LINEBURG


<< . .

 6
( 18)



. . >>

be better understood by considering what advantages kin-
ship and ethnicity may offer the individual during social life.
The most elementary forms of social life among non-human
species arise where individual organisms increase their own
reproductive ¬tness by interacting with other members of their
Order and anarchy
56
own species (Trivers 1985: 41“65). Recognition of kinship is
fundamental to interaction in human society (the importance
of kinship among the Nuer and Somali was mentioned in
chapter 1). The great advantage of kinship as a means of
organising social relationships is that it has the potential to
shape interaction automatically, from birth. A person is a
child of their mother and father, a sibling to their brothers
and sisters, and so forth. If membership of a social group is
assigned by patrilineal or matrilineal descent, then the person
automatically becomes a member of their father™s or mother™s
group. The organisation of social relationships that provide
co-operation and reciprocity seems to unroll almost without
human intervention, and society to be reproduced of its own
accord. Even in our own society, where kinship is relatively
unimportant, we commonly know the names of relatives up
to the level of second cousin.1
There are also good biological reasons why people should
prefer to maintain relationships with their kin, rather than
strangers. Biological and social scientists have thus inves-
tigated the importance of kinship from different angles.
Darwinian biological evolution is about success in transmit-
ting genes. Two evolutionary theories to account for the sig-
ni¬cance of kinship in human social behaviour have been put
forward by William Hamilton and Robert Trivers. William
Hamilton (1964) developed the ¬rst theory to explain how
social interaction can contribute to an animal™s reproductive
success. Hamilton is famous in biology not only for having
demonstrated, with mathematical precision, how genetically
determined altruistic behaviour toward kin can be favoured
through natural selection, but for having his PhD thesis on
In a practical conducted with 149 ¬rst year anthropology students between
1

2002 and 2004, 51 per cent could name a great-grandparent and 49 per cent
a second cousin, but fewer than 5 per cent a more distant relative.
Self-interest and social evolution 57
the subject rejected by examiners “ who failed to appreci-
ate its signi¬cance “ as substandard (Trivers 1985: 47). In
response, Hamilton published two linked papers (Hamilton
1964) that revealed the originality of his ideas. Hamilton™s
predictions have been supported by subsequent research that
shows some animal species have evolved ¬ne-tuned abilities
to recognise their relatedness to other individuals and modify
their behaviour accordingly (Trivers 1985: 129“35).
Assuming that social behaviour was entirely under genetic
control, Hamilton asked two questions:
Under what conditions may it be advantageous, in evolu-


tionary terms, for animals to co-operate with other mem-
bers of the same species, or to forgo resources to bene¬t
another?
What genetic mechanisms favour the spread of such


behaviour through natural selection?
In the narrow sense, each individual is competing with every
other one for reproductive success, that is, the successful pro-
duction of children who will grow up carrying the same genes
as the parent, and in turn transmit these genes to grandchil-
dren. However, the individual does not share genes only with
their parents and children. Genes are also shared with brothers
and sisters, uncles and aunts, ¬rst cousins and second cousins.
But the proportion of genes held in common decreases with
genealogical distance. The evolutionary biologist Hamilton
argued that if we were more willing to help (even to die for) our
close relatives than our more distant ones, the gene responsible
for such behaviour could increase in frequency over succes-
sive generations. ˜A gene may receive positive selection even
though disadvantageous to its bearers if it causes them to con-
fer suf¬ciently large advantages on relatives™ (Hamilton 1964:
17). This is called ˜kin-selected altruism™, and explains why
Order and anarchy
58
worker ants and bees have evolved to give their lives to save
the rest of the hive from attack. In a bee or ant nest, all the
workers are children of the same queen. If, therefore, a few
sacri¬ce their lives to save the colony, the survivors are likely to
carry the same ˜altruistic™ gene, or gene complex, and transmit
it to successive generations. This extension of the concept of
reproductive success is known as ˜inclusive ¬tness™: sacri¬cing
one™s life for the colony does not increase one™s personal ¬t-
ness, but it does ensure one™s genes are transmitted to the next
generation. Hamilton predicted that second cousins would lie
at the limit of kin-selected altruism: beyond that, the percent-
age of genes shared between the altruist and the bene¬ciary
would be too low to justify risking death.
In some cases, Hamilton™s theory ¬ts well with human social
behaviour. The Sarakatsani shepherds of northern Greece
behaved as his theory predicts, by refusing to behave altru-
istically to anyone more distant than a second cousin. John
Campbell (Campbell 1964) found that every individual recog-
nised kin on his mother™s and father™s side to the level of second
cousins. For any individual, the total Sarakatsani community
was divided into two categories: those who were kin, and
those who were strangers (relatives by marriage formed a
third, intermediate category). Con¬dence, trust and a gen-
uine concern for the other™s welfare only existed between
kin. Campbell calculated that an individual™s kindred would
contain about 250 people, about half of whom were second
cousins.
As a way of talking about social relations, however,
human ˜kinship™ often extends beyond biological relatedness
to include adopted children and close friends of one™s par-
ents (as ¬ctive ˜aunts™ and ˜uncles™). Hamilton™s theory cannot
therefore provide the whole answer. In small-scale societies,
social kinship (that is, the cultural interpretation of biological
Self-interest and social evolution 59
relatedness) locates people from birth into social relationships
that allow them to call in speci¬c ways upon each other™s
labour or resources during subsistence activities. Small-scale
societies frequently have devices for extending kinship beyond
biological relatedness. The hunter-gatherers of the Kalahari
have a limited number of personal names. Anyone sharing
the same name (a ˜namesake™) is treated as a brother or sister
(Marshall 1957). In Aboriginal Australia, strangers such as
white anthropologists or community workers must be assim-
ilated into the local kinship system in order to acquire a social
identity. Once integrated, the stranger must avoid all women
classed as ˜mother-in-law™, support his ˜brothers™ in disputes,
respond generously to requests for gifts from his ˜brothers-
in-law™ (to whom he is notionally in debt for a potential wife)
and so forth.
The anthropologist Marshall Sahlins (1976) criticised socio-
biology™s use of kin selection to explain human social
behaviour. He pointed out that behaviour toward socially
recognised kin often varies independently of the degree of bio-
logical relatedness. A very good example of this phenomenon
can be seen in the custom known as cross-cousin marriage.
Cross cousins are the children of a brother and a sister; chil-
dren of two brothers or two sisters are referred to as parallel
cousins, because the gender of the linking parents is the same.
Cross-cousin marriage is practised by a number of small-scale
societies in Australia, Southeast Asia and North and South
America. Suppose that, in a society where group membership
is traced through men (patrilineal descent), and women have
to leave their group at marriage, two men in different groups
form an alliance by exchanging their sisters in marriage. If the
alliance is to be sustained in future generations, cross cousins
(mother™s brother™s children and/or father™s sister™s children)
are ideal marriage partners because they are born into the
Order and anarchy
60




Figure 2.1 Development of a Yanomam¨ marriage alliance.
o

allied lineage, whereas parallel cousins are born into one™s own
lineage (Figure 2.1). Parallel cousins are therefore classed as
brother and sister, while genetically equally close cross cousins
are classed as brother/sister-in-law or wife/husband.
The ¬‚exibility of social kinship has also enabled human soci-
eties to adjust their behaviour toward kin in ways appropriate
to the subsistence economy. Most recent hunter-gatherer soci-
eties live in marginal environments where freedom of move-
ment is essential. In order to sustain far-¬‚ung social networks
Self-interest and social evolution 61
the majority of hunter-gatherer societies therefore make no
distinction between kin on the mother™s and father™s side liv-
ing elsewhere in the region and have equal visiting rights
with both. Among nomadic pastoralists, on the other hand,
men must co-operate to defend livestock against raids, and
the majority of pastoral societies practise patrilineal descent.
Rights to cattle are shared by male relatives and a man™s rights
are inherited by his sons. Among horticulturalists such as the
Iroquois (chapter 1) much of the work is done by women, and
mother“daughter co-operation has been argued to explain the
relatively high incidence of matrilineal descent in such soci-
eties. A woman™s rights to land are inherited by her daughters
(Goldschmidt 1979, Holden and Mace 2003). Where descent is
patrilineal women join their husband™s group; where descent
is matrilineal men join their wife™s group.


s o c i a l e vo lut i on an d ga m e t h e o ry
The theoretical background
The ¬‚exibility of human social behaviour toward kin demands
a theory to explain why different strategies are adopted in dif-
ferent contexts. Social anthropologists long regarded this vari-
ability within a single human species as a trump card in their
arguments against genetic determinism of the kind implied by
Hamilton. Game theory has provided a useful way forward
in the reintegration of social and biological theory. Game
theory provides explanatory models that are vital to appre-
ciating the relationship between order and anarchy, showing
why people work hard to construct social relationships in cer-
tain contexts, yet repudiate them if conditions change. The
modern theory for the evolution of co-operation originated
in John von Neumann and Oskar Morgenstern™s Theory of
Order and anarchy
62
games(1944),atreatiseoneconomics.Morgensternconsidered
that economic theory treated economic actors as autonomous
decision-makers. It failed to take proper account of the fact
that economic actors are dependent on one another™s decisions;
that they operate in a social milieu (Nasar 1998: 84). Treating
economic negotiations as a game between two players would
help overcome this weakness. Such a micro-approach seems
well suited to the analysis of small-scale social interaction in
local communities but it has also been applied to international
relations (cf. Locke 1960: 98“9). According to Sylvia Nasar,
Morgenstern persuaded von Neumann to set out a mathemat-
ical framework for the study of games. The best-developed
part of the theory concerned ˜zero-sum two-person games™. In
a zero-sum game the winnings are ¬xed, and the two players
are therefore in competition to see who can gain the largest
share. Hargreaves and Lispy Jones were playing a zero-sum
game; one™s survival would be secured at the cost of the other™s
death. When they were proposed, such games of total con¬‚ict
did not seem particularly relevant to the social sciences, but
the model was taken up by post-war military strategists. Air
battles were represented as duels between a pair of opposing
planes. There was a trade-off between two con¬‚icting plans:
waiting until the opponent approached, so as to have a better
chance of hitting him, and ¬ring ¬rst to avoid being hit. Kemp
describes how ˜Lispy spurred forward his horse at breakneck
speed, shooting a staccatic volley. Hargreaves moved more
slowly.™ (Afterwards Lispy revealed that Hargreaves™s belt
buckle, glistening in the sun, had made an easy target.)
As nuclear weapons grew more destructive, however,
strategists in the United States came to appreciate that the duel
model was inappropriate and co-operation advantageous. The
United States and the Soviet Union now shared an interest in
avoiding mutual disaster. This dilemma posed sociologically
Self-interest and social evolution 63
more interesting questions. Co-operation, negotiation and
disarmament could bene¬t both, if the other could be trusted.
In a non-zero-sum game, the winnings can be increased
through co-operation. The problem, as Nasar (1998: 117)
explains, was that co-operation appeared to demand an
umpire, a Hobbesian sovereign, a World Government who
could enforce disarmament on both sides, or (perhaps) a medi-
ator trusted by both sides. This was not attractive to the United
States, whose government was determined to live, in Locke™s
terms, in a state of nature.
John Nash solved the problem of co-operation in the
absence of a sovereign by demonstrating that even where there
is no umpire to enforce an agreement, non-zero-sum games
can reach an equilibrium point, ˜a situation in which no player
could improve his or her position by choosing an alternative
available strategy™ (Nasar 1998: 97). If players can calculate
their best strategy, they do not need an umpire to reach agree-
ment. The problem identi¬ed by Hobbes and Locke had been
solved, and the conditions under which reciprocal altruism
will bene¬t both parties can potentially be identi¬ed.
A year after the theory™s publication, the model of the
Prisoner™s Dilemma was devised to exemplify Nash™s theory,
and to explore when reciprocal altruism becomes a stable strat-
egy. This famous parable uses the model of two suspects who
have been arrested and are being interrogated in different
rooms. The prisoner wonders whether he can trust the other
to remain silent. Each is told that, if they alone implicate the
other in the crime, they will be rewarded. If both confess, both
will receive a moderate sentence, since their confession helped
the police solve the crime. If one refuses to confess (i.e. refuses
to ˜defect™), even though the other has done so, his sentence
will be heavier. If the other prisoner is suspected of having
confessed, it will therefore be better to take the same course
Order and anarchy
64
oneself (Trivers 1985: 389“90).2 At ¬rst sight, the most rational
plan seems to be to defect rather than trust the other prisoner
to remain silent. Mutual defection is however more costly than
co-operating with the other prisoner so both should remain
silent. Each prisoner faces the dilemma that, although defec-
tion is less risky than co-operation, if both defect they will both
do worse than if they had co-operated with each other. If Nash
were correct, the game should produce an equilibrium point in
which both prisoners use the strategy that gives them the best
outcome available in the circumstances. If the other prisoner
cannot be trusted, that would seem to be for both to confess.
Nasar (1998: 119) argues that the Prisoner™s Dilemma
refutes Adam Smith™s claim that individuals pursuing their
private interests will inevitably bene¬t the collective (in this
case, the pair of prisoners as a group). The dilemma shows
that if each prisoner pursues their immediate private inter-
est every time they are arrested they do not achieve the best
long-term outcome for themselves, let alone for the other
prisoner. Here lies a solution to the war of each against all.
Robert Axelrod realised that, even though co-operation may
be preferable to short-term self-interest, it would only develop
if the prisoners can anticipate each other™s intentions (Axelrod
1990). Since they are secluded from one another in the cells,
anticipation must be based on prior knowledge. If the game
is played once, the stable strategy will be to defect, but if it
is played repeatedly by the same players the stable strategy
may be to co-operate through remaining silent. To rely on

The actual rewards and punishments can be set at different values (compare
2

Dawkins™s 1976 account with Trivers™s). The crucial feature is that they
must be ranked so that the temptation to defect alone gives the highest
˜payoff ™, followed by the reward of mutual silence, the punishment for
mutual defection and last, the ˜sucker™s payoff ™ for remaining silent when
the other defects.
Self-interest and social evolution 65
co-operation, the prisoners must already have interacted with
each other in ways that test their loyalty to one another. They
must, in other words, have evidence of the other™s commit-
ment to reciprocal altruism. This provided a clear explana-
tion for the desire to perpetuate social relationships out of
self-interest, the condition envisaged by Locke. By simulat-
ing the game on a computer, Axelrod (1990: 42) found the
most stable long-term strategy was one called ˜Tit for tat™. In
˜Tit for tat™ the player begins by anticipating that the other will
co-operate (not confess) and then, in subsequent moves, does
what the other player did in their previous move. In this way
other players who co-operate are rewarded, but those who
defect are punished. The cumulative bene¬ts of co-operation
are greater than those of always confessing to the jailer, since
mutual betrayal eliminates the reward for confession. If they
are able to identify and pair off with trustworthy partners,
those playing ˜Tit for tat™ can isolate those who play ˜Always
defect™ and refuse to play with them.
This discovery can be extended to real life situations of
reciprocal exchange where the return is delayed, and the
dilemma is whether one can trust the other to make a return
gift when one is in need. Make a gift on the ¬rst occasion, but
withhold it on the second if it is not reciprocated. It explains
how co-operation can evolve in a ˜state of nature™, even when
it is in competition with sel¬shness.
The ideas developed by Nash, Axelrod and others show
how people who renege on their reciprocal obligations, or fail
to contribute to co-operation, are a threat to self-regulated
social order. In the terms of commons management, they
are the ˜free-riders™. In a realistic situation, where people
make mistakes, a strategy a little more forgiving than ˜Tit
for tat™ may score even more highly. In fact, since the success
of each strategy varies according to those played against it,
Order and anarchy
66
increasingly generous strategies do better than others until
eventually they themselves succumb to the short-term strat-
egy ˜Always defect™, allowing ˜Tit for tat™ to succeed once
more (Ridley 1996: 75“8). Social networks may thus expand
or shrink according to the competing strategies in play at any
time. Stable social situations, where people frequently inter-
act with each other, provide the best context for trust and
co-operation to develop.
The costs and bene¬ts of each strategy depend on the
resources being played for. Bruce Winterhalder (1996) empha-
sises that there may always be more than one strategy in play
at any time. ˜Tolerated theft™ or ˜scrounging™, which has been
observed in non-human species, can take place if an individual
has so much of a resource that the surplus is not worth defend-
ing and scroungers are allowed to take some. The success of
this strategy will depend on the proportion of scroungers
to productive individuals (Vickery et al. 1991). The higher
the frequency of scroungers the more are incentives to work
productively threatened. Scrounging was rampant in Johan-
nesburg and had become a real threat to social order. ˜The
easiest way to get the gold dust was “ hop in and take it
from some other fellow. Get him drunk ¬rst, or get him in an
argument. No one cared what happened to him™ (Kemp 1932:
24). Bonny McCay and James Acheson (1987) and Elinor
Ostrom (1990) showed that, where the management of com-
mon property is concerned, self-regulation only works when
scroungers (˜free-riders™) who over-exploit the commons can
be detected and punished. There are two options. Everyone in
the community may police their own partners in exchange, or
authority (sovereignty) may be delegated to representatives
of the community. Those who co-operate can probably never
entirely eliminate free-riders who scrounge or renege on their
obligations, but it is in their interests to minimise competing
strategies if they are to get the fullest bene¬ts from reciprocity
Self-interest and social evolution 67
(Nowak and Sigmund 1998: 575). This is a strong incentive
to maintaining social order, even in the simplest societies. If,
however, conditions change so as to reduce people ™s expecta-
tion that they will need each other in future, or if the rewards
of scrounging increase, the social order may break down.

Game theory and altruism
Robert Trivers™s theory of reciprocal altruism built on the dis-
coveries of game theory to explain the evolution of altruistic
exchange between individuals who are not necessarily closely
genetically related; to move, in other words, beyond the purely
kin-selected form of altruism explained by Hamilton. If one
individual temporarily has more of a resource than she or he
needs, they may choose to share the surplus with another indi-
vidual who is temporarily suffering a shortage. Trivers argues
such behaviour will bene¬t both individuals if the debt is sub-
sequently repaid. This pattern is most likely to develop where
there is a risk of death, such as from starvation, and where it
is impossible to predict which individual will be successful on
any one occasion, yet those who are successful in obtaining
food get more than their immediate need. Both partners will
therefore survive whereas, on their own, both would proba-
bly soon have died. They are playing a non-zero-sum game.

<< . .

 6
( 18)



. . >>

Copyright Design by: Sunlight webdesign