Patrick Bateson - Specialized Research AI at MIT

Patrick Bateson

The origins of human differences

With increasing frequency the media

report the discovery of genes for dis-
tinct human characteristics, such as ath-
letic prowess or male promiscuity. Yet
it is obvious that experience, education,
and culture make a big difference in how
people behave, whatever their genetic
inheritance. Why is it that behavioral
and psychological development are so
often explained in terms of the exclusive
importance of one set of factors, either
genetic or environmental?

Oversimpli½ed opinions may derive
from a style of advocacy that is common
in many academic debates. If Dr. Jones
has overstated her case, then Professor
Smith feels bound to redress the balance
by overstating the counterargument.
The way scientists analyze complex pro-
cesses further ampli½es the confusion.
When somebody has conducted a clever
experiment demonstrating an important
long-term influence on behavior, that
person has good reason to feel pleased. It

Sir Patrick Bateson is a Fellow of the Royal Soci-
ety of London and a professor of ethology at the
University of Cambridge. He is the author (with
Paul Martin) of “Design for a Life: How Behav-
ior and Personality Develop” (2000).

is easy to forget, however, about all those
other influences that a competent scien-
tist contrives to keep constant or to play
no systematic role.

Even if the debates are seen for what
they are–irritating examples of advoca-
cy–is it not the case that complex hu-
man behaviors have come from some-
where? In some instances, surely, they
will be inborn and in other instances
they will be acquired by experience.
But the apparent good sense of this view
leaves out of account the ways in which
the inborn can be changed by experience
and the ways in which the gathering of
experience is itself inborn. Even so, it is
worth looking at some straightforward
examples.

Studies of animal behavior do, indeed,

tell us that much complex behavior can
develop without opportunities for prac-
tice. The European garden warblers that
have been hand-reared in cages never-
theless become restless and attempt to
fly south in the autumn, the time when
their wild counterparts migrate in that
direction. The warblers continue to be
restless in their cages for about a couple
of months, the time it would take them
to fly from Europe to their species’ win-
tering grounds in Africa. A similar rest-
lessness during the following spring sim-

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

The origins
of human
differences

ulates their return flight north. This
migratory response occurs despite the
birds’ rearing in social isolation, with no
opportunities to learn when to fly, where
to fly, or for how long. Cases like these
are marvels of developmental biology.
They involve the construction of a ner-
vous system that can express the full
complexity of the behavior. But the prin-
ciples involved are no more dif½cult to
understand than those involved in the
construction of, say, a kidney.

Continuing with this general line of
argument, many aspects of human be-
havioral development recur in every-
body’s life despite the shifting sands
of cultural change and the unique con-
tingencies of any one person’s life. Indi-
vidual differences among humans seem
small when any human is compared
with any chimpanzee. All humans have
the capacity to acquire language, and the
vast majority do. With few exceptions,
humans pass the same developmental
milestones as they grow up. Most chil-
dren have started to walk by about eigh-
teen months after birth, have started to
talk by two years, and have reached sex-
ual maturity by their late teens.

Similarly, human facial expressions
have characteristics that are widely dis-
tributed across cultures. The emotions
of disgust, fear, anger, and pleasure can
be easily deciphered in facial expressions
in any part of the world. Toward the end
of his life, Charles Darwin wrote The Ex-
pression of the Emotions in Man and Ani-
mals, a book that provided the stimulus
for observational studies of animal and
human behavior that have continued in-
to modern times. Darwin would show
his friends and colleagues pictures of
people expressing various emotions and
ask them, without further prompting,
to describe the emotions. In one case
he showed a picture of an old man with
raised eyebrows and an open mouth to

twenty-four people, only one of whom
did not understand what emotion that
expression indicated. Such research as
well as his extensive correspondence
with travellers and missionaries con-
vinced Darwin that humans from all
round the globe express the same emo-
tion in the same way. Darwin concluded:
“That the chief expressive actions, ex-
hibited by man and by the lower ani-
mals, are now innate or inherited–that
is, have not been learnt by the individ-
ual,–is admitted by every one.”

Subsequently, an enormous photo-
graphic archive of human expressions
from different cultures at different
stages of economic development was
formed. The similarities in the appear-
ance of the smile or the raised eyebrows,
for example, are striking. The cross-
cultural agreement in the interpretation
of complex facial expressions is also re-
markable. People agree not only about
which emotions are being expressed, but
also about which expression of a partic-
ular emotion is the more intense.

All of this might seem straightfor-
ward; the argument that some human
behavior is instinctive seems to be cor-
rect. However, the concept of instinct
is riddled with confusion. For some, it
means a distinctly organized system of
behavioral patterns, such as those in-
volved in searching for and consuming
food. For others, ‘instinct’ simply refers
to behavior that is not learned, that is
present at birth (the strict meaning of
‘innate’), or that emerges at a particular
stage in the life cycle. Another sugges-
tion is that ‘instinct’ refers to behavior
that, once developed, does not change,
or to behavior that develops before it
serves any biological function, like some
aspects of sexuality. At the same time,
some de½ne ‘instinct’ as those behav-
ioral patterns shared by all members of
the species (or at least by members of

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Patrick
Bateson
on
human
nature

the same sex and age), while others de-
½ne it as the behavioral difference be-
tween individuals caused by a genetic
difference. None of this would matter if
it were always the case that all examples
of supposed instinct had all the charac-
teristics variously attributed to it. That,
unfortunately, is not the case; many ex-
amples have some of the characteristics
but not others.

One aspect of the unitary concept of
instinct that has unravelled on further
inspection is the belief that learning
does not influence instincts once they
have developed. Learning modi½es many
cases of apparently unlearned behavioral
patterns after they have been used for
the ½rst time. Human babies who have
been born blind, and consequently never
see a human face, nevertheless start to
smile at around ½ve weeks–the same
age as sighted babies. But while sighted
children learn to modify their smiles ac-
cording to their experience, producing
subtly different smiles that are charac-
teristic of their particular culture, blind
children become less responsive and less
varied in their facial expressions. Experi-
ence can and does modify what started
out as apparently unlearned behavior.
Conversely, some learned behavioral
patterns are developmentally stable and
virtually immune to subsequent modi-
½cation. The songs some birds learn
early in life may be extremely resistant
to change once they have been acquired.
Similarly, modes in humans of perceiv-
ing language and articulating particular
sounds, once acquired, are extremely
dif½cult to change in adulthood.

The idea that one meaning of instinct,
‘unlearned,’ is synonymous with anoth-
er, namely, ‘adapted through evolution,’
also fails to stand up to scrutiny. The de-
velopment of a behavioral pattern that
has been adapted for a particular biologi-
cal function during the course of a spe-

cies’ evolutionary history may nonethe-
less involve learning during the individ-
ual’s life span. For example, the strong
social attachment that young birds and
mammals form to their mothers is clear-
ly adaptive and has presumably devel-
oped through evolution. And yet the
attachment process requires the young
animal to learn the distinguishing fea-
tures of its mother. An important point
is that Darwinian selection, by acting on
mechanisms that regulate changes in be-
havior in response to challenges from
the environment, can increase plasticity
and behavioral diversity.

In short, many behavioral patterns
have some, but not all, of the de½ning
characteristics of instinct, and the uni-
tary concept breaks down under closer
scrutiny. The various theoretical conno-
tations of instinct–namely, that it is un-
learned, caused by a genetic difference,
adapted over the course of evolution,
unchanged throughout the life span,
shared by all members of a species, and
so on–are not merely different ways of
describing the same thing. Even if a be-
havioral pattern is found to have one di-
agnostic feature of instinct, it is certain-
ly not safe to assume that it will have all
the other features as well. Perhaps for
that reason Darwin wisely refused to de-
½ne ‘instinct.’ In The Origin of Species he
wrote:

An action, which we ourselves require
experience to enable us to perform, when
performed by an animal, more especially
by a very young one, without experience,
and when performed by many individuals
in the same way, without their knowing
for what purpose it is performed, is usual-
ly said to be instinctive. But I could show
that none of these characters are univer-
sal. A little dose of judgment or reason . . .
often comes into play, even with animals
low in the scale of nature.

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

The origins
of human
differences

Should we be worried about the confu-

sion with how terms are used? Only if
we suppose that we can easily divide be-
havior into the two categories ‘innate’
and ‘learnt.’

One of the triumphs of behavioral

biology in the latter part of the twentieth
century was to relate differences in mat-
ing systems, parental behavior, foraging,
and many other aspects of adult behav-
ior to differences in ecology. A compa-
rable coherence can be brought to the
great variation in the ways in which
adult behavior can develop. In partic-
ular, the role of experience is likely to
vary considerably from one behavioral
system to another. In predatory species,
the successful capturing of fast-moving
prey requires considerable learning and
practice. The osprey does not learn to
snatch trout from water overnight. Such
animals that rely upon highly sophisti-
cated predatory skills suffer high mortal-
ity rates among their young, and those
that survive are often unable to breed
for years. This is because they have to
acquire and hone their skills before
they can capture enough prey to feed
offspring in addition to themselves. In
such cases, a combination of different
developmental processes generates the
highly tuned skills seen in the adult.
The developmental processes that
make learning, like behavioral imprint-
ing, easier at the beginning of a sensitive
period are timed to correspond with
changes that the individual will encoun-
ter as it develops under natural condi-
tions. The processes that bring the sen-
sitive period to an end are often related
to the gathering of crucial information,
such as the physical appearance of the
individual’s mother or close kin. In the
unpredictable real world, the age when
the individual can acquire crucial knowl-
edge is variable; the design of the devel-
opmental process reflects that uncer-
tainty.

In contrast to those processes ½ne-
tuned by experience, cleaning the body
is not generally something that requires
special skills tailored to local conditions.
Indeed, grooming among mammals has
almost all the various de½ning character-
istics of the old-fashioned notion of in-
stinct. Rodent grooming is, for example,
a species-typical, stereotyped system of
behavior that develops before it is of any
use to the individual.

In other words, biologists expect varia-

tion in the way behavioral patterns and
their underlying structure develop. At-
tempts to shoehorn each example into
one of the two categories are ridiculous.
The muddled use of ‘instinct’ (and
with it, ‘innate’ and ‘inborn’) does not
mean that the expression of each be-
havioral characteristic is, what Salman
Rushdie called in another context, a
p2c2e–a process too complicated to
explain. Nor does it mean that such ex-
pressions cannot be subject to evolution
when critical environmental conditions
are stable from one generation to the
next. And it certainly does not mean that
all adult behavior is totally dependent on
the environment. What we must con-
clude is that if we want to understand
developmental processes then we have
no alternative but to study them.

Plant and animal breeders know well

that many of the characteristics that
matter to them are inherited. Long be-
fore genes were postulated and dna was
discovered, breeders took this as a boun-
tiful fact of life, even though they had no
idea how inheritance worked. To take
just one example, dogs have for many
centuries been bred for their behavioral
characteristics as well as their appear-
ance. The sheepdog is especially sensi-
tive to the commands of humans, wait-
ing until the shepherd gives it a signal to
start herding the sheep. The pointer is
especially attentive to the presence of

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Patrick
Bateson
on
human
nature

certain game, stopping in its tracks
when it detects the smell of a species
such as grouse. Valued behavioral char-
acteristics such as these are clearly in-
herited and are quickly lost if breeds are
crossed with others.

Humans may also reveal through their

children how particular characteristics
are inherited. Two healthy parents from
a part of the world where malaria is rife
may have a child who develops severe
anaemia. Both parents carry a gene that
has some effect on red blood cells, pro-
tecting them against the malarial para-
site that enters those cells during part of
its life cycle. However, a double dose of
this recessive gene leads to the red blood
cells collapsing from their normal disc
shape into strange sickle-like shapes.
The child who receives this genetic lega-
cy has sickle-cell anaemia.

Few behavioral characteristics are in-
herited in as simple a fashion as sickle-
cell anaemia, and when they are, the ef-
fects are usually damaging and perva-
sive. A well-known case is the disabling
disease phenylketonuria (pku). If a
child inherits two copies of a particular
recessive gene from both parents, the
child cannot produce a crucial enzyme
required to break down phenylalanine,
an amino acid that is a normal compo-
nent of the average diet. The resulting
accumulation of phenylalanine in the
body poisons the child’s developing
brain and causes severe mental retarda-
tion–unless the condition is diagnosed
and the child is given a special diet.
Evidence for genetic influences on
human behavior is usually indirect. It is
bound to be so, because naturally occur-
ring breeding experiments are rare, and
deliberate breeding experiments in the
interest of genetic research would obvi-
ously be prohibited in most societies.
However, the study of twins has cast
some light on the links between genes
and behavior.

Research into the inheritance of hu-
man behavior has been greatly helped
by comparing genetically identical twins
with nonidentical twins. Identical (or
monozygotic) twins are genetically iden-
tical because they develop from the split-
ting of a single fertilized egg; they are
naturally occurring clones. Nonidentical
(or dizygotic) twins, in contrast, develop
from two fertilized eggs. Consequently,
they are no more similar to each other
genetically than any two siblings born at
different times. If identical twins are no
more alike than nonidentical twins in a
given behavioral characteristic, then the
genetic influence on that characteristic
is presumably weak. Conversely, when
identical twins are substantially more
alike than nonidentical twins (or sib-
lings) in a behavioral characteristic, then
the mechanism of inheritance is likely
genetic.

Another way of exploring how genes
influence behavior is to compare twins
who have been reared apart with twins
who have been reared together. The
thought behind this approach is that
separation in early infancy removes the
influence of the shared environment,
leaving only the inherited factors. The
thought is not wholly correct, however,
because even twins who are separated
immediately after birth will have shared
a common environment for the ½rst
crucial nine months after conception,
while they are together in their mother’s
womb. This obvious truth can add to the
dif½culties of sorting out the sources of
individual distinctiveness. Moreover,
being separated at birth does not pre-
clude the possibility that the different
environments in which the twins are
raised may in fact have many important
features in common.

Nevertheless, the appearance, behav-

ior, and personality of identical twins
who have been reared apart are often
startlingly similar. In one documented

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

case, for example, a pair of twins had
been separated early in life, one growing
up in California, the other in Germany.
Yet when they met for the ½rst time in
thirty-½ve years, they both arrived wear-
ing virtually identical clothes and with
similarly clipped moustaches; both had
a habit of wrapping elastic bands around
their wrists; and both had the idiosyn-
cratic habit of flushing lavatories before
as well as after using them.

Accounts such as these are sometimes

greeted with skepticism, because it is
suspected that in the interest of a good
story only the startling matches have
been reported while the discrepant twins
have been ignored. Nevertheless, some
properly conducted statistical surveys
have revealed that, on a range of mea-
sures of personality, identical twins
who have been reared apart are more
like each other than nonidentical twins
also reared apart. When making such
comparisons, it does not matter wheth-
er, as has often been argued, the mea-
sures of behavioral characteristics are
crude and relatively insensitive. The in-
escapable conclusion is that some ob-
servable aspects of individual behavior
are influenced by inherited factors.

Even the most cursory glance at hu-

manity reveals the enormous impor-
tance of each person’s experience, up-
bringing, and culture. Look at the aston-
ishing variation among humans in lan-
guage, dietary habits, marriage customs,
child-care practices, clothing, religion,
architecture, art, and much else besides.
Nobody could seriously doubt the re-
markable human capacity for learning
from personal experience and learning
from others.

Early intervention can bene½t the dis-

advantaged child, but in ways that had
not been fully anticipated. In the 1960s,
great efforts were made in the United

States to help people living in dif½cult
and impoverished conditions. The gov-
ernment program known as Head Start
was designed to boost children’s intelli-
gence by giving them educational expe-
rience before starting school. But the
program did not seem to have the sub-
stantial and much hoped-for effects on
intelligence, as measured by iq. Chil-
dren who had received the Head Start
experience displayed an initial modest
boost in their iq scores, but these differ-
ences soon evaporated after a few years.
The fashionable response was to dispar-
age such well-meaning efforts to help
the disadvantaged young.

Later research, however, has revealed

that some of the other effects of the
Head Start experience were long-lasting
and of great social signi½cance–greater,
in fact, than boosting iq scores. Several
long-term follow-up studies of people
who had received preschool training
under Head Start found they were dis-
tinctive in a variety of ways, perhaps the
most important being that they were
much more community-minded and less
likely to enter a life of crime. Head Start
produced lasting bene½ts for its partici-
pants and for society more generally, but
not by raising raw iq scores. Evidence
for the long-term bene½ts of early educa-
tional intervention has continued to ac-
cumulate. Studies like these raise many
questions about how early experiences
exert their effects, but they do at least
show how important such experiences
can be.

Even relatively subtle differences in
the way children are treated at an early
age can have lasting effects on how they
behave years later. One study compared
the long-term effects of three different
types of preschool teaching. In the ½rst
type, three- and four-year-olds were
given direct instruction, with the teach-
ers initiating the children’s activities in

The origins
of human
differences

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Dædalus Fall 2004

Patrick
Bateson
on
human
nature

a strict order; in the second, the teachers
responded to activities initiated by the
children; and in the third, known as
High/Scope, the teachers involved the
children in planning the activities, but
arranged the classroom and the daily
routine so the children could do things
that were appropriate to their stage of
development. Striking differences were
found between the children as they
grew up. When followed up at the age
of twenty-three, the individuals who
had been in the direct instruction group
were worse off in a variety of ways than
those in the other two groups. In partic-
ular, they were more likely to have been
arrested on a criminal charge and more
likely to have received special help for
emotional impairment. In comparison,
the individuals who had received the
more relaxed type of preschooling were
more likely to be living with spouses and
much more likely to have developed a
community spirit.

The importance of both genes and en-

vironment to the development of all ani-
mals, including humans, is obvious. This
is true even for apparently simple phy-
sical characteristics–take myopia, or
shortsightedness, for example. Myopia
runs in families, suggesting that it is in-
herited, but it is also affected by individ-
ual experience. Both a parental history
of myopia and, to a lesser extent, the ex-
perience of spending prolonged periods
studying close-up objects will predispose
a child to become shortsighted.

A more interesting case is musical
ability, about which strong and contra-
dictory views are held. Dissociation be-
tween general intellectual ability and
musical ability is strongly suggested by
the phenomenon of the musical idiot
savant–an individual with low intelli-
gence but a single, outstanding talent for
music. Such individuals are usually male

and often autistic, and their unusual gift
(whether it be for music, drawing, or
mental arithmetic) becomes apparent at
an early age and is seldom improved by
practice. One typical individual could
recall and perform pieces of music with
outstanding skill and almost perfect
pitch; he had poor verbal reasoning, but
that was to some degree offset by high
levels of concentration and memory.
Children who are good at music, on
the other hand, also tend to be good at
reading and to have a good sense of spa-
tial relations. The main factors fostering
the development of musical ability form
a predictable cast: a family background
of music, practice (the more the better),
practical and emotional support from
parents and other adults, and a good re-
lationship with the ½rst music teachers.
Practice is especially important, and
attainment is strongly correlated with
effort. A rewarding encounter with an
inspirational teacher may lock the child
into years of effort, while an unpleasant
early experience may cause the child to
reject music, perhaps forever. Here, as
elsewhere, chance plays a role in shaping
the individual’s development.

Research on identical and nonidenti-
cal twins has shown that the shared fam-
ily environment has a substantial influ-
ence on the development of musical
ability, whereas inherited factors exert
only a modest effect. Genetically identi-
cal twins are only slightly more alike in
their musical ability than nonidentical
twins or siblings. A study of more than
six hundred trainee and professional
musicians analyzed the origins of per-
fect pitch, the ability to hear a tone and
immediately identify the musical note
without reference to any external com-
parison. Heritable factors appeared to
play a role, as musicians with perfect
pitch were four times more likely than
other musicians to report having a rela-

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

The origins
of human
differences

tive with that skill. But the same study
also found that virtually all the musi-
cians with perfect pitch had started
learning music by the age of six. Of
those who had started musical training
before the age of four, 40 percent had
developed perfect pitch, whereas only 3
percent of those who had started train-
ing after the age of nine possessed the
ability. So early experience is also impor-
tant.

Like many other complex skills, mu-
sical ability develops over a prolonged
period; and the developmental process
does not suddenly stop at the end of
childhood. Expert pianists manage to
maintain their high levels of musical
skill into old age despite the general
decline in their other faculties. They
achieve this through copious practice
throughout their adult life; the more fre-
quent the practice, the smaller the age-
related decline in musical skill. Practice
not only makes perfect, it maintains per-
fect.

Is it possible to calculate the relative

contributions of genes and environment
to the development of behavioral pat-
terns or psychological characteristics
such as musical ability? Given the pas-
sion with which clever people have ar-
gued over the years that either the genes
or the environment are of crucial impor-
tance in development, it is not altogeth-
er surprising that the outcome of the
nature-nurture dispute has tended to
look like an insipid compromise be-
tween the two extreme positions. In-
stead of asking whether behavior is
caused by genes or the environment,
the question became: How much is due
to each? Within a single individual this
question cannot be answered, but it can
be posed for a population of individuals
as follows: How much of the variation
between individuals in a given charac-

teristic is due to differences in their
genes, and how much is due to differ-
ences in their environments?

The nature-nurture controversy ap-
peared at one time to have been resolved
by what seemed like a neat solution to
this question about where behavior
comes from. The suggested solution was
provided by a measure called heritabili-
ty. The concept of heritability is best il-
lustrated with an uncontroversial char-
acteristic such as height, which clearly is
influenced by both the individual’s fami-
ly background (genetic influences) and
nutrition (environmental influences).
The variation between individuals in
height that is attributable to variation
in their genes may be expressed as a
proportion of the total variation within
the population sampled. This index is
known as the heritability ratio. If people
differed in height solely because they
differed genetically, the heritability of
height would be 1.0; if, on the other
hand, variation in height arose entirely
from individual differences in environ-
mental factors such as nutrition, then
the heritability would be 0.

Calculating a single number to de-

scribe the relative contributions of genes
and environment has obvious attrac-
tions. Estimates of heritability are of un-
doubted value to animal breeders, for ex-
ample. Given a standard set of environ-
mental conditions, the genetic strain to
which a pig belongs will predict its adult
body size better than other variables
such as the number of piglets in a sow’s
litter. If the animal in question is a cow
and the breeder is interested in maxi-
mizing its milk yield, then knowing that
milk yield is highly heritable in a partic-
ular strain of cows under standard rear-
ing conditions is important.

Behind the deceptively plausible ratios
lurk some fundamental problems. For a
start, the heritability of any given char-

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

Patrick
Bateson
on
human
nature

acteristic is not a ½xed and absolute
quantity–tempted though many scien-
tists have been to believe otherwise. Its
value depends on a number of variable
factors, such as the particular population
of individuals that has been sampled.
For instance, if heights are measured
only among people from affluent back-
grounds, then the total variation in
height will be much smaller than if the
sample also includes people who are
small because they have been under-
nourished. The heritability of height will
consequently be larger in a population of
exclusively well-nourished people than
it would be among people drawn from a
wider range of environments. Converse-
ly, if the heritability of height is based on
a population with relatively similar
genes–say, native Icelanders–then the
½gure will be lower than if the popula-
tion is genetically more heterogeneous;
for example, if it includes both Iceland-
ers and African Pygmies. Thus, attempts
to measure the relative contributions of
genes and environment to a particular
characteristic are highly dependent on
who is measured and under what condi-
tions.

Another problem with the heritability

ratio is that it says nothing about the
ways in which genes and environment
contribute to the biological and psycho-
logical processes involved in an individ-
ual’s development. This point becomes
obvious when considering the heritabili-
ty of a characteristic such as ‘walking on
two legs.’ Humans walk on less than two
legs only as a result of environmental
influences such as war wounds, car acci-
dents, disease, or exposure to terato-
genic toxins before birth. In other
words, all the variation within the hu-
man population results from environ-
mental influences, and consequently the
heritability of walking on two legs is
zero. And yet walking on two legs is

clearly a fundamental property of being
human, and is one of the more obvious
biological differences between humans
and other great apes such as chimpan-
zees or gorillas. It obviously depends
heavily on genes, despite having a heri-
tability of zero. A low heritability clearly
does not mean that development is unaf-
fected by genes.

If a population of individuals is sam-
pled and the results show that one be-
havioral pattern has a higher heritability
than another, this merely indicates that
the two behavioral patterns have devel-
oped in different ways. It does not mean
that genes play a more important role in
the development of the behavioral pat-
tern with the higher heritability. Impor-
tant environmental influences might
have been relatively constant at the stage
in development when the more heritable
pattern would have been most strongly
affected by experience.

The most serious shortcoming of heri-
tability estimates is that they rest on the
spurious assumption that genetic and
environmental influences are indepen-
dent of one another and do not interact.
The calculation of heritability assumes
that the genetic and environmental
contributions can simply be added to-
gether to obtain the total variation. In
many cases this assumption is clearly
wrong.

One surprising conclusion to emerge

from studies of identical twins is that
twins reared apart are sometimes more
like each other than those reared togeth-
er. To put it another way, rearing two
genetically identical individuals in the
same environment can make them less
similar rather than more similar because
one of the twins is dominant to the oth-
er, entering the room ½rst and speaking
for them both. This fact pleases neither
the extreme environmental determinist
nor the extreme genetic determinist.

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

The origins
of human
differences

The environmental determinist sup-
poses that twins reared apart must have
different experiences and should there-
fore be more dissimilar in their behavior
than twins who grew up together in the
same environment. The genetic deter-
minist does not expect to ½nd any be-
havioral differences between genetically
identical twins who have been reared to-
gether. If they have had the same genes
and the same environment, how can
they be different?

Siblings are less like each other than
would be expected just by chance. The
child picks a niche for him or herself, not
on the basis of his own characteristics
but on what his siblings have done. Indi-
vidual differences emerge because chil-
dren are active agents in their own de-
velopment; children seek out their own
space. When Mary did well at art, her
younger sister Susan would not have
anything to do with drawing or painting,
even though she would probably have
been good at both. When Henry devel-
oped a flair for history and languages,
George inclined toward math and sci-
ence. Most parents with more than one
child can tell such stories.

Such interplay between siblings proba-

bly accounts for some of the influences
of birth order. Other things are also at
work, of course. Parents treat their suc-
cessive children differently–sometimes
deliberately, sometimes unwittingly.
They often have a more taut relationship
with their ½rst child than with their
later-born children; they are usually
more relaxed, positive, and con½dent
with their subsequent children, and
their preoccupation with every detail
of their children’s behavior and appear-
ance lessens. These examples emphasize
how important it is that we look careful-
ly at the transactions between the devel-
oping child and the social and physical
worlds in which he or she lives.

Any scienti½c investigation of the ori-

gins of human behavioral differences
eventually arrives at a conclusion that
most nonscientists would probably
have reached after only a few seconds’
thought: genes and the environment
both matter. How much each of them
matters de½es an easy answer, and we
have to accept that no simple formula
can solve that conundrum. We also have
to wean ourselves away from the con-
fused and utterly false idea that genes
give rise to instincts and experience
gives rise to acquired behavior. The an-
swer to the question of where knowl-
edge comes from will not emerge from
the conventional opposition between
nature and nurture. The answer requires
understanding of the biological and psy-
chological processes that build a unique
adult from a fertilized egg.

As attention is focused on develop-
ment of behavior, more and more will
be learned about the underlying process-
es. My own view is that many of these
have regularities that will be amenable
to analysis. But it does not follow that as
these regularities are uncovered human
behavior will become more predictable.
To understand why, consider a rule-
governed game like chess. It is impossi-
ble to predict the course of a particular
chess game from a knowledge of the
game’s rules. Chess players are con-
strained by the rules and the positions
of the pieces, but they are also instru-
mental in generating the positions to
which they must subsequently respond.
The range of possible games is enor-
mous. The rules may be simple but the
outcomes can be extremely complex.

The adult human brain, on which its
owner’s behavior depends, has around
one hundred thousand million (1011)
neurons, each with hundreds or thou-
sands of connections to other neurons. A
diagram of even a tiny part of the brain’s

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3

rules influence the course of a life, but
they do not determine it. Like chess
players, children are active agents. They
influence their environment and are in
turn affected by what they have done.
Furthermore, children’s responses to
new conditions will, like chess players’
responses, be re½ned or embellished as
children gather experience. Sometimes
normal development of a particular abil-
ity requires input from the environment
at a particular time; what happens next
depends on the character of that input.
The upshot is that, despite their un-

derlying regularities, developmental
processes seldom proceed in straight
lines. Big changes in the environment
may have no effect whatsoever, whereas
some small changes have big effects. The
only way to unravel this is to study what
happens.

Patrick
Bateson
on
human
nature

connections would look like an enor-
mously complex version of a map of the
New York subway system. The brain is
organized into subsystems, many of
which are dedicated to different func-
tions that may run separately but, if the
behavior of the individual is not to be a
mess, must be integrated with each oth-
er. The products of genes, the impact of
experience, and the resulting activities
of neurons are all embedded in elaborate
networks.

The idea that genes might be likened

to the blueprint of a building is hope-
lessly misleading because the corre-
spondences between plan and product
are not to be found. In a blueprint, the
mapping works both ways. In a ½nished
house, the position of each room can be
found on the blueprint, and the blue-
print indicates where every room will
be. This straightforward mapping is not
true for genes and behavior, in either di-
rection. The language of a gene for a par-
ticular behavior pattern, so often used
by scientists, is exceedingly muddling to
the nonscientist (and, if the truth be
told, to many scientists as well). This is
because the phraseology seems to imply
that the gene determines the characteris-
tic of the behavior without anything else
being important. What the scientists
mean (or should mean) is that a genetic
difference between two groups is associ-
ated with a difference in behavior. They
know perfectly well that other things are
important and that, even in constant
environmental conditions, the develop-
mental outcome depends on the whole
‘gene team.’

Nevertheless, it is likely that order
underlies even those learning processes
that make people different from each
other. Knowing something of the under-
lying regularities in development does
bring an understanding of what happens
to the child as he or she grows up. The

Dædalus Fall 2004

D
o
w
n
o
a
d
e
d

f
r
o
m
h

t
t

:
/
/

d
i
r
e
c
t
.

i
t
.

e
d
u
d
a
e
d
a
r
t
i
c
e
–
p
d

f
/

1
3
3
4
3
6
1
8
2
8
8
9
2
0
0
1
1
5
2
6
0
4
2
3
6
5
6
3
6
p
d

b
y
g
u
e
s
t

o
n
0
8
S
e
p
e
m
b
e
r
2
0
2
3
Download pdf