The Extended Phenotype (EP) mainly describes
instincts which allow a creature to manipulate its environment
in a way which benefits its genes. I think, though, that some of the
most interesting and wide-ranging applications of this perspective
lie in the realm of human cultural innovation.
At present, humans are Earth's most important staging-points
for the transmission and organisation of cultural information; without
humans, the vast cultural edifice we've built would not exist - not,
that is, until machines get a lot smarter, more dynamic and independent.
Some of the intelligent decisionmaking and information-processing functions
of a human are contained within its head. Some lie in other human's
brains (parents, bosses and role models for example), and others within
computers. Memory is contained in the same places; because it's a static,
non-dynamic form of information*, though, it can also be contained in
forms which aren't actively involved in information-processing - for
example in libraries and in files on computers.
Traditionally, we have thought of human nervous
systems as being wholly contained within our body; but as Marshall
McLuhan perceptively observed, we partially externalized our nervous
system by inventing technology which transmits electricity**.
So, for example, I can think of my nervous system as being in two stages
when I think up this crap and post it on the Internet to be read. First
of all, my brain sends signals to my fingers which make them type out
what I want to say; next, the message travels down phone lines to your
computer. The point here is that (to simplify a little in order to keep
my head from exploding) my brain is controlling the transmission and
behaviour of information; it is using various levers and channels, some
of which involve my body, some of which don't.
* This is only one possible definition
of memory; neurologists tend to think of it as something more dynamic.
But I'll use the 'static' definition for the purposes of exposition.
** Strange as it may sound,
I got chills down my spine when I read that. It was that observation
that got me started on this article.
Well, that's one extension of my phenotype. Are there
others? Very much so - the list is almost endless. Before going on,
though, I think I should explain more clearly what I mean by 'genotype',
'phenotype', and 'extended phenotype'.
The cells in your body (or at least the surprisingly small
number which are 'yours' rather than those of 'foreign' organisms) each
contain your full 'DNA library'. This library is known as the genome
((the Human Genome Project is concerned with describing all of the genes
which make up the genome; of course, describing them and knowing what
they do are very different things. This field will grow and grow, for
the simple reason that the workings of animals and plants are so dumbfoundingly
complex.)). In this article, I will take the word 'genome' to be synonymous
with 'genotype' (professional biologists, wince away).
When the genotype is put into practice (via activators
such as RNA and enzymes), what emerges
is the phenotype - your body. The phenotype is the most integrated expression
of your genes; natural selection has designed it to carry out the primary
function of terrestrial evolution,
which is the perpetuation of specific gene sequences*. The various genes
in our genome are deeply symbiotic with each other; what is good for
one is usually good for the others.
* If you think I'm saying that everything
you do is aimed at promulgating your genes, stop. I freely acknowledge
that not everything you think or do is hard-wired by your genes. Evolution
has found that making a creature which can think for itself can be
an excellent - though indirect and occasionally ineffective - way
of spreading genes.
The extended phenotype performs exactly the same function
as the phenotype. The only difference is that the tools it uses are
not the heart, brain cells, fingers and so on; they are outside the
body. The EP is a subsection of the environment; it is that part of
the environment whose form is influenced by the phenotype in ways
which comply with the phenotype's goals. The less intelligent an animal
is, the more obvious the correspondence between the goals of the phenotype
(the body) and those of the animal's genes. Thus, the beaver's dam,
and the manipulated brains of the birds which host (European) cuckoo
chicks, are quite precisely programmed by beaver and cuckoo genes
to take forms which aid in the propagation of those genes. With humans,
however, the brain's spectacularly rich analysis of the environment
make the relationship between genotype and EP far more roundabout;
these analytic powers allow the EP to take a much greater variety
of forms than the EP of a cuckoo or beaver.
|
SPECIFIC
EPs
For just about every human sense, intellectual
interest or behaviour, there is a corresponding EP.
|
PART
OF PHENOTYPE:
|
• |
SOME
OF THE CORRESPONDING PARTS OF EXTENDED PHENOTYPE: |
Nervous system
|
• |
Electricity,
which is increasing the reach of growing numbers of our
senses, in a growing variety of ways. |
|
Vision
|
• |
Glasses; telescopes;
microscopes; infrared goggles. |
|
Legs
|
• |
Cars. |
|
Arms
|
• |
Cranes. |
|
Hearing
|
• |
Radio; hearing aids. |
|
Memory
|
• |
Books. |
|
Thought &
ideas
|
• |
Memes |
|
Sexual desires
|
• |
Pornography;
that subset of another person's sexual behaviour which
is influenced by your behaviour. |
|
Fist
|
• |
Nuclear bomb. |
|
Food
|
• |
Cooking utensils;
fast food; cookbooks. |
As you can see, all I'm
doing here is describing some of the tools which have supplemented
what we can do by using just our bodies. That's what the extended
phenotype is: tool use (or to be more precise, use of tools
other than those of which the body consists). So why not just
call it tool use? Because it emphasizes the continuity between
bodily tools, external tools, and the tools known as people.
(Other people and lifeforms are not always part of the EP,
because obviously one person can't constantly influence or
control everyone else's behaviour. It is only those tools
which are being used at a given time which may be seen as
influenced by your brain in the same way as your own body's
behaviour is).
You can see what the trend
is; the more technologically rich the EP becomes, the more
it supplements - and sometimes obviates the need for - specific
parts of the phenotype. The EP replaces the phenotype with
varying degrees of facility. For example
• a
good calculator far outstrips the number-crunching performances
of even the most brilliant and/or well-trained maths prodigies;
• books
contain far more accurate memory than we are capable of;
• pornography
is (at present) a fairly anaemic substitute for good sex
(the behaviour of potential and actual sexual partners is
itself an EP - one which is rather difficult for most of
us to manage as we would like).
And of course, some of our EP now
gives our mind tools which the rest of our phenotype (our
bodies) just doesn't have. Among these are:
• the ability to
see microorganisms, distant stars, body heat and ultraviolet
markings on flowers;
• the
ability to hear the subsonic rumblings of elephants and
the ultrasonic squeakings of bats;
•
the ability to visit previously inaccessible environments,
such as the great pressures of the ocean depths or the near-vacuum
of space.
Why do some technologically-based EPs work well and others
poorly? There are many reasons.
•
First of all, the manipulability
of other people's behaviour is not increasing at the same
rate as the manipulability of non-human items. This is for
the very good reason that people often object to being psychologically
influenced or technologically altered to suit another person's
goals. If you want to manipulate other people's behaviour,
there are several ways to go: surreptitious, consensual, and
by force.
•
Surreptitious:
where someone is being manipulated by you without knowing,
without having even been given a deceptive account of how
you will use them, and to their own detriment. For example,
I, who have poisonous attitudes to synchronized swimmers,
realize that you've been coaching a team of them, and covertly
stitch an explosive device into your swimming trunks. The
bomb is rigged to detonate after a certain number of dumb
stylized movements. This isn't just my lame attempt at humour;
I genuinely can't think of many examples of surreptitious
trickery in humans. Email
any suggestions.
•
Coercive:
as with police states, bullies, domineering parents, etc.
You've seen plenty of this yourself. Today, coercion does
not usually alter behaviour by directly manipulating the brain.
But technologies to do this are increasingly available: truth
serums, disabling operations such as lobotomies, pacifying
narcotics, and so on. Will more of these be used in the future?
That's not a question I can answer.
• Consensual:
this is today's EP of choice. Picture a woman who works
as an advertising copywriter. From her job, she wants the
following things; success, money, fame, intellectual stimulation.
By working for an advertising company, whose overriding
goal is financial success, she narrows the techniques which
she can use to create a compliant EP; if her job is interesting
and lucrative, that is incidental to the company. What she
must do is make her target audience behave in certain ways,
i.e. buy the product. This process is about achieving the
goals of the company and the copywriter; but it must seem
to be about achieving the goals of the target audience.
I am reminded of the sci-fi film 'They Live', in
which special glasses enable the hero (played by wrestler
Rowdy Roddy Piper) to see that many of the people around
him are in fact horrendous-looking aliens, who have covered
buildings with messages that he's only now able to see (although
they've influenced his subconscious all along): "don't
think", "worship television", "serve
your alien masters", and the like. I'd enjoy having
x-ray specs that enabled me to effortlessly perceive ads
as deceptive EPs; instead of "Our sole aim in life
is to make your shopping experience truly joyous",
I'd see "Let's pretend we're friends with our customers,
and make our workers smile inanely at total strangers to
stop them giving their money to anyone else". Missin'
ya already!
In advertising (which includes much
of politics and interpersonal relations), the goals of the
two parties may indeed be compatible; one needs food, the
other sells it; or a politician needs a job, and he'll help
out whoever elects him. But they aren't always compatible;
and the "seller" 's goals
are only rarely explicit. More often than not, the 'consensus'
approach beloved of our society is an ill-informed one. Everywhere,
Olga is trying to impregnate Patti with memes which favour
Olga. If Patti thinks the memes will help Patti, so much the
better. And if they actually do help Patti, well, the
best lies incorporate some truth. Useless skin lotions, charismatic-and-corrupt
politicians, and 'ab shapers' are not created for your benefit;
the Olgas of this world are trying to make you part of their
extended phenotypes. Or perhaps not; Olga may mean well, and
may herself be part of the EP of the company she works for.
Sometimes, everyone means well. This can just be because they're
just nice people who want to help (from whatever dark biological
motive!); but it can also be because they're part of the phenotype
of moneygathering blocs. Commerce is a Darwinian
process; companies which are successful will have 'assembled'
a favourable phenotype, which can include things such as people
who are obsessed with professionalism and/or service. I've
met one or two of these programmed businesspersons - they
scare me.
I'd say that, other things being equal,
'consensual' behaviour manipulation isn't as bad as coercion;
but it's probably just as bad as surreptitious
EP manipulation. I'm not claiming that such behaviour
originates from the capitalist and democratic systems; it
will take root in any human social system. It will probably
be the most durable and widespread of the three ways of behaviour
manipulation, for two reasons: firstly, there will always
be some failures to recognize misrepresentations for what
they are; secondly, some situations - i.e. sexual encounters,
social work and friendships - don't (believe it or not) always
require constant misrepresentation in order for the goals
of both parties to be achieved.
•
Here's another reason why some
of the technologically-based EPs work well and others poorly:
some abilities are more difficult to integrate into the EP
than others are. These abilities often involve cognitive function,
which in humans is by far the most complex way of organizing
and transmitting information.
•
we are currently unable to use our technology
to create well-rounded sexual experiences, because sex involves
all our senses; the feel is particularly difficult to recreate,
for two reasons: firstly, that top scientists are probably
uninterested in creating what might seem to be essentially
be a porn machine (it's a lucrative opportunity! And anyway,
other sensual experiences could also be mimicked); secondly,
because creating a warm, three-dimensional, behaviourally-correct
(in the physical sense, never mind intellectually) 'body'
is very tough. Bodies which move, cling, and thrust in the
right sections and at the right time are very complex feats
of engineering; so are other devices such as bodysuits which
mimic these sensations. Our genes (and all right-thinking
people) must be breathing a sigh of relief, for the moment
- but their future is still imperilled.
Speaking of bodysuits, I was interested
to hear of a recently-developed tactile computer mouse, which
mimics sensations such as running water, warmth etc.* We are
heading in the direction of an EP which we can manipulate
exactly as we please from the comfort of our beds or chairs
(though I don't know whether we will continue in this direction,
or veer off somewhere else).
* I'm
guessing at the details of what it mimics.
•
as for the externalization of intellect,
progress is incredibly fast - but the mountain to be climbed
is enormous. Computers are still only good as go-betweens
(despite talking
computer programs such as Eliza, which you can have stilted
or highly tangential conversations with - the older conversation
programs are stilted and repetitive, the newer ones are weird).
The difficulties of replicating the functions of Earth's most
brilliant machine - the human being - are obvious. Humans
move around a variety of terrains with great ease, and use
their brains and five senses to examine the outside world
and come up with massively flexible behaviours. I don't think
this preeminence will last, though - elsewhere on this site,
I discuss why I think computers
will eventually become vastly more intelligent than us,
and what place humanity will have
in their world. Computer intelligence is the part of our
EP that will, I think, have the most dramatic effects; not
just insofar as they relate to us, but also in their own right
- as unimaginably intelligent entities with thoughts and goals
(and perhaps emotions, consciousness, and aesthetic sensibilities)
far beyond those which humans experience. Once this happens
(I think it's a question of when rather than if), part of
our extended phenotype will, to all intents and purposes,
have become a wholly new kind of phenotype. Our tools
will become the greatest tool-users our planet has spawned.
|
What is happening is that patterns of information (DNA) are seeking to
perpetuate and (sometimes to) spread themselves; to avoid being obliterated,
they need to harness energy and materials, and to favour the production
of other types of information (contained in brains, individual cells,
and bodies) which will allow them to survive. We can therefore call DNA
'manipulative information'. Some of the thoughts and instincts in our
minds can also be called manipulative information (MI) - even though there
are some major differences, notably in terms of goals (more on this later
in the article). Even our hearts and blood vessels could be thought of
as MI; but I will not use the term except for the more powerful information-manipulating
centres such as DNA and the brain, each of which directs a vast range
of activities in pursuit of its goals.
Why has one form of MI (DNA) given rise to another (the mind)? Particular
forms of MI arise in particular contexts: in a carbon-based twisted
ladder (DNA), or as alliances of cells in brains and nervous systems.
Each context has its advantages and limitations - if all Bill Gates
had to work with was rubber bands, he would have a hard time making
functional computer programs (cue crap remarks: 'he doesn't make functional
computer programs, ha ha!', etc.). MI patterns are not good at doing
everything themselves; most twisted ladders cannot fight all comers
or climb over obstacles. For this reason, they develop phenotypes which
serve their goals - for instance, cellular membranes to sustain a favourable
chemical environment, or things (such as muscles) which contract to
facilitate movement. Using other MIs (for example DNA in unrelated bodies,
as with the European cuckoo chick) has severe limitations - those MIs
have their own goals, and will usually not willingly be manipulated
by foreign MI.
Sometimes - and here's the bit that interests me - MI develops a phenotype
which is itself MI. In the two (Earthly) cases to date, this has happened
when when changes in the MI's form allowed its sphere of influence to
undergo a dramatic expansion. With DNA, this expansion took place around
the Cambrian 'Explosion' of 544 million years
ago, when multicellularity took off. With the brain, it is taking place
now, during the 'cultural explosion' which is following hot on the heels
of the human brain's recent growth. The result of multicellularity was
the new MI contained in brains; the result of the cultural explosion is
the nascent MI in computers.
Creating a 'daughter' MI can give the 'parent' certain advantages.
It circumvents the problem of manipulating an MI (such as another animal's
brain or DNA) which has goals of its own, because the new MI's
goals are designed (at least until it learns how to revamp its own design)
to serve those of the old one (i.e. one of our brain's goals is eating,
because that allows our genes' "survival vehicle" - us - to
survive in anticipation of spreading those genes). A 'subordinate MI'
is only useful if it is capable of processing and organising information
in ways which its parent is incapable of. Thus, genes gave rise to humans
which can think; and humans have given rise to computers which process,
organise and present some types information better than we can. And
it's only fifty (or so) years since computers came into being. Compare
this with the brain, which has existed for at least 540 million years
- and has certainly undergone some very impressive alterations during
that time. Computer intelligence has only just begun what will be a
stunning, unprecedented (on Earth) growth in mental abilities.
One of DNA's children was the brain. And one of the brain's
children is the computer. How similar are these two 'parent-child inter-MI'
relationships?
• When DNA gave rise to another MI, the new MI could
do whatever its design permitted it to, but only so long as the sum
of the brain's activities allowed its creator to survive. That is an
important limitation. If, for example, a man is born who decides to
dedicate himself solely to thought and not to sex and family, his genes
will die, and with them the genetic code that (given the right environment)
could create such a man. If there were types of abstract thought that,
when highly developed, guzzled enough energy to damage gene survival
prospects, then those types of thought would not prosper. And so on.
At present, there is no doubt that machine MI (such as it is) is
subservient to the MI in human brains. Thus far, the satisfaction
of human goals has always been the reason for improvements in computer
information-processing. But those goals are not the same as DNA's.
Here are some human goals: love, sex, children, strong relationships
with friends, popularity, power, and the satisfaction of intellectual
curiosity. Much of the best work on Artificial Intelligence (AI)
is being done because of intellectual curiosity; this means that there
is not always as much pressure on computers to serve other human interests
as there is on humans to (indirectly) serve the interests of DNA.
• Another
important difference is that, unlike DNA, humans have foresight. This
means that if the intermediate steps in the move to a given form of
computer cognitive ability do not produce much of use, the quest can
still continue. Unlike humans, DNA would never reach point 'D' if all
possible points 'B' and 'C' were not profitable. But for us, costly
Bs and Cs are perfectly acceptable if we think they'll lead us to point
D. I'll use a very common analogy to clarify this. Imagine that DNA
and humans are both mountain climbers. They're on the side of a mountain
which is surrounded by other peaks, some of which are higher. DNA is
blind, so it can't see the higher peaks. It just trudges upwards; when
it reaches the top of the mountain it's on, it can go no further. Humans,
on the other hand, can see some (but not all) of the higher peaks; they
can make short-term sacrifices (i.e. climbing down into the valleys
and losing altitude) so as to attain their long-term goal of climbing
the highest mountains around.
• Yet another
difference is that computers and their programs can be (even if they
usually aren't) designed from scratch; rather than adding a branch to
what came before, a wholly new tree can be built, using only the most
useful bits of its predecessors.
I find it interesting that each new MI has been more flexible and brilliant
than the last (assuming as I do that computer intelligence will be vastly
greater than ours), and that the gap between later generations (i.e. the
brain and the computer) is shorter. Is this a coincidence, or something
intrinsic to younger generations of MI? Are there generations of MI which
haven't appeared yet?
The capacity to manipulate information is accelerating into something
quite spectacular - something which, in large part, we probably cannot
foresee. To be conservative about it, all an MI needs of its subordinate
MI is that it can perform different functions (or the same functions,
but better) to those carried out by the parent. But that's a worst-case
scenario, and only applies to MIs (such as DNA) that lack foresight
and a strong ability to think up and build tools; MIs with foresight
won't just stumble on vastly better MIs, they will design
them if they are willing and able to do so.
We are not the result of foresight; all of our abilities had to be
compatible with previously-evolved abilities; and we are still subservient
to our DNA in many ways. If you remove these constraints, what becomes
possible is incredible: entities which get smarter and smarter with
stunning (and perhaps accelerating)
speed.
|
