Matt Ridley - Recherche en IA spécialisée au MIT

Matt Ridley

The dnabehind human nature:
gene expression and the role of experience

The idiosyncrasies of one person can-

not be human nature, nor can a feature
of human behavior that is merely typi-
cal of many animals, such as hunger.
Human nature must be the product of
a uniquely human, but near species-
universal, combination of dna se-
quences suitably refracted through typ-
ically human environmental experi-
ences.

Those sequences do not have to be
only genes. En effet, recent evidence sug-
gests that regulatory sequences, rather
than coding sequences, may be the best
place to search for ‘human nature dna.’
As Steven Pinker has pointed out, it is a
historical accident, and the source of
much confusion, that genes are equated
with the genome by lay people but strict-
ly de½ned as protein-coding regions by
molecular biologists.

Right up until the sequencing of the
human genome, a piece of conventional

Matt Ridley is the author of “The Red Queen:
Sex and the Evolution of Human Nature” (1993),
“The Origins of Virtue” (1996), “Genome: Le
Autobiography of a Species in 23 Chapters”
(1999), and “Nature Via Nurture: Genes, Expe-
rience, and What Makes Us Human” (2003).

wisdom was con½dently repeated as
truth by scientists, journalists, and com-
mentators: there were a hundred thou-
sand genes in the human genome, à propos
half of which were unique to the brain.
So widely was this ‘fact’ disseminated
that it is hard now to discern its original
source.1 But it was about as wrong as a
scienti½c assertion can be. We now know
that human beings have approximately
twenty-½ve thousand genes (humiliat-
franchement, that is ½fteen thousand less than a
rice plant has); that most are expressed
in both the brain and the body; et ça
very few indeed, perhaps none, sont
unique to the human species. Not only
do mice also have twenty-½ve thousand
genes, but they have essentially the same
twenty-½ve thousand.

Yet mice are not men. Something must
be different. The sequencing of genomes
has suggested a new hypothesis: that an-
imal evolution usually works not by in-
venting new protein-coding genes (ce
appears to be commoner in plants), mais
by altering the timing, intensity, and lo-
cation of the expression of preexisting

1 Voir, Par exemple, J.. Madeleine Nash, “Fertile
Minds,” Time 149 (5) (3 Février 1997): “There
are only 100,000 genes in human dna. Même
though half these genes–some 50,000–appear
to be dedicated to constructing and maintain-
ing the nervous system.”

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

Matt Ridley
sur
human
nature

genes.2 For instance, the Hoxc8 gene es-
sentially tells a developing fetus where to
grow a thorax with ribs. Hoxc8 is turned
on farther back in a chicken than in a
mouse, giving a chicken a longer neck.
It is turned on throughout the body of a
python, which is almost all thorax. Yet it
is essentially the same gene: equivalent
Hox genes can be swapped between ani-
mal species and still work. Somewhere
in each species’ dna are sequences that
cause slightly different tissue-speci½c
expression of Hoxc8.3

A literary analogy is helpful. Any two

novels, say, David Copper½eld and The
Catcher in the Rye, are written using
roughly the same set of words. Some
words appear in one but not the other–
‘caul’ and ‘pettish’ appear in Dickens;
‘crap’ and ‘elevator’ appear in Salinger–
but they are very few. The difference be-
tween the books’ plots lies in the order
and pattern of the words, not in the
words themselves. The difference be-
tween a man and a mouse lies in the or-
der and pattern of gene expression. Et
that difference is achieved by variations
in the regulatory sequences of the ge-
nome (hereafter referred to as promot-
ers), of which more shortly.

In this context, the genes that lie be-

hind human nature are universal to
mammals, possibly to all animals, mais
the pattern and timing of their expres-
sion during normal development results

2 Sean B. Carroll, “Endless Forms: The Evolu-
tion of Gene Regulation and Morphological
Diversity,” Cell 101 (6) (2000): 577–580.

3 Heinz-Georg Belting, Cooduvalli S. Shashi-
kant, and Frank H. Ruddle, “Modi½cation of
Expression and Cis-Regulation of Hoxc8 in the
Evolution of Diverged Axial Morphology,” Pro-
ceedings of the National Academy of Sciences 95 (5)
(1998): 2355–2360; Martin J. Cohn and Cheryll
Tickle, “Developmental Basis of Limblessness
and Axial Patterning in Snakes,” Nature 399
(1999): 474–479.

in typical human behavior. This has an
unexpected bonus for scientists interest-
ed in human nature. It means that the
discovery of a gene’s function in an ani-
mal will almost certainly lead directly to
the discovery of the same gene’s func-
tion in a human being. As our knowl-
edge of the genes that affect behavior is
deepened by experiments in mice, comme
well as in dogs and other species with
behaviorally distinct breeds, that knowl-
edge will quickly and inevitably improve
our understanding of human behavior,
aussi. Bien sûr, there will be differences,
but discovering these differences will
itself be both easy and instructive.

De même, the source of genetic vari-
ability in human nature among individ-
ual people will be found mainly in se-
quence differences that affect gene ex-
pression. It has been known since the
work of Jacques Monod and François
Jacob in the late 1950s that a gene is ex-
pressed, or transcribed into messenger
rna, by the binding of a protein called a
transcription factor to a promoter, a spe-
cial sequence of bases usually found im-
mediately upstream of the gene itself.
En outre, a gene may be switched
off by the binding of another protein to
another sequence nearby. In some cases,
more than one protein must bind to the
dna before a gene is expressed, et le
regulatory sequences may be spread out
over long stretches of dna–even longer
than the gene itself. Par exemple, le
‘eve’ gene in fruit flies, whose job is to
control other genes during development,
is switched on at least ten separate times
during development, and it has eight
separate regulatory sequences attached
to it, three upstream of the gene and ½ve
downstream. Each of these sequences
requires ten to ½fteen proteins to attach
to it to switch on expression of the eve
gene, and together they cover thousands
of letters of dna text. In different tis-

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

The dna
behind
human
nature

sues, different promoters and enhancers
(distant or downstream switches) sont
often used to express the gene.4

This implies that many, perhaps most,
of the interesting differences between a
human being and a mouse, or between
one human being and another, will be
found in the sequence of bases in pro-
moters, rather than in protein-coding
genes. Intriguingly, this hypothesis
opens the door for cultural and environ-
mental influence, because the ef½cient
binding of transcription factors, et
therefore the expression of genes, can in
some cases be altered by factors extrinsic
to the organism–by, in a word, experi-
ence. Steroid hormones, Par exemple,
once they have formed a complex with
their receptors, act as transcription fac-
tors, activating or suppressing the ex-
pression of genes. So elevation of corti-
sol–following the sensory detection of
a stressful experience–can alter gene ex-
pression, particularly in the immune sys-
tem. It indirectly reduces expression of
Interleukin-2 and turns down the activi-
ty, number, and life span of lympho-
cytes.5

Even more strikingly, it is now clear
that a genetic mechanism underlies the
very un-hereditary process of forming
new memories by associative learning.
Such learning in flies, mice, et les gens
consists mainly of changes in the expres-
sion of creb genes in response to expe-
rience. These changes result in shifts in
the strength of particular synaptic con-
nections between neurons–and these
shifts are the manifestation of new
memories. It is clearly misleading to call
the creb gene a determinant of human
4 Mark Ptashne and Alexander Gann, Genes
and Signals (Cold Spring Harbor, N.Y.: Cold
Spring Harbor Laboratory Press, 2002).

5 Paul R. Martine, The Sickening Mind: Cerveau, Be-
haviour, Immunity and Disease (Londres: Harper-
Collins, 1997).

nature, because what it determines
depends on what the organism experi-
ences, and yet it is human nature to have
a responsive creb gene that enables us
to learn.6
The twenty-½ve thousand genes in a

mammalian genome, played like a great
piano by their many thousand promot-
ers, and probably able to express at least
three times as many proteins through al-
ternative splicing, are amply capable of
encoding a subtle and complex human
nature throughout the tissues of a hun-
dred-trillion-cell body, even without
supposing a role for experience. Là
is no reason to assume that the ‘higher’
and more peculiarly human faculties
such as intelligence, langue, and social
empathy are less influenced by genes
than are features we normally think of
as more primitive, such as aggression or
hunger.

Here follow three genes that distin-
guish human beings from other animals,
not by their existence, but by their se-
quence–in either the coding or the reg-
ulatory region–and by their pattern of
expression. The ½rst bears on intelli-
gence, the second on language learning,
the third on pair-bond formation–all
‘higher’ human faculties.

The ½rst gene concerns the place
where human anatomy meets human
nature: brain structure. Having an un-
usually large brain for its body size is
characteristic of the human being. Si le
gene expression theory is correct, ce
feature should result from the differen-
tial expression or activity of a gene or
genes in human beings. One candidate
gene is already known, thanks to the
study by Geoffrey Woods and colleagues

6 Josh Dubnau and Tim Tully, “Gene Discov-
ery in Drosophila: New Insights for Learning
and Memory,” Annual Review of Neuroscience 21
(1998): 407–444.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

Matt Ridley
sur
human
nature

of inherited microcephaly in a group of
inbreeding Kashmiri immigrants in
Bradford, England. Microcephaly is the
development of a small but otherwise
normal brain.

Four separate mutations in the same
gene were found to be one cause of the
condition. The gene, ½rst isolated in
drosophila, is called aspm, for abnor-
mal spindle protein. Found on chromo-
quelques 1, it is a gene that varies consider-
ably in length between species, produc-
ing a protein that is 1,186 amino acids
long in nematode worms; 1,861 in fruit
flies; 3,123 in mice; et 3,477 in human
beings. This elongation is caused mainly
by extra repetitions of a 75-base-pair
calmodulin-binding motif, which is
repeated seventy-four times in human
beings, sixty-one in mice, twenty-four
in flies, and twice in nematodes. (Le
motif, by a happy accident, is called the
iq motif, after the ½rst two letters of its
amino acid sequence, isoleucine and glu-
tamine.) It appears that the longer the
protein, the more effective it is at assist-
ing mitosis in neuronal stem cells in the
developing brain. Since stem cells multi-
ply only for a set period during develop-
ment, faster mitosis will yield more neu-
rons and a bigger brain.7

aspm is not in itself suf½cient to ex-
plain the expansion in human brain size
over the past ½ve million years, because
all higher primates have approximately
the same number of iq repeats in the
gene.8 The gene may have altered to

7 Jacquelyn Bond, Emma Roberts, Ganesh H.
Mochida, Daniel J. Hampshire, Sheila Scott,
Jonathan M. Askham, Kelly Springell, Meera
Mahadevan, Yanick J. Crow, Alexander F.
Markham, Christopher A. Walsh, and C. Geof-
frey Woods, “aspm Is a Major Determinant of
Cerebral Cortical Size,” Nature Genetics 32 (2)
(2002): 316–320.

8 C. Geoffrey Woods, personal communica-
tion, Mars 2004.

make primates brainier than other
mammals, but not to make human
beings brainier than other primates.
The search for the source of that differ-
ence has now turned to other genes af-
fecting brain size. Yet the aspm story
serves as a strong reminder of just how
simple it might be for a species to ac-
quire an increased brain size merely by
lengthening one gene with extra copies
of a motif. Dans ce cas, the change is
not in a promoter but in the gene itself,
resulting in a more active protein from
a longer gene.

The second gene affecting higher hu-

man function concerns language. Hu-
man beings are not just chimpanzees
with bigger brains; they also have quali-
tatively different natures. Some differ-
ences of degree between human beings
and all other mammals are so wide that
they qualify as differences in kind. Un
such is language. The human capacity
for learning languages shows all the hall-
marks of an instinct underpinned by
genes: it emerges unbidden and shows
universal similarities in all people.

Encore une fois, the study of people with
linguistic defects has led to insights into
which genes are especially important in
differentiating human language skills
from other primates’ communication
talents. By examining an extended fami-
ly in which speech and language de½cits
are plainly inherited as a dominant al-
lele, Simon Fisher and Cecilia Lai found
a candidate gene, called foxp2, on chro-
mosome 7.9 Other cases now con½rm
that lack of a functional form of foxp2
seems to impair learning that uses senso-
ry feedback to alter the circuitry in the
brain to lay down new sequences of

9 Cecilia S. L. Lai, Simon E. Pêcheur, et coll., “A
Forkhead-Domain Gene is Mutated in a Severe
Speech and Language Disorder,” Nature 413
(2001): 519–523.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

The dna
behind
human
nature

orofacial gestures and new memories
thereof.

How might foxp2 do this? Fox, ou
forkhead box, genes are transcription
factors whose job seems to be to activate
or repress transcription of other genes.
They are universal to animals and fungi.
In mammals, which have at least forty
Fox genes, foxp2 shows remarkably
little variation between species. Of 136
nucleotide substitutions in the gene
between chimpanzee and mouse, only
one alters the amino acid sequence of
the protein; the rest are synonymous.
Since the common human-chimp ances-
tor, cependant, there have been two
amino-acid-altering changes, making
the human foxp2 protein stand out
from all other mammal versions so far
studied. And all but a very few human
beings have identical versions of
foxp2.10

De plus, a study by Svante Pääbo
and colleagues of the number and pat-
tern of silent substitutions in noncoding
dna nearby seems to show that the two
mutations were involved in a selective
sweep about two hundred thousand
years ago, during which they elbowed
aside all other versions of the gene, pre-
sumably as a result of natural selection.
This date is intriguing because it does
not predate by much the Upper Pale-
olithic Revolution in Africa and, là-
fore, possibly the beginning of symbolic
communication and modern language,
according to physical anthropologists.
After a million years of technological
stasis, there was sudden and cumulative
cultural change–new tools, artifacts,
pigments, trade–some time before one

10 Wolfgang Enard, Molly Przeworski, Simon
E. Pêcheur, Cecilia S. L. Lai, Victor Wiebe, Ta-
kashi Kitano, Anthony P. Monaco, and Svante
Pääbo, “Molecular Evolution of foxp2, a Gene
Involved in Speech and Language,” Nature 418
(2002): 869–872.

hundred thirty thousand years ago, le
time when long-distance trade was de½-
nitely established.11 Some small band of
African human beings apparently took
over the world starting at this time, et
we are all their descendants.

Bien sûr, there is no direct evidence
that foxp2 was anything other than a
fortunate bystander at this revolution,
and even if it did play a role, it has plenty
of other functions in the body besides
facilitating language–it is expressed in
the lung, Par exemple. But it is also ex-
pressed during early development in
those parts of the brain crucial to
speech. People with mutated foxp2
genes show under-activation of Broca’s
speech area when engaged in linguistic
tasks, implying that some de½ciency in
the structure of that part of the brain re-
sulted from that mutation.12

Birds have a FoxP2 gene that is surpris-

ingly similar to that of mammals, given
the evolutionary distance between the
two classes, which suggests extreme
conservation or convergent evolution.13
This gene (and another, FoxP1) is ex-
pressed especially strongly in male song-
birds in the striatal nucleus known as
Area X–part of the ‘song circuit.’ For
instance, expression here rises during
the period when young zebra ½nches
learn their songs and during the season
when adult canary songs become unsta-

11 Sally McBrearty and Alison S. Brooks, “The
Revolution That Wasn’t: A New Interpretation
of the Origin of Modern Human Behaviour,»
Journal of Human Evolution 39 (2000): 453–563.

12 Gary F. Marcus and Simon E. Pêcheur,
“foxp2 in Focus: What Can Genes Tell Us
About Speech and Language?” Trends in Cog-
nitive Sciences 7 (2003): 257–262.

13 Note that the current convention is to ex-
press the names of human genes in uppercase
letters, mouse genes in lowercase letters, et
bird genes in a mixture of the two. This absurd
system cannot last.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

Matt Ridley
sur
human
nature

ble. Both these results hint that FoxP2
expression is somehow vital to the laying
down of new vocal procedures in birds’
brains.14

Assuming human foxp2 does alter
the development of Broca’s area in such
a way as to facilitate the learning of lan-
guage, there is an obvious problem. Le
½rst human being with a modern foxp2,
somewhere in Africa two hundred thou-
sand years ago, would have been in the
same ½x as Victor of Aveyron, Kaspar
Hauser of Nuremberg, or Genie of Los
Angeles–children reared largely in iso-
lation from spoken language until their
teens and who thus missed the critical
period when the brain is most open to
language learning. He or she would have
developed little of his or her full linguis-
tic potential. Cependant, once there were
several children with the new gene, un
sort of bootstrapping might have been
possible as they practiced their language
skills among themselves; something
similar happened in Nicaragua in 1979
when deaf children were suddenly
brought together in one school and
spontaneously developed their own Cre-
ole sign language. Then each generation
would have added to the complexity, et
within a few generations this chattering
group of people would have been capa-
ble of feats of planning and organization
foreign to their fellow human beings.
I repeat: it is highly implausible that
changes in foxp2 alone made language
possible. Plutôt, it was probably one
of many genetic changes that helped
improve the emerging communication
skills of proto-people. But the principle,
that species-wide changes in single

genes can affect ‘higher’ behavioral traits
in predictable ways, is well supported.

The third gene that affects a human

trait concerns love. One way in which
human beings differ markedly from
their closest evolutionary relatives, le
chimp and the bonobo, is in habitually
forming long-term pair bonds. Ceux-ci sont
so intrinsic to human nature that they
form even in libertarian communes that
expressly try to outlaw them. Bien sûr
not all human beings form long-term or
exclusive pair bonds, but human beings
show all the hallmarks of a long-bonding
species: sexual jealousy, paternal care,
sexual division of labor, etc.. Chimpan-
zees and bonobos, on the other hand,
maintain only brief pair bonds that do
not last longer than the estrus period of
the female, if that.15

A cet égard, human beings resem-

ble prairie voles and chimpanzees re-
semble montane voles, two equally
closely related species that also differ
in mating systems. The control of pair-
bonding in voles is now well understood.
In both sexes in both species of vole, sex-
ual intercourse stimulates the release of
the small peptide hormones oxytocin
and vasopressin in the brain. Injecting
the hormones into the brain brings on
pairing behavior in prairie voles but not
in montane voles. Increasing the expres-
sion of the receptor genes also makes
prairie voles quicker to form pair
bonds.16

15 Matt Ridley, The Red Queen: Sex and the Evo-
lution of Human Nature (New York: Manchot,
1995).

14 Sebastian Haesler, Kazuhiro Wada, UN.
Nshdejan, Edward E. Morrisey, Thierry Lints,
Eric D. Jarvis, and Constance Scharff, “FoxP2
Expression in Avian Vocal Learners and Non-
Learners,” Journal of Neuroscience 24 (2004):
3164–3175.

16 Expressing the prairie-vole version of the
gene in a mouse markedly increases its social-
af½liation behavior. See Thomas R. Insel and
Larry J. Jeune, “The Neurobiology of Attach-
ment,” Nature Reviews in Neuroscience 2 (2001):
129–136.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

The dna
behind
human
nature

Receptors for these hormones are dis-
tributed differently in the brains of the
two species. In prairie voles, the recep-
tors are found in the nucleus acumbens
(oxytocin) and the ventral pallidum (va-
sopressin). These brain areas contain a
dopamine system that is responsible for
addictive behavior. A prairie vole there-
fore becomes ‘socially addicted’ to its
mate following sex. A montane vole does
pas. De même, when human beings who
are in love are asked to contemplate a
picture of their beloved, the area of the
brain that is active is a dopamine region
implicated in cocaine addiction.17

The different distribution of the recep-
tors is in turn caused by the presence (dans
prairie voles) or absence (in montane
voles) of a long segment of highly repeti-
tive dna text in the promoter upstream
of the gene. Inserting this text into the
promoter of a promiscuous vole species
essentially monogamizes the rodent.18
Human beings also have a repetitive seg-
ment in this region, though it is shorter
than that in prairie voles. As of this writ-
ing, the equivalent region of the chim-
panzee genome has not yet been looked
à. I predict it will be shorter than the
human one.

These three cases illustrate very graphi-

cally that it is possible to isolate genes
that have disproportionate influence on
behavior, and to do so in features rele-
vant to ‘higher’ human nature, tel que
intelligence, langue, and love. In the
1960s, the idea of ½nding ‘behavior

17 Andreas Bartels and Semir Zeki, “The Neu-
ral Basis of Romantic Love,” NeuroReport 11
(2000): 3829–3834.

18 Miranda M. Lim, Zuoxin Wang, Daniel E.
Olazabal, Xianghui Ren, Ernest F. Terwilliger,
and Larry J. Jeune, “Enhanced Partner Prefer-
ence in a Promiscuous Species by Manipulating
the Expression of a Single Gene,” Nature 429
(6993) (2004): 754–757.

genes’ at all would have been astonish-
ing, not to say heretical, but people such
as Benson Ginsburg working with mice
and Seymour Benzer with flies soon es-
tablished that behavioral mutants could
be produced just as easily as anatomical
mutants. The unexpected similarity of
human and animal genomes has now
made it possible to study in other spe-
cies the evolution of genes relevant to
human intelligence, langue, and love.
The development of behavior, in other
words, proves to be just as amenable to
genetic reductionism as anatomy and
physiology.

Human nature, cependant, is not identi-
cal in all people, and much of that diver-
sity in behavior is a consequence of the
fact that we are not clones. Studies of
identical and fraternal twins raised
à part, but in similar social settings, have
unambiguously revealed that different
people have different personalities large-
ly because they have different genes,
rather than because they have different
upbringings. Cependant, these studies,
which prove so powerful in showing the
influence of genes, have been largely in-
capable of shedding light on precisely
which genes influence personality. Depuis
the other end of the telescope, cependant,
genetic differences among individuals
are emerging that correlate with differ-
ences in how people behave. The hay-
stack is revealing its ½rst few needles.

One example is the gene on chromo-

quelques 11 for a protein called brain-
derived neurotrophic factor (bdnf).
The gene spells out the recipe for a pro-
tein that acts as a sort of fertilizer in the
brain, encouraging the growth of neu-
rons, and that probably does much else
besides. In most people, the 192nd letter
in the gene is G, but in about one-quar-
ter of people it is A. This causes a slightly
different protein to be built–with me-
thionine instead of valine at the 66th

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

Matt Ridley
sur
human
nature

(out of 247) codon. Since everybody has
two copies of each gene, there are three
kinds of people in the world: those with
two methionines in their bdnfs, ceux
with two valines, and those with one of
chaque. Personality questionnaires reveal
que, at least in one population, the met-
mets are noticeably less neurotic than
the val-mets, who are in turn noticeably
less neurotic than the val-vals.19

Cependant, this kind of single-nucle-
otide polymorphism (snp), while fre-
quently found to cause rare hereditary
diseases, is proving to be the exception
rather than the rule in the study of nor-
mal human variation. It is much com-
moner to ½nd a polymorphism that con-
sists of different lengths of sequences of
promoters upstream of genes. To return
to the vasopressin receptor gene, for in-
position, it appears that the repetitive box
in the promoter is highly variable in
length in wild prairie voles. Its length
va de 350 à 550 base pairs in a
typical sample of the rodents. De même,
in a sample of 150 human beings, là
were seventeen different lengths of the
equivalent box next to the same gene. Il
is perhaps too simplistic, and possibly
unethical, to ask if those people with
longer boxes generally form more lasting
pair bonds. But note that divorce rates
show surprisingly high heritability in
studies of twins raised apart.20

Entre-temps, the study of twins shows
that the same upbringing does not nec-

19 Srijan Sen, Randolph M. Nesse, Scott F.
Stoltenberg, Sheng Li, Lillian Gleiberman,
Aravinda Chakravarti, Alan B. Weder, and Mar-
git Burmeister, “A bdnf Coding Variant Is As-
sociated with the neo Personality Inventory
Domain Neuroticism, a Risk Factor for Depres-
sion,” Neuropsychopharmacology 28 (2003):
397–401.

20 Judith Rich Harris, The Nurture Assumption:
Why Children Turn Out the Way They Do (Lon-
don: Bloomsbury, 1998).

essarily produce similar personalities in
two different people, whereas the same
genome often does. A possible explana-
tion of this surprising result is that genes
do not decide personality directly, mais
they do decide how an individual will re-
spond to a particular upbringing. Hard
evidence for this hypothesis is now be-
ginning to accumulate. Perhaps the best
example is the study of childhood mal-
treatment and genotype in a New Zea-
land cohort.

In a study of 442 young men from
Dunedin born in the year 1972–1973,
Terrie Mof½tt and her colleagues found
evidence that an abusive upbringing
does predispose a boy to later antisocial
behavior (including getting into trouble
with the law), but much more strongly if
the boy has a particular genotype: a low-
activity version of the monoamine oxi-
dase A gene on the X chromosome. Dans
the promoter upstream of the gene there
is a 30-base pair phrase repeated three,
three and a half, four, or ½ve times.
Those genes with three or ½ve repeats
are much less active than those with
three and a half or four repeats. About
one-third of men have low-activity ver-
sions of the gene (femmes, having two X
chromosomes, present a more compli-
cated picture). The low-activity allele
itself does not appear to cause antisocial
behavior, nor does childhood maltreat-
ment alone, but together they have a
marked effect.21

The correlation between parental
abuse and antisocial behavior in the
Dunedin study cannot be assumed to be
causal. It may be that another undiscov-
ered gene causes both the abuse and the
antisocial behavior in combination with

21 Avshalom Caspi, Joseph McClay, Terrie E.
Mof½tt, Jonathan Mill, Judy Martin, Ian W.
Craig, Alan Taylor, and Richie Poulton, “Role of
Genotype in the Cycle of Violence in Maltreat-
ed Children,” Science 297 (2002): 851–854.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

The dna
behind
human
nature

the low-activity mao-a gene. A long his-
tory of fallacious assumption teaches
us to be cautious before presuming that
parents cause effects in children by
their actions rather than by passing on
genes.22

This precaution, cependant, does not
apply to a similar result in another gene.
Again using the Dunedin cohort, Mof½tt
found that a functional polymorphism
in the promoter region of the serotonin
transporter (5-htt) gene affects the
way people react to stressful life events.
Stressful life events are less likely than
abusive treatment to be even indirectly
caused by genes. People with one or two
copies of the short allele of the 5-htt
promoter showed more symptoms of
depression following at least three
stressful life events than people with
two copies of the long allele.23

Considering that genes influence de-
pression by altering people’s ability to
cope with life events, can anybody doubt
that the genes that influence personality
and intelligence work this way–that
they are genes for responding differen-
tially to experience? A person with high
intelligence is a person whose genes en-
able him to react ef½ciently to the expe-
rience of learning. A person with an ath-
letic talent is one whose genes enable
her to respond easily to practice and
entraînement.

Notice, in passing, how important the
length of, rather than the sequence of, un
promoter often proves to be. This is a
general principle that is emerging from
many studies of gene function. The di-

22 Harris, The Nurture Assumption.

23 Avshalom Caspi, Karen Sugden, Terrie E.
Mof½tt, Alan Taylor, Ian W. Craig, HonaLee
Harrington, Joseph McClay, Jonathan Mill,
Judy Martin, Antony Braithwaite, and Richie
Poulton, “Influence of Life Stress on Depres-
sion: Moderation by a Polymorphism in the 5-
htt Gene,” Science 301 (2003): 291–293.

versity in the human population is start-
ing to be explained at least as much by
variations in the number of repeats of a
genetic phrase in the regulatory region
of the gene as by single-nucleotide poly-
morphisms. The phrase may be two or
three letters long (as in the case of the
vasopressin receptor), twenty-two let-
ters long (5-htt serotonin transporter
gene), thirty letters long (monoamine
oxidase gene), or seventy-½ve letters
long (aspm gene). Varying the number
of repeats of a phrase has a much subtler
effect on gene function than does chang-
ing a single nucleotide in a codon, lequel
tends to shut the gene down. It seems to
be the principal way in which natural
selection alters the intensity, and per-
haps the pattern, of gene expression.
Nor is this phenomenon con½ned to
the regulatory regions of the genome.
At least six neurological diseases are
now known to be caused by excessively
long polyglutamine runs–most notably
Huntington’s disease, whose severity
depends on the number of repeats of a
three-letter phrase (cag) in the gene for
the huntingtin protein.

Precisely how does a gene open the or-

ganism to experience? A nice example
of how, paradoxically, the capacity for
nurture can be genetically programmed
comes from features that show critical,
or sensitive, periods in development.
There are many features of animal and
human behavior that are sensitive to en-
vironmental influences only during a
limited period in youth. Language learn-
ing is one. Filial imprinting in birds is
another. The best-studied case, cependant,
is that of ocular dominance, or the sort-
ing of cells in layer 4c of the visual cortex
into those that take their signal from the
right eye and those that take it from the
gauche. Ocular dominance emerges in re-
sponse to experience soon after a mam-

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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

Matt Ridley
sur
human
nature

mal’s eyes ½rst open and is thereafter
irreversible. Experiments have revealed
that the gene for a protein called gad65
must be switched on for the sorting to
occur, and that another protein, bdnf,
brings the sorting to an end. Genetically
modi½ed mice with no gad65 gene nev-
er enter the critical period; those with
overactive bdnf genes close down the
critical period prematurely.24

Both genes regulate the activity of
gaba, a neurotransmitter that has also
been shown to be vital to ½lial imprint-
ing in chicks. This ½nding hints at a gen-
eral genetic mechanism, based in gaba-
ergic neurons, for opening the organ-
ism’s brain to calibration by experience
during a narrow window in infancy. If
individuals’ critical periods differ in
length or openness, this may be because
they differ in sequences of promoters
attached to gaba-related genes. These
variations, à son tour, would produce a dif-
ferent pattern of learning in different in-
dividuals. Ainsi, even the acquired dif-
ferences between people in skills and
interests might be partly caused by se-
quence differences at promoter sites.
A good tennis player is the product of
much practice, but the ability to bene½t
from practice could prove to be innate.
Nurture, in that sense, is a form of
nature.

There was an old joke, ½rst told by Jane
Gitschier, that we would one day be able
to ½nd out where on the Y chromosome
lie the male tendencies to flip between

24 Z. Josh Huang, Alfredo Kirkwood, Tom-
maso Pizzorusso, Vittorio Porciatti, Bernardo
Morales, Mark F. Bear, Lamberto Maffei, et
Susumu Tonegawa, “bdnf Regulates the Matu-
ration of Inhibition and the Critical Period of
Plasticity in Mouse Visual Cortex,” Cell 98
(1999): 739–755; Michela Fagiolini and Takao
K. Hensch, “Inhibitory Threshold for Critical-
Period Activation in Primary Visual Cortex,»
Nature 404 (2000): 183–186.

channels on the television, to sit on the
john reading, and to be incapable of ex-
pressing affection over the telephone
(the me-2 gene). It was a joke that ex-
posed not only the absurdity of men, mais
also the absurdity of speci½c genes for
speci½c behaviors–the old Daltonian,
particulate, ‘blueprint’ model of a ge-
nome, in which one gene corresponds to
one attribute of behavior. Genes are not,
bien sûr, like that. As Pat Bateson has
argued, they act more like recipes than
blueprints. Attributes of an organism no
more map directly to single genes than
pieces of a cake map directly to lines in a
recipe: they are the product of a transac-
tion between many genes and the envi-
ronment in which they ½nd themselves.
Néanmoins, it was widely assumed
in the heyday of the blank slate in the
1950s–1970s that speci½cally behavioral
mutations would not be found, et ça
therefore behavior would remain a
p2c2e (a process too complicated to
explain), at least in genetic terms. Le
studies of twins raised apart, et le
discoveries of dna sequence changes
that cause predictable changes in behav-
ior, even in ‘higher’ behavior, demolish
this assumption. The magnitude of that
paradigm shift has yet to dawn on many
social and even biological scientists.
A different hypothesis is needed if
we are to reconcile the evident fact that
there is an innate human nature with the
equally evident feeling that experience
molds individual lives. That hypothesis,
I suggest, must hold that human nature
is speci½ed in species-typical dna se-
quences, that many of those sequences
determine the expression rather than the
protein product of genes, and that the
expression of many of these sequences is
actually ‘designed’ (by natural selection)
to be affected or calibrated by expected
kinds of environmental experience.

Dædalus Fall 2004

D
o
w
n
o
un
d
e
d

F
r
o
m
h

t
t

:
/
/

d
je
r
e
c
t
.

je
t
.

e
d
toi
d
un
e
d
un
r
t
je
c
e
–
p
d

F
/

1
3
3
4
8
9
1
8
2
8
8
7
8
0
0
1
1
5
2
6
0
4
2
3
6
5
5
4
6
p
d

b
oui
g
toi
e
s
t

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