teidän yksi viittauksenne minulta poisti ahdistuksen.
Nyt minä myöskin voin puhua suoraan teille, kun näen, että teidän
rauhanne ei häiriydy."
"Mikä sitte muka saattaisi häiritä minun rauhaani?"
Lehtori jätti siihen vastaamatta. Vasta hetkisen kuluttua virkkoi

hän:
"Minä olen nyt tänä kesänä huomannut, mikä suuri ero on teidän
ja minun vaimoni välillä."
"Tuskin suurempaa kuin teidän ja minun mieheni välillä. Minä olen

huomannut enemmänkin. Silloin kuin teidän rouvanne laittoi meille
kaupunkilais-illalliset, joita te ette näyttäneet hyväksyvän, sanoi
Viktor kotiin palattuamme, kun niiden pitojen johdosta kiistelimme,
että te ja minä sopisimme paremmin yhteen kuin hän ja minä.
Minusta hän siinä arvasi ihan oikein enkä minä sitä siis
vastustanutkaan, vaan sanoin mitä nytkin sanon, että mikä kerran on
mahdotonta, miksipä sillä rupeisimme sotkemaan rauhaamme, ehkä
vielä toistenkin."
Lehtori vaipui ihmettelemisen äänettömyyteen. "Saatammehan

siltä pysyä ystävinä", jatkoi rouva. "Minä huomasin teidän
yrityksenne totuttaa rouvaanne toisiin tapoihin ja rupesin mielelläni
auttamaan; toivoin näette samalla, että Viktorikin hiukan luopuisi
muutamista tavoistaan. Kun se ei onnistunut täydellisesti teille —
Viktor näyttää olevan kyllin onnellinen sellaisenaan kuin on — olen
sitte tarkastellut, miten te tästä asiasta selviäisitte. Minua
pakenemisenne näytti olevan mielessänne jotakin, johon en tahtonut
antaa teidän eksyä. Sen tähden piti minun saada teidät ennen
puhumaan ja muistamaan velvollisuutenne, kuin kokonaan jouduitte
harhateille."
Lehtorin ihmettely yhä kasvoi. Miten tarkka aisti tuolla rouvalla!
"Niin, pysykäämme ystävinä", virkkoi hän viimein. "Näenhän, että

on turha teiltä mitään salata ja että teihin myöskin voi kuka hyvänsä
luottaa. Teillä on aina oikea pohja. — Ja nyt poimikaamme marjoja,
että kotiinkin joudumme."
"Minun korini on kohta täysi. Saanko auttaa teitä?"
"En minä tiedä minkään sitä estävänkään, ell'ette tahdo lähteä

edeltä kotiinne. Kyllä minun koriini mahtuu." Pian täyttyi lehtorin kori
kahden poimiessa.
Kotimatkalla ryhtyi rouva Streng vielä puhetta jatkamaan.
"Minun pitäisi oikeastaan huomauttaa teille vielä yksi asia, joka

arvattavasti on teillä pysynyt epäselvänä, ell'ette vain pahastu minun
tunkeutumisestani teidän asioihinne."
"Mistäpä minä voisin teihin pahastua. Huomaanhan, että te

kuitenkin näette kaikki oikeassa valossa."
"Katsokaas, mieheni uskoo itseään niin eteväksi kaikin puolin, että

hänen ei tarvitse peljätä jättäessään minua kenen seuraan hyvänsä.
Sen tähden minä voinkin olla aivan vapaa, ei hän tule
mustasukkaiseksi. Mutta tiedättekö varmaan, että teidän rouvanne
yhtä lujasti luottaa valtaansa teidän ylitsenne?"
Lehtorilta ikään kuin silmät kerrassaan aukesivat. Hän edelleen

havahtui suloisesta unesta.
"En tosiaankaan tiedä. Kiitos viittauksesta. Nyt kyllä arvaan, mitä
tahdotte ja mikä on velvollisuuteni. Mutta auttakaa, jos horjun."
"Sitä teidän ei tarvitse pyytääkään."
*****
Rouva Lajusella oli myöskin salaisuutensa ja sitä hautoessaan

hän vähitellen oli melkein sairastunut. Umpimielisyys esti häntä sitä
ilmasemasta edes miehelleen, kaikkein vähimmin se toki sai tulla
hänen tietoonsa; ennemmin joku kaupunkilaisystävistä voi sen kuulla
ja ruveta apulaiseksi tutkimusten teossa.
Mutta maalla oloa kesti liian kauan. Rouva alkoi nukkua

levottomasti ja nähdä puoli valveillakin asioita, joita sitte uskoi tosiksi
ja jotka vain pahensivat hänen tilaansa. Hänen rakkautensa ei tosin
ollut varsin syvällinen laatuansa, vaan tuo romaaneista opittu
mustasukkaisuuden tauti tuntui niin erittäin sopivalta syyltä tekeytyä
kärsimyksen uhriksi ja sen nojalla taipumuksensa mukaan kiusata
itseään ja piikoja, joita hän taas alkoi juoksuttaa ja komentaa yhä
enemmin.
Oli rouvassa sentään syvimmällä kätkössä myöskin oikeuden ja

tosi ihmisyyden siemen, vaikka se oli tähän asti pysynyt
kehittymättä, osaksi lehtorin itsensä tähden, hän kun oli avioliittonsa
alussa liian pian luopunut rouvansa kasvatuksesta taikkapa ollut liian
jyrkkä siinä, osittain myöskin sopimattomain
kaupunkilaisystävättärien tähden, he kun rouvan turhamaisuutta yhä
kiihdyttelivät. Nyt kun nämä ystävättäret olivat poissa ja päin vastoin
oli usein tilaisuus kuulla oikeampia tapoja ja ajatuksia, alkoi se
ihmisyyssiemen itää ja samalla kasvattaa kiivasta ristiriitaa rouvan
mielessä entisen nukkemaisuuden ja vasta itävän uuden ihmisen
välillä.
Tämän ristiriidan tähden tuntui mustasukkaisuustauti vielä

rasittavammalta. Mutta vanha puoli oli vielä niin paljon voitolla, että
rouva ei omin neuvoin siitä jaksanut selvitä.
Nyt kun lehtori tänä äsken kerrottuna aamuna nousi, oli rouvakin
ollut sen verran liikkeellä, että sattumalta huomasi rouva Strengin
lähdön kori kädessä. Hän oli nukkuvinaan uudestaan; mutta kun
lehtori peseydyttyään ja pukeuduttuaan läksi ulos, riensi hän salaa
ikkunasta katsomaan, menikö hänen miehensä samanne päin. Ihan
oikein: kori hänelläkin kädessä ja samaa tietä. Eiköhän tuo ollut vain
edeltä päin sovittua!
Vielä oli liian aikainen nousta. Rouva siis palasi vuoteelleen ja

valvottuaan hyvän aikaa vaipui horrostilaan, joka ei vähääkään
virkistänyt, vaan sen sijaan antoi mielikuvitukselle vapaan tilaisuuden
kiusata hermostoa ja luoda ilmi eläväksi kaikenlaisia kuvia, joita ei
avoimin silmin nähdä eikä huolitakaan nähdä.
Ja mielikuvitus tietysti jatkoi ja kehitteli kernaimmin niitä kuvia, joita

mieli oli täysimpänä. Se loi makaajan nähtäväksi kauniin marja-ahon
ja siihen lehtorin ja rouva Strengin marjoja poimimaan. He ensin
pysyivät etempänä toisistaan, mutta lähenivät vähitellen ja viimein
pistelivät mehukkaita mansikoita toinen toisensa suuhun ja poimivat
niitä toinen toisensa koriin.
Tämän kaiken oli rouva Lajunen näkevinään pensaston takaa,

johon muka oli kätkeytynyt heitä tarkastelemaan.
Viimein poimijat muka nousivat seisomaan, syleilivät ja suutelivat
toisiansa.
Se toki jo oli liikaa.
Lehtorin asunnossa kuulivat Katri ja Mari kyökkiin, jossa

odottelivat rouvan heräämistä, kovan parahduksen. Mari juoksi
sisään.
Omasta äänestään heränneenä istui rouva sängyssään märkänä

hiestä ja silmät oudosti leimahtelevina. Tottunut lääkäri olisi niistä
heti nähnyt kuumeen oireita.
"Missä lehtori on?" kysyi hän.
"En minä tiedä, jo hän oli poissa, kun minä heräsin ja yritin kahvia
viemään."
"Ette te tiedä mitään, mutta tiedän minä itsekin. Hän suu…"
"Oikeaan aikaan hän toki muisti, että Mari olikin vain piika ja
keskeytti syvälle kätketyn salaisuutensa tunnustuksen.
"En minä nyt jaksa nousta," sanoi hän sitte selvemmällä äänellä,
"tuo vain kahvia ensin."
Mari riensi toimittamaan.
Rouva painui takaisin makuulle. Kuuma vuode melkein poltti,

mutta ruumis oli niin hervakka, että tuntui ikään kuin vajoavan
lävitse. Putoamista peljäten hän ummisti silmänsä ja hiukan vapisi.
Katri kuitenkin paraiksi toi kahvin, että rouva pääsi vaarasta. Hän
joi kupin, jopa toisenkin, ja vähän virkistyi hetkiseksi, kunnes kahvin
lämpö tuotti uutta kuumetta. Ei ollut ajattelemistakaan ylös
nousemista.
Hetkisen kuluttua saapui lehtori kotiin.
Rouva kuuli hänen askeleensa ja huomasi niistä, että hän astui

suoraa päätä hänen luoksensa.
Lehtori asetti mansikkakorinsa pikku pöydälle sängyn viereni,

jossa kahvitarjotin äsken oli ollut.
Rouva aukasi silmänsä nähdäkseen, rohkeniko tuo mies enää

katsoa häneen suoraan. Niin varma oli rouva alkavan kuumeen
käsissä äsken näkemänsä unen todellisuudesta.
"Joko olet hereillä? Katsos, miten kauniita mansikoita!"
Sekä ääni että käytös olivat ihan suoravaiset eikä katseessakaan

näkynyt mitään arkuutta.
Rouva ensin oli vähällä muistaa, että unta hän kaiketi olikin
nähnyt; vaan toinen tunne oli vahvempi. Hän ajatteli: "teeskentelijä!
mutta osaan sitä minäkin teeskennellä!"
"Kalinoita tosiaankin", virkkoi hän ääneensä ja alkoi syödä hyvällä

halulla, kuten näytti.
Lehtori hyvillänsä, että hänen äsken tehdyn päätöksensä toimeen

panon alku onnistui niin hyvin, viivähti vähän kammarissa, ja rouva
sen aikaa söi syömistään marjoja.
Mutta mansikat ovat petolliset. Tervettä ne virkistävät, vaan lapsille

ja heikkohermoisille tuottavat kuumetta. Rouva Lajunen kun ei vielä
ollut syönyt tukevampaa ruokaa, vaikuttivat marjat tyhjässä vatsassa
ihmetyttävän nopeasti. Taudin alku kehittyi kovaksi kuumeeksi;
sairas alkoi houraella.
Lehtori istui taukoamatta sairaan luona. Kuultuaan hänen

houreistansa asioita, joita oli tarpeeton antaa palvelijain kuulla,
päästi hän heitä sisälle ainoastaan välttämättömille asioille, vaan
muuten teki itse kaikki, mitä sairaan hoitoon kuuluu.
Ei rouva kuitenkaan aina ollut houreen tilassa. Selkeämpinä

hetkinään hän kyllä huomasi miehensä valppauden, huomasipa
myöskin, että rouva Streng usein kävi siellä ja kuuli miehensä ja
hänen keskusteluistaan, miten perin väärä koko hänen
luulevaisuutensa oli ja miten kerrassaan jalo nainen tuo
naapurirouva oli, jota hän tähän asti oli halveksinut.
Siitä huomiosta hänen hourehetkensä alkoivat vähitellen lyhetä ja

harveta. Kuitenkin hän mielellään makaeli silmät ummessa
selvänäkin välistä vain katsellen salaa miestänsä, ja nähtyään hänen
huolestuneen muotonsa päätti hän koettaa ansaitakin miehensä
rakkautta.
Täten ei tauti tullut pitkälliseksi. Kuudentena päivänä sairas jo

arvaeli olevansa terve, vaikka tosin vielä heikko.
"Annatko anteeksi, että olen sinua väärin käsittänyt?" kysyi hän

silloin mieheltään.
Lehtori hämmästyi tätä muutosta. Hänen mielessään välähti

autuaallinen tunne, ensi kerran siitä asti, kuin hän oli vihille astunut
rikkaan ja äitinsä puolelta ylhäissukuisen neiti Savolinin kanssa,
jonka oli oppinut seuroissa tuntemaan viehättäväksi tytöksi.
Heti hän avioelämänsä ensi päivinä oli huomannut pahasti
pettyneensä, mutta kuitenkin nurisematta kantanut rangaistusta siitä,
että oli osittain rikkauden houkutuksesta antautunut huolettoman
toimeen tulon toivoon.
Nyt lehtori tarttui siihen vähäiseen toivon korteen, jonka vaimonsa

hänelle ojensi.
"Etkös enää ole mustasukkainen?" kysyi hän iloisesti.
"Rouva Strenginkö tähden? En, vaikka sinulla olisikin kyllin kyllä

syytä luopua minusta ja rakastua häneen; hän sen ansaitsisi monta
vertaa paremmin kuin minä."
"Jos tunnet käsittäneesi minua väärin, niin sillähän se jo on

sovitettukin. Mutta olisiko se tosiaankin mahdollista?"
"On se ihan totta. Olenhan tässä maatessani kuullut ja huomannut

paljonkin. Ja sitte ajattelin, että tuskinpa vain minä olisin sinua niin
hoidellut, kuin sinä nyt hoidit minua. Johtui sitte myöskin mieleeni,
että tosiaankaan minä en tähän asti ole tehnyt mitään sinun
huviksesi enkä myöskään ole osannut olla kiitollinen siitä, mitä sinä
olet tehnyt minulle. Jos et ole liian malttamaton, etkä vaadi yht'aikaa
liian paljoa, niin koetan muuttua kokonaan toisenlaiseksi."
Lehtori ei kyennyt vastaamaan. Hän vain ääneti kyynelsilmin syleili

vaimoansa.
"Arvaatko, mistä tämä mieli nyt on minulle tullut?"
"En, mahdoton minun on sitä arvata. Houreistasi luulin ihan toista."

"Minä unissani näin sinun juuri näin hellästi syleilevän, arvaat toki,
ketä, ja hän sitä paremmin ansaitsisikin. Mutta kyllä nyt tiedän, että
uni oli sula valhe. Sinä et rakasta häntä, vaan minua."
"Älä sano niin; en minä ole sinua oikein rakastanut ennen kuin
nyt."
Pitkät selitykset kasvoivat molemmin puolin ja niiden johdosta

lehtori vaimoinensa alkoi nyt vasta viettää lempiviikkojansa luonnon
helmassa, mutta eivät ne siltä olleet vähemmin suloiset. Lehtorilla
tosin täytyi olla hyvä varasto kärsivällisyyttä; sillä eipä vanhoista
tavoista niin äkisti päästä; mutta hän osasi aina vähitellen saada
tahtonsa noin vain kiertoteitä huomatuksi ja karttaa jyrkkiä
vaatimuksia, jotka vaimoa olisivat suututtaneet ja kiusantekoon
ärsyttäneet.
Hyvänä apuna oli rouva Lajusella naapurirouvan esikuva. Heistä

tulikin hyvät ystävät ja eroamattomat kumppanit. Väsymätön oli
rouva Streng opetuspuuhassaan ja iloksensa huomasi hän
oppilaansa hyvin edistyvän.
Rehtori yksin pysyi osattomana näistä tapauksista ja muutoksista.

Mutta kun rouva Lajunen alkoi yhä enemmän seurata toisten tapoja
ja olla mukana kävelyretkillä, katsoi hänkin vaikka vastahakoisesti
olevansa velvollinen mukautumaan. Ainoastaan aamu-unestansa
hän ei jaksanut luopua.
Kaksi Kesävierastamme oli löytänyt onnen luonnon helmasta;

kolmas oli onnellinen omasta mahdistaan ja oman täydellisyyden
tunnostaan; neljäs olisi kyllä ansainnut olla onnellisempi, mutta hän
tyytyi siihen onneen, jonka velvollisuuksien täyttäminen ja hyvä
omatunto tuottavat.
HAAKSIRIKKO.
"Kun on huono pää, pitää lukea ahkerammin", lausui eräänä

keväänä koulun loputtua kunnallisneuvos Curtén pojallensa
Florianille, joka oli töin tuskin päässyt ylemmälle luokalle, vaikka
kyllä oli koettanut koko vuoden oikein voimainsa takaa lukea
pinnistää.
Tuo säännöllisesti joka kerran koulusta arvosanoja tuodessa

lausuttu kehoitus oli oikeastaan tarpeeton, sillä ei Florianilta
suinkaan ahkeruutta puuttunut, mutta hirmuisen kova pää hänellä oli
aivan pienestä pitäin. Mielellään hän oli ensi alussa kirjaimia
katsellut, mutta nimet ne eivät ruvenneet, pysymään muistissa.
Sittemmin koulussa oli niin paljo kaikenlaista opittavaa, että kävi
vielä mahdottomammaksi muistin ahjoon sovittaa niitä kaikkia.
Kernaimmin olisi Florian viimein jättänyt koko lukemisen sikseen,

varsinkin kuin kumppanitkin alkoivat pilkkaella hänen
osaamattomuuttaan. Vaan silloinpa sanottiin kotona, että kovinhan
näyttäisi rumalta kunnallisneuvoksen pojan tyhmäksi jäädä, vaikka
tosin ei kaunista ollut tuokaan huonojen todistusten saanti.
Tämä tarttui mieleen ja siitä vähitellen kasvoi kiihkeä kunnianhimo,
joka nytkin mainittuna keväänä oli Florianin melkein saanut itkuun
tillahtamaan tuossa ihan kumppanien nähden, kun todistus tuli
kouraan ja hänet sanottiin päästettävän viidennelle luokalle
ainoastaan ahkeruutensa tähden.
Isän lausuessa kertoiluksensa päätti Florian lukea vielä,

ahkerammin ja pani päätöksensä heti toimeen. Hänen vanhempansa
kun asuivat lyseokaupungissa, ei hänellä ollut kotimatkastakaan
huolta, Sentähden hän jo tutkintopäivän iltapuolella alkoi kirjan
ääressä istumisensa. "Ahkera, ole ahkera!" kaikui alinomaa hänen
korvissaan. Niinpä kumppanit häntä tuskin koko kesänä saivat
liikkeelle, taikka jos saivatkin, oli hänellä ajatukset kirjoissa, joten
hänestä ei kellekään hauskuutta ollut.
Kerran hänet puoli väkisin vietiin onkimatkalle. Hän istui

muutamain kumppanien kanssa veneessä ja katseli tosin onkensa
poloa; mutta siitä johtuivat hänen ajatuksensa siimaa myöten alas
kaloihin, jotka muka piirustelivat siellä vanhan Kreikanmaan karttaa
ja asettelivat sen maakuntia paikoillensa. Vaan nepä eivät
osuneetkaan oikein, vaikka kalat miten olisivat koettaneet; sen
tähden niiden kesken syntyi kinaa ja ne siinä mellastellessaan
nykivät hänen siimaansa ja poloansa.
"Johan ne kalat vievät sinut koko miehen!" ivasivat kumppanit.
"Missäs kohti se Fokis-maa on?" kysyi Florian, voidakseen neuvoa

kaloille ja siten toimittaa sinne rauhaa.
Kumppanit tietysti nauramaan ja ivaamaan. Florian joutui

luonteensa arkuuden tähden tuosta hämilleen, ja yhä harvemmin
kumppanit häntä enää saivat mihinkään lähtemään.
Kova pää vaati kovaa lukua ja luku ikäänkuin sulatti uhrinsa
kirjoihin ja teki, kuten muista näytti, hajamieliseksi.
Viime luokalle päästyään joutui Florian kaikkein ihmeeksi kihloihin.

Mutta omituisestipa se sekin tapahtui.
Kunnallisneuvoksella oli pidot ja Florian oli väkisin saatu

franseesiin. Hän istui parinsa vieressä. Soitosta johtuivat ajatukset
äänioppiin, mitenkä eri äänet saadaan eri pituisista ja paksuisista
pianonkielistä eri lailla pingoittamalla ja miten eri äänet tekevät eri
suuria aaltoja ilmassa.
Niinpä Florian ei huomannut pitää vaaria tanssivuorostansa.

Vastatanssijat tulivat aivan eteen ja herra ivaten virkkoi: "no,
Florian, suutele jo!" tarkoittaen, että Florian parinsa kanssa olivat
niin rakastuneet, että muuta eivät enää olleet vailla kuin suutelua.
Äänioppi haihtui kerrassaan, vaan sen sijaan eksyi Florian toiselle

taholle. Hän luuli nyt oltavan "kaivoleikkisillä", jossa sormusta
kuljetetaan suusta suuhun. Ja hän ikään kuin havahtuen alkoi
huulillansa etsiä sormusta parinsa huulilta.
Tytön veli, vastatanssija, se juuri oli pannut tuon kujeen toimeen,

uskomalla siitä kuitenkaan tuollaista tulevan. Mutta kun vieraat eivät
ottaneet oikein tajutaksensa moista leikkiä eikä asianomaisten
vanhemmilla ollut kummallakaan puolen mitään esteitä lähemmän
sukulaisuuden rakentamista vastaan, julistettiin Florian ja Lisette
kihloihin, johon he itsekin sekä taipumuksesta että häpeätä
välttääkseen pitkittä puheitta suostuivat.
Mitään puutetta ei Florianilla koskaan ollut, sillä hänen

vanhempiaan ainakin luultiin rikkaiksi. Mutta vanhempain ja
palvelijain täytyi aina pitää huoli hänen pienimmistäkin tarpeistansa;
niitähän ei sopinut jättää Florianin ajateltavaksi, se olisi häirinnyt
lukua. Näin oli hän vähitellen vierastanut kokonaan tosi elämästä ja
alkoi yhä useammin ajatuksissaan johtua kaikenlaisista
jokapäiväisen elämän pikku asioista joko etäiseen muinaisuuteen
taikka muihin hullunkurisuuksiin.
*****
"Ahkeruus kaikki voittaa." Niinpä Floriankin viimein pääsi

ylioppilaaksi.
Mutta silloin ne vasta huolet ja vastukset alkoivat.
Isä oli varmuuden vuoksi ollut saattamassa ensi lukukauden

alussa Florianiansa Helsinkiin ja hankkinut hänelle asunnon
ravintoineen kaikkineen, että poika saisi oikein rauhassa lukea. Isä
hänet myöskin oli jo keväällä määrännyt historiallis-kielitieteelliseen
tiedekuntaan, josta helpoimmin sitte voi siirtyä mille uralle hyvänsä,
kun pojalla itsellään ei kumminkaan ollut edes aavistustakaan, miksi
hänen pitäisi ruveta.
Näistä siis ei ollut huolta. Mutta Florian ei ollut tottunut vapaasen

elämään eikä vapaasen, omintakeiseen työhön. Hän oli tähän asti
tehnyt työtä käskettyä, lukenut, mitä opettajat luettavaksi määräsivät.
Nyt hän ei itse osannut järjestää lukuansa, kun ei enää niin tarkkaa
ohjetta eikä ohjaajaa ollut.
Hän luki vain, luki ahkerasti aamusta iltaan milloin mitäkin ja

toisinaan kävi luennoilla kuulemassa konsa ketäkin. Mitäpä hän
niistä luennoista hyötyi, ajatukset ne kun vain aina

The Origins of Unfairness Social Categories and Cultural Evolution 1St Edition Cailin Oconnor Full Chapter

Uploaded by

Document Information

Original Title

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

The Origins of Unfairness Social Categories and Cultural Evolution 1St Edition Cailin Oconnor Full Chapter

Uploaded by

Copyright:

Available Formats

The Origins of Unfairness: Social

Categories and Cultural Evolution 1st

So many people have contributed to the production of this book. Many

on so many of the papers described here. And thanks to Nicole Bourbaki

1.1. Payoff table for a simple, correlative coordination game 28

4.2. Proportions of simulations that go to the jointly less

6.5. Phase diagrams for two populations playing a Nash

7.2. A network with two types represented by black and white

described above, in these real problems members of a group have to take

emphasize exploitation or coercion by one segment of society” (118).

inequity emerges in these models, it takes very little for it to persist

resources between social groups more generally. In Chapters 8 and 9, I

0.2 Explanation and Models of Cultural

1 These models can be especially important in directing us toward future empirical

research (Rosenstock et al., 2017).

the models discussed play an important epistemic role by outlining the

it is so well studied in the social sciences, and in part because it is a

1.1 Gender and Gendered Division of Labor

1 In philosophy, Haslanger (2000) provides a discussion of different types of definitions

shape the development of bodies (Butler, 2011a), but such consider-

1.1.2 Gendered division of labor

cross-culturally (Bird and Codding, 2015).

1.2 Coordination Problems

The second class of coordination problems I will call complementary

these jobs for gathering, hunting, toolmaking, etc.

1.3 Coordination, Convention, and Norm

10 Though discussion of social convention by philosophers goes much farther back.

Hume, in The Treatise on Human Nature, describes a convention with a coordination

stipulates a requirement of common knowledge, which is approximately

as the typical solution to a coordination problem—that actors have mutual expectations,

in Anderson (2000), who discusses several approaches to understanding social norms,

“binding” custom, where failure to meet it will lead to “sanctions of dis-

using evolutionary models to understand the emergence of behavioral

of labor. Let’s now turn to game theory to start building models of

1.4 Coordination Games

sometimes assumed about game theory).

Figure 1.1 Payoff table for a simple, correlative coordination game

convention character and toward a functional character, as discussed in

sort of game as a “mixed-motive” game because it represents a “mixture of

Figure 1.3 Payoff table for the dancing game

strategic character of the situation. Consider the games in Figure 1.4.

"Kas sinua veitikkaa!" Luunapilla hän singautti sen oikeaan

Lehtori tarkasteli katseillaan myöskin saalia, jolla rouva Streng

Rouva Lajunen raotti silmiään. Hän ei ensin käsittänyt, missä oli.

Ei hän kuitenkaan enää uudelleen nukkunut. Tuulen henki hiukan

Kuului vähän oksain raksetta. Rouva avasi silmänsä ja näki

"Mitä sinä siellä puuhasit?"

"Karkoittelin muurahaisia", selitti lehtori viattomasti.

"Onko täällä niitä?" — Rouva kauhistui omasta puolestaan.

"Näkyy olevan maassa, vaan ei puissa."

Samassa tuli Mari tuomaan kahvia, kun oli nähnyt herrasväen

Siihen keskeytyivät rouva Lajusen arvelut ja utelemiset. Mieli

Tähän puheen hälinään ja kahvikompeiden kilinään toisetkin

Nytpä kahvilla olikin maku, jollaista rouva Lajunen harvoin muisti

"Mitäpä täällä on puuhaamista", vastusti häntä rouvansa.

"Ehk'et ole ihan väärässä", myönsi viimein rouva. Mitäpä siinä

Vielä selvemmäksi varttui tämä kokemus parin kolmen viikon

Omituinen oli todellakin rouvain luonteiden vastakkaisuus. Yhä

Mikäli tämä asia hänelle selvisi, mitä suloisemmaksi, etevämmäksi

Jonkun ajan jaksoivat rehtori ja rouva Lajunen kestää uutta

Mikäli entiselleen palaaminen toisilla tapahtui, jäivät lehtori ja

Vähitellen kasvoi lehtorin mielessä omituinen kaipaus, ystävän

Jos rouva Lajunen viihtyi rehtorin seurassa, niinkuin näytti