Without Miracles: Universal Selection Theory and the Second Darwinian Revolution

I The Need for Selection 1 We see beautiful adaptations everywhere and in every part of the organic world . - Charle...

Author: Gary Cziko

105 downloads 1177 Views 74MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

I The Need for Selection

1

We see beautiful adaptations everywhere and in every part of the organic world . - Charles Darwin1 Are you awed by the exquisite fit between organism and environment , and find in this fit a puzzle needing explanation ? Do you marvel at the achievements of mod ern science, at the fit between scientific theories and the aspects of the world they purport to describe ? Is this a puzzling achievement ? Do you feel the need for an explanation as to how it could have come about ? - Donald T . Campbell2

Much about the world we live in captures our attention der . Majestic

snow -crowned

the surrounding

demand our gaze as they rise above

plain . Raging storms hurl rain , wind , and jagged spears of

fire . The normally the continents

mountains

unyielding

continue

ground

on which we walk shakes violently

march

dictability

and waning

of the seasons and tides provide

of the moon , and the

welcome

for

understanding

of these and other physical

provided

and religious

a naturalistic

events as the mechanical

conse -

of physics and chemistry . The sun is no longer driven

across the sky in a chariot , and serpents no longer

continue

and pre -

gods , goddesses , and a host of spirits to

suc ~ happenings , science has now

quences of principles

Myths

rhythm

to our lives and activities .

Whereas our ancestors invoked account

as

their slow drift over the earth 's molten core . The ris -

ing and setting of the sun , the waxing orderly

and elicits our won -

interpretations

for such natural

to exist , but few expect miraculous

consume

the moon .

phenomena

certainly

explanations

to be offered in

our schools and universities . But there is another Darwin

mentions

side to the world

in the epigraph

at . This living world

we inhabit , the organic

side that

above . And here is even more to marvel

includes both microscopic

bacteria and giant redwood

4

Without Miracles

trees, thewildly colorful birdsandinsects oftherainforests, thelarge mammalsoftheAfrican savannah, thefishes andbizarre denizens oftheoceans depths,aswellas thecreature whowrotethesewordsandtheonenow

reading them. These andother living organisms evoke ourawebecause they

appearsomarvelously designed tofittheparticular environment inwhich

theyexist. Thisfitoforganism toenvironment involves notonlythestructureandbehavior oforganisms, butoftenalsotheinanimate products of theseorganisms. Whereas biology nowrecognizes thatliving organisms

mustobserve thematerialistic lawsofthephysical world,allformsoflife appearsomehowto haveobtainedknowledge of theirenvironment that

transcends theblind, uncaring forces ofphysics, andtheignorant, inevi-

tableprinciplesof chemistry.

Thecommon web-spinning spider provides manyexamples ofwonderful instances offit.Itseightlegs,moving ina coordinated fashion, allowit to travel rapidly across theground, scale vertical surfaces, andevengowith easeupside downacross ourbedroom ceilings. Itcandropquickly andsafelyfromhighlocations byriding down onitsinternally produced silkdragline,excreted byspecial organs called spinnerets, andifitshould change its

mind, itcanhoistitself upagain byreeling inthelineithasjustlaid.Being apredator, thespider must catch insects tosurvive. Thisisaccomplished by

constructing awell-designed webwhose strands ensnare andquickly entangleanyedible insect unfortunate enough totouch them. When thisoccurs, thespider rushes overandinjects a deadly venom intothehelpless victim. Thenit mustdecide whether toconsume themealnow,drawing outthe nutritious bodyfluids andperhaps spraying itwithdigesting juices tomake iteven more palatable tothespiders sucking stomach, orsaveitforlaterby wrapping the lifeless preywithsilk.Spidersalsoengagein elaborate courtship behaviors, withthecustoms ofsome species requiring themaleto presentthe femalewitha gift-wrapped flyso that he can matewithher

without being eaten himself. Some females protect theireggs untilthespiderlings emerge andcasttheirdraglines intothewindtogoballooning off

to new locations to establishresidence.

Thespiderthusprovides us withmanyinstances of fita remarkable meshing of its body,behavior, andproductswithits environment. Its appendages andtheirmovements fitthedemands oflocomotion. Thanksto

Puzzlesof Fit

5

a specialoil covering its body, it can move with easeover its sticky web. Its web is designedto fit the spacein which it is built and to enmeshunsuspecting insects. Its venom is chemically suited to kill its prey. And we have not evenmentioned the intricately designedre,spiratory, circulatory, digestive, nervous, and reproductivesystemsthat function together to give life to the spider and permit its reproduction. Theseand other instancesof organic fit are quite unlike anything in the inanimate world . They attract our attention and elicit our wonder because they seemto have a purpose. Legs are for walking, mouths are for eating, and webs are for snaring prey. In biology, the word adaptation (from the Latin adaptare, " to fit " ) is used to describesuch instancesof fit between organism and environment (with the environment understood to include other organisms as well , as in the fit between a male 's and a female's sexual organs ).

When one looks at the fit and function of theseadaptations, it doesnot seem unreasonable to consider them as forms of knowledge . In an impor -

tant sense,the spider's legsand feet know somethingabout the terrain that the spider considershome. Its venom has a kind of knowledge of the physiology of its prey. And the male spider knows what it must do to find a female and enticeit to mate. Of course, in the caseof spiders, we are referring to a type of built -in biological knowledge, not the conscious, reflective knowledgethat we associatewith human thought. But insofar asbiological adaptationsare functional and aid in an organism's survival and reproduction in what is usually an uncaring and competitive world , theseinstances of fit clearly reflect knowledge of the world that the organism inhabits. But the fit of body parts, behavior, and instinctive products (such as the spider's web) are not the only adaptationsin the organic world . More is to be found by looking deeperusing the instruments of modern science. The cells themselvesand the precisely designedorganellesthey contain reveal remarkable fit in form and function to their surroundings within the organ -

ism. Mitochondria produce energy, ribosomes create new proteins, and DNA of the nucleus is an archive of all the genetic information the organ -

ism requiresto function and reproduce. Also at the level of cells is the exquisite fit achievedby the mammalian immune

system . Ever since it was discovered

late in the nineteenth

century

that animals are able to produce antibodies that provide protection from

6

Without Miracles

disease , the mechanism of this processhas been the subject of extensive research, leading to discoveriesthat have earned 12 Nobel prizes since 1901 .

An intriguing puzzle motivated this research. Since many different types of antibodies were found , each effective against a particular antigen (virus ,

bacterium, chemical toxin ), it was thought that information for the construction of each antibody was carried in the individual '8 genes. However, Austrian-born Karl Landsteiner discoveredearly in the twentieth century that the immune systemis capableof producing antibodiesthat are effective against artificially synthesizedantigens, that is, foreign substanceswith which the animal had no contact either during its own lifetime or during its evolutionary past. The creative ability of the immune systemto produce antibodies to novel antigens is even more amazing when it is realized just

how well an antibody fits its target antigen, a degreeof fit comparableto that a key must have to open a lock. One also must considermore abstract types of fit , particularly when one reflects on the knowledge that humans require to survive in the world 's many different environments . Indeed , the human species lives in a wider range of habitats than any other vertebrate on earth , from the steaming jun -

gles of the tropics to the numbing Arctic tundra, from low-lying coastal areasto the Andean altiplano 4000 meters above sealevel, from sparsely settled wildernessto teeming cities. We have so far made only short visits away from the surfaceof our home planet, but it may not be too long before we establish permanent

settlements in outer space and on the ocean floor .

The desert nomads of the Saharawear long, loose-fitting garments to protect them from the sun, dry wind , and blowing sand. The Eskimosof the frozen Arctic wear warm garmentsmadefrom the skins of the animalsthey hunt. Peopleliving in remote tropical rain forests often wear little or no clothing at all. Astronauts working in spacewear pressurized, air-conditioned spacesuitswith visors coatedwith a thin transparent film of gold to reflect the intense, damagingrays of an unattenuatedsun. Other aspect~ of how these remarkably diverse people live- what they eat and how thev .

obtain and prepareit , how many spousesand children they have, how their living quarters are made and maintained- provide clear evidenceof fit to features of their particular environment , without which human life would not be possible in its wide range of habitats .

Puzzles of Fit

7

But if human knowledgefits aspectsof the world , it must also be the case that our brains, the seat of our knowledge, are adapted to aspectsof the world . Arguably the most complex object yet discoveredin the universe, the human brain is a rather soggy, irregular sphere, about the size of a large grapefruit. Within its modest volume, however, it has enough " hardware" to make eventhe most sophisticatedsupercomputerappearcrude by comparison, including about 20 billion electrochemicalswitching units, with each unit having between 1000 and 10,000 connectionsto other similar units. The pattern of connections and activation of these neural circuits underlie just about everything we do, from walking, eating, and making love to talking, cooking, and making scientific discoveries.The plasticity of our brain allow~ us to acquirenew skills and knowledge asrequired by our environment, suchaslearning to speakanother languageor learning to ride a bicycle, drive a car, or fly a spaceship. Other instancesof fit can be found in the material products of human scientific and technological knowledge. A good example is the modern passengeraircraft, which has helped to make today's world a much smaller place. Able to accommodatehundredsof travelersin relative (if often somewhat cramped) comfort, today's passengerjet is jammed full of sophisticated technology. Its sleek aluminum skin resists corrosion and combines strength with low weight. The configuration of its wing and tail surfaces provides the lift necessaryfor flight . Its enginesdeliver amazing thrust for their sizeand weight. Sophisticatednavigational and radar systemspermit the jet to avoid both bad weather and other aircraft and keep a steady course for its final destination. Flying 10,000 meters above the ground at 1000 kilometers per hour, passengersare provided with in-flight movies, meals, lavatory facilities, and even telephonesto keep in touch with their earthbound families, friends, and businessassociates . The sophisticatedtechnology representedb)Tthe jet is evenmore impressive when one takes into account the planning and coordination that is required to construct it . Considering the complexity of theseremarkable flying machines, it is hard to imaginehow any singlepersoncould everhave all the expertise necessaryto design one, let alone the time, skill, and strength to build one. But despitethesedifficulties, air travel in the industrialized countries hasbecomealmost as routine astaking the family car on a trip , with accidentsmuch lower on a passenger -distancebasisthan for any

8

Without Miracles

other form of transportation. Clearly, the modern jet airplane is an amazingly fit vehicle for the purpose of transporting passengerssafely, quickly, and comfortably over distancesthat not very long ago would have taken months to cover. And its successin doing so reflects well on the scientific knowledgethat is applied to its design, construction, and operation. As

these

ucts

examples

havior

,

web

. The

precisely

ferent

brain

the

's

sex

,

health

science

jet It

its

,

, gene

could

be

.

that

makes

life

ever

Fit

,

teria

here

we

details function

here

a

,

wa

to

to

has

spurred

technological

is

a

future

lose

of

it

hu

-

do

finding

ever

food

, shelter

the

,

growth

of

like

of

the

the

fit

might

to

have

life

delicate

and

balance

visitors

artifacts

which

-

. The

and

characteristic

ever

extraterrestrial

to

of

dif

.

universal

for

world

for

-

very

achievements

computer

should

variety

the

on

needs

be

that

experience

our

-

,

predatory

. The

insistence

satisfy

personal fit

degree

may

be

y

takes

observe

" any

the . " 3

parts

it

or

system

'

structure

of

is

able

it

present

?

account

to

our

then

determine

been

composed

and

It

achieve

fit

some it

is

a

of

arrangement

part

many

,

when

if

-

life

accompanied

,

as

a

judge

see

it

. What

appear

that

purpose

by

relevant

only

insofar

useful

way

. We

but we

must

this

important

is

in which

formally

know

to

behavior

contributes

created

to

fit

is

into

or

define

seem

and

structure

and

to

we

that

purpose

that

A

environment that

difficult ,

determine

some

.

of

to

look

the

to

its

to

us

prod

body

antibodies

the

permits

entertainment

earth

,

case

use

for

environment

we

the

of

produce

adaptation

stubborn

led

the

possible

pornography

do

system

and

certainly

and

used

a

designed

its

and

such

if

"

's

.

as

in

,

would

existed

in

,

its

their

spider

throughout

Our

has

that

Indeed

intelligence

said

therapy

argued

planet

had

,

and

The

antigens

for

ways

and

to

- producing

" wiring

.

.

demands

able

conditions

things

fit

the

is

disease

enjoyable

technology

for

remarkable

social

companionship

products

dead

of

of

organisms of

system

intricate

more

living

equipped

is

exceedingly

and

airplane

well

and

, and

of instances

structure

variety

, easier

world

immune

knowledge

remarkable

faster

ing

it

environmental

man

by

make

cultural

the

remarkable

mammalian

match

human

,

countless

and

lifestyle

a

indicate

provides

to

it

the

recognize

interacting

suggest

is

it

-

of

its

to

organism

or fit

parts

where

to

-

was

adapted

such

design

cri

function

aspects

as

once

fill

when

the

some

Puzzlesof Fit

9

Suchinstancesof fit also demonstratt:~a degreeof complexity that is highly unlikely to be due to chance. This is certainly the casewhen we observe any living organism, even one as simple as a common bacterium such as Escherichiacoli. The probability that bj.1lionsof different organic molecules would by pure chancejust happento assemblethemselvesto form the complex arrangement required to produc(~ a cell that is able to take in and metabolize nutrients, eliminate wastes~,move about, and reproduce seems (and is) too tiny to consider seriously. :It is the astronomical improbability of such functional arrangementsof many components that makes them puzzling in the first place. Such achievementsof design are instancesof what will be called " adapted comple:x:ity " throughout this book. When considering examplesof adapted complexity in the biological world , such asthe spider's prey-catching behavior, it appearsasif knowledgehad somehow beenobtained by an organism abc)ut someaspectof its environment. But there is something evenmore puzzling about the many instancesof adapted complexity among living org;anismsand their products. In cases such asthe mammalian brain and imrrLunesystem, a continuing processof " fit making" results in the emergenceof further achievementsof adapted complexity. The ability to fashion novel and more impressiveinstancesof fit is itself an instanceof adapted complexity. It is at the sametime a very specialkind of adaptedcomplexity tha't can perhapsbe bestdescribedusing the related but distinct term adaptivecomplexity, with the descriptor adaptive intended to indicate a continuing processof fit making in contrast to adapted, which describesalready achit~ved fit . The goal of this book is to explore (explanationsfor both adaptedcomplexity (already achievedfit ) and adaptive complexity (ability to achieve new fit ) in our world - existing puzzlesof fit and the emergenceof new ones that are just about everywherewe look, and that we have yet to find anywhere elsein the universe. A quite simple and compelling explanation for the puzzles of fit demonstrated by t]tle structure and behavior of living organismswas proposedover 130 yeaJrs ago and is still regardedasthe central unifying principle of biology. No comparable explanation, however, is generallyacceptedfor all other puzzlesof fit of the type mentioned above. The purposeof this book is to presenta casefor just such an explanationand one that works without recourse1:0 miracles.

II The Achievements of Selection

2 The

Fit

of

Biological

Structures

Providence Every and

organised

these -

natural

propagation purposes

William

body , in the provisions

, testifies

a care , on

which

the part

it contains

of the

Creator

for

its sustenation

, expressly

directed

to

.

Paley

lnstrnction All

that

nature

stances the

influence

through

their

of

acquired

duced

these

new

to gain

has been

in the

changes

J ean - Baptiste

individuals

race

a predominant

generations

these

-

has caused

to which

use

new

are

or

for

a long

disuse

individuals

common

individuals

or lose by the influence

exposed

of

an

organ

descending

to the

two

or

from

sexes

of the circum

time " and " consequently

-

" by

part , is conserved

them , provided

or to those

which

that

have

pro -

.

Lamarck2

Selection Slow

though

powers

the

ty and

infinite

another

and

course

of time

-

Charles

of

with

either

and

for

complexity with

the

and

puzzles

may

be , if feeble

, I can see no limit

conditions

"'s power

man

to the amount

of the coadaptations physical

fit

their of

of

structure

adapted

for the

their

by nature

the

scientists

count

of selection

selection

between

do

all organic

of life , which

of selection

can

much

of change

may

by

his

, to the beau -

beings , one with

be effected

in the long

.

Darwin3

Most

sity

process

of artificial

fit

mentioned

complexity to

of our

expend

existence organisms

in

or behavior

planet

. considerable . This to

their

the

of living

chapter

preceding

chapter

organisms

. The

' s life time

forms

and

presents

environments

three .

have

energy types

have striking

led

do -

philosophers

attempting of

to diver

to

explanations

ac -

14

Without Miracles

Fit by Providence The oldest, most intuitively appealing, and still most widely held explanation for the adapted complexity shown by living organisms is that of a knowledgeable designerwho expresslycreated the remarkably fit forms and behaviors we observefor the very purposesthey serve. Thus we find Aristotle (384- 322 B.C.) asking how it is that " the front teeth [of humans] come up with an edge, suited to dividing the food, and the back ones flat and good for grinding it ," and that " the swallow makes her nest and the spider his web, and that plants make leavesfor the sake of the fruit and strike down (and not up) with their roots in order to get their nourishment." 4Aristotle's answer is that nature createdall thesethings for a purpose, and so they consequentlyreflect nature's goals and knowledge. This philosophy, which attributes the adapted complexity of living entities to somehigher sourceof purposeand knowledge, becamea major component of Judeo-Christian thought that eventually replaced the notion of a purposeful Mother Nature with that of an all-knowing, all-powerful, and personal God. It is within this tradition of Christian thinking that we find William Paley, an English archdeacon, theologian, and philosopher who lived from 1743 to 1805. In his Natural Theology, first published in 1802, Paley used the adaptedcomplexity found in the designof plants and animals as powerful argumentsfor the existenceof God. His thesisis quite simple and, on first hearing, quite appealing. Paleyasksus to considerour reaction to finding a watch lying on the ground and being askedto explain its origin. On opening the watch's case, we find an elaborateand complex mechanismconsisting of finely interlocking wheels, cogs, and springs. We also notice that when wound, the watch's mechanismfunctions for a particular purpose, that is, to mark the passageof time accurately. Clearly, it would not occur to us that the watch had always lain at that spot on the ground, or that it had somehow beenconstructed and depositedthere by the blind and random forces of nature in the way that we might explain the presenceof a nearby stone. It is absolutely clear to Paley that " the watch must have had a maker: that there must have existed, at sometime, and at someplace or another, an artificer or artificers who formed it for the purpose which we find it actually to answer: who comprehendedits construction, and designedits use." 5

The Fit of Biological Structures

15

Paleythengoeson to providenumerousexamplesfrom plants, insects , andanimalsthat showevenmoreimpressive achievements of adaptedcomplexity than that of the watch. Theseincludebonesand their joints, muscles, circulatory systems , internal organs, senseorgans, and instinctive behaviors . To seethat Paley's ideaof adaptedcomplexityis consistent with the conceptof fit introducedin the previouschapter , we needonly to look at theheadings listedunderchapter17in thetableof contentsof a lateredition of his book. Therewe find "wings of birds- fins of fish- air and water; earto the air; organsof speech - voiceandrespirationto air; eyeto light; sizeof animalsto externalthings; of the inhabitantsof the earthand seato their elements ; sleepto night." For all of thesepuzzlesof fit therecan be, for Paley , only onepossibleexplanation - a granddesignerin the form of an all-knowing and benevolentGod who fashionedeveryremarkable contrivanceof the living world for the verypurposefor which it is seento, befit. Theargumentfrom design , asPaley'sreasoningis now known, hasprobably beenusedin someform or anotherin everyhumansocietythat ever developed to thepoint of beingableto marvelat theliving world of nature and ponderits origin. All humansocieties , eventhosewithout advanced technology , are ableto build sheltersand fashionclothingfor protection againstthe elements , and producehunting and farming tools to obtain food. No society , however , not eventhemosttechnologically advanced , has developedthe technologyand tools for creatingeventhe simplestliving organismout of nonlivingmatter.It thereforeseems onlyreasonable to concludethat theexistence of life, in all its diversity , mustbethecreationof an intelligenceand powerthat far surpasses that of humankind.This type of explanationis attractivein that it canreducemanymysteries(theexistence of millionsof differentliving plantsandanimals)to just one(their creation by a supernaturalpower). If a major purposeof religionis to makesense out of theworld, it is hardlysurprisingthat mostif not all religionsexplain the existenceof the earth's flora and faunaas beingprovidedby a supernaturalbeing. This canbereferredto asa providentialexplanationfor the fit of living thingsto their environments .6 Theargumentfrom designto divineprovidence , however , losesits appeal when examinedmore closely . If God is responsiblefor the origin of the adaptedcomplexityof all living organisms , what (or who), may we ask, is responsible for the origin of God? A typical answeris that God hasno

16

Without Miracles

origin, that He always existed and will continue to exist for all eternity. This, however, is tantamount to an admissionthat an explanation for the existenceora complex being is not necessary . If this is indeedthe case, the argumentfrom designeffectivelycontradicts itself, for all it doesis passthe buck of the origin of adaptedcomplexity back to a beingwho must be even more complex than any of his creations but for whom no explanation is needed!71f a being as knowledgeableas God (who, by all accountsfar surpassesthe complexity of a watch) requires no explanation, then why does the watch? And why do the earth's living creatures, including our own species? We can only conclude, therefore, that this oldest, most intuitively appealing, and probably still most widely held explanation is seriously flawed. That is not to say, however, that we can confidently discount the idea that living organismsare the result of specialcreation by a supernatural being. Indeed, it could be arguedthat an explanation doesexist for a creative deity, but that we are simply unable to grasp it . Nonetheless, when examinedcritically, the argument from designshould fail to convinceus of the existenceof God in the way that it so compellingly convincedPaleyand other natural theologiansof his and our time. Fit by Instruction

Jean-Baptiste Lamarck lived in France from 1744 to 1829 , a period of in -

tenseinterest in natural history in Europe. Lamarck, who first gainedfame as a botanist for his practical guide to Frenchplant life, went on to develop a new scienceto account for the origin, structure, and interrelationships of all living organisms. He called his new sciencebiologie.8 Lamarck's biological theory had two central tenets. First, although early in his careerLamarck believedthat all specieshad originally comeinto existencein much the sameform ashe saw them during his lifetime, he eventually cameto acceptthe mutability of species , that is, that over the courseof long periods of time, organismscould changeenough to evolve into new species . Second, he saw the complexity of organismsnot as the direct work of God , but rather as a natural

outcome of the " power of life " and the inter -

action of organisms with their environment . We can therefore appreciate

that Lamarck's theory of evolution was primarily a constructiveor creative one, in contrast with the purely providential perspectivesof Aristotle and Paley.9

17

TheFit of Biological Structures How

did

Lamarck

organisms

and

ments

without

served

example

In

, the

to

to

the live

in

the

to

the

neck

being

head

raised

of

organs

that

,

"

by

would

" the

disuse

which

individuals

in

transmitted changes

This

principle it

is

this

, that

reality

throughout

to

so

, on

thereby

peculiar

a long

time

than

all

in

a

and

continually the

hind

remain

arid

to

by the

form

, is known

always and

its

of

its

at

all

members

legs

hindlegs

cave

.

In

to

would

and

the

, with

its

and

completely

other

slowly

de

of

-

and

" use

power

,

to

atrophy

of

their

words

continue

' s principle

and

dis

action

-

are

. " 12 " characters

by

, is

eyes

stimulated

organs

"

its

or

acquired

offspring

in

through

spring

by

an

organism

:

changed

preserved

which

be

known

that

on

to

lifetime

its

the

organization

the

from

of

reproduction

those

is

their

. " 14 Indeed

observer

, the

of grief

the

is

who

, and

have

the is

undergone

is

theory

biology

century .

true

has

was

today

well

that

been

,

so

been

the

,

to

in

few

such

,

organism

'8

" law

in

striking

accepted

. Yet

an

acquired

unable

characters .

during

wrote

has

so

acquired

remembered

he

that

who

of

acquired

that

all

nineteenth in

still

changes

parents

no

inheritance

offspring

receive

of

there

the

Lamarck that

individuals

more

as

convinced

passed

most caused

individuals

theory

new

the

facts

new

was

were

during

the

the

the

life

trees

Lamarck

, outlined

their

the

is nearly

on

use

not

inherited

acquired of

came

for

which

been

course

.

, would and

mammals

standing

be

organs

that

were

-

. 13

Lamarck

lifetime

has the

to

these

its

believed

and

the

or

these

ob

environment

in

the

of

longer

to

was

of

of

for

would

seldom

use

-

, he

. 11

deal

development

the

them

earth

, without

. This

their

, stretch

of

leaves

becoming

lose

great

used

, Lamarck

Everything

and

a

the the

, maintained

meters

of

largest

advantageous

disappear

to

on

eventually

used

those

use

it

a product

where

six

finding

were

which

Second

have

,

of

trees

environ

. First

vegetation

the

, the

-places

giraffe

to

of

their

:

animal

forelegs

complexity to

processes

ground

from

browse

the

a height

response

observe

habit

the that

lizard

related

through

in to

. This

lengthened

light

to

it

in

in edible

legs

. . . . This

Africa

them

ultimately

always

of

to

whereas

perhaps

by

of

resulted

, a

velop

giraffe

reaches

Conversely

by

the

up

so

and

increasing behavior two

change little

the and

invoked

sustenance

neck

, obliging

, has

devoid

its

interior

reach

race

its

form

He

to

, finding

find

explain

their ?l

, it is interesting

of

pasturage

strive

use

to

habits

height

without

of

giraffe

elongate

regard

and

God

seemed

upward

permanently

of

to

organisms

straining

to

correlation

recourse

that

For

propose

the

so

of

nature

organization

much

attested

convince

Britain

apparent

himself

and

Europe

" facts

"

18

WithoutMiracles

Although Darwin convincingly argued in The Origin of Species , published in 1859, that there was more to evolution than the inheritance of acquired characteristics, it was not until the 1880sthat the transmissionof acquiredcharacteristicsfrom parentsto offspring was seriouslychallenged. German embryologist August Weismann(1834- 1914) made an important distinction betweencells of an organism that passgenetic information to the offspring (germ cells) and all the other cells that have no direct role in reproduction (somatic cells). According to Weismann, it was simply not possiblefor changesto the somatic cells (for example, those in the giraffe's neck) to be transmitted to the germ cells so that offspring could inherit the acquired characteristicsof its parents. He set out to demonstratehis point by amputating the tails of severalgenerationsof mice and showing that successivegenerationsfailed to grow shorter tails. In contrast, Austrian biologist Paul Kammerer appeared to have obtained considerable successin producing what was thought to be clear evidenceof the inheritance of acquired characteristics. However, in 1926 it was announcedthat Kammerer's last remaining specimenof Lamarckian inheritance, the so-called nuptial pads of the midwife toad, showedno evidenceof the claimed evolutionary changeand had instead been doctored using injections of India ink . Six weeks later on an Austrian mountain path Kammerer put a bullet through his head. Unfortunately, this personal tragedy did nothing to dissuadeRussianbiologist Trofim Denisovich Lysenko from his doomed attempt to increasethe productivity of Soviet agriculture basedon Lamarckian principles. Stalin's receptivity to and imposition of Lysenko's Lamarckian ideas were to cripple the development of Soviet biology and geneticsuntil the 1960s.15 Weismann's separation of germ from somatic cells remains today in biology an almost universally acceptedbarrier to Lamarckian inheritance of acquired characteristics.16Lamarck's principle of use and disuse also encounters seriousdifficulties. Let us take the blacksmith's arm as an example.1?We know that the blacksmith is required to do much of his work with a heavy hammer, and we therefore expect this activity to make him strong and muscular, at least in the arm he usesto swing his hammer. The same expectation leads countlessfitness-consciousindividuals to subject themselvesto various forms of discomfort in gylllSand health clubs in an attempt to firm up musclesthat would otherwise be flab. This would seemto be a clear exampleof the acquisition of a characteristicfrom interacting with the


19

environment. But we must stop and ask ourselveswhy it is that the increaseduse of musclesmakes them bigger and stronger. Indeed, from our observationsof nonliving objects we should expect just the opposite. Our shoesdo not grow thicker solesthe more we walk in them, nor do they becomethin by the disuseof being left unworn in the closetfor long periods of time. On the contrary, their soleswear thin from extendeduseand maintain their original condition only if not used. Automobile enginesdo not develop more horsepower with increased use, or atrophy and shrink away when left to sit in a garage or museum for extended periods of time. Why then should the' blacksmith's arm not also be worn away by its heavy work load? How can the environment instruct the arm's muscles to grow in size, strength, and endurance in responseto the work they do? And what is it that tells the retired blacksmith's arm to give up its hard-earned strength? In an attempt to provide an explanation for such adaptive changes, Lamarck made reference to a providential-sounding " plan of nature" and the " power of life " by which organismscould somehow bring about necessaryadaptivechangesby their own efforts.Is Lamarck's theory fails not once but twice for its relianceon environmental instruction. First, according to the principle of useand disuse, the environment must in someway transmit to the organism instructions to make the required adaptivechanges. The cold must somehowinstruct the beaver to grow thicker fur. The stretching of the giraffe's neck must instruct it to grow longer (and also strongerto support its increasedsize). And the heavy workload of the blacksmith's arm must instruct its muscles to gain in strengthand endurance. Second, sinceonly a singleliving cell is passedfrom parent to offspring, the changedbody of the parent must somehow transmit instructions to the germ cell (sperm or egg) for these changesto be passedon. Yet biological sciencehas been unable to discover any mechanism that would make either mode of instructive transmissionpossible.19 So today Lamarck's theory of the inheritance of acquired characteristics is rejectedby virtually all biologists. It did, however, enjoy wide popularity throughout most of the nineteenthcentury, and this despitethe total lack of any clear experimental support. It even seemslikely that it would have retainedits popularity much longer if a competing, noninstructionist explanation for the adapted complexity of life forms had not beenproposed in the secondhalf of the nineteenthcentury.

20

Without Miracles

Fit by Selection

Any student preparing for the bachelor of arts degree examination at Cambridge University in 1831 was expected to be thoroughly familiar with the argumentsfor the existenceof God provided by Paley. One such individual who was particularly impressedwith the bishop's examplesof fit was a divinity student and naturalist named Charles Robert Darwin (1809- 1882 ).

Although Darwin remained impresseduntil the end of his days by the puzzlesof fit shown by nature's living things, he eventually came to reject Paley's providential explanation for their origin. And although he was never able totally to reject Lamarck's theory of the inheritance of acquired characteristics(nor did he understand why such an instructionist theory was genetically problematic), he found it incomplete. Consequently, he developedan alternative theory for the evolution and growth of adapted complexity that did not rely on Lamarckian instruction. Darwin 's theory, published in 1859 as On the Origin of Speciesby Means of Natural Selection , was the major scientific achievement of the nineteenth century .

Darwin discoveredan explanation for the emergenceof adapted complexity in nature that required neither a supernatural provider nor an instructiveenvironment. He proposeda thoroughly natural theory to account for organic fit , the basic principles of which are so easyto grasp that once they are, the discoveryitself seemsmuch lessremarkable than the fact that so many great minds

before him had failed

to come up with

it first . As

Darwin 's closefriend and defenderThomas Huxley wrote after learning of the theory of natural selection: "How extremely stupid not to havethought of that ! " 20

The theory of evolution has three primary components. First, Darwin observed that all speciesshowed considerable natural variation in the forms

and behaviors of individual organisms, due primarily to the fact that offspring usually differed, if only slightly, from their parents. Second, he recognizedsuperfecundity, that is, speciestypically producemany more offspring than can be supported by the environment. Thesetwo conditions being the case, Darwin 's great insight was to realizethat the individuals that by their particular variations are better suited to survival and reproduction must leavemore offspring (which eventually come to dominate the population)


21

than those that, for whatever reason, are lessfit to survive and reproduce. This process Darwin

referred to as natural

selection .

The theory of natural selectionseemsstrikingly simple when so stated, yet it was Darwin 's geniusto realizethat such a process, operating cumulatively and given enough time, could account for the evolution of the wondrous diversity and adapted complexity of all living things, including our own species , starting with the simplestliving organismas a common ancestor. For the samereason that the theory becameand remains the cornerstone of biology (it relies on neither instruction nor providence), it has evokedmore controversyand condemnationthan any scientific theory ever proposed. We will not discussherethe details of the theory, or give examplesof evidence supporting it , since many excellent readable accounts are available .21

However, since a selectionistaccount of adapted complexity as first proposed by Darwin is the major theme of this book, a few important points haveto be made before proceeding. First, although considerablecontroversy concerningDarwin 's theory of evolution still exists among nonscientists, the basic principles continue to constitute the core of modern biology. The observation of evolution on a small scaleboth in nature and the laboratory, the patterns of similarity and diversity revealed by the scientific classification of organisms, and fossil records all provide strong support for natural selection.22And although biologists still do not agreeon all the details of how evolution occurs, no seriousrival theorieshave beenproposedto challengethe process. Second, since Darwin 's theory explains evolution as a natural process that does not entail the existenceof a supernatural designer, his emphasis was quite different from Paley's. Paley emphasizedwhat he saw to be the perfect designof the earth's living things; Darwin repeatedlymade a point of describinghow nature's design, although complex and well adaptedto its environment, is not necessarilyoptimal. This is because" ideal arrangement is a lousy argument for evolution, for it mimics the postulated action of an omnipotent creator. Odd arrangementsand funny solutions arethe proof of evolution- paths that a sensibleGod would never tread but that a natural process, constrained by history, follows perforce. No one understood this better than Darwin ." 23It is for this reasonthat Darwin provided so many examplesof uselessorgans and inelegant, although still functional, solutions to the unending challengesof survival and reproduction.

22

Without Miracles

It is crucial to realize that the variation considered by Darwin to be the fuel for natural selection and evolution was not required to be in any way intelligent or foresighted . Confessing ignorance as to the source of variations in nature , he made clear that it was not the presence of varia tions alone that resulted in adapted complexity , but rather the cumulative selection of those relatively few variations that by blind chance gave the organism fortunate enough to possess them even the slightest advantage for survival and reproduction . Thus Darwin concluded chapter 5 of The Origin (" Laws of Variation " ) with the statement that " whatever the cause may be of each slight difference in the offspring from their parents . . . it is the steady accumulation , through natural selection, of such differences , when beneficial to the individual , that gives rise to all the more important modifications of structure , by which the innumerable beings on the face of this earth are enabled to struggle with each other, and the best adapted to survive . " 24 This " steady accumulation " of beneficial differences is an essential part of the theory . Since variations are blindly and ignorantly produced , it is overwhelmingly likely that any given variation will be detrimental to the survival and reproduction of the organism , particularly

since the current

standards of form and behavior for that species have obviously so far proved successful, or the species would no longer exist . But organic evolu tion does not abide by the maxim , " if it ain 't broke , don 't fix it ." Through genetic mutation and the sexual reshuffling of genes, new variations are constantly being tried out . And whereas it is a rather rare event that anyone of them will prove to be advantageous , given enough variations and enough time , an adaptive change will eventually occur, and to this improvement (as measured by reproductive success) more can be added later. These adaptive changes will tend to be gradual in their evolution since more drastic changes will be even less likely to result in better adaptations . This does not mean that a sudden and dramatic mutation cannot be beneficial to a species, but simply that a very large adaptive change (such as the appearance in one generation of a complex , functioning eye in a species that had been com pletely blind ) is astronomically improbable , and therefore does not pro vide a reasonable explanation for the puzzles of fit achieved by evolution . The process of adaptive evolution can therefore be summarized as " cumu lative blind variation and selection .,,25


23

Neither is natural selection guided by purpose or planning. Many of Darwin 's readers misconstrued just such a purposeful, foresighted, and providential view, imagining nature to be guided by a supernatural intelli gencein the production and selectionof variations most usefulto the organism. It was for this reasonthat Darwin later adoptedHerbert Spencer's term " survival of the fittest," sinceasAlfred RusselWallacenoted in a letter writ ten to Darwin in 1866, survival of the fittest " is a plain expressionof fact; Natural Selectionis a metaphorical expressionof it , and to a certain degree indirect and incorrect, since, even personifying Nature, she does not so much selectspecialvariations as exterminatethe most unfavorable ones."26 It may in some ways be more appealing and certainly kinder to think of nature as selectingout and preservingthe most useful variants. However, selectionis in fact achievedby a processof elimination, a processthat may be kept in mind more easily by thinking of it as Darwin 's hammer.27 It should not be overlooked, however, that despite the immediate and continuing successof Darwin 's theory, his view of evolution sufferedfrom a number of seriousdefects. Ignorant (as was the rest of the world ) of the geneticbasisof inheritance as revealedby the work of Austrian monk and pea gardenerGregor Mendel (1822- 1884), Darwin had no explanation for how heredity operated or how heritable variation was produced. In addition, he wrote, " I think that there can be little doubt that usein our domestic animals strengthensand enlargescertain parts, and disuse diminishes them; and that such modifications are inherited." 28Thus like Lamarck, he believedin the inheritance of acquired characteristics. The realization that natural selectionalone was the mechanismunderlying all biological evolution had to wait until Weismann's work in the 1880s, and it was not until after the tragic Kammerer affair in 1926 that the theory of the inheritance of acquired characteristicsbecamealmost universally rejectedin biological circles. Nonetheless, Darwin did make a convincing argument that increases in adapted complexity can arise naturally without instruction, without prior design, and without the miraculous intervention of a supernatural creator or guide. Where Paley's ultimate watchmaker was an all-knowing and benevolent God, Darwin 's watchmaker was totally blind. Where Lamarck had seenheritable adaptations as the direct instructive effectsof environment on organism, Darwin saw them as the results of the selection

24

Without Miracles

of purposelessand accidentally advantageousvariations. Darwin dared to seedesign without a designer and fit without instruction, and the world has yet to appreciatefully the implications of his vision. Much has happenedin biology since Darwin 's time. Many more species have been discoveredand catalogued, and many others have disappeared forever. Much more is known of the lifestyles of organismsranging from protozoa to gorillas. Great strides have been made in unlocking the secrets of the molecular foundations of life responsiblefor the processesof metabolism, reproduction, and heredity. The languageof the geneticcode has been deciphered. Yet despitethese developments, Darwin 's theory of natural selectionremains as the central pillar of modern biology. This is not to say that the theory has remained unchangedsince 1859. Indeed, many biologists today are more likely to refer to " neo-Darwinism " or the " synthetictheory of evolution" than to " Darwinism " or " Darwinian theory." But despitethesenewer terms, which reflect updatings basedon recent advancesin evolutionary theory, population genetics, and molecular biology, the fundamental features of Darwinian theory remain essentially unchanged. An exception is the almost universal rejection today of the inheritance of acquired characteristics , a belief that Darwin himself was

neverable to reject completely. Darwin 's theory has beenchallenged, but as yet it has no serious rivals . Not " neutralism , " 29" mutationisrn

, " 3oor " mole -

cular drive" 31can account for the adaptiveevolution of organisms.32(Other challengesto natural selectionwill be consideredin chapter 16.) Furthermore, since current evolutionary theory rejects all forms of Lamarckian instruction

and considers

natural

selection

to be the sole mechanism

for

adaptive evolutionary change, neo-Darwinian theory can be consideredin this respect to be more Darwinian

than Darwin

was himself .

The theory of evolution by natural selectionremainscentral to the field of modern biology, but it has not met with similar successamong the general population, at leastnot in the United States. In a poll conductedby the Gallup Organization in 1991, 47% of Americans expresseda belief that " God created man pretty much in his present form at one time within

the

last 10,000 years," and only 9% agreedthat " man has developedover millions of years from

less advanced

forms

of life . God had no part in this

process." 33Perhapsevenmore troubling , a surveyof 387 United Stateshigh

TheFit of Biological Structures

25

school biology teachersconducted in 1987 found that about one-third of them held beliefs incompatible with Darwinian evolution, such as that " Adam and Evewere the first human beingsand were createdby God" and that " the Bible's account of creation should be taught in public schoolsas an explanation of origins." 34 It

is

cans

a

,

hard

not

the

major

full

of

in

the

fit

plants

selection

was

You

obtain

a

that

prediction

that

true

the

predictions

;

your

is

The

A

from

you

cuts

;

for

sort

of

the

con

tion

not

man

is

picture

larger

picture

Of

that

course

for

an

,

offer

also

no

of

see

possibilities

immeasurably

only

is

being

for

an

Similarly

its

to

plausible

a

real

( and

weaken

than

the

of

complementary

Only

since

;

only

we

out

think

test

when

of

the

Paley

' s

we

only

. "

Because

,

he

,

evolu

look

at

-

the

the

which

be

fit

respon

.

if

that

have

.

that

such

could

were

But

also

making

for

the

admitting

Darwin

,

If

it

-

fabricated

,

suckers

. 36

produced

is

,

faith

conclusion

a

thinks

we

organisms

us

see

at

who

spiritual

argument

"

trial

looks

ever

could

)

;

proofs

free

complexity

living

from

damned

"

he

organisms

being

a

that

when

adapted

presence

the

list

the

sent

can

' s

the

half

mailing

two

designer

earth

half

received

suckers

,

.

argument

the

"

are

producing

omnipotent

as

the

nothing

more

.

conned

no

hiding

:

's

supernatural

all

evolution

selection

does

win

guessing

has

have

a

. 35

Dennett

biological

such

pay

in

one

that

which

is

a

they

to

we

other

of

announce

evolution

variation

create

in

,

now

you

of

will

is

implementing

Indeed

natural

one

that

to

brace

have

biological

against

successful

.

he

admitted

able

but

for

:

send

the

half

second

; "

in

team

nature

that

proof

,

and

and

a

design

clear

list

other

see

Dennett

,

,

We

made

mailing

predictions

random

be

evidence

of

see

and

be

-

a

like

wrong

which

not

how

indeed

plans

fossil

is

sucker

must

doubt

its

believers

these

a

is

definite

were

part

it

we

planning

entity

the

it

the

advance

but

can

can

will

sible

in

all

them

will

half

-

plays

generated

is

week

the

successes

are

in

your

Discard

they

we

half

suckers

"

occurs

about

,

of

prediction

,

,

send

more

survive

knows

.

pool

few

next

know

charge

your

a

it

,

therefore

the

Daniel

divide

Ameri

and

be

view

philosopher

later

and

a

of

not

.

blind

that

next

week

,

phenomenon

purposeful

larger

of

again

After

which

does

free

half

your

organisms

sucker

-

A

the

must

such

championship

.

of

says

of

,

or

reasonable

folly

gamblers

majority

evolution

little

theory

by

the

lose

down

.

the

the

serious

win

will

in

this

over

same

those

ry

team

clairvoyance

period

of

will

the

Darwinian

seems

But

for

considered

very

described

A

remainder

but

therefore

.

list

team

prediction

divide

It

advance

mailing

of

,

that

.

,

explicitly

rejection

experiment

prediction

that

whether

animals

work

in

thought

conclusion

,

variation

to

planned

simple

the

and

unfit

natural

see

to

design

widespread

overwhelmingly

of

come

from

role

world

to

argument

's

the

theo

simple

-

26

WithoutMiracles

fact that " Darwin 's theory does not require positing things for which we have no evidence. The argument from designinvolves a being of unimaginable powers, while the theory of natural selection involves forces and mechanismsthat we observe today and that are easily explained." 37In short, natural selectionprovides a nonmiraculous explanation of puzzlesof

fit .

We should now be better able to appreciatethe power and appeal of Darwin '8 account of adaptive biological evolution. It alone provides a nonprovidential and noninstructionist account of the emergenceof adapted complexity. Indeed, its power and appeal are such that Oxford evolutionary zoologist Richard Dawkins proposed " universal Darwinism." He believes that if life were ever discovered

elsewhere in the universe , it too

would almost certainly be the product of a long seriesof gradual increases in adaptedcomplexity brought about by cumulative variation and natural selection .38This , he argues , is because natural

selection turns out to be the

only known theory that, in principle, is able to offer a natural explanation for the many puzzlesof fit we observein living things. It might be diffi cult to find a better test of the intelligenceof any advancedextraterrestrial beings we might someday encounter than to inquire as to whether they have yet discovered

the cumulative

process of blind variation

that is responsible for their own existence.

and selection

3 The Emergence of Instinct

Providence

Be it that thoseactions of animals which we refer to as instinct are not gone about with any view to their consequences , but that they are attendedin the animal with a

presentgratification,andarepursuedfor thesakeof thatgratificationalone; what doesall this prove, but that the prospection, which must be somewhere, is not in the animal , but in the Creator ?

- William Paley1 I nstntction

As the penchants that animals have acquired through the habits they have been forced to contract havelittle by little modified their internal organizations, thus rendering the exerciseof [thesepenchants] very easy, the modifications acquired in the

organizationof eachraceare thenpropagatedto newindividualsthroughgenera ~ tion. Indeed, it is known that generationtransmits to thesenew individuals the state of organization of the individuals that produced them. It resultsfrom this that the samepenchantsalready exist in the new individuals of each race even before they have the chanceto exercisethem, so that their actipns can only be of this one kind. - Jean-Baptiste Lamarck2 Selection

To my imagination it is far more satisfactory to look at such instincts as the young cuckoo ejecting its foster-brothers,- ants making slaves,- the larvae of echneu~ monidae [wasps] feeding within the live bo,dies of caterpillars,- not as specially endowedor createdinstincts, but assmall consequences of all organic beings, name~ ly, multiply, vary, let the strongestlive and the weakestdie. -

Charles Darwin3

Among the most marvelous puzzlesof fit are those observedbetweenthe behavior of animals and their environments , environments that include not

only a world of objects and forces but also other organisms. Indeed, many of the highly specializedstructuresof animalswould be quite uselessif they

28

Without Miracles

werenotcoupledto correspondingly specialized behaviors. Themarvelous-

lydesigned wings ofa hawkwouldbeoflittleusewithout theability to movethemin sucha wayas to generate liftforflight.Therattlesnakes

rattlewouldbea frivolous accessory without thesnakesability toshake it vigorously to soundits warning.Andthe adroithumanhandwithits

opposable thumb would bea worthless appendage ifitwerenotcoupled withthecoordinated movements thatallow ittoplantcrops, throw spears, constructcathedrals,andperformpianosonatas.

When weconsider animal behavior, wearefirststruck bywhatappear to

betwoquiteseparate categories ofactions. Onecategory consists ofcomplexbehaviors thatalltheindividuals ofa givenspecies aresomehow able to perform without firstexperiencing thebehaviors performed byothers andwithoutbeinginanywayguidedorinstructed inthem.Thusa mother ratwillbuilda nestandgroomheryoungevenifsheisraisedintotalisola-

tionfromotherfemale rats. 4 Usually, theevolutionary significance ofsuch behavior isquitecleatInthecaseofmotherrat,it isnotdifficult to seehow

building a nesttokeepherpupswarmandsecure andkeeping themclean increases theirchances ofsurvival, thereby enhancing herownreproductive success andthesurvival ofhergenes. Thebehaviors involved inthespiders spinning a web,thebeavers constructing dams,andthehoneybees sculpt-

inga honeycomb areadditionalexamples.

Theothertypeofbehavior consists ofthoseactsthatappearto result

fromananimalsparticularexperiences, andit isherethatwenoticestrikingdifferences acrossindividuals ofthesamespecies. Thecircusshowsus

whatdogs, bears, horses, lions, tigers, andelephants candogiven thespecialenvironment provided bytheanimal trainer. Dogsdonotnormally walkupright ontheirtwohindlegs,andbearsarenotto beseenriding

motorcycles throughthewoodsor sealsbalancing beachballson their

nosesintheArctic.Yetthesecreatures canperformtheseandotherunnatu-

ralactsifprovided witha special typeofenvironment. Similarly, whereas allnormal children manage to walkandtalkwithout formal instruction, thisisnotthecaseforreading, writing, andmathematics skills, thedevelopment ofwhichnormally requires manyyearsofformal schooling. Ofthesetwotypesofbehavior, thefirstistypically referred to asinstinc-

tive,innate, orinherited, andthesecond learned oracquired. Bothprovide

countless instances of puzzlesof fit.Wewillexamine theevolution of ex-

planations forthefitofinstinctive behavior in thischapter, andaddress

The Emergence of Instinct

29

learnedbehavior in chapter 7. Thesetwo chapterswill also show how traditional

theories

of behavior

fail to account

for the full extent of the ob -

served fit .

InstinctThroughProvidence Our prehistoricancestors ' survivaldependedon their understanding the waysandhabitsof the animalsthat sharedtheir surroundings . Theyhunted the animalsusefulasfood, and avoidedandrepelledthosethat would havehumansastheirfood. Thisdelicaterelationshipledto an appreciation of thewaysin whichanimalbehavioris adaptedto survival.Thebird'sability to fly, the frog's skill in catchinginsectswith its long, stickytongue, the beaver 'sconstructionof darnsandcanalsystems - theseandcountless other examples provideevidence of apparentlybuilt-in abilitiesthat emerge without long studyor labor, andthat all normalmembersof the species share. This adaptednatureof behaviorledmanytraditionalsocieties to venerate certainanimalsasembodiments of divinespirits. Two interrelatedquestionsmustbe consideredin attemptingto understandinstinctivebehavior . The first dealswith the origin of the behavior itself andthe secondwith the propagationof the behaviorto newgenera tions. It is importantto addressboth of thesequestionsseparately , but the mostsatisfactoryanswerto eachturnsout to beverymuchthe same. The providentialview of instinctiveanimal behaviorcameto us in Westernphilosophyprimarily throughthe influenceof Aristotle, Thomas Aquinas, and Descartes , and it remainedpopular and virtually unchallengedthroughthe eighteenth century.As ThomasAquinasreasoned : Although dumbanimals donotknowthefuture,yetananimalismovedbyitsnaturalinstinctto something future,asthoughit foresaw thefuture.Because thisinstinct isplanted in thembytheDivineIntellect thatforesees thefuture.s The viewsof the Aristoteliansand Cartesians of the eighteenthcentury differedin manyrespectsconcerninganimalbehavior . Nonetheless , they agreedthat "complexanimalbehavior(e.g., birds buildingtheir nestsand beestheir cells) shouldbe explainedby appealto instincts, which they understoodasblind, innateurgesinstilledby theCreatorfor thewelfareof hiscreatures ."6 ReverendPaleywas alsointerestedin usingthe instinctivebehaviorof animalsas evidencefor the existence , goodness , and wisdomof God. He

30

Without Miracles

continued hisargument fromdesignby emphasizing thosebehaviors that couldnotpossibly havebeentheresultof anyinstruction providedduring thelifetimeof theorganism . Thushedescribed howmothsandbutterflies

TheInstructionof Instinct

It is in the work of Charles Darwin 's grandfather , Erasmus Darwin (1731 - 1802 ), thatwefindanearlyalternative to theprovidential viewof instinct . Erasmus Darwin 's annoyance withthisviewcanbeseenin his observation thatfromthisperspective , instinct"hasbeen explained to bea divinesomething , a kindof inspiration ; whilstthepooranimal , thatpossesses it, hasbeen thought littlebetter thanamachine !"8Heandother"sen sationalists " of thetimeemphasized theroleof sensory experience in the development of behavior . Theybelieved thatallbehavior wasbased onthe experience andintelligence oftheindividual organism anddescribed ways in whichapparently instinctive behavior couldbeexplained assuch . But thisexplanation faredlesswellwithbehaviors demonstrated immediately afterhatching or birth. HereLamarck 's concept of theinheritance of acquired characteristics seems atfirstconsideration to domuchbetter . Lamarck 's understanding of instinctive behavior wasintimately tiedto histheoryof evolution . Whereas todaybiological evolution is usedto

The Emergenceof Instinct

31

explain theemergence ofinstinctive behavior , Lamarck sawbehavior asthe motorof evolution .9Forhim, changing environmental conditions forced organisms to change theirhabitsto survive , andchanged habitsinvolved theincreased useof certain bodypartsandsystems accompanied bythe decreased useof others . According to histheory , suchorganic changes wouldbepassed onto succeeding generations . Since behavior is clearly influenced bybiological organs , including appendages , theinheritance of suchmodified organs wouldresultintheaccompanying instinctive behav iordependent ontheorgans insucceeding generations . InthiswayLamarck attempted to provide explanations bothfortheoriginandtransmission of newinstinctive behaviors . Thistheorydepends crucially onmec :hanisms of instruction in at least threeways . Thefirsttwowehavealread 'y considered andrejected ; namely , thetransmission ofinstructions frombehavior tobiological structure withoutprevious selection , andtheinstructive transmission fromthesomatic (body)cellstothegerm(sperm andegg ) cells . Wenowhavetoaskhowit is thata changing environment results in theanimalassuming justthose habitsthatareadapted to thenewenvironment . If a particular source of foodisnolonger available , whatisit in theenvironment thatinstructs the animal to adoptaneweffective behavior to feeditself?Certainly , atypeof instruction isimplied , although Lamarck andothers ofhisdaydidnotconsidertheproblems inherent in understanding hownewadapted behaviors couldinitiallyarise . Particularly problematic in thisregard arebehaviors thatcannot beimagined astheresultof individual learning , asin theegg layingbehavior ofthemothandbutterfly (asPaley describes inthefirstepigraphto thischapter ) andtheegg -layingbehavior ofthewasp(mentioned below ). Soevenif Lamarck 'stheorycouldaccount forthetransformation oflearnable habits intoinstinctive behavior (whichit does not), it stilloffers noexplanation for theoriginof thoseinstincts thatcouldnothavebeen acquired ashabits duringthelifetime ofanyindividual organism inthefirst place . TheSelection ofInstinct Charles Darwin 'sinitialattempt to explain instinct hadmuchin common withLamarck 'stheoryof theacquisiti ~)n of acquired characteristics . Darwin"supposed thathabitsananimalrnightadoptto copewitha shifting

32

Without Miracles

environment would,duringthecourse ofgenerations, slowly become instincts,thatis,innatelydetermined patternsof behavior. Instinctsin turn

would gradually modify theanatomy ofanorganism, adapting thecreature toitssurroundings. Thus,habits thatpersisted overmany generations,

suchaseatingberries froma certainbush,andwerebeneficial to thesurvivalandreproduction ofthespecies wouldinscribethemselves intheheritable structures of the brain. 11

Gradually, however, Darwin became dissatisfied withtheideaofinherit-

edhabits asthesoleexplanation forinstinctive behaviors, particularly

whenherealized (astamarckapparently hadnot)thatmanyofthese behaviors could nothaveoriginated ashabits. Another example wasprovided byBritish natural theologian Henry LordBrougham, writing in1839 aboutthe femalewaspwho providesgrubsas foodfor the larvae

(worms)thatwillhatch fromitseggs andyetthiswasp never sawanegg produce awormnoteversawawormnay,istobedeadlongbefore the wormcanbeinexistenceand moreover sheneverhasinanywaytasted

orusedthese grubs, orusedtheholeshemade, except fortheprospective

benefit oftheunknown wormsheisnevertosee. WeknowthatDarwin

wasintrigued bythisobservation, since headded thecomment extremely hardtoaccount byhabittohiscopy ofBroughams work. 2 Itwas,infact, morethanextremely hard sincean actperformed onceina lifetime,

without relevant experience, andhaving agoalofwhich theanimal mustbe ignorantthiskindofbehavior couldnotpossibly havearisenfromintelligentlyacquiredhabit. 13

ItwascleartoDarwin thathabitcouldnotbetheexplanation forsuch

instinctive behaviors. Therefore he considered themto be not the result

ofinherited useful habits,butrathertheselection ofindividuals withusefulhabits,although henevercompletely abandoned theformeridea.Thus naturalselection provided an explanation forinstinctive behaviors that never could haveoriginated ashabits, suchasthewaspsegg-laying behavior;however, oneparticularly thornyproblem remained, thatoftheevolution and behavior of the neuter insects. 14

TheHymenoptera orderofinsects includes beesandantstogether with some wasps andflies. Many ofthese insects liveinawell-structured society where theirsurvival depends ona specialized division oflaboramong its

members, reflected indifferent castes suchasthequeen, drones, andworkersina beehive. Particularly intriguing andtroublesome forDarwinstheo-


33

ryofnatural selection wasthefactthattheworker caste inthese colonies oftencomprise sterile insects thattherefore cannot genetically passontheir instinctive behaviors tothenextgeneration ofworkers. Thisposednosmall threat.Indeed, it maywellbethatthe caseoftheneuterinsects presented the mostseriousobstacleto hisgeneraltheoryof evolution. 15

Darwinsanswercameto himafterhe learnedhowcattlewereselected

forbreeding toproduce meatwithdesirable characteristics. Asdescribed in a bookbyWilliam Youatt published in1834andreadbyDarwin in1840, animals fromseveraldifferent families wouldbeslaughtered andtheirmeat

compared. When aparticularly desirable typeofmeatwasfound, itwas,of course, impossible to breedfromtheslaughtered animal. Butitwaspossibleselectively to breedcattlemostclosely relatedto thoseslaughtered to

produce thedesired meat. Inlikemanner, acolony ofinsects thatproduced neuters thathelped thesurvival ofthecommunity (bytakingcareofyoung,

providing food,ordefending thecolony against enemies) would benaturallyselected tocontinue toproduce suchneuter insects. AsDarwin noted, this principle of selection, namely notof theindividual whichcannot breed, butofthefamily which produced suchindividual, hasI believe been followed bynature inregard totheneuters amongst social insects. 6 The

concept ofselection bycommunity orkinrather thantheindividual later

became a powerful ideaintheunderstanding oftheevolution ofaltruistic behavior, 17 andit provided Darwin withanexplanation forcomplex and usefulinstinctive behaviors thatcouldnotbeexplained byLamarckian inheritance ofhabits.Butwherethetransmission of inherited habitsseemed

conceivable, particularly where Darwin could seenoselective advantage for thebehavior, hemadeuseofLamarckian principles. Andsincehewasunableor unwilling to seeanysurvival or reproductive advantages accruing fromtheexpression ofemotions, heexplained theseasinherited habits. 18 Themixofnaturalselection andinstruction mayseemlikean oddcom-

binationto us today,butwhenseenfromDarwinsownperspective it makes rathergoodsense. Certain instinctive behaviors (aswellastheexistenceofneuterinsects) couldnotbeexplained bytheLamarckian theory

simply because hehadnowaytoconceive howtheycould havedeveloped originally ashabits.Forthesebehaviors, naturalselection ofindividuals possessing behaviors leading to reproductive success wastheoniywayto

explain theirappearance. However, other instinctive behaviors, suchasthe expression ofemotions through facial orpostural means, appeared tohave

34

Without Miracles

nofunctional value. Since natural explanation isnoexplanation atallfor thedevelopment oftraits thatprovide nosurvival andreproductive bene-

fits,Darwin reasonably concluded thattheymusthaveresulted fromthe inheritance ofuseless habitsaccompanying moreuseful ones. The Ultra-Darwinians

Others, however, bothduring andafterDarwins lifetime, completely re-

jected theLamarckian ideathatacquired characteristics couldbeinherited.

Consequently, theydared to explain allofevolution, including thatof

complex instinctive behaviors, solely through natural selection. Themost prominent oftheseultra-Darwinians, asDarwins young disciple anddefender George J. Romanes called them,wasfellow British naturalist and codiscoverer of thetheoryof naturalselection, AlfredRussel Wallace (18231913). Andsincetheywereultra-Darwinian onlyinthesensethat

theydiscounted allmechanisms ofevolution other thannatural selection,

theyareprobably moreaccurately referred toasultraselectionists.

ButWallaces viewofevolution wasmorethanjusta rejection ofall

Lamarckian instruction. Heandhisfellowultraselectionists viewed eachbitofmorphology, eachfunction ofanorgan, eachbehavior asan

adaptation, aproduct ofselection leading tobetterorganisms. They held adeep

belief innaturesrightness,intheexquisite fitofallcreatures to theirenvi-

ronments. in a curioussense,theyalmostreintroduced thecreationist notionof

natural harmony bysubstituting anomnipotent forceofnatural selection fora

benevolentdeity. 19

Wallaces ownwords express thisviewquiteclearly: Noneofthedefinite factsoforganic selection, nospecial organ, nocharacteristic

form ormarking, nopeculiarities ofinstinct orofhabit, norelations between species

orbetween groups ofspecies, canexist butwhich must nowbe,oronce have been,

useful totheindividuals orraceswhich possess them. 2

Butthisconcept poseda special problem whenWallace considered the

human intellect. This isbecause hebelieved, incontrast toother European intellectuals ofhisday,thatthebrains ofsavageswere notintrinsically inferior tothose ofcivilized Europeans. Heknew ofnon-European military bands thatvery capably performed Western music andwasmuch impressed bythevery fine singing ofAfricans who hadlearned thismusic inEngland. 2

Sohereasoned thatifnatural selection canselect onlybehaviors andabili-


35

tiesthatareadvantageous to thespecies, naturalselection couldnotbe

responsible forsuchlatent intellectual andmusical abilities sincethey served nopurpose intheirnative environment. Therefore, thehuman brain andtheintellectual andmoralqualities thatit confers couldonlybethe work of a divine provider.

Soalthough bothDarwin andWallace wereconcerned withthesame problem, theyarrived at verydifferent conclusions. Darwin sawcertain behaviors that to himservedno usefulfunctionandexplained themby

inheritance ofacquired characteristics. Wallace sawabilities thatserved no usefulfunction andexplained themastheworkofGod.It wasonlymuch laterthatDarwins selectionism andrejection ofprovidentialism wouldbe combined withWallacesselectionism andrejectionofLamarckian instruc-

tiontogivebirthtothemodern neo-Darwinian viewofbiological evolution that we embrace today.

Despite theenormous impact thatDarwins workhadonthelifesciences during hisownlifetime, it hadrelatively littleimmediate impact onthe comparative study ofanimal behavior. Twoofthereasons forthisarethe methodological difficulties ofbothnaturalistic andexperimental research onanimalbehavior, andtheheavyuseofanecdotal evidence andanthropomorphic 22 interpretation practiced byRomanes, whowroteextensively aboutanimalbehavior andmindfroma Darwinian perspective whilemaintaininga beliefintheinheritance ofacquired habits. 23 The Beginningof an EvolutionaryEthology

It wasnot untilthe 1930sthat a seriousattemptto conductresearchon animalbehavior fromevolutionary andselectionist perspectives wasbe-

gun.Konrad Lorenz (19031989) grewupsharing hisfamilys estatenear Vienna withdogs,cats,chickens, ducks, andgeese. Inthissetting hisobservationsof animalbehaviorledto the foundingof the fieldof ethology, whichhedefined asthe comparative studyofbehaviour. . . whichapplies

to thebehaviour of animals andhumans allthosequestions askedand methodologies usedasa matter ofcourse inallotherbranches ofbiology since Charles Darwins time. 24

Assuggested byhisdefinition ofethology, Lorenz wasprimarily interestedinfinding anevolutionary explanation fortheinstinctive behavioral

36

Without Miracles

patterns characteristic of a species. For example, it was brought to his attention that greylaggeesethat were rearedby humanswould follow them in much the sameway that naturally hatchedgoslingswaddled after their mother. Lorenz confirmed thesefindings for the greylaggooseand extended them to a number

of other birds as well . This pattern of behavior , result -

ing from a type of bonding with the first moving object seenby the bird, he called " imprinting ," and it is for this finding that Lorenz is still bestknown today.

Byextending Darwin 's theory of natural selection to behavior , Lorenz posited agenetic basis forspecific behaviors thatwassubject tothesame principles of cumulative blindvariation andselection thatunderlie the adapted complexity of biological structures . In thecase of thegreylag goose , goslings thatmaintained close contact withthefirstlarge moving object theysaw(usually themother goose ) would beinabetter position to enjoy herprotection andnurturance . Consequently , theywouldbemore likelytosurvive andtohave offspring thatwould similarly show behavioral imprinting . Those goslings thatlacked thisbehavioral characteristic would beless likelytosurvive tomaturity andreproduce . Inmuch thesame way thatweunderstand howatreefrogcanbecome sowellcamouflaged over evolutionary timethrough theelimination bypredators ofthose individuals whoarelesswellcamouflaged , wecanalsounderstand howinstinctive behavior canbeshaped through theelimination of individuals whose behaviors areless fit totheirenvironment . Another example of Lorenz 's conception of instinctive behavior isthe egg -rolling behavior ofthegreylag goose .25 When thegoose sees thatanegg hasrolledoutof hernest , shestands up, moves to theedge of thenest , stretches outherneck , androllstheeggback intothenestbetween herlegs , pushing it withtheunderside ofherbill. Thisbehavior depends onwhat Lorenz called a "fixedmotorpattern ," thatis, apattern ofactivity in the central nervous system ofthegoose thatisreleased ortriggered bythesight oftheeggoutside thenest . In otherwords it isa specific fixedresponse released byaspecific typeofstimulus . Thepurpose ofthisinstinctive act is ---clearly toreturn theeggtothesecurity ofthenest , andit iseasy toappreci ateitsvalue forthesurvival ofthespecies ,&. . Buta problem withLorenz 's conceptualization of instinctive behavior becomes apparent whenoneconsiders thatexactly thesame pattern of


37

behavior willnot besuccessful in returning thewaywardeggto thenest unlessall environmental conditions are exactlythe samefor eachegg-

rolling episode. Instead, forthegoose tobesuccessful inkeeping hereggs

nestbound, shemustbeabletomodify herbehavior notonlyfromepisode toepisode butalsowithin eachepisode tocompensate forthevariability in conditions anddisturbances thatsheinevitably encounters, suchas differencesinthedistance between herandtheeggatthebeginning ofthebehav-

ior,andirregularities intheterrain between theeggandthenest.As

American philosopher andpsychologist William Jamesnotedover100

years ago,thebehavior ofallliving organisms ischaracterized bytheattainmentofconsistent endsusingvariable means. 26 Anditisforthisreason that

Lorenzs analysis failstoexplain thetypical success (andtherefore thefitto environment) of instinctive behavior. Thisessential aspectof adapted behaviorwillbe consideredin moredetailin chapter8.

Butdespite thisshortcoming, Lorenz mustbeacknowledged asbeing the firsttoattempt toprovide a trulyDarwinian (or,moreaccurately, a purely selectionist) account ofspecies-specific behavior patterns, andhewasrec-

ognized forhisachievement in1973 when heshared aNobel prize withfellowethologists Nikolaas Tinbergen andKarlvonFrisch. Inthesameway thatbiologists constructed evolutionary trees(orphylogenies) bycompar-

ingtheanatomical similarities anddifferences among living animals and fossils, Lorenz proposed patterns ofinstinctive behavior forthesamepur-

pose. 27 Indeed, hebased hiscomparative study ofanimal behavior on the

factthat thereare mechanisms of behaviorwhichevolvein phylogeny

exactly as organs do. 28 Hisevolutionary perspective alsoledhimto emphasize thatunderstanding animal behavior required oneto appreciate itspurposefulness inpreserving thespecies, itsroleintheentirerepertoire of theanimalsactivities, anditsevolutionary history. 29 Forthesereasons,

among others, Lorenz regarded withsuspicion research onanimal behavior being conducted byexperimental psychologists using domesticated animals inartificially contrived environments ontheothersideoftheAtlantic. But to understand howanimals areableto modify theirbehavior duringtheir ownlifetime, wemustturntotheworkoftheseAmerican animal psychologists later in chapter 7.

4

The Immune System Selection by the Enemy

FromProvidence to Instruction The extraordinarynumberof specificantibodies , includingthoseagainstartificial antigens , defiedthe geneticorigin originallypropoundedby Paul Ehrlich. A somatic , custom-madetemplatemechanism seemed morelogical. - DebraJanBibel1 Selection It follows that an animalcannotbestimulatedto makespecificantibodies , unless it hasalreadymadeantibodiesof this specificitybeforethe antigenarrives. It can thusbeconcludedthatantibodyformationisa selective process andthatinstructive theoriesof antibodyformationarewrong. - NielsJerne2

The previoustwo chaptersdescribedhow Darwin's theoryof cumulative variationandselectionprovidesa naturalisticexplanationfor thefit of biologicalform andbehaviorobserved in living organisms . But althoughthefit thusachievedis striking, naturalselectionis a verywastefulprocessin the sense that it depends on eliminatingentireorganisms , thoselessfit, from the evolutionaryprocess . Workingon largepopulationsof organisms overlong periodsof time involvingmanygenerations , Darwiniannaturalselectionis what is referredto asa phylogenetic process , sinceit resultsin newphyla, or newbranchesof theevolutionarytreeof life. It would thusseemto bein an organism 's own interestif it could somehowincreaseits fit to its environmentduringits lifetime, that is, becapableof ontogenetic3 adaptation. Indeed, individualorganismsundergoontogeneticadaptationin many ways. Musclesgrow strongerwith increaseduse, and animaland human behaviorschangeover time in functional, adaptiveways. Thesechanges

40

Without Miracles

arereadilyobservable , butthebestunderstood ontogenetic adaptation involvesthemammalian immunesystem . Probablythebestexample of a puzzleof fit at themicroscopic levelis demonstrated bytheantibodies produced bythemammalian immunesystem. Fortheimmunesystem to beableto rid thebodyof antigens - foreign invadersin theform of chemical toxins, viruses , bacteria ~cells- andtissues - theremustbeaprecise fit between antibodyandantigen . Attempts to explainthisfit gobackover100yearsandresulted in oneof themoststrikingtriumphsof therelatively newfieldof molecular biology . A (Genetically ) Providential Theoryof AntibodyProduction Afterthegeneral acceptance of LouisPasteur 's germtheoryof disease late in thenineteenth century , anumberof scientists became interested in under standing themechanisms responsible for theappearance of antibodies in an animal 's bloodafterit hadbeeninfectedwith disease -producing bacteria . Antibodies wereof specialinterestsincetheywereknownto .orotectthe L animalfromsubsequent infectionbythesamepathogenic bacteria . PaulEhrlich(1854-1915)wasa German whoin the1890sdeveloped a technique for estimating thequantityof antibodies in blood. He wasintriguedbytheexplosive increase in antibodyproduction afterexposure to anantigenandattempted to account for thisphenomenon byformulating hisside-chaintheory . According to thistheory , thesurface of whiteblood cellsis covered with manysidechains , or receptors , that form chemical linkswiththeantigens theyencounter . Foranygivenantigen , atleastoneof thesereceptors wouldbind, stimulating thecellto producemoreof the sametypeof receptor , whichwouldthenbeshedintothebloodstreamas antibodies . According to Ehrlich'stheory , anantibodycouldbeconsidered an irregularlyshaped , microscopic , three -dimensional labelthat would bindto a specific antigenbutnot to theothercellsof theorganism .4This analogy of antibodies aslabelsis usedheresincetheantibodies themselves usuallydonot destroytheantigen , butlabelit, providinga molecular kiss of deathfor destruction of the antigenby complement proteins , macro phages , or otheragents thateitherperforate theantigen 'scellmembrane or completely engulftheantigen .5

41

The Immune System

The

major

assumption

providential

cally

of

theory

specified

tered

,

side

antigen

.

possibly

,

entire

set

That

is

is

of

early

1900s

,

apparently

mal

could

Karl

not

produce

an

theories

count

for

task

of

new

An

The

tern

in

tion

of

that

's

line

an

any

:-

encoun

~

of

genes

.

all

For

this

referring

to

organism

fit

or

difficulty

range

of

the

pair

of

the

Such

but

that

response

that

to

it

.

of

not

adaptive

completely

~

animal

only

to

This

led

to

constructivist

theories

was

ani

in

the

antigens

was

an

required

consideration

system

,

there

produced

information

constructivist

immune

be

revealed

the

the

that

even

necessary

to

During

that

antibodies

could

possibly

and

.

demonstrated

substances

.

in

major

clearly

genome

all

antibodies

of

)

the

finite

theory

the

fit

Instructionist

Theory

first

well

'8

to

with

theory

, 6

produce

had

adapted

as

well

,

unpredictable

to

ac

to

the

able

and

~

to

novel

produced

themselves

was

by

likened

to

would

stick

able

be

label

an

innately

to

to

for

,

given

seemingly

Ehrlich

' s

set

antigen

,

,

By

of

the

labels

template

instruc

Nobel

proposing

that

,

range

designed

theory

that

saw

..

antibodies

theory

so

anti

template

of

- chain

they

the

the

-

explain

because

production

..

prize

to

antigen

side

~

appeared

by

limitless

innatist

sys

template

attempted

novel

antibody

immune

antigens

advocated

any

antigens

for

the

the

theory

with

the

if

novel

of

and

with

account

for

to

template

models

determined

any

response

bind

contact

analogy

The

account

introduction

developed

. 7

made

as

to

's

further

direct

to

the

attempted

in

Pauling

served

was

pursue

that

Haurowitz

Linus

could

gens

Production

antibodies

and

which

antibodies

Antibody

theory

Breinl

chemist

theory

of

~ known

ability

1930

were

To

animal

of

.

winning

that

~

production

germ

antibodies

artificial

production

puzzles

antigens

to

its

number

's

1943

that

in

Ehrlich

fact

typ

genetio

to

the

from

a

-

limit

finding

novel

antibody

the

passed

a

numerous

the

the

encountered

( 1868

possess

producing

create

soon

His

infinite

of

by

,

it

bind

for

theory

is

makes

would

provided

that

that

possessed

which

essential

)

discernible

.

of

one

.

completely

rejection

of

- line

genome

and

cells

one

germ

Landsteiner

possibly

,

blood

least

was

a

theory

no

to

could

theory

information

offspring

~ line

produce

response

the

its

germ

was

( or

's

white

at

the

as

genes

the

,

known

to

But

,

antibodies

this

organisms

that

chains

necessary

reason

was

Ehrlich

.

could

at

the

least

immune

be

one

42

Without Miracles

system as supplied withtheknowledge of.a procedure whereby labels wouldbecustom-made foreachantigen inthesamewaya tailormakesa suitofclothes usingthecustomer asa template. Thetemplate theory could therefore beconsidered instructionist since itdidnotrequire innate, germ-

lineinformation fortheproduction ofallantibodies, butratherrequired onlya general label-construction procedure forbuilding antibodies using

theantigens themselves astherequiredsourceofinstructions.

Although oneversion oranother oftheinstructionist template theory

remainedinfluential for over20 years,it too encountered a numberof

serious difficulties. Inthe1950sDanishimmunologist NielsJernenoted several immunological findings thatthetheory couldnotexplain. 8 Firstof all,itcouldnotaccount fortheincreasing rateofantibody production duringtheinitialimmune response. Iftheantigen itselfwasnecessary forthe production ofeachantibody, howcouldtheantibodies soquickly outnumber thetemplates? Second, it couldnotexplain thememory ofthe immune system bywhicha second exposure to a givenantigen resultsina muchmorerapidproduction of antibodies thandoestheinitialcontact.If

theantigen itselfserved asa template, itstotalelimination bytheimmune system wouldalsonecessarily entaildestruction oftemplates, soa second exposure to thesameor similarantigenshouldbe accompanied byan immune response nodifferent fromthefirst.Third,sinceit wasthought thatantibody cellswerequiteshortlived,thetemplate theorycouldnot account forthefactthatantibodies continued to bepresentlongafterthe antigenhadbeeneliminated fromthebody.Finally, thetheorydidnot explain thefactthattheantibodies produced during thelatterstages ofan immune response areusually moreeffective in binding withtheantigens (arebetter-fitting labels) thantheantigens initially produced. It wasthereforeclearthat in attempting to accountfor the formation

ofnewantibodies, theinstruction-based template theoryranintoa num-

berofratherserious problems thatmotivated scientists to pursue other

explanations.

A Selectionist Theoryof AntibodyProduction

Jernespaperprovided notonlyarguments against thetemplate theorybut

alsoanalternative theory thatinsome respects resembled Ehrlichs original side-chain theory. Jernesnatural-selection theory ofantibody production

The Immune System

43

statedthat a mammalinitially possesses a relativelysmallnumberof antibodies. The successful binding of an antibodyto an antigentriggersthe antibodyto producea largenumberof copiesof itself. In this way, a preexistingantibodyis effectivelyselected by theantigen,whichstimulatesthe chosenantibod'y to producea multitudeof clones.Australianvirologistand immunologistSirFrankMacfarlaneBurnet(1899- 1985) furtherdeveloped thetheoryusingthetermclonalselectionto describeit.9 Much researchstill has to be done to understandfully the complex dynamicsof the immunesystem , but the jerne-Burnet clonal-selection theory of antibodyproductionis generallyaccepted . Althoughits major characteristics are not difficult to understand , we haveto go into a bit moredetailconcerningthe immuneresponse to appreciateits selectionist functioning. First, insteadof thinkingof an antibodyaseitherbindingor not binding to an antigen,we mustappreciate that antibodieshavea verywiderangeof affinity (that is, attractionor binding power) to a givenantigen.1OThus whereasa well-fitted antibodywill almostalwaysbind to an encountered antigenfor whichit iswell fit, anotherlesswell-fitted antigenmayalsobind with thesameencountered antigen,althoughits rateof bindingmaybeconsiderablylessthan 100%. In contrast, an antibodythat is completelyunfit for bindingwi,th a givenantigenwill seldom,if ever,bind to the antigen. Second , insteadof everyantigenhavinga single" handle" (calleda determinant or epitopeby immunologists ) for the antibodyto grasp, they all havemanysuchdeterminants , eachof which is differentand thus can be boundby a di.fferenttype of antibody. Thesedeterminantscorrespondto smallpatternsof molecularstructureon the surfaceof the antigen.I1 The largenumberof determinants on eachantigeneffectivelyincreases the likelihoodthat the immunesystemwill beableto producean antibody that will bind to anyantigenintroducedinto the body. It still hasto create a staggering numberof differentantibodiesto ensureits effectiveness , however,a diversi1 :y that includesup to 10 billion B lymphocytecells, eachable to producemc)re than 100million differentantibodyproteins.And sincea personhas only about 100,000 genes , thereis simplyno way our genes couldspecifyeachandeveryoneof theseproteins.12 The answerto this enigmarequireda radical reconceptualization of genesandho'\ivtheyfunction. Until themid-1970sit wasa generallyaccept ed principlec.f biologythat onegenealwaysleadsto the synthesisof one

44

Without Miracles

andonlyoneprotein, andthatthegenetic makeup ofanorganism remains constant throughout itslife.(Relatively rareandusually harmful mutations

dooccur duetogenetic copying errors during celldivision orexposure to

irradiation or otherenvironmental mu tagens.) Butin 1976at theBasel

Institute forImmunology, Susumu Tonegawa discovered thatantibody

genesarenotinherited complete, butratherasfragments thatareshuffled

together toforma complete gene thatspecifies thestructure ofa given B

lymphocyte andtheantibodies it produces (Tonegawa received a Nobel prizein1987forhisdiscovery). Inaddition, astheDNAsegments arecom-

bined toformthecomplete Blymphocyte gene, newDNA sequences are added atrandom totheends ofthefragments, ensuring even more antibody

13 diversity. In this manner:

Thereceptors usedintheadaptive immune response areformed bypiecing together gene segments, likea patchwork quilt. Each cellusestheavailable pieces differently tomake a unique receptoi enabling thecells collectively torecognize

theinfectious organisms confronted duringa lifetime. 4

Thisactive, random reshuffling ofimmunoglobulin genes, together with theinsertion ofrandom DNA sequences during therecombination process, isresponsible forthediversity ofantibody receptors attached toeachBcell.

Such diversity virtually ensures thatatleast oneantibody, although perhaps

notfitting perfectly, willbeabletobindwithatleastoneofthemany deter-

minantspresentedby a newantigen.

Onceanantibody isselected byanantigen bybinding, itstimulates theB

lymphocyte towhich itisattached todivide andmake exact copies ofitself. Some cftheselected clones remain ascirculating Blymphocytes andassuch serve astheimmune systems memory. Increased numbers ofthese cells pro-

videfora fasterimmune response tosubsequent infections andestablish the immunity thatfollows manyinfections andvaccinations. Otherselected

clones stopdividing, growlarger, andturnintoplasma cellswhosesole function is to roduce large numbers of free antibodies to fightthecurrent infection. Theclonal-selection theory explains thegreatdiversity ofantibodies and

theability oftheimmune system tobindwithcompletely novel antigens. It alsoprovides anaccount forJernes firstthreefindings, noted above, inhis criticism ofthetemplate theory. Thetheory asdescribed sofar,however, stillfails toaccount forthefinding thattheantibodies produced during the

The Immune System

45

latterstages oftheimmune response aremoreeffective inbinding withthe antigens thantheantibodies initially produced. Thisfine-tuning of antibodiesis accomplished by anothermechanism that alsochangesthe ge-

neticmakeup oftheantibodiesthe random mutation ofthegenes within the B cellclones.By alteringindividual nucleotide basesthe mutations fine-tunethe immuneresponse,creatingimmunoglobulin geneswhose products better match the antigen. 15

Wethereforeseethat the clonal-selection processof antibodyproduction

hasa numberof noteworthy characteristics. First,it is constructive in that the actualstructureof the antibodiesis not explicitlyincludedin the ge-

nome.Thus,antibodies arequiteunlikeotherphysiological structures (such aslivers,eyes,andnoses) whosebasicdesigndepends onwhatisbelieved to be a fixed set of genes. 6

Second, thestructure ofantibodies isdetermined bywhatappears to be anessentially blindprocess. Thisblindness showsitselfatthreelevels: the randomrecombinationof immunoglobulin genesas Blymphocytecellsare

formed, therandominsertion ofDNAsegments intotherecombined gene, andtheconsequent blindhypermutation of the cloneB cellsto fine-tune themto theantigen. Thus,theimmune systemdoesnotattemptto predict theantibody structure thatwillbindwithanantigen, butratherusesa type of shotgun approach thatsendsina diverse armyto meettheinvaders. Almostalloftheseproducedantibodies willturnoutto bequiteineffective

inbinding withtheantigens, butthediversity ofthisarmyvirtually ensures that at leastoneof themwillbeeffective. Indeed,studieshaveshownthat

ifantibodies areproduced blindly, theprobability thata novelantigen will berecognized isvirtually assured ifverymanyantibody typesarepresent. 17

Finally, theimmune system isdesigned sothatonlythoseantibodies that areableto bindwiththe antigenarereproducedandremembered thenext timethe sameor similarantigeninvadestheanimal.Antibodies that arenot successful leaveno offspring andthereforesoonbecomeextinct,to bere-

placed bytheestimated onemillion newBlymphocytes produced inthe bone marrow every second. 18

Butwe stillhaveto accountfor anotherimportantabilityof the immune

systemthatit doesnotattackthecellsandproducts of itshostbody. Although it wasfirstthought thattheimmune system wasprovided this information in the germline,researchhas nowdemonstrated that the

46

Without Miracles

immune systemin fact learns to distinguish self from nonself. The process by which it doesso will not be describedin detail here, but it is virtually a mirror image (with an important reversal ) of how it produces antibodies to

recognizeinvaders. That is, mature lymphocytesare triggeredto reproduce and mutate when they encounter a foreign antibody; however, immature lymphocytesthat form while the animal is still in utero or shortly after birth " go through a stagewhen binding of their receptors causesthem to die. Self-reactive cells are killed before they have a chanceto proliferate and damagetheir host. " 19This is the clonal-deletion theory of how the immune systemlearnstoleranceto self, and was first proposedby JoshuaLederberg

in 1959.

Tosummarize , theclonal -selection theory states thataverylarge number of unique B lymphocytes to whichareattached antibody receptors are always circulating throughout thebody . Theirgreat diversity results from therandom recombination ofimmunoglobulin gene fragments andrandom insertion ofDNAsequences astheBcells develop . Thisblinddiversity ofB cellsvirtually ensures thatatleast onewillproduce anantibody thatwill bindwithanyantigen thatmakes itswayintotheorganism . Thebinding of aBcell'santibody withanantigen stimulates thecelltodivide andproduce clones , withsuccessive generations of reproducing clones resulting in an exponential riseovertimein thenumber ofcirculating antibodies of the selected type . Some oftheBcellclones remain incirculation to formthe immune system 'smemory oftheantigen . Others terminally differentiate , forming plasma cells thatproduce large numbers ofantibodies thatfightthe current infection . Finally , astheBcellclones reproduce theyundergo ahigh rateofsomatic mutation that,when combined withthecontinued selection pressure exerted bytheantigen , fine-tunes thefit oftheantibodies to the antigen . A similar process of variation andselection of immature Bcells (although nowwithselected cellseliminated rather thanreproduced ) accounts fortheimmune system 'sabilitytotolerate thecells andproducts of itshostbody . AntibodyProduction asaMicrocosm of DarwinianEvolution Eventhismuchabbreviated andsimplifiedaccountof theclonal-selection functioning of theimmunesystem reveals it to bea remarkable microcosm of Darwinianevolutionwith thethreemajorprinciples of superfecundity ,

The Immune System

47

variation, and natural selection each playing an essential role. Superfecundity is evident in that the immune system produces far more antibodies than will be effective in binding with an antigen . In fact , it appears

that the majority of produced antibodies do not play any active role whatsoever in the response of the immune system. Natural (and blind ) variation

is providedby thevariablegeneregionsresponsible for theproductionof a highly diversepopulation of antibodies. And selectionoccurs, as only antibodies.able to bind with an antigen reproduce. The similarity betweenadaptive biological evolution and the production of antibodiesis evenmore striking when one considersthat the two central processesinvolved in the production of antibodies, geneticrecombination and mutation, are the sameonesresponsiblefor the biological evolution of sexually reproducing species.We have seenthat the recombination of immunoglobulin genesunderlies the large diversity of the antibody population , and the mutation of these genes serves as a fine -tuning mechanism .2O

In sexuallyreproducing species , the sametwo processesare involved in providing the variations on which natural selectioncan work to fit the organism to the environment lection , which

. Thus cumulative

over many millions

blind

variation

of years resulted

and natural

se-

in the emergence of

mammalian species , remain crucial in the day-to-day survival of these species in their ceaselessbattle against microscopic foreign invaders . A final similarity between the functioning of the immune system and bio -

logical evolution is worth noting- the evolution of our knowledge of how eachoperates. Our understandingof the fit of organismsto their environment has progressedfrom a providential explanation to an instructionbased(Lamarckian) one to a purely selectionist(neo-Darwinian ) account. The samestagesof thought can also be seenin biology's attempt to account for the puzzleof fit of antibody to antigen. Ehrlich's side-chain theory can be considered providential in that the organism's geneswere believedto provide all the knowledge necessaryto construct antibodies that would be able to fit all the antigens the organism

would ever encounter. Sincethis knowledgewas consideredto be the result of past evolutionary selection, the geneticallyprovidential theory doesnot lead to the sameproblem of ultimate origins that the supernaturally providential theory of the origin of species encounters . However , it did run into

difficulties when it was found that the immune system could produce

48

Without Miracles

antibodiesthat fit novel antigensneverencounteredbefore, eitherin the organism 's own lifetimeor in that of its ancestors . The appreciationof the immunesystem 's ability to adaptto a changing environmentof novelantigensled to the instruction-dependenttemplate theoryof antibodyformationin whichtheenvironment(in theform of antigens) somehowtransmittedinstructionsfor the formationof close-fitting antibodiesto theB lymphocytes . Thistheory, however , did not accountfor manyof the characteristics of the immunesystemthat wereobservedsubsequently . In addition, no transmission of informationfrom antigento antibodywaseverobserved . Giventhe problemof accountingfor adaptivechangewithout recourse to instructionisttheories , it shouldperhapsnot besurprisingthat the field of immunologywouldeventuallyhit on thesamesolutionthat Darwinhad discovered , evenif it took an additionalcentury. By combiningthe basic principlesof superfecundity , blind variation, andselectionthat explainthe adaptationof organismto environment , the clonal-selectiontheory providesan understanding of how theimmunesystemcanproduceantibodies adaptedto its environmentof novelantigens . Thisit doeswithout recourse to providentialor instructionistexplanations . But whereasadaptivebiologicalevolutionproceedsby cumulativenatural selectionamongorganisms, researchon the immunesystemhas now providedthe first clear evidencethat ontogeneticadaptivechangecan be achievedby cumulative blind variationandselectionwithin organisms .

Brain Evolution and Development The Selectionof Neurons and Synapses

Instruction

versus Selection

The 1O~OOOor so synapsesper cortical neuron are not establishedimmediately. On the contrary~they proliferate in successivewavesfrom birth to puberty in man. . . . One has the impression that the system becomes more and more ordered as it

receives"instructions" from the environment. If the theory proposedhereis correct, spontaneousor evoked activity is effective only if neurons and their connections already exist before interaction with the outsideworld takesplace. Epigeneticselec-

tion actson preformedsynapticsubstrates . To learnis to stabilizepreestablished synaptic combinations, and to eliminate the surplus. - Jean-Pierre Changeux1

The

most

organ

complex

that

1300

to

11

billion

ing

,

1500

and

grams

nized

and

with

other

Although share

soma

of

cells

's

.

are

the

functioning

cells

they

's

of

and

throughout

rarely

as

the

of

genes

The

that

to

body

reproduce

types

of

is

. In

the

these

new

in

almost

turn

by

essential

neurons

, neurons

that

.

all

orga

-

them

body

a

, or

complete

cytoplasm

,

to

are

unlike

of

cell

houses

organelles

fact

are

communicate

surrounded

the

they

of

.

5 . 1 . The

respects

for

,

,

number

neurons

figure

-

sensations

sheer how

neurons

in

nucleus

except

, but

over

, process

our

the

the

about

contains

receiving

structure

, which

contains

,

brain

is

only

all

not

typical

portrayed

neuron

metabolism

the

divide

.

cell

the

is

how

their

universe

brain

of

which

it

understand

different

nucleus

organism

"

to

on

the

consider

features

the

" soup

And

many

human

. 2 But

about

the

weighing

, capable

pulses

must

common

, contains

) , the

neurons

depend striking

first

there

chemical

most

in

. Although

pounds

, or

emotions

is

we

certain

ears

electrochemical

and

that

four

cells

interconnected

each

other

the

anywhere

our

to

nerve

,

discovered

between

( three

thoughts alone

yet

space

relaying ,

neurons

ron

the

specialized

actions

set

object

fills

the

the

neu

similar

most

-

to

other

50

Without Miracles Dendrites

Cell body or soma

Figure5.1 A typicalneuron(afterChurchland , 1990). Dendrites

Jt/ ~""""""~---~ Axons from other neurons

@ = Synapse Axon to another

~

~

neuron

Figure 5.2 Neural synapses (after

Churchland , 1990).

51

Brain Evolution and Development

The

ways

nicative

in

they

sport

to

a

a

dendrite

that

bushy

is

,

its

axon

system

Thus

its

the

considerable

one

of

one

,

, 000

a

junctions

typical

wiring

system

even

nections

that

gives

by

within

the

,

the

ability

to

of

How

this

have

be

ally

,

,

most

has

three

primitive

made

and

.

comprehend

,

regions

we

to

con

-

and

the

planet

the

a

single

in

it

is

is

us

us

.

over

the

thriving

,

)

outer

space

possible

able

for

egg

cell

to

continue

of

the

to

the

-

is

perceive

with

a

and

.

on

the

organ

that

the

tropics

other

intricate

to

planets

. 4

structure

,

to

-

adaptable

the

Finally

years

modifying

from

-

actu

of

an

be

may

puzzle

most

and

the

,

what

The

into

become

,

to

millions

evolved

living

all

walk

.

presents

fertilized

brain

that

to

-

today

reasoning

how

con

muscles

believed

able

unintentionally

soon

how

mature

-

and

of

and

widely

encountered

to

complexity

organs

neurons

ancestors

about

complex

the

allows

species

of

organization

is

this

having

mathematical

yet

early

perhaps

from

it

and

synaptic

exceedingly

being

consider

human

understand

develop

how

a

.

dendrites

philosophers

the

have

must

intentionally

,

must

,

environment

we

our

on

this

organization

we

of

is

from

and

fit

the

organism

,

brain

First

for

( both

polar

an

giraffe

These

the

sense

even

among

our

of

system

powerful

Second

control

possible

environment

of

the

,

) .

magnitude

It

Indeed

-

scientific

puzzle

aspects

it

and

contains

existing

and

nervous

has

the

,

other

one

brain

an

distant

.

. 2

human

of

more

complex

striking

the

orders

,

brain

5

with

constitute

to

psychologists

connections

feel

of

connected

( figure

of

abilities

unfathomably

think

the

out

length

another

be

neuron

.

and

abstract

the

signal

of

the

neck

of

When

attached

dendrites

the

must

one

many

amazing

human

perform

function

of

cortex

by

skull

its

neuroscientists

knowledge

is

a

,

-

likened

.

it

body

firing

two

supercomputers

the

brain

the

therefore

advanced

both

sending

the

in

synapses

the

synapses

surpasses

most

,

' s

the

,

axon

in

The

that

the

influence

which

by

neuron

meters

called

the

neuron

.

up

the

be

neurons

to

fire

commu

appendages

can

other

neuron

picked

of

another

connect

synapses

dendrites

may

several

can

influence

by

usually

another

to

The

from

the

their

the

.

to

accomplished

,

be

size

up

.

and

may

neuron

away

neuron

junctions

of

,

distance

For

it

small

out

of

signals

way

polarity

the

firing

siderable

receives

where

carry

examination

axons

certain

electrical

axon

be

a

to

closer

and

that

in

Considering

can

specialized

by

dendrites

stimulated

single

. 3

are

evident

their

changes

along

the

made

antenna

is

neurons

10

neurons

is

,

suddenly

is

which

function

we

modify

of

must

try

its

to

own

52

Without Miracles

structure so that it can acquire new skills and information to continue surviving and reproducingin an unpredictable, ever-changingworld . In this chapter we will consider the researchand theories that are beginning to provide answersto thesequestions. Indeed, the 1990shas beenreferred to as " the decadeof the brain," as scholarsand scientistsin fields from philosophy to molecular neurobiology focus their energieson understanding humankind's ultimate inner frontier.

The Evolutionof theBrain Neurons are quite distinct from other body cells in ways that make them suited to their specializedrole of signal processingand communication, but it is not too difficult to seehow they could have evolved from lessspecialized cells. All living cellsare surrounded by a cell membranethat separates the special chemical composition of its interior from that of the external world . This differencein chemicalcomposition results in a small electrical potential betweenthe inside and outside of the cell, in much the sameway that a voltage exists betweenthe two sidesof a battery. When a part of a cell's membraneis disturbed in a certain way, it losesits electricalpotential, becomingdepolarizedat the site of the disturbance. This suddenchangein electrical potential can itself be a disturbance, causingadditional depolarizations along the membrane. In most cells, such depolarization would not spreadfar, certainly not to neighboring cells. But a few changesin the shape and arrangementof cells (in just the way that neurons are fashioned) permits depolarization to propagatequickly from one neuron to the next, and allows it to travel quickly as an electrochemicalsignal from one end of an animal to the other. An example of a simple nervous systemis provided by the jellyfish (or Medusa). The jellyfish's nervous systemforms an undifferentiated network and servesprimarily to coordinate the animal's swimming motions. Since the jellyfish's skirt must open and contract in a coordinated manner for the animal to move through the water, its nervous systemservesas a simple communications network making it possible for all parts of the skirt to open repeatedlyand then contract at the sametime. Worms are the simplest organisms to have a central nervous system, which includesa distinct brain that is connectedto groups of neuronsorga-


53

nized as nerve cords running along the length of its body. This more complicated nervous systemallows worms to exhibit more complex forms of behavior. An anterior brain connectedto a nervecord is the basicdesignfor all organisms with central nervous systems, from the earthworm on the hook to the human on the other end of the fishing rod. But although we can discerna separatebrain in worms, it is not the casethat the brain is the sole " commander" of the animal that the rest of the nervous systemand body obeys. Indeed, even with its brain removed, worms are able to perform many types of behaviors, including locomotion, mating, burrowing, feeding, and evenmazelearning.5 As we move to insectswe find increasedcomplexity in all aspectsof the brain and nervoussystem. So-called giant fiber systems(also found to some extent in worms and jellyfish) that allow rapid conduction of nerve impulsesconnect parts of the brain to specificmusclesin legsor wings. Such connections permit the cockroach to dart away as soon as it sensesthe moving air precedinga quickly descendinghuman foot . The brain itself is typically divided into three specializedsegments , the protocerebrum, the deutocerebrum, and the tritocerebrum. In addition, insectspossessa greater variety of sensoryreceptorsthan any other group of organisms, including vertebrates, that are sensitiveto the odors, sounds, light patterns, texture, pressure, humidity, temperature, and chemical composition of their surroundings. The concentration of thesesensoryorgans on the insect's head provides for rapid communication with the tiny yet capable brain located within . Although minuscule by human standards, the range of abilities made possibleby insect brains is impressive. Thesecreaturesshow a remarkable variety of behaviors for locomotion, obtaining food, mating, and aiding the survival of their offspring. They can crawl, hop, swim, fly, burrow, and evenwalk on water. The female wasp hunts down a caterpillar, paralyzes it with her venom, and then lays its egg on the motionlessprey so that her offspring will havea freshand wholesomemeal immediatelyafter hatching. Leafcutter ants harvest leavesand bring them into their nest where they use them to cultivate indoor gardensof edible fungus. Honeybeeslive in socialcommunitieswhere there is a strict division of labor, and where foodgatheringworker beesperform a specialdanceto communicatethe location and richnessof food sourcesto their hivemates. It is the evolution of their

54

WithoutMiracles

brains, together with the complementary evolution of their other body parts, that make insectsthe most abundant multicellular organismson our planet. The brain becomesboth much larger and still more complex as we move to vertebratessuchas fish, amphibians, and reptiles. The spinal cord, now protected within the vertebraeof the backbone, has becomeprimarily a servant of the brain, a busy two-way highway of communication with fibers segregatedinto descendingmotor pathways and ascendingsensory ones. The brain itself is now composedof a seriesof swellings of the anterior end of the spinal cord (the brain stem), the three major onesmaking up the three major parts of the vertebrate brain: the hindbrain, midbrain, and forebrain. From the hindbrain sprouts a distinctive structure, the " cerebellum" (Latin for " little brain" ). Among mammals, the brain keepsits three major components, but with two new structures. The neocerebellum(" new cerebellum" ) is addedto the cerebellum, looking much like a fungal growth at the baseof the brain, and the neocortex ("new cortex" ) grows out of the front of the forebrain. In most mammals, thesenew additions are not particularly largerelative to the brain stem. In primates they are much larger, and in the human they are so large that the original brain stem is almost completely hidden by this large convoluted mass of grey neural matter. In keeping with this remarkable increaseof neocerebellarand neocortical tissue, humans enjoy the largest ratio of brain weight to body weight of any of earth's creatures. It is not possible to know exactly why the human brain evolved as it did, but considerationof the structural evolution of the brain and resultsof comparativeresearchon human and nonhuman brains provides someuseful clues. It is now believedthat during the long evolution of our brain, nervous systemschanged in four principal ways. First, they becameincreasingly centralized in architecture, evolving from a loose network of nerve cells (as in the jellyfish) to a spinal column and complex brain with impressiveswellingsat the hindbrain and forebrain. This increasinglycentralized structure also became increasingly hierarchical. It appears that newer additions to the human brain took over control from the previous additions and in effect becametheir new masters. Accordingly, the initia tion of voluntary behavior aswell asthe ability to plan, engagein conscious thought, and use languagedepend on neocortical structures. Indeed, the


55

humanneocortexcanactuallydestroyitselfif it wishes,as whena severely

depressed individual usesa guntoputa bulletthrough hisorherskull. Second,therewasa trendtowardencephalization, that is,a concentration of neuronsand senseorgansat one end of the organism.Byconcen-

tratingneuralandsensory equipment inonegeneral location, transmission timefromsenseorgansto brainwasminimized. Third,the size,number, andvariety ofelements ofthebrainincreased,, Finally, therewasanincrease inplasticity, thatis,thebrainsability tomodify itselfasa resultofexperienceto makememory andthelearning ofnewperceptual andmotorabilities possible.

Onewayofunderstanding theevolution ofthehumanbrainistoseeitas theaddition ofhigherandhigherlevelsofcontrol.Wewillseeinchapter8 thatthefunctionof animalandhumanbehaviorcanbeunderstood asthe controlof perceptions, with perceptionscorresponding to importantas-

pectsoftheenvironment. Fora sexually reproducing organism to survive andleaveprogeny, it mustbeableto controlmanydifferent typesofperceptions, thatis,sensed aspects ofitsenvironment. Ata minimum, it must beableto findfood,avoidenemies, andmate.Butaslifeevolved, theenvironmentof our ancestorsbecamemorecomplexdueto increasingnumbers

ofcompeting organisms. Soit wouldhavebeenofconsiderable advantage to beabletoperceive andcontrolincreasingly complex aspects ofthisenvironment.ThebacteriumE. colicancontrolits sensingof foodand toxins

onlyina primitive way;organisms withmorecomplex brainsareableto senseand controlmuchmorecomplexaspectsof theirsurroundings.

Thiscapacity forincreased environmental control isnowhere morestrikingthaninourspecies. Using theadvanced perceptual-behavioral capacitiesof ourbraintogetherwithourculturally evolved knowledge of science

andtechnology, wecanvisitoceanfloors,scalethehighest peaks,andset footonotherworlds.(Therolethatlanguage isbelieved to havehadinthe evolutionof the humanbrainwillbeconsideredseparatelyin chapter11.)

But can the mostcomplexhumanabilitiesand mentalcapacities be explained bynaturalselection? Ourbrainhascertainly notchanged appre-

ciably overthelastcouple ofhundred years,andyetwecansolve mathematical, scientific, technological, andartisticproblems thatdidnoteven exist.a hundred yearsago.Sohowcouldnaturalselection beresponsible for thestrikingabilities oftodaysscientists, engineers, andartists?

56

Without Miracles

Thisisactually thesame problem thattroubled Alfred Russel Wallace, as mentioned inchapter 3.Itwillberecalled thatWallace, despite being an

independent codiscoverer ofnatural selection, could not,forexample, imagine hownatural selection could account forAfricans ability tosing

andperform European music,sincenothing in theirnativeenvironment couldhaveselected forsuchanability. Consequently, forhimthebrain couldonlybea creation provided tousbyGod.Wenowknowthatinhis embrace of thisprovidential explanation, Wallace failedto realizethat

natural selection canlead tonewabilities unrelated tothose thatwere origi-

nally selected.

Touseanexample fromtechnological evolution, thefirstpersonal com-

puterswereusedto performfinancial calculations intheformofelectronic

spreadsheets. However, thesesamemachines withthepropersoftware

could alsobeusedforwordprocessing, telecommunications, computer

games, andmanyotherpurposes, eventhough theywerenotoriginally

designed withthese functions inmind. Aclassic example ofthispheno-

menonoffunctional shiftin biological evolution isthetransformation of

stubby appendages forthermoregulation ininsects andbirds intowings for flight. 6 Inthesameway,selection pressure wasundoubtedly exerted on

earlyhominids to become betterhunters. Theability to understand the behavior ofotheranimals andorganize hunting expeditions musthavebeen veryimportant intheevolution ofourspecies. Andtheincreasingly com-

plexandadapted brain thusselected would have made other skills possible, suchasmaking toolsandusinglanguage, traitsthatinturncouldbecome

targets forcontinued natural selection. Thistransformation ofbiological

structures andbehaviors fromoneusetoanother wasgiven theunfortunate

nameofpreadaptation byDarwin, unfortunate since itcantooeasily be

misunderstood toimply thatsomehow evolution knows whatstructures

willbeuseful forfuture descendants ofthecurrent organisms. American evolutionary paleontologist Stephen JayGould provided abet-

tertermforthisphenomenon-.--exaptation. Hemadea majorcontribution

toourunderstanding ofevolution byinsisting thatwedistinguish adaptation,theevolutionary process through which adaptedly complex structures

andbehaviors areprogressively fine-tuned bynaturalselection withno

marked change inthestructures orbehaviors function, from exaptation,

through which structures andbehaviors originally selected foronefunction


57

becomeinvolvedin another,possiblyquiteunrelated,function.Exaptation

makesit difficult ifnotimpossible to understand whyourbrainevolved as

it did.Although thebrainallows usto speak,sing,dance, laugh,design computers, andsolvedifferential equations, theseandotherabilities may wellbeaccidental sideeffectsof itsevolution. AsGouldandhisassociate Vrba cautioned:

current utility carries noautomatic implication about historical origin. Mostof

whatthebrainnowdoesto enhance oursurvival liesinthedomainofexaptation

anddoesnotallowustomakehypotheses abouttheselective pathsofhuman his-

tory. Howmuch oftheevolutionary literature onhuman behavior would collapse

if weincorporated theprinciple ofexaptation intothecoreofourevolutionary thinking? 7

Butalthough wemaynever knowtheactual events andspecific selection pressures responsible forourbrainpower, wehavenoscientific reason to believe thatevolution couldnothavefashioned ourbrainthroughnatural selection. Thefactthatlivingorganisms todayhavenervoussystems and

brainsranging fromquitesimple to amazingly complex iscompelling evidencethat our brain evolvedthroughforgottenancestorsin progressive

stages fromsimple tocomplex. Andsomehow, asa partofthisevolutionaryprocess, thatmostremarkable andmystifying ofallnatural phenomena came into beinghuman

consciousness.

The Development of the Brain

Fromthisevolutionary perspective, onemightbeledto conclude thatour braininallitsstriking adapted complexity isaninherited legacy ofbiologicalevolution. Thatonceevolved it isthereafter provided to eachindividual

bygoodoldnatural selection, specified inallitsfinedetail inthegenome andtransmitted throughthegenerations fromparentto offspring.

Thistypeof genetically providential thinking of courseis selectionist fromtheviewpoint ofbiological evolution, butnonetheless providential at theleveloftheindividual organism. It canbeseeninthepioneering work onbraindevelopment andfunction ofRogerSperry forwhichheshareda

Nobel prize in1981. Thisresearch inthe1950s involved disturbing thenormallocationof nervefibersin thedeveloping brainsof fishandrats.For

example, nerve fibers thatnormally connect thetoppartofthefishsretina withthebottompartofthebrain,calledtheoptictectum, weresurgically

58

Without Miracles

removed

and reconnected

modification Similar fibers

that

innervate

muscles their

specified

Gazzaniga

explains

back

out

part to

of the optic its

by other

also

proper

to conclude

completely

In the original

grew

carried

and made

led Sperry

to the top

nerve

experiments

attached This

, the

normal

researchers

" knew

" to which

connections

that

despite

the connections

in the organism

tectum

. Despite

position

in

the

this brain .

on rats indicated muscle

they

surgical

should

be

disturbances

of the nervous

's genes . As his former

that

system

student

. are

Michael

:

Sperry view of the nervous

system , brain and body developed

under

tight genetic control . The specificity was accomplished by the genes ' setting up chemical gradients , which allowed for the point -to -point connections of the nervous system .8 But there

is a vexing

complete

information

and other

mammals

about

1015 (one

synapses

by

acters on Changeux

plus

unlike

similar

a computer noted ;

one

information

that

can

provides of humans alone

million

3 .5 billion

has

million

,)

(3 .5 x 109 )

30 % to 70 % of these ap -

capacity

information

have

concluded

to specify

on the location

that

all of these and type

for the rest of the body . The prob -

to save a document disk

about

scientists

storage

system neocortex

thousand

base pairs ), with and molecular

to including

trying

the genome

of the nervous

has only

do not have enough

, in addition

of each neuron lem is not

genome

neural

that

that just the human

15 zeros , or

(nucleotide

silent , l some

our genes simply

the notion

for the construction

.9 Since the human

connections

with

. It is estimated

followed

bits of information pearing

problem

hold

made only

up of 100 million

1 .4 million

characters

char . As

Once a nerve cell has become differentiated it does not divide anymore . A single nucleus , with the same DNA , must serve an entire lifetime for the formation and maintenance of tens of thousands of synapses . It seems difficult to imagine a differ ential distribution of genetic material from a single nucleus to each of these tens of thousands of synapses unless we conjure up a mysterious " demon " who selectively channels this material to each synapse according to a preestablished code ! The dif ferential expression ficity of connections Additional

of genes cannot alone explain between neurons . 11

understanding

nervous

system

referred

to as the water

familiar

to many

of the relation

can be gained

by considering

flea or daphnid

aquarium

owners

the extreme

between Daphnia

, this small

since

diversity

the genome magna

fresh -water

it is relished

and speci -

and

the

. Commonly crustacean

by tropical

is

fish . But


59

whatmakesthedaphnidinteresting forourcurrentpurposes isthatwhen thefemale isisolated frommales,shecanmostconveniently reproduce by theasexual process ofparthenogenesis, giving birthtogenetically identical clones. Inaddition, thedaphnid hasa relatively simple nervous system that

facilitates itsstudy. Ifitsgenome completely controlled thedevelopment of itsnervous system, it should bethecasethatgenetically identical daphnids shouldhavestructurally identical nervoussystems. However, examination

of daphnid eyesusingtheelectron microscope reveals thatalthough genetically identical clones allhavethesamenumber ofneurons, considerable variationexistsin the exactnumberof synapsesand in the configurations

ofconnections leading toandawayfromthecellbodyofeachneuron, that is,thedendritic andaxonal branches. Aswemovetomorecomplex organ-

isms, thevariability oftheirnervous systems increases. Thisprovides clear evidence thatthestructure andwiringof thenervoussystemarenotthe resultoffollowing a detailed construction program provided bythegenes. Howthenisthebrainableto achievetheveryspecificandadaptedwiring

required tofunction insomany remarkable ways? Forexample, howdoes a motorneuronknowto whichparticularmusclefiberit shouldconnect?

Howisa sensory neuroninthevisualsystem ableto joinitselfto thecorrectcellinthevisualcortexlocatedintheoccipital lobeofthebrain?Ifthis detailedneuron-to-neuron connection information is notprovidedbythe genes,whence does it come?

The first cluesto solvingthis puzzlego backto 1906whenit was

observed thatin embryonic nervetissue,someneurons didnotstainwell

andappeared to bedegenerating anddying. 12 Since it hadbeenassumed thatin a developing embryo, nervecellsshouldbeincreasing in number andnotdyingoff,thisfinding wassomewhat surprising. Butnervecell deathin the developing nervoussystemhas sincebeenobservedrepeat-

edly.Theextent to which it occurs wasdramatically demonstrated by

ViktorHamburger. Hefoundthatina certainareaofthespinalcordof thechicken embryo over20,000neurons werepresent, butthatintheadult chicken onlyabout12,000, or60%,ofthesecellsremained. 13 Muchofthis

neuronaldeath occursduringthe earlydaysof the embryosexistence. Nervecellscontinueto expirethereafter,albeitat a slowerpace.

Aparticularly striking example ofneuronal elimination indevelopment involvesthe death of an entire group of brain cells:

60

Without Miracles

Mostfrequently, neuron death affects onlysome oftheneurons inagiven category.

However, inonecase.. . a whole category ofcellsdies.These particular neurons of

layerI, themostsuperficial layerofthecerebral cortex, characteristically have axonsanddendrites oriented parallel tothecortical surface ratherthanperpen-

dicularto it,likethepyramidal cells.Thesecellswerefirstobserved inthehuman

fetusbuthave since beenfound inother mammals. Purely andsimply, theydisappear in the 14 adult.

Butthedeathofobviously useless braincellscannot account forthespecific connections thatareachieved bytheremaining neurons. Forexample,

thevisualcortexofcatsandmonkeys haswhatarecalledoculardominance

columns within a specific region known ascortical layer4.Inanyonecolumnofthisbrainareaintheadultanimalwefindonlyaxonsthatarecon-

nected totherighteye,while intheneighboring column arelocated only

axonswithsignals originating fromthelefteye.Sonotonlymusttheaxons

findtheirwaytoa specific region ofthebrain, which canbequitefarfrom wheretheircellbodies arelocated, theymustalsofinda specific address

withina certainneighborhood.

Theabilityofaxonsto connect to theappropriate regions ofthebrain during development hasbeenstudied incareful detail since thebeginning of thiscentury. Axons growinthebrainlikethestemofa plant.Attheendof

thegrowing axonisfound a growth conewhich wasdescribed bySpanish neuroscientist RamnyCajalin1909asa sortofcluborbattering ram, possessing an exquisite chemical sensitivity, rapidamceboid movements, anda certain driving forcethatpermits it topushaside,orcross,obstacles initsway.. . untilit reaches itsdestination. 15 Although theexactmechanismsbywhichthisisaccomplished arestillunknown, it appears thatthe growthconeissensitive tocertainchemicals alongitspaththatarereleased byitstargetregion. Inthiswayvisual system axonsoriginating inthelat-

eralgeniculate nucleus findtheirwaytocortical layer4intheoccipital lobe ofthebraininmuchthesamewaythata policebloodhound isableto sniff out the escapedprisonerhidingin an Illinoiscornfield.

Butalthough thesegrowth cones leadtheiraxons totheproper region of

thebrain(ormuscle inthecaseofmotorneurons), theycannotleadthem

totheprecise target addresses. Foraparticular growth cone, itappears that

anycellofa particular typewillserveas a target.Indeed,inthenewborn cat,ocularcolumns receive axonsfrombotheyes,notjustfromoneorthe

other, asintheadultbrain.Forthisfinalandimportant fine-tuning to be


61

62

Without Miracles

wouldhaveto waituntilthebirthof theanimal whenit isdelivered from thecomforting warmdarkness ofthewombto thecoldlightof day.However , recent evidence suggests thatthisfine-tuningactually begins to take place inutero . Prenatal development appears todepend onspontaneous firingofretinalcellsthatdonotdepend onlightstimulation fromtheexternal world.Similar endogenous patterns of activitymayalsoexistin thespinal cord,andmayrefinethesynaptic connections ofmotorsystems aswell.I? Nonetheless , interactive postnatal experience of theexternal worldis required for normaldevelopment of senses andnervous systems in mam mals . Catswhohaveoneeyesewnshutat birthloseallabilityto seewith thiseyewhenit isopened several months later . Thesame applies tohumans . Before thewidespread useof antibiotics , eyeinfections leftmanynewborn infantswithcloudylenses andcorneas thatcaused functional blindness , eventhoughtheirretinas andvisualnervous systems werenormalat the timeof birth. Years latera number of these individuals underwent opera tionsto replace theircloudy lenses andcorneas withclearones , butit was toolate.Contrary to initialexpectations , noneofthese people wasableto seeafterthetransplant .18It wassimplynotknownat thetimethatearly visualexperience wasessential to thenormalmaturation of thebrain's visualcircuitry . Similarly , somechildren arebornwitha wandering eye thatdoesnotfixatethesame partof thevisualfieldasthenormaleye , andotherchildren haveoneeyethatisseriously nearsighted orfarsighted ; in bothcases , theretinaof theabnormal eyemustbeprovided withclear visualstimulation , usually byagefouryears , orit will become functionally blindsince itsconnections tothebrain 'svisioncenters will beeliminated in favorofthenormal eye . Wethusseethatthenormaldevelopment of thebraindepends on a critical interaction between genetic inheritance and environmental experi ence. The genome provides the general structure of the central nervous system, and nervous system activity and sensory stimulation provide the

meansby which the systemis fine-tuned and made operational. But this fine-tuning does not dependon adding new components and connections in the way that a radio is assembledin a factory, but rather it is achieved by eliminating much of what was originally present. It is as if the radio arrived on the assemblyline with twice as many electricalcomponentsand connections

as necessary to work . If such an overconnected

radio

were

plugged in and turned on, nothing but silence, static, or a hum would be


63

heard from its speaker. However, careful removal of unnecessarycomponents and judicious snipping of redundant wires would leave just those componentsand connectionsthat result in a functioning radio. This snipping is analogousto the elimination of synapsesin the human brain as part of its normal development. The processby which brain connectionschangeover time as maturing animals interact with their environments has been studied in detail by psychologist William Greenough of the University of Illinois at UrbanaChampaign. Using sophisticatedtechniquesfor determining the numbers and densitiesof neurons and synapsesin specificregions of the rat's brain, he and his associatesfound that during the first months of the rat's life a rapid spurt in the growth of synapsesoccurs regardlessof the amount or type of sensory experience.19This period of synaptic " blooming" is followed by a sharp declinein the number of synapses . That is, an elimination or " pruning" of synapsesthen takesplace basedon the activity and sensory stimulation of the brain, and ultimately results in the configuration of connections characteristic of the mature rat's brain. Greenoughrefers to this initial blooming and pruning of synapsesas " experience-expectant" lea~ning, sincethe initial synaptic overproduction appearsto be relatively independentof the animal's experiences . It is as though the brain is expecting important things to be happeningduring the first weeksand months of life, and is prepared for these experienceswith an overabundance of synapses , only a fraction of which, however, will be selectivelyretained. The work of Greenoughand his associatesis limited to rats and monkeys, but their findings have much in common with those of PeterHuttenlocher of the University of Chicago who counted the synapsesin specific regions of the brains of humans who died at various ages. Huttenlocher found that: The increasein synapticdensityplus expansionof total corticalvolumeleaveno doubtthatthepostnatalperiodis oneof veryrapidsynaptogenesis in humanfrontal cortex. By age2 years, synapticdensityis at its maximum, at aboutthe sametime whenothercomponents of cerebralcortexalsocease growingandwhentotal brain weightapproaches that of the adult. Synapticdensitydeclinessubsequently , reaching by adolescence an adultvaluethat is only about60% of themaximum.2O This wealth of synapsesis thought to be responsiblefor the striking plasticity of the immature brain that permits the learning of skills that can be learned only with much greater difficulty or not at all by the already

64

Without Miracles

prunedadultbrain.Wealreadysawhowimmature animalsandchildren

areunable todevelop normal vision iftheyarenotexposed toa sharply focused visual worldduring thisperiodofbraindevelopment. It hasalso beenrepeatedly observed thatalthough manyadultsinitially maymake quiterapidprogress inlearning aforeign language, young children appear

to havean important advantage overadultsin beingableto masterthe

sounds oflanguages. Canadian childlanguage researchers JanetWerker andRichard Tees observed thatchildren younger thanoneyearappear able

todistinguish between thespeech sounds usedbyanyhuman language. By age12 months,however, theybeginto losethe abilityto discriminate

between sound contrasts thatarenotusedinthelanguage theyhearevery

day.Sowhereas allnormalinfantscandistinguish between thetworelated

butdistinct sounds represented bytheletter tinHindi, those whohearonly

English quickly losethisability, andHindi-speaking children retainit. 21 The workofWerker andTeestherefore provides important humanbehavioral

evidence thatisconsistent withtheviewthatnormal braindevelopment

involves thelossofsynaptic connections, whichresultsinthelossofcertain skillsas the brainapproachesits adultform.

A sensitive periodfortheacquisition of a firstlanguage wasdemon-

strated bytheplight ofGenie, anAmerican girlwhowasbrutally isolated fromallnormal human interaction untilshewasfound atage13years, andwhoneversubsequently developed normallanguage abilities. 22 There is strikingevidence that the immature, overconnected brainis alsobet-

tersuitedthana mature oneto acquiring second languages andsign

languages. 23

Takentogether, thesefindings painta picture ofthedeveloping brain

thatcontrasts sharply fromthegenetic providentialism favored bySperry. Instead ofthebrainunfolding according to a genetically specified blueprint,weseeinstead a process ofselection bywhichoverly abundant neu-

ronalconnections areeliminated through a weeding-out process, leaving onlythoseconnections thatpermit theanimal tointeract successfully with its environment. LearningandMemory:RewiringtheBrain

Themammalian brainappears mostadaptive during theearlypostnatal period, andcontinues toadaptandlearn fromnewexperiences throughout


65

its adult life. During the 1960s and 1970s a seriesof studies offered impressiveevidencethat rats grew thicker brains and new synapseswhen they were placed in complex and challengingenvironments. Thesefindings were consistent with the then-popular belief that learning and memory in mature mammals (as opposed to the brain development of immature animals) were additive processesinvolving the formation of new synaptic connectionsor the strengtheningof already existing ones. The influential Canadian psychologist Donald Hebb assumedthat " the changedfacilitation that constitutes learning" was the result of " the growth of synaptic knobs." 24Similarly, Sir John C. Eccles, who shareda Nobel prize in 1963 for his researchon the transmissionof nerveimpulses, believedthat memory and learning involved " the growth . . . of bigger and better synapses ." 25 However, it was also suggestedthat more than just adding synapses was involved in learning. One of the first to propose that subtractive brain changescould be involved in adult learning and memory was J. Z . Young, who in 1964 posited that such learning could be the result of the elimination of neuronal connections.26Severalyears later J. S. Albus hypothesizedthat " pattern storage must be accomplishedprincipally by weakening synaptic weights rather than by strengthening them," 27and Richard Dawkins speculatedthat the selective death of neurons could underlie the storageof memories.28 But how could a subtractiveprocessof neuron elimination be involved in learning and memory? It is particularly difficult to understand how the learning of a new skill, such as riding a bicycle or speakinga foreign language, or acquisition of new memories, such as learning the words to a poem or song, could be made possibleby loss of synapses . We saw in the developmentand maturation of the brain that synapticconnectionsthat are rarely used are weakened or eliminated, whereas those in active neural pathways are retained and perhapsstrengthened. This subtractive process makes sensewhen dealing with an overwired, immature brain that may havecloseto twice asmany synapsesasit will haveasan adult. But how can it work for a mature adult brain that has already been substantially whit tled down by synapticpruning? To illustrate this problem, imaginean adult Spaniardlearning English. To do this, the Spaniardwill have to learn to hear and produce certain sound distinctions that are not used in Spanish, such as the contrasts involved in

66

Without Miracles

ship

versus

Werker tially

, sue

versus

Tees

would

lead

be able

guage

he

we

to

would used

how

any

further

the

Instead

, or

learning

to

take

modify

place

! Surely

Spaniard

to , and of

the

? At

trick

have

to tryout

to select the

the

initial

Jean

possible

of

the

there

In

tion

per

proliferate

effect

when , he

occurring of

necessarily

him

away

to

the

lan -

he was

learn

born

,

since

they

not

clear

is therefore

the

this

aspect

new

of changes

in the

that

the and

of variation

is necessary

connections

present

the

con will

the

do

somehow

best

ones

, perhaps

in

the lan -

synaptic

would

select

add

other

synapses

brain

of

to allow

and

of

this

problem

should

English

combination

appear

of new for

synapses

changes learn

result new

addition

thorny

which

synaptic adults

combinations

synaptic

the

be necessary

equally

know

of

it would

involving

into

ever

which

to

this

in

1983

scheme

activity

its

, culture

makes

and

birth

synapses at the

the

not

. But

unlike

immature

from

, over

-

birth

that

, but same

of

all

that time

brain

and

the

progressively

. The . On

in man . With . This

causes

10 , 000

the

or

contrary

each

a series

,

wave , of criti -

effect .3O

adaptive

puberty

the

immediately

stabilization

(published a " Darwinism

.

to puberty

its regulatory

and

Neuronal

development

established

neurobiologist

proposed

environment

selective

French

L "'Homme

its impression

are not

exercises

by

) , Changeux

cultural

waves

suggesting

between

present

for

offered

book

Man

neuron

redundancy

was

. In his

account within

cortical

was

riddle

as Neuronal

" 29 to

preexisting all

run

be the

just

in successive

is transient

cal periods

those

1985

to this

so synapses

pruned . It

ini -

connections

when

ones , would

many

of new

of

it undergoes

According

in

of

not

.

synapses

learning

we

ones , a source

Changeux in

made

synaptic

permit

a process

could

must

least

solution

- Pierre

English

then

, since

a number

brain

in

they

very

best

research

would

not

present

promptly

current

combination

. But

variation

connected A

brain

the

life . The

would

that

brain

learning

the

are

environment

synapses

. But the

English

such

nections

his

. The

Spaniard

they

were

been

of

likely

, some

learn

have

reorganizing

how

his

wash

the

.

more

at least

to

of

language

understanding

guages

language

, it seems

synapses

or

them

in

pruning

English

all

discriminations

expect

versus

that since

spoken

these

the

watch

predict

distinctions

and

making

not

zoo , and us to

these

heard

for

were

of

make

has

necessary but

sheep

and

brain

changes

in humans

these . From

, involve

preexisting birth

synapses to

puberty

, or the

at

least

elimina were

, Changeux

not


67

hypothesizedthat waves of synaptic growth would occur, with subsequent experienceserving to retain the useful ones and eliminate the uselessand redundant ones. These waves of synaptic overproduction would provide the source of variation on which synaptic selection could operate. Such learning resulted in an absolute increase in synaptic growth and numbers over time. This growth was not constant, but was rather envisionedas analogousto repeatedlytaking two stepsforward- randomly adding new synapses - followed by one step backward- eliminating the uselessones just added. Changeux provided no hard evidencefor his hypothesis that synaptic variation in the form of overproduction would precedethe elimination of synapsesas part of the brain's restructuring to permit the learning of new skills and acquisition of new knowledge. But suchevidencewas found a few yearsafter the publication of his book. William Greenoughand his associates, whose work on the maturational developmentof the rat's brain was noted earlier, also conducted research on changesin the brain induced by placing adult rats in special, enriched environments. In one study this resulted in a 20% increase(roughly 2000) in the number of synapsesper neuron in the upper layersof the visual cortex.31Later researchshowedthat such dramatic increasesin synapseswere not restricted to the rat's visual cortex.32 Theseand other similar findings led Greenough's group to proposethat the waves of synapse proliferation first described by Changeux could be elicited by the complex demandsplaced on the adult brain in a new, challengingenvironment. Theseresearchersreferred to this processas " experience-dependent" developmentsinceit dependson the environmenttrig gering the formation of new synapticgrowth on which the selectiveprocess can act.33 Greenough's conceptionof how the adult brain is able to learn new skills and form new memoriesoffers an appealing solution to the problem concerning the additive and subtractiveprocessesunderlying the adult's brain adaptation to new environments. According to this theory, experiencedependentlearning combinesboth additive and subtractive processes . The additive component involves the blooming of new synapsesin responseto the animal's attempt to control aspectsof a new, complex environment. Although the brain does appear to know what part of itself has to be


69

formed synapsesin the adult brain in responseto environmental changes. Such a finding would place the brain alongside the immune system as another striking example of how cumulative variation and selectionprocessesduring the lifetime of an organism make it possibleto adapt to complex, changingenvironments.

Wehavenowseenthatunderstanding boththeadapted andadaptive complexityof thehumanbraininvolves findinganswers to threequestions : how didthebrainoriginateasa biologicalorgan?; howdoesit develop froma fertilizedeggintoamaturebrain?; andhowisit ablewhenmatureto rewire itselfto learnfromandadaptto changes in itsenvironment ? Muchmoreworkmustbedonebeforewehavedetailed answers to these questions . Butsubstantial progress hasalready beenmadeaswemovemidwayintothe"decade of thebrain." Toa largeextentthisprogress hasconsistedof rejectingprovidentialandinstructionist explanations for these puzzlesof fit, andfindingconsiderable evidence andreasonin favorof selectionist explanations . Thepowerfulprocess of cumulative blindvariation andselection workingovermillionsof yearsis not onlytheonlyreasonable theoryfor thebiological evolutionof theb~ain, butwefindthatit hassurfaced againin a differentbutstill recognizable formasanexplana tion for thebrain's embryonic growthandcontinued development during itsrelatively brieflifetime . It is here,asChangeux remarked , that "theDarwinismof synapses replaces theDarwinism of genes ."38Toclosethecircle,it shouldbenotedthat a strikingconsequence of thejoint effectsof among -organism genetic and within-organism synaptic selection is thebrain's understanding bothitself andtheprocess of selection thatisresponsible for itsextraordinary abilities .

III The Promise of Selection

6 The Origin and Growth of Human

Knowledge

Providence

Now

,

to

try

the

truth

boldly

present

-

if

of

to

-

that

things

is

inquire

is

,

always

into

what

you

do

in

and

recollect

not

remember

our

soul

,

the

what

soul

is

you

do

immortal

not

.

happen

So

it

to

is

right

know

at

.

Plato1

Instruction

Let

us

then

without

store

:-

the

variety

answer

,

one

mind

How

busy

?

in

John

the

ideas

which

endless

-

suppose

any

to

and

has

,

,

it

from

as

we

to

boundless

Whence

word

be

comes

say

be

fancy

it

all

the

,

paper

of

?

man

materials

experience

white

furnished

has

of

,

void

Whence

of

painted

reason

all

comes

it

on

and

it

characters

by

with

an

knowledge

,

that

?

vast

almost

To

this

I

.

Locke2

Selection

Our

categories

adapted

is

already

the

-

and

to

fish

is

Konrad

the

limits

of

of

of

to

adapted

to

time

the

of

human

perception

for

of

water

",

exactly

ground

the

fixed

prior

the

the

to

same

steppe

before

Socrates

inquiry

,

fish

individual

reasons

before

the

to

today

and

can

.

for

the

hatches

as

experience

the

horse

is

hoof

of

born

and

,

the

are

horse

the

fin

of

.

,

as

knowledge

origin

.

major

for

the

,

the

extent

instructionist

validity

,

philosophers

to

the

,

nature

to

continues

account

explore

providential

,

interest

examine

to

and

the

much

, 4

will

attempt

universe

of

of

epistemology

we

to

understood

human

study

been

as

Here

taken

our

be

the

has

referred

have

knowledge

,

knowledge

philosophy

explanations

of

world

adapted

philosophers

proaches

forms

external

Lorenz3

Since

field

the

be

.

a

major

focus

approaches

puzzle

to

of

which

,

and

and

This

that

fit

between

these

selectionist

ap

-

WithoutMiracles

74

Knowledgeby Recollection Plato

, who

provides knowledge

in

with

the

. His

knowledge our

possess

seen

during

first

to

senses

knowledge

never

Athens

well

. For

example

about

) a perfect

one , since

any

thus

and

touch

be the

sale

can

imagine

perfect

object

or

quality

. The

our

event

to

Chomsky

sensory

experiences 's problem

problem between

of

how

will

put

forth

you

ever

know

Meno are

you

what

already

that

fact

world must

we

have , re -

although , no

and

we

of

such

so much events

American

,

particular

instance

know

objects

that

the

see , hear , smell

, since

can

belief

that

justice

a perfect

we

of

examination

we

, or

influential

this

a

, given

, has

linguist

been Noam

for , how

will

knowledge

you

the

for

it ? Socrates

recognizes

and

paraphrases

it thus

I know

, Mello

ducing

. You that

, what argue

which

if not , he cannot enquire .?

; for

fully kinds

of

when

did

concerning

utility

you find

not

of their

question

the

essence

do not what

know

have mustn

and and

know

you

's the

inquiry

? What

will

, how

will

want

?6

possess

you

, then

in

mean ; but

he

in Plato

the

found

knowledge it ? And

' t it be the

therefore importance

if you

case

have

you

no

of Meno

that

do you

need

to

's ques -

:

a man

he does not

appreciate two

you

that

the

which

' t already

seeking

knowledge

conversation

doubt

if you

you

don

know are

look

that

? And

which

, if you

you

their to

appears

: , into

is the thing

of knowledge

. In

begins

of enquiry

asking

source

Meno

, Socrates

subject

tion

between

the

dilemma

effect

possess

To

be

particular

the

' friend

enquire

as the

is in

looking

about

of

and

, Socrates

an intriguing

you

know

ever

his

therefore

close

knowledge

B .C .,

origin

about

the

world

, beauty

how

" by

concerning Socrates

virtue

presents

And

can

by

( and

, on

the

our

understanding

the

.5

dialogue

and

experience of

will

that

centuries

learn

despite

circle

of

can

imagine

circle

source

fifth

influenced we

can

, goodness

as " Plato

A related

nature

our

problem

limited

referred

in

, we

reasoned

circularity

and

concerning

much what

particular

. Plato

cannot

very

beyond

imperfections

fourth

discussions

was

go

the

written

philosophy

seems

through

veal

lived us

cannot know

does

Meno knowledge

not

just

see what

enquire ; for

a tiresome

either

if he knows

know

's dilemma . Clearly

the

very

, we , one

about

dispute that

, he has no need subject

have can

to

you

which

about

make

be ignorant

are intro

he knows to enquire which

a

, or ; and

he is to

distinction of

a certain

-

TheOriginandGrowthofHumanKnowledge 75

factorpiece ofinformation, makeanattempt tofinditout,and,ifsuccessful,bequitecertain thatwhatwasfound outwaswhatonehadwanted to knowinthefirstplace. Forexample, itiseasytoimagine notknowing the

telephone number ofanacquaintance, looking itupinthetelephone directory,andthenknowing thatthenumber sofound isindeed whatonewantedto know.In thiscasewhatweseemto bedoingis simply fillingin a

specific pieceoffactual knowledge. However, sinceMenosquestion arose inthecontext ofdiscussing thenatureofvirtue, itisobviously notthiskind offactual knowledge heandSocrates foundtroublesome. Thatproblem hasto dowithconceptual knowledge. 8 Ifyoudonotalready knowwhat virtue is,howwillyoueverdiscover itsmeaning? Now,ifvirtue issimply somecombination ofotherconcepts already wellunderstood, saya mixof one-third goodness, one-third fairness, andone-third moralstrength, itmay bethecasethatyoucouldsimply betoldwhatvirtue iswithrespect tothese otherconcepts andthereby acquire a newconcept. However if,asisusual-

lythecaseforconcepts, thereisnogenerally agreed uponwaytodefine a newconcept assome easily understood combination ofoldones, thefull force of Menos dilemma becomes apparent.

Letuslookat anotherexample inwhatisperhapsthemorefamiliar context of a child learning to add.

Children inthepreschool yearshavebeenfoundtomakea transition, without any

instruction, fromacrude addition algorithm toa moreefficient one.. . . According totheearlier algorithm, theproblem of4+3=?issolved byaprocedure analogous tocounting outfourblocks, thenthreeblocks, andthencounting thecombined set. Amoreadvanced algorithm consists ofstarting withfourandcounting threemore.

Akeystepinthistransition iseliminating thecounting outofthefirstaddend. To

attribute thissteptosome kindofinsighttorealizingorseeingthatcounting isunnecessary because theresulting number isalready givenis,ofcourse, totum-

blerightintothelearning paradox [thatis,Menos dilemma]. Forsuch aninsight presupposes anunderstanding ofthemore sophisticated procedure inadvance of discovering it. 9

Thispatternofbehavior wasdemonstrated bymysonwhenhewas aboutfouryearsoldinhisinteraction witha computer program designed toteachaddition. Toteach4 +3 =7,thecomputer program woulddisplay images offourobjects (say, puppies) ina boxontheleftsideofthescreen withthreelikeobjects ina boxontherightside.Above theimages wasthe equation 4 +3 =? withthenumber 4 directly above theboxcontainingthefourpuppies andthenumber 3 directly above theboxcontaining

76

WithoutMiracles

the

three

puppies

puppies

and

efficient

looking

three

see

he

method

he

constitutes

Bereiter

own

the

crux

's

is

this

a

of

and

therefore

to

able

dilemma

the

here

,

issues

simply

to

question

solve

seemed

to

the

by

efficient

a

somehow

few

days

the

corre

-

managed

this

new

above

and

note

that

,

prescience

's

we

to

knowledge

a

an

( "

in

origin

making

to

.

was

growth

appeal

to

know

It

to

chapter

and

insight

before

would

"

)

imply

available

knowledge

solution

consider

the

already

new

as

will

dilemma

somehow

be

he

concerning

Meno

was

less

of

acquire

but

obtained

the

within

quotation

foreknowledge

nothing

that

understanding

to

,

that

.

in

now

as

of

pleased

realize

to

that

was

insight

for

an

cardinality

he

to

same

not

set

strategy

seem

returned

could

initially

spontaneously

with

interest

,

what

back

time

important

us

he

was

of

the

first

efficient

' t

the

own

move

and

more

didn

was

his

this

's

this

I

the

-

to

instead

or

that

the

child

,

actually

old

.

But

this

acter

is

of

knowledge

,

,

been

exactly

)

he

but

area

of

results

have

of

known

learned

acquired

through

is

a

as

of

during

some

'

life

previous

as

evidence

,

and

recollected

questions

the

response

the

,

view

that

,

argues

's

known

only

concern

Socrates

must

already

were

must

have

.

all

Thus

knowledge

in

-

takes

these

soul

a

never

problem

they

-

of

having

boy

that

is

not

since

char

pursuit

,

.

slave

boy

the

boy

a

that

slave

the

provides

slave

of

and

of

a

He

solve

uneducated

,

to

to

that

geometry

an

's

series

existence

demonstration

or

Meno

able

answering

( through

devoted

.

how

nonetheless

demonstration

Plato

unanswered

of

by

present

was

Meno

fundamentals

his

remembered

leave

,

that

Socrates

demonstration

square

the

Socrates

essence

a

answer

Since

geometry

this

during

.

not

a

in

unsatisfactory

offers

could

first

educated

the

the

Socrates

response

in

Let

in

knowledge

method

,

particular

of

I

in

count

,

count

.

,

instead

to

solutions

to

the

first

puppies

such

need

he

the

continuing

provided

use

on

few

no

that

how

raises

does

knowledge

only

a

was

the

then

all

used

all

After

make

left

Meno

foresight

clairvoyance

the

the

of

.

type

of

. "

counting

initially

counting

number

new

of

word

knowledge

as

to

So

Just

dismissal

problem

how

.

7

the

ordinality

his

,

he

and

there

discover

method

between

on

the

,

efficient

spondence

acquire

!

understood

more

6

difficulty

,

equation

,

use

this

bjects

5

no

answer

of

the

"

able

using

0

had

right

that

to

obtained

the

as

was

the

of

just

son

all

"

him

dismayed

answer

to

4

right

to

son

the

above

"

could

my

quickly

counting

ing

the

out

that

with

first

on

since

then

of

the

pointing

objects

my

up

described

at

objects

resist

to

Although

procedure

just

9

.

coming

philosophy

not

been

Plato

,

is

TheOriginandGrowthof HumanKnowledge as the doctrine

of recollection

77

, or to use the Greek word , anamnesis . As

Plato explains : Now, if the truth of things is always in our soul, the soul is immortal . So it is right to try boldly to inquire into and recollect what you do not happen to know at present - that is , what you do not remember .1O

Plato's view that all knowledge is simply recollectedmay have seemed reasonableat the time. We seetoday, however, that it is seriouslydeficient in a number of important respects, of which we will consider just two . First ,

the doctrine of recollection doesnot adequatelyaddressthe core of Meno's original question, sincesubstituting the words recollectand recollection for inquire and inquiry in the original shows that the essenceof the original dilemma remains. In other words, if you don't already know what you are trying to remember, how will you know when you have indeed remembered it ? And if you do know what you are trying to remember, then it must be the case that you already have access to the knowledge in question and therefore

have no need to remember

it !

Second, and perhapsof greater importance, the doctrine of recollection doesnot addressthe issueof the origin of knowledge. If Meno had wished to continue his discussionwith Socrates, he might have asked where the soul obtained its knowledge in the first place. From Plato's other writings, it is quite clear that Socrates would have answered that the soul is immor -

tal and therefore its knowledge simply has no beginning and no end. Sowe

seethat Plato'sproposedsolutionto thequestionof theoriginof knowledge is a providentialone, with no beginningandno endin muchthe sameway that religions view the existenceof God.

KnowledgeProvidedby a Benevolent God The scientific revolution that began in Europe at the end of the Middle Ages

was accompaniedby increasinginterest in the problem of knowledge, particularly the problem of the reliability of our perceptionsof our surroundings. Is the universe actually as it appearsto be? Can we trust our senses that fire is hot , that water

is wet , and that rocks are hard ? Or are these

impressionsmerely illusions, perhapsin the way that our dreamsappearto be? Although sciencewas beginning to make remarkable progressat this

time, its achievements wereaccompanied by a fair amountof philosophical

78

WithoutMiracles

skepticism concerning thetruth andaccuracy of thelawsof naturebeing discovered . Arethesescientific lawsto becompletely trusted , or arethey illusionsor at bestonlylikelyto betrueor onlyapproximately true?The first personin Renaissance Europeto consider theseissues seriously was ReneDescartes , wholivedfrom1596to 1650. Descartes began withaskeptical viewof humanknowledge andhopedto provebyhis"methodof doubt" thatwecanindeedtrustourimpressions of thephysical world. Histhinkingonthismattercanbesummarized asfollows. First, herealized thatall hisknowledge wassubject to doubt,except theknowledge that, to doubtin thefirst place , theremustcertainlybea doubter . Fromthisreasoning arosehisfamousdictum,cogito , ergosum("I think, therefore 1 am"). Second , Descartes convinced himselfthat some thingexternalto himselfexists . He reasoned that the causeof anything musthaveasmuchperfection aswhatit hascaused . Therefore , sincehewas ableto imagine anentitythatistheperfectinstantiation of all thatisgood, suchanentity(thatis, aperfectly goodGod) musttrulyexist.Thethirdand final stepin Descartes 's methodwasto conclude that a benevolent God wouldnot deceive . Therefore wecanbeconfident thatourimpressions of theworlddoindeedcorrespond to whatisreallythere. Descartes consequently believed thatGodplayedanindispensable rolein humanknowledge . NotonlydoesGodprovideassurance thattheworldwe experience corresponds to theworldthattruly exists , He alsoprovides us with certaininnateideasthatarenot in anywaybasedon experience , for example , theveryideaof Goditself, andotherabstractconcepts suchas beauty , goodness , justice , and virtue. Because Descartes 's epistemology depends cruciallyon the knowledge and assurance providedby an allknowingandbenevolent God, it is, likePlato's, a providential view.Butto theextentthatourknowledge alsodepends oninformation gathered byour senses , it isalsoaninstructionist viewin whichthereliabilityandvalidityof sensory instruction arecertifiedandguaranteed byGodHimself .II Descartes 's arguments havesincebeenmuchcriticized , especially his proofof theexistence of a benevolent andall-knowingGod, buttheyare widely regardedas markingthe beginningof modernphilosophical thought . Thebeliefthatthebasisof humanknowledge cannotresidesolely in sensory experience isstillconsidered a keyinsight . Andthequestto provideajustifiable andrationalbasisfor knowledge isverymuchalivein philosophytoday.

TheOriginandGrowthofHumanKnowledge 79 KnowledgeInstructedby the Senses

ThreeBritishphilosophers wholivedduringthe seventeenth andeighteenthcenturies developed theories thatareinstriking contrasttotheprovi-

dential epistemologies ofPlatoandDescartes. Since theyallemphasized the roleofsensory experience intheacquisition ofknowledge, theyarereferred to as the British empiricists. 12

Thefirstofthese,Englishman JohnLocke(16321704),completely re-

jected theconcept ofpreexisting innate ideas andargued thatallknowledge hasitsorigin insensory experience. Tohimthemindat birthwaslikea blankslate,a tabularasa.TorepeatpartofLockes epigraph usedat the beginning of this chapter:

Letusthensuppose themindtobe,aswesay,whitepaper, voidofallcharacters, without anyideas:How comes ittobefurnished? .. .Whence hasitallthematerialsofreasonandknowledge? TothisI answer, in oneword,fromexperience.

It was thereforethe sensoryexperiences providedby vision,hearing,

smell, andtouchthatwroteonthemind, leaving impressions thatwould thenbeourknowledge oftheworld.SinceLockesawtheroleof sensory

experience asthatoftransmitting knowledge tothemindfromtheoutside world,hisepistemology isessentially nonprovidential andinstructionist. Butalthough hebelieved thatthesenses weretheoriginofallhuman knowledge, hedidnotassume thattheyprovide accurate knowledge con-

cerning allaspects oftheexternal world. Hemadea distinction between whathecalledtheprimaryandsecondary qualitiesof things,theprimary

including shape, weight, number, andmovement, andthesecondary includingcolor,taste,smell, texture, andtemperature. Hebelieved thatour senses provide accurate knowledge abouttheprimary qualities sothat, forexample, ifweseeorfeela roundobject, it isthecasethattheobject actually isround. Butthisisnotthecaseforthesecondary qualities (such as the tasteor colorof a lemon)sincethesearesensations producedby an objectanddonotreflect theproperties oftheobjectitself. Lockesideaofsecondary qualities wasanimportant recognition thatnot

allourperceptions oftheexternal worldnecessarily indicate theactualstate oftheworld.Butinmakingthisadmission, heopenedthewayforothersto doubtthatourperceptions correspond toanything ina real,material world at all.Onesuchphilosopher wastheIrish-born Anglican BishopGeorge

Berkeley, wholived from1685to1753. LikeLocke, hebelieved thatall

80

WithoutMiracles

knowledge wasbasedon sensory experience . However , whereas Locke insisted thatourperceptions of primaryqualities provideknowledge about whattheexternal , physical worldisreallylike, Berkeley logicallyconclud edthatit isnotpossible to checkwhetheranyof ourperceptions accurate ly correspond to a realworld. If all wecaneverknowabouttheworldare ourperceptions of it, howcanweeverknowwhetheranyof ourpercep tionsactuallydocorrespond to physical objects andevents ? Berkeley 'sconclusion wassimplyto denythatanindependently existing , physical worldexists . Thiswasin somerespects aquitelogicalcontinuation of the train of thoughtbegunby Locke , andthis immaterialist , idealist philosophy makesevenmoresense whenit is realizedthat theprimary purpose of thebishop 'swritingswasto argueagainst"skepticism , atheism , andirreligion ." Accordingto Berkeley , whatwe experience throughour senses isdueto thedirectactionof God, andsoit isthroughthesenses that Godcommunicates with usandinformsusasto whatthingsaregoodfor usandwhatareharmful . As onecontemporary criticof Berkeley put it, "Roughlyspeaking , [Berkeley 's] immaterialism iswhatyougetif youstart offwithLocke 'spictureandreplace matterbyGod." 13 Wetherefore seethat Berkeley 's epistemology is bothprovidential , in thatall knowledge is providedby God, andinstructionist , in thatall knowledge istransmitted to us throughthesenses . Scottish philosopher DavidHume(1711-1776)wasthelastof thethree Britishempiricists andarguably themostinteresting , troubling , andinfluential. He wasverymuchimpressed by Newton's success in discovering lawsof physics , andattempted to applyNewton'8experimental methodto understand thecontentandabilitiesof the humanmindandto createa methodfor discovering truth. However , Hume's attemptto bescientific andempirical in hissearch for truthledhimto aninescapable anddistressing conclusion . Lockebelieved thatfromourexperience of theworldwecanknowthatanexternal world existsandknowat leastsomeof its actualcharacteristics . Berkeley concludedthatwecanknowonlythatourideasandGodexist.Hume,almost in spiteof himself , reasoned thatwecannotclaimto knowanythingabout anexternal world(or evenif it exists ), aboutGod, or aboutourownmind. LikeBerkeley , Humerecognized that we canneverknowtheexternal worlddirectlysinceall thatweknowof sucha worldareoursensory perceptionsof it andthe ideasthat theseperceptions generate . But unlike

The Origin and Growth of Human Knowledge

81

Berkeley whoconsequently rejected beliefinamaterial world,Humeinsist edthatbyourverynature wearecompelled tobelieve thatexternal objects existdespite thefactthatsuchbeliefisclearly irrational . Hisconclusion is therefore essentially thathumans areatthecoreirrational in theirbeliefs thatsensory experiences canreveal to themanything of anexternal , materialworld.14 Letusbeabitlessskeptical foratleastamoment andaccept forthesake ofargument ourintuitions thataworldofrealobjects exists independent of ourexperience ofthem , andthatourperceptions provide generally accurate information abouttheworld.If wegrantthismuch , canwenotestablish an empirical basis forknowledge ?Unfortunately not, asHumedemonstrated inpointing outtheproblem ofinduction . Induction refers totheprocess by whichwederivegeneral knowledge based on observations of a limited number ofinstances .Forexample , afterhaving eaten bread afewtimes , one mightbeexpected to believe thatbreadis anedibleandnourishing substance , andnothesitate to eatit inthefuture . Thatis, if thebread thatone haseaten sofarhasbeen nourishing , onemightwellconclude thatallbread isnourishing . Thisseems reasonable enough untilonerealizes thatnomatterhowmuchnourishing breadoneeats , thiscanin nowayguarantee (or evenmakeit probable ) thatsomebreadsomewhere is notnourish ing. Indeed , thebreadbaked bythelocalbakertomorrow couldpossibly becontaminated withtheergotfungus , resulting in bread thatcouldcause illness . Theproblem ofinduction asraised byHumehad,andcontinues tohave , asignificant impact onepistemology , particularly onthephilosophy ofscience . In essence , Humerealized theimpossibility of aninstruction -depen dentexplanation of knowledge in thatnoamount or kindof sensory experience (which , evenif trustworthy isalways limitedto aparticular time, place , andcontext ) couldeverresultincertain , justifiable knowledge inthe formofuniversal generalizations orlawsofnature . Inessence , hecouldnot escape fromtheconclusion thatallhuman knowledge mustbefallible , and thatnokindor amount of experience , logic,orreasoning couldbetrusted to eliminate thepossibility of error . In thissense , heeffectively destroyed empiricism byrevealing theirrationality of human beliefbased onsensory experience andreflections onthisexperience . AsBertrand Russell , British philosopher , mathematician , socialreformer andNobellaureate in litera turenoted , "thegrowthofunreason throughout thenineteenth century and

82

WithoutMiracles

whathaspassed of thetwentieth century is a naturalsequel to Hume 's destruction ofempiricism ."15 Although thischapter ismeantto bephilosophical , it shouldbenoted before leaving thissection thatmuchpsychological research hasrevealed theinadequacy oftheempiricist viewthatoursenses provide uswithtrustworthyinformation abouttheworld.Perceptions of thesame objectcan varyfromperson to person , andevenwithinthesame person . Thechemi calphenylthiocarbamide (PTC)hasanunpleasant tastefor somepeople andnotasteatallforothers . Plunge yourrighthandintoacontainer ofhot waterandyourlefthandintooneof coldwater . After15seconds or so, place thembothintothesame container ofwarmwaterandtheverysame waterwillfeelsimultaneously cold(bytherighthand ) andwarm(bytheleft hand ). Observe anoarthathasbeen placed atanangle in waterandit will appear bent , although youknowthatit isstraight . Thethreevisual illusions in figure6.1 showobjects thatappear to beof different lengths andsizes although arulerindicates theyareactually thesame . Manyotherexamples ofhowwecanbefooled byoursenses couldbegiven , andofcourse amagi cian'slivelihood depends onsuchmisperceptions . Wewill alsoseeinchap ter11howourunderstanding ofspoken andprinted wordsmaydepend as muchon ourexpectations andpriorknowledge asontheactualwords

. .

Figure 6 . 1 Three visual illusions . The three horizon lines appea to be of diffe len , although they are all the same length . The top solid squa seem larg tha th bottom one ,but they are both of identica size .The left arrow sha loo lon than the right one ,but they are equal length

TheOriginandGrowthof HumanKnowledge

83

themselves . Addtothistheproblem ofinduction raised byHume , andit should notbesurprising thatfewphilosophers consider sensory experience tobeanabsolutely reliable source ofknowledge oftheexternal world . The Importance of Prior Knowledge

experiences .17 In our pursuit of a naturalistic , nonmiraculous account of the knowled thatappears tobeprior tosensory experience ,weonce again meet Konrad Lorenz ,who , inaddition tohisimportant contribution totheunderstand ingoftheevolution ofanimal behavior ,made animportant contribution to epistemology . Inapaper firstpublished in1941 , heargued thattheneces sary apriori knowledge , including concepts ofspace , time , and causation , isactually theproduct ofthebiological evolution ofthehuman nervous sys tem .isAssuch , thisknowledge does notresult from thelimited experience ofanindividual butrather isthehard -won product ofthelong and arduous

84

Without Miracles

This admittedly cursory discussion ofhowphilosophy hasdealt withthe puzzle offitofhuman knowledge toitsenvironment cannot pretend todo justice tothevast amount ofthought devoted tothisissue . Nevertheless ,

The Origin and Growth of Human Knowledge

85

we have touched on the major themesand approachessincePlato's time,

and detecteda trend from providentialto instructionistto selectionist epistemologies . BeforeDarwin, philosophyhadessentially two waysof accounting for human knowledge- providence and instruction. Plato's doc-

trine of recollectionis a ratherpureprovidentialepistemology . Locke's and Hume's empiricism emphasizessensory-based instruction of knowledge from the environmentto the individual . Hume recognizedthe essentialirra-

tionality of suchknowledge . And theepistemologies of Descartes , Berkeley , and Kant are basedon a mixture of both providentialismand sensory instruction

.

It was only after Darwin 's revolutionary theory becameknown that a third perspectivewas imaginable; namely, that human knowledge owed its origin and developmentto somethingother than providenceor instructive

sensoryexperience . Although certainlyno philosopherhimself, Darwin made possiblea reconceptualizationof knowledge as a type of adaptation of the brain to its environment , an adaptation resulting from the same processes of cumulative

blind

variation

and selection

that

underlie

the

adaptation of other biological structures and behaviors. Both Lorenz and Popper present such an evolutionary epistemology, although this Dar-

winian perspective is embracedby only a smallminority of philosophers today . As evolutionary biologist Ernst Mayr observed:

Mayrmakesan interesting pointhere,but hegoestoo far in implying that manyphilosophers find Darwin's selectionist theoryof evolution incomprehensible . Instead , the philosophers who takethe timeto reject explicitlyanevolutionary epistemology areinvariably wellacquainted with Darwinianselectionism . However , theyadvance reasons whytheybelieve thattheprocesses underlying thegrowthof humanknowledge areverydifferentfromthethoseunderlying adaptive organicevolution(someof their reasons will beconsidered in thelasttwochapters ).

86

Without Miracles

It doesappear thatanevolution -inspired epistemology is resisted by manyphilosophers because it isinconsistent withtheirattempts toestablish aninfallible , justifiable foundation forhuman knowledge . Inthissense , the continually reappearing themes of providentialist rationalism andinstruc tionistempiricism canbeseen asattempts tofindsome bedrock , some firm base onwhichto base ourknowledge , whether it beinfallible priorknowl edge , God,orcompletely trustworthy sensory experience . Anevolutionary , selectionist epistemology cannot provide suchafoundation since selection istprocesses arenotforesighted andgivenoguarantee oferrorless fit, espe ciallynotwithfutureenvironments notyetencountered . Yetjustsuchaselectionist perspective isthebasis foranalternative epis temology thatavoids theproblems ofprovidential andinstructionist episte mologies whileatthesame timeaccounting fortheincreasingly better fit of ourknowledge totheworld.Because ofthis,aselectionist , Darwin -inspired epistemology hasgained manyproponents overthelastcentury .22This trendtowardaselectionist account ofknowledge growthwilllikelycontin ueasphilosophers become moreinterested in andfamiliarwiththeevolu tionanddevelopment of thehumanbrainandtheselection of synapses that, asdiscussed in theprevious chapter , is nowbelieved to underlie all memory andlearning .23Suchcomplementary selectionist perspectives on learning andthinking arepresented in thenextthreechapters .

The Adaptive Modification

of Behavior

The environment made its first great contribution during the evolution of the species , but it exerts a different kind of effect during the lifetime of the individual, and the combination of the two effectsis the behavior we observeat any given time. Any available information about either contribution helps in the prediction and control of human behavior and in its interpretation in daily life. To the extent that either can be changed, behavior can be changed.l In certain respectsoperant reinforcement resemblesthe natural selectionof evolutionary theory. Just as genetic characteristicswhich arise as mutations are selected or discardedby their consequences , so novel forms of behavior are selectedor discarded through reinforcement.2 - B . F. Skinner

In addition to the inherited instinctive behaviors demonstrated by animals as discussed in chapter 3, we cannot fail also to notice behaviors that are modified to fit the circumstances of each animal during its own lifetime .

Biological evolution can account for the emergenceof adapted instincts through the natural selection of organisms having useful behaviors. It is simply too slow, however, to generatenew behaviors adapted to rapidly occurring changes in the environment . To keep pace with these environ mental changes, organisms must be able to learn or acquire new behaviors during their lifetimes .3

The ability to modify behavior adaptively is most impressiveamong the more complex animals such as birds and mammals. In temperate forests these creatures must search for new sources of food as the seasons change

and learn to avoid enemiesand physical dangers. We have all 'seenhow

-

dogs, cats, and birds learn new behaviorsthat allow them to adapt better to the artificial world of their owners. The ability of humans to modify their

88

WithoutMiracles

behaviorto acquirenew job-relatedand leisureabilities, from computer programmingand'speakingforeignlanguages to bicycleriding and piano playing, is striking.

Oneof themajortasksthatthefieldof psychology hassetfor itselfis to discover thewaysin whichananimal 's experiences leadto theacquisi tion of newbehaviors . Countless worms , snails , rats, pigeons , monkeys , humans , andotheranimals havebeensubjected to awidevarietyof experi mentaltreatments to helpusunderstand underwhatconditions andhow thisadaptive modification takesplace . Andalthoughrelativelylittle consensus existstodayin'thefieldof psychology concerning themechanisms of learning , especially for themostintelligent animals suchasdolphins , apes , andhumans , it will beinformative to consider a shorthistoryof psycho logicalresearch andtheorythathaveattempted to analyze learning intoits basiccomponents . .

PavlovianConditionmg

Thefirstscientific attempts to studychanges in behavior beganduringthe 1890sat theInstituteof Experimental Medicinein St. Petersburg . There , IvanPavlov(1849 - 1936)wasdirectorof whatwasat thetimetheworld's bestequipped physiology laboratory , withfacilities to supportalargenumberof dogs . Workingwith dogsintowhichstomach tubeshadbeeninserted to collectgastricjuices , Pavlovand his assistants observed that the animals wouldsecrete gastricjuicesnotonlywhenfoodwasplacedin their mouthsbutalsoat themeresightof foodandevenat thesightof anyone whoregularlyfedthem. Thiswasfollowedby theobservation of Stefan Wolfsohn , a studentin Pavlov 's laboratory , thata dogthathadrepeatedly hadsandinjectedinto its mouth(causingsalivationto removethesand ) beganto salivateat the meresightof sand. AntonSnarsky , anotherof Pavlov 's students , thendemonstrated thata dogcouldlearnto salivate in response to completely arbitrarystimuli. Oneof Snarsky 's experiments involvedcoloringanacidblackandallowingthedogto seeit beforeintroducingit intothedog'smouth.Aftera fewsuchrepetitions , thedogwould salivate profusely atthesightof anyblackliquidin a jar. In thiswayresearch began onthetypeof learning thatis stillreferred to todayasPavlovian conditioning .4It issaidto haveoccurred whena neutral

TheAdaptiveModification of Behavior

89

stimulus, forexample a soundoralightthatatfirstelicits nostrong behavioralresponse, is pairedwithan unconditional stimulus thatnormally always results ina specific response, thatis,theunconditional response. In Snarskys experiment, theblackacidplaced intothemouthwouldbeconsideredtheunconditional stimulusandthe secretionof salivatheuncondi-

tionalresponse. Theoriginal unconditional response to theunconditional stimuluswasnot consideredto betheresultof anypreviouslearningexperi-

ences(hence thetermunconditional), butdueinsteadto aninherited prewiredreflexarcconnecting theperception of theparticularstimulusto a

specific behavioral response. Byrepeatedly presenting a neutral stimulus suchasthesounding ofa bellimmediately beforetheunconditional stimulussuchasplacing meatpowderinthemouth,Pavlovs dogssoonlearned

toproduce theresponse (inthiscasesalivation) atthepresentation ofthe previously neutral stimulus. Inthisway,thedogswouldlearntosalivate at thesoundofa bellif thesoundhadregularly preceded theplacingoffood in their mouths.

It isinteresting tonotethatalthough Snarsky attempted to explain this changein behavior by appealing to the dogshighermentalprocesses involving feelings, expectations, andthoughts, thiswasresisted byhis professor. 5 Pavlov wished to remainin theroleof a purephysiologist, thatis,an objective observer andexperimenter. 6 Thisledhimto reject anysuchmentalistic interpretations, preferring to consider theobserved change inbehavior astheresult ofthemodification ofasimple reflex. After Pavlovreceived a Nobelprizeforhisworkonthephysiology ofthediges-

tivesystem in 1904(thefirstRussian andthefirstphysiologist to beso honored), heshifted hisattention awayfromdigestion andfocused his researcheffortsalmostexclusively onthelearningphenomenon discovered in his laboratory.

WhilePavlovrestricted hisresearch to dogs,American psychologist John B.Watson(18781958)appliedPavlovstheoryof conditioning to under-

standing theemotional development ofhumaninfants. Watsons observationandexperiments ledhimto believe thatduringthefirstmonthoflife babiesshowed onlythreeemotionsfear, rage,andloveand thatthese emotions couldbeelicitedonlybyspecific unconditional stimuli,suchas a loudsoundto evokea fearresponse. Todemonstrate howan initiallyneutral stimuluscouldelicitemotionalreactions,he performedwhatremains

90

Without Miracles

oneofthebest-known psychological experiments withan 11-month-old

boy referredto eversinceas LittleAlbert. 7

LittleAlbert waspresented witha number ofliveanimals andshowed no

fearashereached outtotouch them. Theconditioning procedure then

began asa whiteratwaspresented tohim.Albert reached outtotouchthe

rat,butashedidso,Watson produced averyloudsound bystriking asteel

barbehind Alberts head. This pairing ofanimal andsound wasrepeated

oncemore.Oneweeklaterwhenbrought backto Watsons laboratory, Albert wasmore cautious toward therat.After fivemore pairings oftherat

withtheloudsound, Albert would cryandattempt tomove away when he

sawtherat.Watson reportsthatAlbertalsoshowed sometransfer ofhis fearto otherfurryobjects, suchasa rabbit,dog,andfurcoat.

Watson usedresults suchasthese toargue thatouremotions arelargely

habits acquired asaresult ofvarious experiences, andbelieved thathisfind-

ings hadimportant implications forpsychological therapy. Asheexplained, if wedopossess, asisusually supposed, many hundreds ofemotions, all ofwhich areinstinctively grounded, wemight very well despair ofattemptingtoregulate orcontrol them andtoeradicate wrong ones. Butaccording

totheview I haveadvanced itisduetoenvironmental causes, thatis,habit

formation, thatsomanyobjects cometocalloutemotional reactions. 8

Pavlovian conditioning andhabit formation hadagreat impact onpsychology, andcontinue toinfluence thepractice ofclinical psychology in

treating individuals suffering fromvarious psychological disorders. But sincethetheorydealsonlywiththebondingof newstimulito oldresponses, it cannotaccountforthedevelopment ofnewbehaviors. Forthis a different theoryoflearning wasrequired. OperantConditioning

Itisa noteworthy coincidence thatthesameyear(1898) inwhich Wolfsohn

submitted hisdissertation inSt.Petersburg onthePavlovian conditioning ofthedogssalivary response, Edward Thorndike (18741949) deposited hisdissertation atNew York Citys Columbia University onlearning incats anddogs.LikePavlov, Thorndikes experimental studies ofanimal behav-

iorwere toconvince himthatalllearning wasdependent onestablishing

connections between environmental stimuli andspecific behaviors. How-

The Adaptive Modification of Behavior

91

ever, the learning task he investigatedwas very different from that studied in Pavlov 's laboratory . Thorndike was interested in the ability of animals to learn and remember

new behaviors. To this end, he constructeda number of " puzzle.boxes" into which he would placea hungry dog or cat. The animal could open the door of the puzzle box only by performing somespecialaction such asturning a catch or pulling on a loop of string. Sincea dish containing a small amount of food was placed in the animal's view just outside the box, the famished animal was quite eager to escape to obtain a morsel .

Thorndike found that a dog or cat madewhat appearedto be many random movementswhen first placed in a puzzle box, but would eventually stumble acrossthe behavior that would allow it to escape. When placed repeatedlyin the samebox, the animal would generally take lessand less time to escapeuntil it was able to perform the specific action required to open the door with no hesitation. Thorndike was also surprised to discover that contrary to previous reports of animal learning , his dogs and cats

were not able to learn by observingthe successfulactions of other dogs or cats, nor were they able to learn from being guided passivelyby Thorndike through the motions that would free them. Basedon this and other animal research, Thorndike boldly concluded that all learning in all animals (including humans ) followed certain laws , the most important being his law of effect: Of severalresponsesto the samesituation, those which are accompaniedor closely followed by satisfactionto the animal will , other things being equal, be more firm ly connectedwith the situation, so that, when it recurs, they will be more likely to recur; thosewhich are accompaniedor closelyfollowed by discomfort to the animal will , other things being equal, have their connectionswith that situation weakened, so that, when it recurs, they will be lesslikely to occur.9

He saw his animals acting with no knowledge whatsoeverof the consequences of their actions , the sole purpose of the reward being to stamp in I

the connection betweentheir perception of the situation and a behavioral response . So like both Pavlov and Watson , he came to the conclusion

that

the formation of connections between stimuli and responses were responsible for learning . But unlike Pavlovian conditioning , which deals with the formation

of

connections

between

new

stimuli

and

old

responses ,

Thorndike's animals demonstratedthe gradual " wearing smooth of a path in the brain of connections

between old stimuli

and new responses . " 10

92

Without Miracles

Thorndikewasthe first psychologist to proposethat all newlearned behavior resultsfromthecombination of randomresponses andreinforce ment.HisfellowAmerican B. F. Skinner (1904-1990)didmostto popular izethistypeof learning . Skinnercalledit "operantconditioning " sinceit dealtwith howanimalscouldlearnnewwaysof operating on theirenvironment . In additionto hisextensive , detailed research onanimallearning , particularly ratsandpigeons , hewrotea numberof popularbooksabout behaviorism andits applications for solvingsocialandeducational problems .11Forthesereasons , heremains amongthebest-knownpsychologists of all time. Hisnameremains mostfirmlyconnected to thetheoryof radicalbehaviorism , a theoretical perspective thatignorestheroleof internal mentalstates , purposes , andthoughtprocesses in behavior , andinstead sees all changes in learnedbehaviorresultingfrom contingencies of environ mentalreinforcement . Limitations of Conditioning Theory

Thecombination of theconditioning of PavlovandWatson with theoperantconditioning of ThorndikeandSkinner mightappear to goa longway towardaccounting for theadaptive changes in behavior occurring during thelifetimeof anorganism asaresultof experience . Since thedescription of Watson 's research with LittleAlbertmightleavetheimpression thatPavlovianconditioning canleadonlyto maladapted behaviors , it shouldbe pointedouthowsuchlearningcanbeadaptive . Toreturnto Pavlov 's setting, learningto salivateat the sightof food, or at the soundof a bell signaling itsarrival,readies themouthwiththemoisture andenzymes necessary for digestion . Fora morestrikingexample , consider theflightreaction of mostwild animalssuchasdeerto a loud, suddensound . If this soundisrepeatedly preceded bythesightof menholdingrifles, thesurvival valueof fleeingat themeresightof huntersbecomes obvious . Pavlovian conditioning cantherefore beunderstood asatypeof stimulus substitution , or theattaching of oldmeanings , suchasdanger , to newexperiences , such ashunters . Previously theloudsoundof a shotgunelicitedanautomatic fleeingresponse ; now themeresightof a hunterwill do thesame . Thus the development of conditionalresponses to previouslyneutralstimuli thatregularly precede anunconditional stimulus wouldallowananimalto anticipate andthereby reactmorequicklyto avoiddanger , locatefood, and

TheAdaptive Modification of Behavior

93

win mates . OnemightinvokePavlovian conditioning to explainhowany previously meaningless stimulicouldcometo acquirea meaning for anindividual , evenin thecaseof learninga language or learning to read. As alreadynoted,however , Pavlovian conditioning cannotaccountfor theemergence of newpatterns of behavior , andit isherethatthetheoryof operantconditioning is relevant . Operantconditioning canbeseenasa wayfor animalsto find andretaincreativesolutions to problems , asdid Thorndike 's dogsandcatsin learning howto openthedoorof theirpuzzle boxesto escape , andSkinner 's ratsandpigeons astheydiscovered howto obtainfoodbypressing alever , pecking atakey,orwalkingin afigure-eight pattern . Thesituations of theseanimals werecontrived andcontrolled , but it is not hardto imagine naturalsettings in whichsuchlearningwouldbe veryvaluable , aswhenananimaldiscovers a newsource of foodor findsa newlocationin whichto find shelter . In naturalsettings , monkeys have learned to washsandfromtheirfood, andbirdshavelearned to gettheir breakfast by sippingfrommilk bottleslefton doorsteps . Ourownspecies wouldappearto bethemostadeptat thistypeof learning , aswe constantly find new and creative ways of feeding , clothing , sheltering , and . entertaining ourselves , andprovidingfor ourfamilies . BothPavlovianandoperantconditioning theoriesof learninggained greatpopularity duringthefirsthalfof thetwentieth century , particularly in theUnitedStates . Seeing behaviorasresponses to stimuli, andexplaining learningastheformationof newstimulus -response connections , madeup thecoreof a behaviorist movement thatattempted to makepsychology scientificby focusingon publiclyobservable stimuliandresponses . Thebehavioristapproach reacted against andcontrasted sharplywith thatof the so-calledstructuralists , whosawpsychology asthestudyof consciousness andusedresearch methods that reliedon thesubjective verbalreportsof subjects . Behaviorists suchasWatson , Thorndike , andSkinner begantheir research with animalsandeventually extended theirtheories to includeall humanactions . In sodoing, theyintentionally disregarded anyrolethat thoughtandotherhighermentalprocesses mighthavein the adaptive change of animalor humanbehavior . Thisneglect of theroleof cognitive processes in humanlearningledto a number of serious difficulties in theapplication oftheories to humanbehav ior. In 1974cognitive psychologist WilliamBrewerpublished areviewof a

94

Without Miracles

largenumber of studies thatweredesigned to determine whether the

change inbehavior demonstrated byadult humans inconditioning experiments could beexplained byunconscious, automatic stimulus-response

connections orifhigher mental processes werenecessary. 12 Forexample, inseveral ofthestudies reviewed, a Pavlovian conditioningprocedure wasemployed thatpaired aninitially neutral lightorsound

stimulus withanelectric shock. Thispairing, aspredicted, resulted ina

response 13 tothelightorsound(nowa conditional stimulus) thatwaslike theoriginal unconditional response totheshock. However, theconditional response tothepreviously neutral stimulus oftenquickly disappeared when

subjects were informed thattheshock would nolonger beadministered, andwhether theconditional response disappeared quickly depended on

whether a particular subject actually believed it.

Otherstudies reviewed byBrewer showed thatimmediate Pavlovian con-

ditioning often occurred when adults were informed ofthepurpose ofthe experiment, andthatitdidnotoccur when subjects wereprevented from

discovering therelationship between theconditional andunconditional stimuli. These anda largenumber ofotherstudies ledBrewer toconclude thatall theresultsofthetraditional conditioning literature aredueto

theoperation ofhigher mental processes, asassumed incognitive theory,

andthatthereisnotandneverhasbeenanyconvincing evidence forunconscious, automatic mechanisms in theconditioning of adulthuman beings. 14

Skinners theory ofoperant conditioning hasalsobeen criticized bycognitive scientists. Perhaps themost important assessment wasprovided by

American linguist NoamChomsky whoin 1959reviewed Skinners at-

tempt toexplain language behavior using operant conditioning theory. 15

Chomskys review willbeconsidered inchapters 9 and11.

Letusnowexamine inabitmore detail howbothPavlovian andoperant theories, although formulated toaccount fordifferent types oflearning, are bothstimulus-response views oflearning. Bythisismeant thataparticular stimulus causes activity insome sensory system thatisconnected bythecentralnervous system (spinal cordandbrain) tomotor neurons, andcauses a reaction insome muscles thatresults inanobservable response. Theview of

learning described byPavlovian conditioning canbeseen asthedevelop-

mentofnewconnections between newstimuli andoldresponses. Inother

TheAdaptiveModificationof Behavior

95

words,iftheorganism isinnately wiredsothatstimulus A(gunshot) isconnectedto response Z (fleeing), thepairingof a new,neutralstimulus B (hunters) withA (gunshot) willcausea newconnection to formbetween stimulusB (hunters)and responseZ (fleeing). Sincethe pairingof the unconditionaland conditionalstimuli(sightof huntersand sound of gun-

fire)isprovided bytheenvironment, andsincenotrialanderrororselectionofresponses isapparentinsuchlearning, Pavlovian conditioning seems to be a form of instructionby the environment. As Britishpsychologist Henry Plotkin remarked:

Whenrespondent [Pavlovian] behavior entersintoa learning relationship it is explained bya process ofinstruction.Thatis,somestimulus or stimulus configuration becomes associated witha reinforcing stimulus andcomes to elicit(in somewaycause)a response similar to thatpreviously elicited bythereinforcing stimulus. 16

Theapparentinstructionist natureof an animalsabilityto makenew, usefulPavlovianconnectionsbetweenstimuliand responsessuggeststhat

notallprocesses resulting in adaptedcomplexity haveto beselectionist in theiroperation. Butalthoughostensibly instructionist initsbasicoperation, Pavlovian conditioning maynonetheless havecertainimportantselectionist

aspects. First,theability tolearninthiswaymustitselfhavebeentheproductof biological evolution, andhencehasitsrootsultimately in selection, thatis,thenaturalselection of organisms whocouldlearnadaptively to associatenewstimuliwith old behaviors.Second,as we willseein chapter

9,theperception ofstimuli, without whichnolearning couldtakeplace, maybebestexplained asa selectionist process. Third,it doesnotappear unreasonableto suspectthat the synapticchangesin the nervoussystem

that underliePavlovian conditioning maydependon a blindvariationand selectionof neurons,as discussed in chapter5. Andit shouldbe keptin mind that the instructionistnature of Pavlovianconditioningimposes severelimitson what can be learned,in contrastto the more creativeprocessof operantconditioning. TocontinuePlotkinsquotation: Whatislearned[in Pavlovian conditioning] isanabsolutely determined association betweenstimuliandreflexive responsesthelearningdoesnot,cannot,gobeyond

theseexplicit events, andthetemporal parameters thatrelatethem.ThisiswhatI mean by instruction. 17

In marked contrast to Pavlovianconditioning,operant conditioning involvesa stimulusthat initiallydoes not elicit any particular response.

96

Without Miracles

Instead , the organism responds with a series of varied , random behaviors ; or to use Skinner 's term , the organism " emits " behaviors . Eventually , one of these creatively fashioned behaviors leads to a reward for the animal (for example , food , water, or reduced discomfort ), and as a consequence, the behavior is more likely to occur in the same or similar situation . Thus stimulus A (the sounding of a tone after the rat is placed in a Skinner box ) originally does not evoke any particular response, but rather the organism emits behaviors X (sniffing the floor ), Y (scratching itself ), and Z (pushing the lever) after A has sounded . If behavior Z (pushing the lever) is followed by a reward (the appearance of a food pellet ), a connection between stimu lus A (sound ) and response Z (pushing the lever ) will be first established and then strengthened by additional reinforcement so that response Z will be likely to occur in the future when stimulus A is again encountered . As described by Skinner, if response Z in the presence of stimulus A resuJ.t ~ In ~ reinforcing stimulus , then response Z will come under the control of stimu lus A . The operant conditioning of Thorndike and Skinner can therefore be considered a stimulus -response theory in that stimuli in the environment come to control the responses of the organism . In contrast to Pavlovian conditioning , the environment now serves only to select the appropriate behavior that must first be emitted by the animal . When an appropriate response is made, the environment will provide reinforcement such as food , warmth , or a mate, which will then increase the probability of this same response occurring the next time similar circumstances (stimuli ) are encountered . The organism may then refine this behavior so that it is more effective or efficient in obtaining the environmental reward by repeated, cumulative rounds of behavioral variation and selection -------.

So in contrastto the seemingly instructiverole of the environment in Pavlovian conditioning , operantconditioning is clearlya selectionist theoryof learning , sincetheenvironment doesnot initiallydetermine the adapted res~onse , butratherselects it, by contingencies of reinforcement , from the manyvariedresponses tried out by the organism . Thorndike , seeing thenewdiscovery of neuronsandtheirinterconnections asadditional evidence consistent with his connection -basedview of learning , wroteof a veryDarwinian -sounding"strugglefar existence amongneuraneconnections ." 18

TheAdaptiveModificationof Behavior

97

Thorndike madeonlypassingreferences to a selectionist viewof learn-

ing.ButSkinner putconsiderable effortintopromoting operant conditioningas accomplishing overthelifetime of theindividual animalwhat biological evolution accomplishes overthelongevolution of a species. 9 Naturalselectionaccountsfor the existenceof remarkablycomplexand

adaptedformsof lifewithoutprovidential recourse to the intentions or

purposes ofa designer. ThusSkinner sawtheselection ofbehavior byits consequences through operant conditioning asanexplanation forthedevelopment ofremarkably complex andadapted forms ofbehavior overthe lifetimeof the individualanimalwithout recourseto the purposeor inten-

tionsof theanimal.Indeed,it is apparentthattowardthelatterpartof his

longandproductive career, Skinners principal objective wastodoforpsychologywhatDarwinhad donefor biology:

Compared withtheexperimental analysis ofbehavior, developmental psychology standsintheposition ofevolutionary theorybeforeDarwin. Bytheearlynineteenth

century itwaswellknown thatspecies hadundergone progressive changes toward moreadaptive forms. Theyweredeveloping ormaturing, andimproved adaptation to the environment suggested a kindof purpose.Thequestionwasnot whether

evolutionary changes occurred butwhy.BothLamarck andBuffon appealed to thepurpose supposedly shownbytheindividual inadapting tohisenvironmenta purpose somehow transmitted tothespecies. It remained forDarwin to discover theselective actionof theenvironment, asit remainsfor us to supplement devel-

opmentalism inbehavioral science withananalysis oftheselective actionofthe environment. 0

SkinnerdiscountsheretheLamarckian viewof biological evolution,and

at firstappearsto dothesamefora Lamarckian viewoflearning. However,

hecuriously abandons Darwin andflirtswithLamarck inhisdiscussion of the learning of human culture: Culturalevolutionis Lamarckianin the sensethat acquiredpracticesaretransmit-

ted.Tousea well-worn example, thegiraffedoesnotstretchitsneckto reachfood whichis otherwise out of reachandthenpasson a longerneckto its offspring;

instead, thosegiraffes inwhom mutation hasproduced longer necks aremorelikelytoreach available foodandtransmit themutation. Butachange ofculture which develops a practice permitting it to useotherwise inaccessible sources offoodcan

transmitthatpracticenotonlyto newmembers butto contemporaries or surviving members of an earlier generation. 21

ThisLamarckianinterpretationof culturallearningappearsfundamen-

tallyinconsistent withSkinners belief thatlearning always results fromcertainspontaneously emitted behaviors beingselected bycontingencies of

98

WithoutMiracles

reinforcement . From theperspective ofoperant conditioning , cultural prac tices cannot besimply transmitted fromoneperson toanother , although it maycertainly appear thatsuch transmission occurs when wesee children adopt thelinguistic andcultural practices oftheirsocial environment provided byparents andpeers . Justasthepattern andcolorofthetreebark appear toinstruct thepattern andcoloroftheback ofthewell-camouflaged treetoad , it alsoappears asif behaviors canbetransmitted fromonegener ationorindividual toanother . ButinSkinner 'stheory , nosuch instruction isttransmission ofbehavior ever takes place . Certainly ,thenatural selection of learning is different in important waysfromthenatural selection of biological evolution . AsSkinner stated above , cultural practices canspread quickly throughout a community in a waythatbiological adaptations cannot . Butthisdifference isnotduetoaLamarckian transmission ofbehavior , butisrather aconsequence oflearning involving selective , psycho logical processes operating withinorganisms onashorttimescale , andnot anevolutionary selection process operating among organisms onamuch longer timescale . (Wewill returnto theproblem of accounting forthe adapted andadaptive nature ofculture andcultural change inchapters 10 and15.) Stimulus -response theories of behavioral change appeared to holdgreat promise during thefirsthalfof thiscentury asobjective , scientific expla nations oftheadapted nature ofbehavior andtheadaptive nature oflearn ing.Butaswenowapproach thecentury 'send , these theories aremuch less popular , particularly asapplied tohuman behavior . Partofthereason for theirdecline hastodowiththeirdisregard ofcognitive processes inlearn ingcoupled withthecontinuing cognitive revolution in psychology that began in the1970s . Wewill seein chapter 9 thatmuchof theadaptive modification ofbehavior inhumans (andinthemore intelligent mammals such asapes )results notfromthecumulative variation andselection bythe environment of overtresponses , asSkinner insisted , butrather fromthe cumulative variation andselection bytheanimal ofmental representations thatserve assubstitutes orproxies forovertactions . Inaddition , byview ingallbehavior asdetermined bytheenvironment , andfailing totakeinto account thepurposeful , goal -directed aspect of behavior , theprinciples espoused bybehaviorists such asSkinner have been found tobeinadequate

99

TheAdaptiveModification of Behavior in

explaining

made and

an

human important

selection

chapter does

that the

imposed hostile

involved he was

selecting by

behavior

in

in learning off

the

as animals

an unpredictably

, environment

and

contribution

.

mark

new

unreliable

changing

and

often

we

what

their

it .22 Skinner

cumulative

. But

both adapt

modifying the

behaviors

concerning

continually

for

emphasizing

will

variation see in the

is selected

behavior

uncooperative

to

and

next what

conditions , even

quite

8

AdaptedBehaviorasthe Control of Perception

What we have is a circuit, not an arc or broken segmentof a circle. This circuit is more truly termed organic than reflex, becausethe motor responsedeterminesthe stimulus, just as truly as sensorystimulus determinesmovement. Indeed, the movement is only for the sake of determining the stimulus, of fixing what kind of a stimulus it is, of interpreting it. - John Dewey!

In chapters3 and 7 we examinedseveralof attempts by philosophers, biologists, ethologists, and psychologists to provide accounts for the fit of organisms ~ behavior to their environment . We first examined instinctive

behaviorand showedhow attemptsto explainits fit progressed from the providentialtheoriesof AquinasandPaleyto the instructionisttheoriesof Lamarck Charles

and Erasmus

Darwin , and then to the selectionist

Darwin , Wallace , and Lorenz . We considered

theories

learned

of

behavior

and saw how the Pavlovian conditioning theory of Pavlov and Watson and the operant conditioning theory of Thorndike and Skinner attempted to explain how behavior can adapt to environmental conditions over the lifetime of an organism through the formation of new stimulus-response connectIons

.

But we also noted one seriousomission in all of thesetheories- they provide no adequateexplanation for the goal-oriented, purposeful or intentional nature of adapted behavior. In addition, Pavlovian conditioning emphasizesa one-way transmission of instructions from stimulus to response, and both Pavlovian and operant conditioning environment

as a sort of behavioral

conductor

see the role of the

that orchestrates

all of the

adaptive changes in behavior . In this chapter we will consider these issues in

102

WithoutMiracles

greater the

detail adaptive

and

introduce

nature

of

a behavior

radically

different

approach

to

understanding

.

The Insightsof] ames , Dewey,andTolman It

was

and

over

years

psychologist

purposeful

ments

100

ago

William

nature

of

of

inanimate

in

1890

James

the

behavior

objects

uliet

that

the

( 1842

of

influential

-

living

1910

things

American

)

strikingly

in

philosopher

characterized

contrast

with

the

the

move

-

:

'

To make it clear that such purposeful behavior was not uniquely human, Jamesprovided an examplefrom the amphibious world : Supposea living frog . . . at the bottom of a jar of water. The want of breath will soon make him also long to rejoin the mother-atmosphere, and he will take the shortestpath to his end by swimming straight upwards. But if a jar full of water be inverted over him, he will not, like the bubbles, perpetually presshis noseagainstits unyielding roof, but will restlesslyexplore the neighborhood until by re-descending again he has discovereda path round its brim to the goal of his desires. Again the

fixed end, the varying means!3

Even though James's Principles of Psychology, from which these passagesare taken, was consideredto be the most important psychological work of its day, the stimulus-response, conditioning theories of Pavlov, Watson, Thorndike, and Skinner all totally discountedpurposeas having a role in a scientific account of behavior. In fact, Skinner repeatedlyusedthe analogy of biological evolution to argue against purpose in behavior. For example:

Evolutionary theorymovedthe purposewhichseemed to be displayed by the humangenetic endowment fromantecedent design to subsequent selection bycontingencies of survival . Operant theorymovedthepurpose whichseemed to bedisplayedbyhumanactionfromantecedent intentionor planto subsequent selection by contingencies of reinforcement . A persondisposed to actbecause hehasbeen

Adapted Behavior as the Control of Perception

103

reinforced for acting may feel the condition of his body at suchtime and call it " felt purpose," but what behaviorismrejectsis the causalefficacy of that feeling.4

Darwin showed how natural selectionleadsto adapted complexity in the structure and instinctive behavior of organisms, without purposeeither on the part of the organism or on the part of a supernatural provider. So Skinner is arguing here that the selectionof behavior by the environment can similarly explain the adapted complexity of learned behaviors with out recourseto purpose. This is accomplishedby forming new stimulusresponseconnectionsstrengthenedby environmental reinforcement. But such a stimulus-responseview of learning is both seriously incomplete and misleading. In addition to the problemsmentionedin the previous chapter, American philosopher and educator John Dewey (1859- 1952)

providedanother . He observed that thestimulus-response interpretationof behavior was flawed since it recognizes that stimuli influence responses, but

it neglectsthe equally important fact that responsesalso influence stimuli . Consequently, he criticized the concept of the stimulus-responsereflex arc by noting that

whatwehaveis a circuit,notanarcor brokensegment of a circle.Thiscircuitis moretrulytermedorganicthenreflex,because themotorresponse determines the stimulus , justastrulyassensory stimulus determines movement . Indeed , themove mentis onlyfor thesakeof determining thestimulus , of fixingwhatkind of a stimulus it is, of interpreting it.s To understandhow behaviorcan adaptto its environmentwe haveto considerjustwhat Deweymeantby hisstatement that behaviordetermines stimulusjust asmuchasstimulusdetermines behavior , an insightthat was almosttotally ignoreduntil the middleof this century. As we haveseen , Pavlovianand operantconditioningtheoriesview all adaptedbehavior asresponses to externalstimuli, includingthosecausedby the behaviorof otherorganisms . Acid is placedin a dog's mouthandit salivates . A hungry rat is placedin a familiar cagewherein the pastfood wasmadeavailable wheneverit pressed thebarwithin two seconds aftera bell sounded , andso it immediatelyproceeds to pushthe bar afterhearingyet againthefamiliar pealof thedinnerbell. To usethemoretechnicaltermsof psychological jargon, the stimulusis consideredto be the independentvariableand the response thedependent variable;that is, theresponse depends on thestimulus, andthe stimulusis independent of theresponse .

104

WithoutMiracles

Another way to conceptualizethis one-way view of the relationship is to seethe stimulus as the sole determining causeof the response, and the stimulus to be isolated from any effects of the response. But Dewey took exception to this view , stating that the response also causes the stimulus . But

how can this be?How can the dog's salivatinginfluencethe presenceof acid in its mouth ? How can the rat 's pressing of the bar in any way cause the bell

to be rung, the sounding of which is controlled by the experimenter? To appreciateDewey's important insight, we have to do somethingthat is rarely done in experimentalpsychology. We haveto abandonthe point of view of an objective outside observerand instead attempt to imagine how things appear from the organism's point of view. This shift in perspective brings with it a realization that a stimulus can have an effect on an organism only insofar as it is experiencedor perceivedby the organism. A dog who cannot perceive that acid has been (or is about to be) placed in its

mouth will not salivatewhen the acid is so placed (or about to be). Oncewe imagine how the world appears to the organism , Dewey 's point about

responseinfluencing stimulus beginsto make sense.Actually, it is not necessaryto consider the world as it may appear to dogs and rats. A cursory look around our own world will quickly and clearly reveal that behavior causeschangesin perception as much as perception causeschangesin behavior . As you move your eyes across this page you cannot fail to notice that as

you do so your perception of the page changes. We might reasonablyconclude that we move our eyesfor the purpose of bringing into view that which we want to read next , and that this in fact is the very reason for the behavior . So response does influence

stimulus

and so behavior

does influ -

enceperception. But this is not to denythat stimulus alsoexertsan influence on response. A poorly written sentenceor complex sentencemay well lead to your returning to it in a secondattempt to decipher its meaning. And a loud sound coming from behind your head will likely have you quickly turning around to seewhat has happened. So proper understanding of behavior has to take into account the reciprocal give-and-take relationship of stimulus and response, something neither Pavlovian conditioning nor operant conditioning proposedto do. The purposeful nature of animal behavior was clearly demonstratedby the researchconducted by psychologist Edward C. Tolman (1886- 1959) and his studentsat the University of California at Berkeleyfrom the 1920s


105

to the 1950s. Among the bestknown of thesestudieswas one conductedby Tolman's student D . A . Macfarlane in which rats learnedto swim through a mazeto obtain a food reward.6After they had learned to do this well, a raised floor

was installed

in the maze so that the rats now had to wade

through the mazeto get to the goal box. It was hypothesizedthat if the rats' learning consistedof acquiring specific swimming behaviors (that is, specific responsesto specific stimuli), they would have to relearn the maze in the wading condition, asthe movementsand stimuli involved in wading are very different from those involved in swimming . It was found instead that

after a very brief period of adjustmentto the new situation (just one " run" through the maze), the rats performed aswell in the new wading condition asthey had in the old swimming condition. This was a clear demonstration that what the rats had learned while swimming the maze could not be

describedas the acquisition of stimulus-responseconnectionsbut rather as more generalknowledge about the location of the goal box, sinceit made little difference to the rats whether they swam or waded to their destination .

Similarly, once a person knows how to get to a specific location by driving a car, he can also get there by bicycle (if he knows how to ride one) or by walking (if not too far), regardlessof the fact that the stimuli and responsesdiffer greatly from one mode of transportation to another. Regardlessof these findings and many others like them, Tolman was never able to eliminate

the concept of stimulus -response connections

from

the very core of his theory of purposeful behavior. Indeed, his attempt to explain how behavior can vary and yet reach a consistent goal involves imagining long, complicated, invisible chains of such connectionsexisting within the organism in the form of intervening variables , and conceiving of

responsesnot as specific muscular contractions but rather as a performance . With respect to the latter Tolman

wrote :

It is to be stressed. . . that for me the type of responseI am interestedin is always to be identified as a pattern of organism -environment rearrangements and not as a

detailed set of muscular or glandular activities. Theselatter may vary from trial to trial and yet the total " performance" remainsthe same. Thus, for example, " going towards a light " is a performance in my senseof the term and is not properly a response (a set of muscular contractions ).7

But substituting the word " performance" for " response" doesnothing to explain how an organism is able to accomplish a repeatable " organismenvironment rearrangement" by respondingto stimuli; it simply statesthat

106

it

WithoutMiracles

somehow

total

happens

' performance

to

vary

its

'

to

rat

's

provided

an

so

he

as

causes of

purposeful

negative life

sciences

by

engineering

An

and

to

Macfarlane

's

the

rally

rat

to

the

same

and flies

the

fly

mal

's

an

ever

's

is

been way careless one

]

nest

she with

filled . The

dirt

is

proprietor instance

, in

blocked

by

minutes

before

the

must

she

fixed

,

we earth cleared

develop

,

and

working

development

of

application

and

-

stimuli

a

their

to

expertise

in

's

both

behind

her

is

out

cases left

been

way

and

to

the

a

had her

it nest

disturbed nest

.9

off

so

, pulls

, and

of the

when

before close

she by

an account

, the

the

the can tunnel wasp

ani

of

:

around

while

in ensure

striking

wasp

that

success

particularly

portion

ground it

clean

A

of

muscular

facilitate

solitary

longer

's

take

turning a

on

opening we

the

then

, where

covers and

.

of

frog

. To

lead

consist of

somehow

-

coopera

they

movements

tongue must

lying

out

which a

wing

natu

to

remain when

the of

the

necessary

pattern

lunging

sometimes it

had

,

is

guarantee

behavior

evidence of

normally

environment

other

that

not

behavior

fly

all

not

pattern

frog

clear

observation

is

will

,

fly

provided Careful

predetermined

unvarying

search

which loose

No

the

. In

kicked

considering

the

es

never

work

of

interests

particularly

unpredictable

, the

was

Tolman

to await

this is , the

feedback

would

and

people

movements

the

to

do ,

- provisioning

earth that

.

some

mandibles

with

.

and

place

avoiding

drops

her

able

solving ; that

. 8 But

system

experiment

. Environments

and

the

a

have

with

animals

reproduction

- changing

Sometimes nest

in and

this

attempt

purposeful

takes

will

success

survival

example

of

tongue

. Nor

the

is

toward

and

engineering

- later

be

organisms

contractions

yet

organism

important

goal

such

pioneers

- wade can

behavior

competing

hunting

- first

conclusion

as

step

tradition

in

and

Systems

behavior

still

other

bold

the

was

final

successful

trial

.

behavior

occurring

tively

few

Control

swim

's

how

systems

psychology

Introduction

that

a

the

would

control

that

initial

behaviorist

first

it

to

?

perceived

just

the

The

trial

feedback it

behavior

- feedback

is

toward

for

of

.

how

important

stimuli

model

out

from

goal

sensory

organism

behavior

, "

vary

desired

an

the

the

broke

of

a

that

explicit

never

model

made

changed

guiding

same

at

realization

behavior

in

the

his

may

the

arrive

Tolman in

sential

" behavior

remains

behavior

Nonetheless problem

us

. If

it

the

in

tunnel

wasp

[ the has

clears

the

way

is

clear

the

store

it

away

. In

was worked

entirely for

ten

AdaptedBehavior astheControlof Perception 107

Countless examples of purposeful humanbehavior areeasyto find. One with whichmanyreaders havedailypersonal experience is drivinga car fromhometo work. Suchbehavior requires a remarkably complex , coordinatedpatternof behavior of thefingers , hands , andarmsoperating onthe steering wheel , shiftlever , andturn indicatorlever , whilethelegsandfeet operate theaccelerator , clutch,andbrakepedals . Manypatterns of behav ior will getthedriverandcarsafelyto work(depending onthespeed of the carandtheparticularroutetaken), butjustslightchanges in anyonemay havefatalconsequences . Andyet, noonesetpatternof behavior will always besuccessful . Trafficmaybeheavyor light, fastmovingor slow.Theroad surface maybedryandfirmorwetandslippery . Theengine mayberespon siveor balky.Roadconstruction andtrafficaccidents maymakedeviations from the normallypreferredroutenecessary . Because of thesechange ableconditions andotherunpredictable disturbances , anexactreplayof the drivingbehaviorthatwassuccessful in gettingyouto work on Monday wouldcertainlynotgetyouto workonTuesday . Whatistruefor thepersondrivinga carto workis alsothecasefor theperson walkingfrombedroomto bathroom , thebeefindingandcollecting nectar , thefox pursuing thehareandthehareavoidingthefox, themonarchbutterflymigrating fromMexicoto Canada , andthegreylag goslingclosely followingthesteps of its mother(or KonradLorenz ). It is difficultto seehowanycomplex behaviorcanremainadapted to somefunctionif theorganism doesnot continually makeadjustments to it whileit isperforming thebehavior . Buthowisthispossible ?Canweimagine thefunctioning of anorganism , madeupof sense organs connected bya nervous system andbrainto muscles,thatis ableto pursuea goalbycontinually modifyingits behavior to adjustfor environmental disturbances thatwouldcauseanyfixedpattern of behavior to missitsmark?Yes,wecan.Theexplanation comes to usnot fromthelife sciences of biology , ethology , or psychology , butfromelectri calandmechanical engineers whoin the1930sbegan to makedevices that couldduplicate thepurposeful behavior of humans . Thesedevices areknownascontrolsystems , andtheywork usingthe same typeof stimulus -response circuit, or loop, thatDeweyfirstmentioned in 1896.Theyalsobehave in thesamevariable -pathbut fixed-goalmannerJamesdescribed overa centuryagoascharacteristic of animaland humanbehavior . Suchdevices arenowcommonly foundin a multitudeof

108

WithoutMiracles

electronic devicesfrom the simple thermostat of a home heating systemto the highly complex guidance system of antiaircraft missiles. To seehow they work , we will consider one that is now commonly found on automobiles- the cruise control that automatically maintains a steady driving speedwith no assistancefrom the driver.1O The cruisecontrol systemis engagedby first turning it on and then pushing the " set" button after the car has reachedthe desiredspeed. This speed then becomes the system's goal , or what control system engineers refer to as the reference level , and the system will then increase or decrease the amount

of fuel it delivers to the motor as necessary to maintain this speed. If the car

begins to climb a hill , the cruise control systemwill sensea reduction in speed(being equippedwith a speedometerthat measuresthe speedof rotation of the wheels) and will provide more fuel to the engine through a mechanicallinkage to the throttle , causingit to increaseits power output to maintain the speeddespitethe hill . As the car beginsto descendthe other side of the hill , the cruise control systemwill sensethe increasedspeed, which will causeit to closethe throttle, reducing the amount of fuel delivered to the engine so that again the desired speed is maintained . Because the

systemrespondsto too-high speedsby reducing the amount of fuel delivered to the motor and to too-low speedsby increasingthe flow of fuel, it is referred to as a negative feedback system.

A clearerideaof the natureand functioningof a simplecontrol system can be obtained by examining figure 8.1. The sensorconvertssomevariable aspectof the environment (for example, light, sound, or speed) into a sensor signal (s), which varies from zero to some higher positive value. This sensorsignal (s) is then compared with a referencesignal (r) in the comparator, which subtractss from r yielding an error signal e. This error signal is then amplified and converted by the activator into behavior (0 for output). This behavior then acts on the environment, changing it in the intended direction, which again provides input (i) to the sensor, thereby closing the loop. However, it is not only the control system's output that influences the input to sensor, but also disturbances (d ) emanating from the

environment. So the feedback resulting from the control system's own behavior and the current disturbance from the environment combine to provide the input to the sensor. A cruise control

system acts very much

like a human

driver , with

the

goal of maintaining a given speed. To do this, the driver must attentively

109


monitor

the

the

accelerator

the

.

called

for

mimics

the

designed

.

pedal

accelerator

is

was

speedometer

.

.

While

It

If

should

the

speed

speed

is

as

a

drops

rises

car

come

of

do

the

the

functioning

to

If

human

,

,

the

moving

no

the

must

driver

at

must

the

surprise

driver

driver

that

so

,

on

pressure

speed

cruise

as

down

reduce

desired

the

well

press

this

,

control

is

on

no

action

system

exactly

what

it

.

Purpose

Reference signal(r)

Com parator

Error signal (e = r - s)

Sensorsignal(s)

sensor

"

;f SYSTEM

Activator

ENVIRONMENT

iii

(Actions on

~

environment ) Output(0)

Input (i)

Feedback

d

Envi ronm ental disturbances

Figure 8control .1 system Abasic (after McClelland ,1991 ).

Uncontrolled events

110

Without Miracles

Thecruise control system hasa number of intriguing aspects thatare shared byallproperly functioning control systems . First , it does notperceive theactual disturbances forwhich it mustcompensate . It hasnoway ofdetermining whether theroadisclimbing ordescending . It cannot tellif there isastiffheadwind ortailwind . It cannot knowif aheavy trailerwas attached tothecaratthelaststop , if atireislosing airandoffering steadi Iyincreasing rolling resistance , orif aspark plughasfouled causing acylin derto failandtheengine to losepower . All it cansense , andtherefore control , isthecar'sspeed . Yetdespite itscomplete ignorance ofthemulti tudeofinteracting influences , it does averygoodjobofmaintaining the desired speed . Second , acontrol system does notcontrol whatit does ; it controls what it senses . Thewordcontrol isused herein itsprecise technical sense of maintaining some variable at or nearspecified fixedor changing values regardless ofthedisturbances thatwould otherwise influence it tovary .The cruise control system canonlycontrol whatit senses tobethespeed ofthe vehicle , andit does sobychanging itsoutput asrequired , thatis, bydeliv ering varying amounts offueltotheengine . Theonlywayit canmaintain itssensing ofthecar'sspeed close tothereference level speed inthefaceof disturbances istovaryitsoutput (change itsbehavior ) asnecessary . Sowe see thatit controls itsinput(whatit senses )andnotitsoutput (orbehavior ). Consequently , using acruise control system tomaintain aconstant speed on atripwillallowyoutopredict accurately howlongit willtaketocover a certain distance . It notletyoupredict howmuch fuelwillbeused ingetting there , because fuelconsumption is notcontrolled , varying asit mustto compensate forunpredictable disturbances . Since acontrol system controls whatit senses , andsince anorganism 'ssensing oftheenvironment isgener allyreferred toasperception inthebehavioral sciences , theapplication of control theory tothebehavior oflivingorganisms isknown asperceptual control theory todistinguish it fromthecontrol theory applied byengineers andphysicists totheinanimate world ---. Finally , it isimportant torealize thatwhereas acontrol system 'sbehavior isclearly influenced bytheenvironment , it isnotdetermined solely bythe environment . Rather , itsbehavior isdetermined bywhatit senses (orperceives ) incomparison withitsinternal goalorreference level . Andit ishere thatwefinda crucial difference between thenonliving control systems

AdaptedBehavior astheControlof Perception 111

designed byengineers andthelivingonesfashioned bybiological evolution . An engineered controlsystem isusuallydesigned sothatitsreference level canbemanipulated bytheoperator , for example , bypushing ~ the - "acceler ~ ate" buttonof thecruisecontrolsystem or by turningup the house thermostat . No suchdirectmanipulation of thereference levelsof livingcontrol systems is usuallypossible . Wecancertainlyaskour taxi driverto drive moreslowlyor ourchildto behomebeforemidnight , butthereis noway to guarantee thattheperson will complywith ourwishes . Controlsystems werespecifically designed to replace humanoperators in jobscallingfor thecontrolof importantvariables (for example , steam pressurein a boiler) andhavebeenin wideusesincethe 1930s . Theywere almostcompletely ignoredby thebehavioral sciences , however , until the appearance of NorbertWiener 's ground -breakingbook Cybernetics in 1948,whichwasquicklyfollowedbyimportantrelatedworksbyW. Ross Ashby .11 WilliamT. Powers , anAmerican controlsystem engineer , wasalso struckby how the behaviorof suchsystems resembled the purposeful behavior of livingorganisms , andit isprimarilydueto hisworkthatacontrol system theoryof humanandanimalbehavior existstoday.12 Perceptual ControlTheory

Perceptual controltheorydoeswhatno behaviorist stimulus -response theory haseverbeen abletodo- it provides anexplicit , working model that accounts for goal-oriented , purposeful behavior . Behavior mustbepurposeful if it isto enable anorganism to survive andreproduce despite the unpredictable disturbances theorganism continually encounters . Indeed , wecouldconsider behavior to beadaptedly complex onlyinsofarasit is ableto achieve itspurposes regardless ofthecontinual challenges posed by environmental disturbances . Theperspective provided byperceptual control theory , however , hasnotbeenreceived enthusiastically bybehavioral scientists . Theprincipal reason foritsneglect appears to bethatit turnsthe traditional analysis of behavior onitshead . Instead of thestill-dominant viewofseeing stimuli(bothpastandpresent ) controlling responses (orperceptions controlling behavior , or environment instructing organism ), the theoryofferstheunorthodox viewof behavior ascontrolling perception throughthe organism 's controlof its environment . Hencethetitle of Powers 'sseminal book,Behavior : TheControlofPerception .13

112

Without Miracles

Diagrams canbeuseful inproviding some basic understanding ofthe functioning ofcontrol systems , butthey have some serious limitations . For one thing , they may easily lead one tointerpret thefunctioning ofacontrol system asaseries ofsequential steps , witheach step waiting forthecom pletion oftheprevious one . Instead , infunctioning control systems , both nonliving andliving , allparts areactive simultaneously sothatbothper ception andbehavior arehappening atthesame time .14 Also ,such diagrams maygivetheimpression thatcontrol systems arelimited toquite simple variables . Infact , complex control systems composed ofmany simple con trolsystems canbedesigned tocontrol quite complex variables , thatis, variables that composites ofvalues oflower -order percep tual variables . arecomputed

movements

114

Without Miracles

toperception . Powers andsociologists Clark McPhail andCharles Tucker developed a computer program thatuses control systems tosimulate justthistypeof collective behavior .16 Figure 8.3shows theresults ofonesuch simulation in which fourindividuals (each indicated bytheletter G)maintain close prox imitytoanother individual (M)whoismoving toward agoal location indicated byalarge circle .Obstacles forallindividuals toavoid areindicated by small circles . Byexamining thepaths of theindividuals (shown bythe meandering lines ), it isseen thatthefourGsaresuccessful inbothmain 0

0

0 0 0

0

0 0

0 00

Figure 8.3 Simulation offour (G)following another (M)(after McPhail ,Powers , & Tucker ,1992 ). individuals

Adapted Behavior astheControlofPerception 115

taming close contact withMandavoiding theobstacles asMmoves toits

destination. Although these computer simulations weredeveloped tomodel

aparticular typeofcollective human behavior, theyalsoprovide astriking simulation ofthebehavior ofa groupofgoslings (Gs)maintaining close contactwiththeirmother(M)assheherself movesto somedestination, withall of themavoidingobstaclesalongthe way.

Astrong argument cantherefore bemade thatitisnotbehavior inthe

formof fixedmotorpatterns asconceived byLorenz thatevolves and

becomes instinctive. Rather, thebasisofinstinctive behavior isaninteract-

ingnetwork ofcontrol systems thatpermits organisms tomaintain certain perceptual goals(reference levels) despite thedisturbances thattheyencounter. Inmuchthesamewaythatweinstinctively knowhowtokeepour

bodytemperature atorneara constant 37Cthrough suchautomatic, unlearned behaviors asshivering andsweating, organisms alsoappearto

inherit othercontrol systems thatunderlie adaptive, species-specific behaviors.It isnotthatthespiderisprogrammed witha fixedbehavior pattern

thatwillresult intheconstruction ofaweb,butratherthespider isableto control itsperception ofitsenvironment tomatch aninherited internal goal specification (reference level)byvarying itsbehavior asnecessary to con-

struct itsweb.Andinkeeping withLorenzs original insights concerning the central roleofnatural selection intheevolution ofinstinctive behavior, itis

only through selecting those organisms with fitcontrol systems andelimi-

nating those withlessfitonesthattheadapted complexity ofinstinctive behaviorcan be explained.As Powersremarked:

itsthecapacity toperceive andcontrol thatevolves, notthespecific actsby change appropriately withevery disturbance, every change ininitial conditions. Theres nowayto inherit behavioral outputs, because theoutputs mustremain adjustable tocurrent circumstances, which never repeat exactly. Allthatcanbe

which control iseffected. Behavioral actsachieve repeatable results onlyifthey

inherited arecontrol systems, andatthehighest existing level perhaps some reference signals. 17

Butwhataboutlearned behavior? Canperceptual control theoryaccount

foradaptive changes inbehavior resulting from experiences during thelife-

timeofanorganism? Toseehowthetheory views learning, wefirstmust takea freshlookatbothPavlovian andoperant conditioning. Withrespect to theformer, wemustreconceptualize theunconditional response to an

116

Without Miracles

unconditional stimulus notasthefunctioning ofastimulus-response connection butrather asthatofa control system. 8 Totaketheexample of

Pavlovs dogintowhose mouth acid(theunconditional stimulus) isintroduced leading tosalivation (theunconditional response), wecanconceive

ofthestimulus asadisturbance toareference level fororalacidity andthe

response ofsalivation asanaction designed torestore oralacidity toitsnor-

mallevel. These andother unconditioned responses (forexample, eyeblinking toapuffofair,thestartle response toasudden loud noise) canbe

seenasverybasic, inherited control systems having obvious survival value

thatevolved through natural selection toprotect theorganism from harm-

ful environmentaldisturbances.

Now, ifacertain neutral stimulus, such asthesounding ofa bell, regu-

larly precedes theunconditional stimulus, itcanbeused bytheorganism as asignal thatadisturbance isabout tooccur. Sobyproducing theresponse

aftertheneutral stimulus (nowa conditional stimulus) butbefore the unconditional stimulus, theorganism canprevent oratleastlessen thedisturbing effect oftheunconditional stimulus. Bysalivating atthesound of

thebellthatprecedes theintroduction ofacid intoitsmouth, thedogcan bufferitsmouthagainsttheeffectoftheacid.

Considering now theoperant conditioning ofThorndike andSkinner, we have seenhowstimulus-response conceptualizations oflearning cannot account forthepurposeful nature ofbehavior asnoted byWilliam James in 1890the ability toachieve fixed ends byvaried means. Construing learningastheacquisition offixed patterns ofbehavior cannot explain how organisms canbesuccessful inachieving important goals, such asfinding

food, mates, andshelter, inthefaceofunpredictable disturbances.

Another problem with operant conditioning theory isthatitprovides no explanation forwhycertain events reinforce theorganisms behavior and

others donot. 9 Whyisitthata hungry butwell-watered ratwillworka lever toobtain foodbutnotwater, while a thirsty butwell-fed onewilldo

theopposite? Perceptual control theory answers thisquestion byseeing the reward asacontrolled variable, thatis,avariable thatiscontrolled bythe

organism byvarying itsbehavior. Ifa hungry ratpushes a lever toobtain food, itisonlytobring hisperceived rateoffoodintake close toitsreference level, which hasbeen chosen through natural selection during theevo-

lutionoftheratasa species.

Adapted Behavior astheControlofPerception 117

Finally, perceptual control theory explains anintriguing pattern ofbe-

havior observed bySkinner thatcontradicts thebasic notion thatreinforcementincreases theprobability oftheresponse preceding thereinforcing stimulus. Skinner foundthathecouldobtainveryhighratesofoperantcon-

ditioned behavior (such asahungry pigeon pecking atakeytoobtain food)

bygradually decreasing therateofreinforcement. Very high rates of behavreinforcement schedule ior could be shapedby startingout with an easy

thatprovided a speckoffoodforeachkeypeck,andgradually moving towardmoreandmoredemanding schedules requiring moreandmore

pecks (2,5,10,50,100) foreach reward. Skinner wasthereby abletoget theanimals to peckthousands oftimesforeachfoodpellet, overlong enough periods towear their beaks down tostubs. They would dothiseven though they were getting only asmall fraction ofthereinforcements initially obtained. 20

Butif,according toThorndikes lawofeffect andSkinners theory of operant conditioning, more reinforcement issupposed tocause more ofthe type ofbehavior thatresulted inthereinforcement, 21 howcould italsobe

thatlessreinforcement couldalsocausemoreofthebehavior? Thisproblemiseffectively solved whenweseereinforcement notasanenvironmentalevent thatincreases theprobability ofthespecific behavior thatpreceded

it,butrather asameans bywhich theorganism canachieve agoalbycontrolling aperception. Ifthecircumstances arearranged sothatthehungry ratmustperform morebarpresses tobefed,andit hasnootherwayto obtain food, theratwilladaptbyincreasing itsrateofpressing toobtain its desired amount offood.Andiftherateofreinforcement isincreased tothe

point atwhich theratcanmaintain itsnormal body weight, acontrol sys-

temmodelofbehavior wouldpredict thatfurtherincreases inreinforcementshould leadtodecreases intherateofbehavior. Indeed, thisisexactly what happens. 22

It shouldnowbeobvious thata perceptual controltheoryinterpretation

ofadapted behavior isradically different froma behaviorist view ofoperantconditioning. Whereas behaviorism sees theenvironment incontrol of thebehavior oftheorganism, perceptual control theoryseestheorganism

incontrol ofitsenvironment bymeans ofvarying itsbehavior. Inother words,to behaviorists, behavior iscontrolled bytheenvironment; to per-

ceptual control theorists, behavior controls theenvironment. Thisisnotto

118

WithoutMiracles

saythattheenvironment hasnoinfluence onbehavior . Rather , behavior canbeadapted onlyifit ispartofalarger control process thatvaries behav iortoproduce theperceptions specified byinternal reference signals leading totheaccomplishment ofgoals important forsurvival andreproduction . Perceptual control theory also makes animportant distinction between changes in performance andlearning thatis notmade bybehaviorist theories . According tothebehaviorist theory ofoperant conditioning , the rat'sincrease inrateofbarpressing inresponse toadecrease inreinforce ment isanexample oflearning - theanimal hassomehow learned that more presses arerequired toobtain apellet offoodandwillconsequently increase itsrateofbehavior inresponse tothisnew demand .Butaccording toperceptual control theory , nolearning hastaken place , asthisisjustthe normal functioning ofacontrol system . Recall fromourearlier description ofanautomobile cruise control system thatcontrol systems varytheiroutputtocontrol theirinputtocompensate forenvironmental disturbances . Norewiring orother modification ofthecontrol system isnecessary forthis tohappen . Consequently , therat'sincrease intherateofbarpressing isa change ofperformance thatcanbeexplained asthefunctioning ofanexist ingcontrol system , inthesame waythatthecruise control system willdelivermore andmore fueltothecar'sengine asexternal disturbances (hill, headwind , the added weight of accumulating ice and snow ) act to slow it down . However , instead ofdecreasing therateofreinforcement fortherat,let usimagine thatthesituation ischanged sothatfoodisdelivered notwhen theratpresses thebardown butwhen it pushes thebarup.Now , therat's existing control system willnolonger prove effective in controlling the amount offoodobtained . Thisisanalogous toreversing certain electrical connections inthecruise control system sothatmore fuel(instead ofless , as before )isdelivered tothemotor when thespeed rises above thereference level speed , changing anegative -feedback closed looptoapositive -feed back one .Astriking difference emerges between theratasacomplex living control system andthecruise control asamuch simpler artificial one .Inthe case ofthelatter , thecarwillaccelerate untilthethrottle iswide open and itsmaximum obtainable speed isreached , amaximum speed thatwillnot becontrolled butwillvaryasafunction oftheexternal disturbances . In other words , thecruise control system willsimply failtodowhat it issup -

AdaptedBehavior astheControlof Perception 119 posed will

to do , and the car 's speed will

act differently

. Although

cessful

in obtaining

so that

after

busily

pressing

Rut how - --- -pressing

food , it will

a while down

it will

start

be busily

just a short

down

is no longer

push the bar up to be fed ? Of course

after

out

. But the rat

it will

its pattern

up the same lever

be unsuc of behavior that

it was

ago .

take place ? How going it

of control

the change

to reorganize pushing

while

does this reorganization ~

the lever

be literally

immediately

is the rat to know

to do any good , and that

cannotknow change

this in advance

that

it must . But the

in the environment

initi persistent , increasederror causedby this atesblind, randomchangesin the control systems . In this casethe reorganizationinvolvesa rathersimplechangein the directionof forceapplied to the lever. But more complicatedsituationscan be imaginedin which changes in theenvironmentrequiremoreelaboratereorganizations involving the perceptionandcontrolof newvariablesandthe resettingof referencelevels.As described by Powers :

Reorganization is a process akinto rewiringor microprogramming a computer so thatthoseoperations it canperformarechanged . Reorganization altersbehavior , butdoesnotproduce specific behaviors . It changes theparameters of behavior , not thecontent . Reorganization of a perceptual functionresultsin a perceptual signal altering itsmeaning , owingto achange inthewayit isderived fromlower-ordersignals.Reorganization of anoutputfunctionresultsin a differentchoice of means ,a newdistribution of lower-orderreference signals asa resultof a givenerrorsignal . Reorganization is anoperation ona system , notbya system .23 Reorganization , then, is alsothefunctioningof a controlsystem , but it is differentin onecruciallyimportantrespectfrom thefunctioningbehavioral controlsystems we haveconsidered up to now. The latter areorganizedso that response to errortendsto removetheerror. In a certainsense , thecruise control systemknows to openthe throttle if the speedof the car drops below the referencelevel, and the rat knows it must producemore bar presses if morearerequiredfor eachreward. But thespecialtypeof control systemdesigned to monitor andreorganizeotherworkingcontrol systems cannotknow what to do whenit beginsto showchronicuncorrected error wheretherewaslittle errorbefore.All it cando in this caseis to modifythe controlsystemblindly in someway. If the changeresultsin a reductionof error, any further modificationswill be delayed . But if the changehasno effecton the error or actuallyincreases it, the next modificationwill come quickly.

120

Without Miracles

In thisrespect , thereorganization system thatrewirescontrolsystems to eliminate theerrorthattheexistingsystems cannotreducemustactvery muchlikeE. coli. Thiscommon microorganism caneitherswimin a more or lessstraightlineor tumbleblindly.If it senses thatit is gettingcloserto foodit will continue alongitsmerryway. Butif it senses thatit isnotdoing so, it will stopin its tracks , tumblea while, andthenheadoff in a new, randomlychosen direction . If thisnewheading is perceived asbetterthan thepreviousone, it will continuemovingin thisdirection , andthenext tumblingactwill beput off for a whilelonger . If, however , theerrorhas notbeenreduced , it will soontumbleagain . Althoughthismethodof locomotionmayinitiallyappearquitecrude , it turnsout to bea remarkably usefulandvirtuallyfoolproofwayfor thebacterium to movearoundits environmentA24 Accordingly , thereorganization of controlsystems ishypothesized to be anevolutionary process dependent oncumulative blindvariationandselec tion. If existingcontrolsystems areineffective in controlling importantperceptions , they will be randomlymodifieduntil the error is reduced . Learning to playthepiano,iceskate , or speaka foreignlanguage requires reorganization of controlsystems thatis achieved bycumulatively modifyingthesystems thatproduce errorandselectively retaining thosemodifica tionsthatproduce lesserror.Andalthough theenvironment certainly plays animportantrolein influencing thestateof thecontrolsystems thatwill be retained , it doesnotdetermine theorganism 's behavior . Instead , theorgan ismisactivelyinvolved in theselection process , withthechosen controlsystemparameters and controlledvariablesdepending on the higher -level goalsthatwereselected asbeingimportant , andtheeffectiveness of these parameters and reference levelsin the organism 's currentenvironment . Thathigherthoughtprocesses playa determining rolein whatappearat firstto besimplePavlovian andoperantconditioning in humans wasmentionedin thediscussion of Brewer 'sreviewin theprevious chapter .25 Althoughfirstdeveloped byPowers andhiscolleagues over30yearsago26 andgivena detaileddescription by Powersin 1973, perceptual control theoryisonlynowbecoming morewidelyknown,appreciated , andapplied asa theoryandresearch tool in the behavioral andsocialsciences . The insightthatbehavior isadapted to itsenvironment onlyto theextentthatit

121


allows

an

reaching

organism

to

implications

rapidly

body

applied

to

, 32

in

its

of

But

of

tasks

but

clearly

with

in

also

not

,

-

very

-

,

uncaring

is

no

and

of

from

of

to

and

,

ironic

selection

irony

fades

often

of

their

led

environment

to

is

the

are

reproduction

,

but

we

of

find

closed

,

the

to

but

also

remain

within

realize

the

that

of

great

the

the

-

purposeful

and

to

an

pur

survival

consistently

despite

.

Rather

of

with

evolution

able

in

.

to

considers

that

and

this

nature

behavior

,

the

stamping

,

next

-

with

comparison

itself

all

one

organisms

survival

hostile

in

has

when

the

behavior

adapt

control

.

of

is

by

,

adapted

behaviors

,

to

the

it

believed

coupled

no

system

of

And

by

as

the

,

patterns

occasion

.

,

perceptual

for

not

allow

of

natural

own

nervous

evolution

useful

,

its

.

organisms

environment

,

environment

the

which

mysteriously

the

the

plausible

instincts

the

evolution

perception

one

advantages

,

in

complex

that

with

accounts

selection

,

as

of

that

by

environment

But

,

a

organism

of

encompassing

only

process

essential

variance

of

organisms

selection

life

within

action

seem

.

-

of

the

,

by

the

course

instruction

loops

,

of

of

and

connections

not

reproductive

goals

There

may

organisms

of

course

known

life

the

brief

changing

poseless

is

under

adaptedly

selected

selection

over

process

continually

What

the

variation

feedback

functional

for

Lorenz

the

behavior

the

is

remain

and

are

during

relatively

response

level

work

provides

during

of

systems

its

reference

now

construction

%

theory

and

behaviors

It

the

Darwin

selected

patterns

.

Darwinian

negative

95

is

a

business

Such

theory

over

,

by

specific

cumulative

.

stimulus

a

fixed

control

behavior

for

it

, 31

the

-

to

, 27

. 33

in

far

addition

ethology

control

become

believed

reward

during

of

as

are

Skinner

perceptual

' s

3O

issues

purposes

can

,

that

of

systems

,

educational

account

our

behaviors

and

organism

In

research

law

demonstrated

that

behavior

case

Thorndike

.

,gical

, 29

perceptual

been

behavior

case

innate

the

of

has

how

the

contingencies

ed

power

has

sciences

sociology

and

for

fixed

resulting

life

environment

. 34

not

useful

its

psycholo

, 28

the

of

the

experimental

importance

for

is

of

philosophical

of

most

explanation

It

aspects

behavior

models

variety

aspects

psychology

,

human

generative

of

and

infancy

standing

all

clinical

administration

still

crucial

for

growing

being

a

control

unpredictable

achieve

,

The Development and Functioning

of Thought

Butthesameterm[thinking ] is alsousedwhenthinkingdoesbecome anachieve ment .) thatis.) whenit isproductive . Thishappens whenit changes ourmentalenvironment bysolvingproblems whichthisenvironment offers . . . . Therangeof such achievements istremendous . It extends fromthesolutionofverysimple problems in everyday lifeto veritable mentalrevolutions suchassometimes occurin theminds ofgreatscientists andmaythenaffectthelivesof humanbeings foreverafterward . - Wolfgang Kohlerl One striking characteristicof our speciesis the degreeto which we use thoughtprocesses to solvethemanyproblemsweencounterdaily. Whether planninga vacation, balancinga checkbook , debugginga computerprogram, or makinga scientificdiscoverythat will ultimatelyaffectthelivesof millions of people, muchmorethan just overt behavioris involved. The behavioristtheoriesof Pavlov , Thorndike, Watson,andSkinnerdominated Americanpsychologyduringmuchof the first two-thirds of the twentieth century. Somepsychologists , however , particularlyin Europe, continued their attemptsto understandthe development and functioningof the animal and human minds, includingthe developmentand functioning of thoughtitself. Whenbehaviorismfinally beganto wanein the secondhalf of this century,therebeganwhat hasbeencalleda cognitiverevolutionas the disciplinesof psychology , linguistics , anthropology , philosophy , neuroscience , and computersciencejoinedforcesin an attemptto shedlight on themysteriousandpowerfulinnerworkingsof the brain.2 But if theprocess of thinkingis adaptedto solvingtheproblemsposedby the physicaland socialenvironments of a personor animal, this presents anotherexampleof a puzzleof fit that demands explanation . How is it that merelythinkingabouta problemcanleadto its solution?If a solutionis not

124

Without Miracles

evident when aproblem isfirstencountered, whatdoesthinking dotofind one?Andhowisitthatanadultisabletoperceive things andsolve problemsthathecould notasa child? These arethetypesofpuzzles offitthat

we will now consider.

TheProblem ofProblem Solving: KhlersChimps

AtthesametimethatThorndike andWatson wereusinga behavioral approach toanimal (including human) learning thatshunned anyconsid-

eration ofmental operations (see chapter 7),three German psychologists MaxWertheimer, FranzKoffka, andWolfgang Kohlerwere taking a

quitedifferent approach tounderstanding animal learning. Ofthesethree, Kohler (18871967) isprobably thebestknown forhisstudyofchimpanzees between 1913and1920onTenerife, oneoftheCanary Islands off the coast of Spain. 3

KOhler devised a number oftaskstoexamine theproblem-solving abili-

tiesofhischimps. Onetaskinvolved suspending abanana highoutofreach

sothattheonlywaytoobtainitwasto stackoneormoreboxesunderneath onwhichtheanimalcouldclimbandgrabthefruit.Asecond taskinvolved

puttinga stickinsidethechimpscageanda bananaoutside, sothatthe bananacouldbehadbyusingthestickto pullit withinreach.Avariation ofthisrequired inserting oneendofa rodintotheendofanother tomake

atoolofsufficient length. IntheUmweg (detour)task,a desired object

wasplaced behind barsora window, making it necessary fortheanimal firsttomove awayfromtheobject tocircumnavigate thebarrier between it

andtheobject. Kohler used thislasttasknotonlywithhischimpanzees but

alsowithdogs,chickens, andchildren ofvarious ages.

Thechimpanzees demonstrated varying degrees ofsuccess onthese problems. Almost allofthem were able tosolve thesingle box-stacking problem, andonewasabletostackuptofourboxes. Theyalleventually discovered howtouseastick topullbananas within reach ofthecage, butonlytwohit onthesolution ofjoining tworodstogether to makea longer one.And

whereas chimps (aswellaschildren anddogs)wereableto solvevarious

Umweg tasksintelligently, thechickens weresuccessful onlywhentheir frantic movements brought thembychance toa spotwhere theycould see

the detour around the obstacle.

TheDevelopmentand Functioningof Thought

125

Theintelligent behavior demonstrated bythechimpanzees appeared very different fromthegradual, trial-and-error solutions tothepuzzle-box problemsdemonstrated byThorndikes dogsandcatsasdescribed inchapter7.

Indeed, theapeswould oftenpauseandappear tothinkabouttheproblem beforesuddenly comingup witha solutionthat wasthenimmediately implemented asa single continuous occurrence, a unity, asitwere,inspace asintime... asonecontinuous run,without a seconds stop,rightuptothe objective. 4 Itwasthispausepreceding a rapidly implemented solution toa

problem thatKhlersawasanindication oftrulyintelligent behavior, as opposed to themechanized behaviordemonstrated bythechickens in hisUmweg task,andthecatsandratsinThorndikes andWatsons puzzle boxes

and mazes.

Khler believedthat for his chimpsto solvetheseand otherproblems

intelligently, theyhadto beabletovisualize theproblem mentally ina new

way.Thatis,aperceptual reorganization hadtotakeplace thatbrought to thechimpsattention certain relationships andpossibilities thatthechimp hadnotnoticedbefore.Seeing a sticklyinginthecageanda bananaoutside

thecagewillnotsuggest asolution unless acertain relationship isperceived between them,namely, thatthestickcanbeusedasa meanstorakeinthe fruit.Indeed,Kdhlerfoundthat suchproblems wereoftensolvedmore

quickly ifthestickwasplaced where itandthebanana could beseenatthe

sametime.Oncethisnewvisionwasmentally realized, thechimpthenonly hadto acton it to solvetheproblem.Thisprocessof perceptualreorganizationfollowed bya recognition thataparticular reorganization provides a solution wasreferred to byKhler asinsight (Einsicht inGerman) andis also known as the aha! phenomenon.

Butcalling suchbehavior insightful doeslittletohelpusunderstand the thought processes involved, asindeed Khleradmitted inhislateryears: Insight isinsight intorelations thatemerge whencertain partsofa situation are inspected. . . . Inthesolution ofa problem. . . wesuddenly become aware ofnew relations, butthesenewrelations appear onlyafterwehavementally changed, amplified,or restructuredthegivenmaterial. 5

Asanexample oftheroleofinsight inhumanproblem solving, Khler offersa problem similar totheoneshown infigure9.1.Here,thetaskisto determine thelengthof linea relativeto someaspectof thecircle.Some

readers mayseetheanswer tothisproblem almost immediately, butothers

126

Without Miracles

The Development and Functioning of Thought

usefulperceptualreorganization . But how can we (or our bram)

127 know

which reorganization will lead to the solution? Indeed, if we knew in advancehow to solve it, we wouldn ~t have had the problem in the first place! This , of course, is just another instance of Meno 's

dilemma

described in

cha pter 6 .

One possible solution would be to postulate that the brain produces a seriesof varied perceptualreorganizationsuntil one is produced that is recognizedasleadingto the solution and therefore selected.If you were unable to solve the problem presentedin figure 9.1 immediately, you may recall experiencing just such a series of variations in your perception of it . Although such thoughts would no doubt be constrained by your previous knowledge of geometryand similar problems,7they would in an important sense be blind variations

in that you could not know beforehand

which par -

ticular perceptual organization would lead to the solution. But this is not the reasoningthat Kohler uses. Instead, he concludeshis discussionof problem solving by simply restatingthe problem, this time usingthe word " revolutions" to refer to perceptualreorganizations: Why do such revolutions which occur in certain brains tend to be the right revolutions? . . . [W]hy do brain processestend to produce perceptual organizations of remarkable

clearness

of structure ? At least this part of nature , the human

brain ,

seemsto operate in a most selectivefashion. It is the direction of its operations which is truly remarkable.8

This last sentencesuggeststhat Kohler believedthat somedegreeof foresight is involved sincethe brain seemsto know in what direction the solution of a new problem lies. But of coursethis conclusion simply begsthe question of the origin of this " directional" knowledge. Kohler and his Gestaltpsychologycolleaguesmademany important contributions to our knowledge of animal and human perception and problem solving. They provided much in the way of evidenceand argumentsthat the problem-solving behavior of animals and humans could not be easily accountedfor by the conceptsof stimulus, response, and reward of American behaviorism that was then so popular. The contrast betweenthe two schoolsof thought was noticed by British mathematician and philosopher Bertrand Russell, who observedthat " animals studied by.Americans rush about frantically, with an incredible display of hustle and pep, and at last achievethe desiredresult by chance[whereasthose] observedby Germans

128

Without Miracles

sit still and think , and at last evolve the solution out of their inner consciousness ." 9However, in their hasteto discard behaviorist, " mechanistic" explanations of problem solving, the Germansmay have also unfortu nately discardedthe solution to the puzzle of problem solving, a solution hinted at by Russell's use of the phrase " evolve the solution out of their inner consciousness ." Piaget's GeneticEpistemology Another Europeanwho had an evengreaterimpact on the study of thought processeswas psychologistJean Piaget (1896- 1980). Prolific in research and writing from the ageof 10 yearsuntil shortly before his death,lOPiaget began his career as a biologist specializingin mollusks such as the snails inhabiting the lakes of his native Switzerland. However, a job in Paris administering intelligenceteststo children sparkeda life-long interest in the development of mental abilities and knowledge. Piaget called this study "genetic epistemology," the word genetic in this casereferring not to the genesbut rather to a conceptualizationof the developmentof thought as a processof internally guided cognitive growth. Piagetemployedan ingeniousmelangeof questioningand simple experiments that led to a number of fascinating discoveriesconcerningchildren's thought and cognitive development. For example, he would spread two rows of eight coins each in front of a young child so that each coin in the bottom row was directly beneaththe corresponding coin in the top row. When askedwhich row had more coins, the child would correctly answer that each row was the same. But if the distance betweeneach coin in the bottom row was increased (without adding any coins) while the child watched so that the bottom row became longer, the child would now answerthat the bottom row had more coins than the top row. Piaget also demonstratedthat children were able to solve certain concrete, hands-on tasks before they could solve the sameproblem at a more abstract, logical, and verbal level. So, for example, given a short red stick (1) and a longer greenone (2), and then shown the samegreenone (2) and a still longer yellow one (3), a nine-year-old would have no difficulty in stating that the yellow stick (3) was longer than the red stick (1). However, when told and not shown that stick 1 is shorter than 2, and.stick 2 is in turn


129

shorter than 3 (that is, 1 < 2 and 2 < 3), and asked whether stick 1 or 3 was

longer, the samechild would have great difficulty arriving at the correct answerthat 3 is longer than 1 (1 < 3). These and many other observations and experiments by Piaget clearly demonstratedthat the thought of the young child is different not only in degreefrom that of the adult, but also in kind, and he concludedthat each child goesthrough an invariant seriesof cognitive stages, with each stage requiring a major overhaul of the precedingone.it For example, for a young infant, an object exists only if it can be presently seen, felt, heard, or smelled. At this age, removing a desired object from view usually results in the infant abandoning all efforts to find and obtain it . But the child soon

develops " object permanence," so that she is now able to seek and find objects that were hidden while she was watching. From a Piagetian perspective, the developingchild is like a little scientistwho is constantly developing and testing new theories about the world , rejecting old theories when a new one proves better suited to making sense of the world and meeting her needs .

But if , as Piaget demonstrated, the child's thought becomesbetter and better adaptedto the world , we must ask ourselveshow this increasein fit

is possible . Piagetwasveryclearin his rejectionof both providentialand instructionistviews. His rejectionof a geneticform of providentialismcan I

be seenin his criticisms of the view that someinnate, preformed knowledge must exist that cannot be explained by a processof cognitive growth . And his rejection of instruction is made clear in his criticisms of empiricist psy-

chological theories that attempt to account for knowledge growth as a process of taking in and internalizing sensory experience. Instead , he con -

sideredmental growth to be a constructiveprocessthat cannot be accounted for by innately provided knowledge or instructive sensoryexperience, either working separatelyor in combination. It might reasonably be expectedthat Piaget's realization of the inadequacy of providential (innatist) and instructionist (empiricist) accountsof mental development, coupled with his early training as a biologist, would have led him quite naturally to a selectionistview of the developmentof thought. He rejectedthe theory of Darwinian evolution,12however, reasoning that : Either chanceand selectioncan explain everything or elsebehavior is the motor of evolution. The choice is betweenan alarming waste in the shapeof multitudinous

130

WithoutMiracles

and fruitless trials preceding any successno matter how modest, and a dynamics with an internal logic deriving from those general characteristicsof organization and self-regulation peculiar to all living beings.13

But what exactly does Piaget mean by his appeal to behavior as the " motor of evolution" guided by " an internal logic" and " generalcharacteristics of organization and self-regulation peculiar to all living beings" ; and where did this logic, organization, and self-organization come from ? Unfortunately, he provided no clear, compelling answersto thesequestions, leaving himself vulnerable, as we will see, to the innatist arguments of Noam Chomsky. Having rejectedDarwin 's selectionisttheory of evolution, it should not be surprising that Piaget saw little role for variation and selection in his

theory of cognitive development. But this put him in a curious predicament in essentiallyrejecting all three major explanations for the growth in adapted complexity of human thought processes : providence (both divine and genetic), instruction , and selection . He described an alterna -

tive theory of developmentusing terms suchas " assimilation," " accommodation," " equilibration," " reflective abstraction," and " autoregulation"terms many psychologistsand students of psychology have struggled to understand .14

Despite anti-Darwinian sentiments, the major themesof his theory can be understoodfrom a selectionistperspective. According to Piaget, the two major ways in which children (as well as adults ) interact with their world

are through assimilation and accommodation. Assimilation refers to an incorporation of somesensoryexperienceinto a preexisting thought structure or schema (we will ignore for now the origin of this preexisting

schema). For example, a child having seensparrows and blackbirds and able to recognizethem as members of the category " bird " would likely assimilatethe sighting of a starling into this samecategory. The child might also attempt to assimilatethe first observedbutterfly into the bird schema since it shares certain

similarities

with

other

members

of this category .

However, calling a butterfly a bird would very likely result in a correction by an adult or older child, " That's not a bird, it 's a butterfly !" This would then require accommodation of the child's thought so that butterflies and birds would be treated as different types of objects, eachwith its own label

anddistinguishing characteristics .15

The Developmentand Functioningof Thought

131

Butan adultcannotdirectlyinstructthechildwithrespectto theconventionaluseof wordssuchas birdand butterfly.Thismeansthat a fathercan-

notsimply transmit themeanings ofnewwordsto hischild.Indeed, the childonlyknowsthatsomesortoferrorhasbeenmadeandthat,accordingtohisfather, thecurrent object inviewisnota birdbuta butterfly. But thefathersremarkdoesnottellthechildwhyit isa butterfly andnota bird. Is it becauseit is yellowand the otherflyingorganismshe has seenare brownandblack(butthenwhatof canaries?)? Is it becauseit stopsto sip nectarfromflowers, andtheotherflying creatures donot(butthenwhatof

hummingbirds?)? Or is it because birdshaveonlybeenseenin theafternoon and it is now morning(but then what of the birdsthat get the

worms?)? Clearly, thechildmustmakesomesortof guessas to howto

modify hisbirdschema andcreate a newbutterfly one.Thisguess maywell beinitially wrong, butbycontinuing touseandtestadditional guesses, the childwilleventually cometo thesamenotionsofbirdandbutterfly thatare

sharedbymostadultsofthespeech community. (Morewillbesaidabout the childs acquisitionof vocabularyin chapter 11.) The ProvidentialInnatism of Chomsky and Fodor

Theperson whoprobably didmorethananyothertochange thecourse of psychologyin the secondhalf of the twentiethcenturywas not evena psy-

chologist. In 1959Chomsky published a devastating reviewof Skinners bookVerbalBehaviorthat is oftencitedas markingthe beginningof the

cognitive revolution in psychology. Thereview legitimized thestudyof thought andcognitive processes afterhalfa century ofbehaviorist dominationinNorthAmericathathadbannedallsuchattemptsasunscientific and subjective.

In thisandotherwritings,Chomsky convincingly maintained that lan-

guagebehavior couldnotbe accounted forbyreinforcement of certain responsesunderparticularstimulusconditions. Rather,any attemptto understandhowwe canacquire,use,andunderstandlanguagemusttake into consideration that mostof languageconsistsof novelutterancesthat thespeakerhasneverheardor utteredbefore.Thustheonlywaya person witha finitebraincanproduceandunderstanda potentiallyinfinitenumberof novelsentences is by usingan internalmentalgrammarof abstract, generativerules and principles.

132

WithoutMiracles

Chomsky

was

language

so

quickly

unsystematic account

It

for

this

and

seems

and

that

about

Hume

child

approaches

child

. They

by

Chomsky goal

to

task

of for

aquiring a child

to

fate

may

have

chosen

But

. In

Fodor is

that

series

to

is

mother

the

debates

him

position perhaps

best

embrace

the

summed

assumptions

. These of

consti

nature

-

,"

in

, to

course

of

his

be

of

the

make

leaving

the

ambitious

would

the

enough

world

lee -

' s languages

powers

, he of

a

and

up

knowledge

Fodor

rhetoric

the

as

in against

,

countering the

behav

-

Jerry

well

theorv

argument by

of

philosopher

selectionist

difficulty basic

of

theories

Piaget

considerable and

still

so

formidable

theory to

-

a rich

of

of

that

instructionist

with

' s constructivist

' s reluctance

con

with

.

and

on

-

the

pursuit

while

tongue

To

. 16

in

or

con

much

endowment in

grammar

5000

,

that

hand

enriched

world

informal .

utterances

biological

and

native

:

original

human

possible

the

their

system

complex the

their

complexity

a language

linguistics

intellect

caused innatist

of

states

: a rich

" from the

as of

attacks

lively

Piaget

Piaget

his

with

more

both

conclusion he

acquiring

, universal easy

the

. As

material

field

any

powerful

of

Piaget

. The of

as

limited

development

assaults tivism

a

as

to

sharpened and

innate

acquire

's

be

attacked

here

tive

to

of

language

Chomsky not

iorists

human

of

master

staggering

of

of

to

despite its

also

come

be

the

system

way

were

the

task and

part to

revolutionized a

innate

that

and

with

find

came

is

structure

constitute

experience

' s interactions

the

knowledge

able

difficulty

, he

place

are

and

the

in

and

our

children

little

feat

already

of

' s phrase

awakened

so

knowledge

structure

parts

how

language

remarkable

the

sound the

with to

linguistic

framework

tute

by

exposure

ceptual

ceptual

fascinated

, and of

it

coQ " ni ...... these

construc

-

:

Thereliterallyisn't sucha thingasthenotionof learning a conceptual system richerthantheonethatonealready has;wesimplyhavenoideaof whatit wouldbelike to getfromaconceptually impoverished to aconceptually richersystem byanything likeaprocess of learning .17

HowcouldPiaget havebestcountered thesearguments ?If wesubstitute thewords"lessadaptedly complex " and "moreadaptedly complex " or "lessfit" and "morefit" for Fodor's "conceptually impoverished " and "conceptually richer ," respectively , wecanseehowanevolutionary process involvingcumulative cognitivevariationandselection couldin orincinle rprovidea selectionist -----answer to Chomsky andFodor , thusallowing Piaget to rejecttheirinnatistviewsof mentaldevelopment for a constructionist one. Piaget 'sfailureto dosoandhisvainattemptto providea Lamarckian

133


explanation

of

genome18

contributed

published

how

of

issues

' s

believe

most

able

influential

evolution

,

and

thought

or

and

fit

counters

serious

vided

principle

.

lead

they

to

would

on

be

their

right

in

more

their

evolution

,

)

is

an

the

animal

important

.

by

The

a

way

additional

of

evidence

reason

to

cannot

may

innate

.

less

adaptive

4

-

the

the

synapses

and

com

that

during

of

,

organic

chapter

evolution

But

new

,

of

Chomsky

be

because

in

-

in

,

two

preexisting

selection

doubt

The

is

discussed

and

pro

systems

.

process

adaptive

variation

the

-

knowledge

innate

from

the

en

could

knowledge

be

that

theirs

of

simpler

must

of

origin

ones

emerge

type

,

adapted

that

in

and

however

the

and

conceptual

cannot

undergoes

to

complex

happens

.

for

child

evolution

position

Piaget

explanation

as

knowledge

of

of

the

,

complex

human

clearly

flawed

brain

of

all

systems

( as

this

-

knowledge

of

biological

adaptedly

such

adapted

obviously

mammalian

time

that

that

systems

inquires

basic

more

innate

concepts

that

of

to

assertion

adapted

one

avoid

knowledge

mechanisms

theories

accepted

their

rise

,

-

selectionist

of

providential

and

an

centu

Darwinian

- organism

preformed

genetically

when

as

of

construction

providential

emergence

argument

complex

plex

the

give

basic

all

they

within

only

this

understood

constructive

language

with

undermining

own

their

If

the

risk

of

potential

the

are

this

three

be

a

for

a

the

difficulties

knowledge

the

and

tolerate

creative

how

convoluted

I

involving

the

either

the

As

in

oversimplified

.

can

of

offered

.

winner

its

development

seriously

thought

account

between

environment

to

the

cognitive

dispute

appreciate

Fodor

remarkable

transmitted

much

and

being

this

requiring

Fodor

Chomsky

his

of

principle

without

and

,

risk

rejection

to

in

be

very

presentation

about

the

looking

thought

cursory

core

failure

could

Chomsky

,

thinkers

of

mechanism

the

the

can

. 19

therefore

a

that

experience

' s

debate

such

consequence

and

Chomsky

.is

and

in

I

' s

language

complex

misunderstood

of

th

concerning

exceedingly

because

organism

to

version

The

ry

an

life

-

provides

Fodor

' s

innatist

. VIews

.

Piaget

as

a

well

also

as

rejected

the

and

Lamarckian

chapter

position

advanced

11

Fodor

. )

of

to

thought

explanations

and

account

innatism

difficult

knowledge

in

evolutionary

providential

particularly

Chomsky

an

for

and

these

of

Chomsky

and

account

for

processes

cognitive

issues

the

,

and

development

in

our

discussion

thought

and

Fodor

.

child

' s

led

knowledge

This

him

.

to

( We

of

left

him

in

construction

will

human

of

flirt

return

language

with

to

134

Without Miracles

EarlySelectionist TheoriesofThought

Regardless oftherejection ofDarwinian explanations ofcognitive develop-

mentandthoughtbyPiaget,Chomsky, andFodorduringthesecondhalfof

thiscentury, selectionist views ofknowledge, thought, creativity, andinventionhavea longandimpressive history. Sinceproponents of suchviews

include some ofthemostinsightful andinfluential psychologists, scientists,

andphilosophers of theirday,it willbe wellworthour timeto become acquaintedwith some of these views.

Alexander Bain(18181903), a Scottish philosopher andpsychologist,

emphasized thegreatnumberoftrialsrequired forscientific discoveries and the importance of thepotentialdiscoverers commitment andfascination withthe subjectof hisproblem:

Theinvention ofDaguerre [of thefirstphotographic process} illustratesby amod-

erninstancetheprobablemethodwhereby someof themostancientinventions

werearrived at.Theinventions ofthescarlet dye,ofglass, ofsoap,orgunpowder, could have come onlybyaccident; buttheaccident, inmostofthem, would probablyfallintothehandsof menengaged in numerous trialsuponthematerials

involved. Intense applicationdays ofwatching, nights ofwaking,went with

ancient discoveries aswellaswithmodern. Inthehistorical instances, weknowas much. Themental absorption ofArchimedes isa proverb. 20

Thenumber oftrialsnecessary toarrive at a newconstruction, iscommonly so

great,that,without something ofanaffection, orfascination, forthesubject, one grows weary ofthetask.Thepatient thought ofthenaturalist desirous ofrising to

newclassifications, growsoutofhislikingforthesubject, whichmakesit to hima sweet morsel rolled under the tongue, and gives an enjoyment evento fruitless endeavours. 21

English economist andlogician W.Stanley Jevons(18351882), in his rejection of SirFrancisBaconsinstructionist viewof science (tobe discussedin chapter10),alsooffereda selectionist accountof the human mindsabilityto providenewinsights intoitsenvironment.

Iholdthatinallcases ofinductive inference wemustinvent hypotheses untilwefall uponsome hypothesis which yields deductive results inaccordance withexperience. Itwould beanerrortosuppose thatthegreatdiscoverer seizes atonceuponthe truthorhasanyunerring method ofdivining it.Inallprobability theerrorsofthe great mind exceed innumber those ofthelessvigorous one.Fertility ofimagination andabundance ofguesses attruthareamong thefirstrequisites ofdiscovery; butthe

erroneous guesses mustbemanytimesas numerous as thosewhichprovewell

founded. Theweakest analogies, themost whimsical notions, themostapparently


135

absurdtheories, may passthrough the teemingbrain, and no record remain of more than the hundredth part. There is nothing really absurd except that which proves contrary to logic and experience.The truest theoriesinvolve suppositionswhich are

inconceivable , andno limit canreallybeplacedto thefreedomof hypothesis .22 Regardlessof Jevons's selectionistinsight, it surelycannot be the casethat " the truest theories involve suppositions which are inconceivable," but rather that suppositions that would be inconceivable if based firmly on already achievedknowledge are conceivablefor selectionist, guess-based thought mechanismsof Jevons's " teeming brain." ChaunceyWright (1830- 1875), an American mathematician and philosopherof the pragmatic school, had an important impact on American philosophy, primarily through his interactionswith William Jamesand Charles SandersPeirce. The social event of Wright's life was his trip to England in

1872to visit CharlesDarwin, andDarwin wassufficientlyimpressed with Wright 's defenseof Darwinism that he reprinted and distributed at his own expense one of Wright 's essaysentitled " The Genesis of Species." Of most

relevanceto the current chapter is that Wright also applied Darwin 's prin ciples of variation and selectionto an attempt to understandthe functioning of the human mind : In further illustration of the range of the explanation afforded by the principle of Natural Selection. . . we may instancean application of it to the more specialpsychological problem of the developmentof the individual mind by its own experiences. . . . Here, then, is a closeanalogy, at least, to those fundamental facts of the organic world on which the law of Natural Selectionis based; the facts, namely, of the " rapid increaseof organisms," limited only by " the conditions of existence," and by competition in that " strugglefor existence" which resultsin the " survival of the fittest." As the tendencyto an unlimited increasein existing organismsis held in checkonly by thoseconditions of their existencewhich are chiefly comprisedin the like tendencies of other organisms to unlimited increase, and is thus maintained (so

long as external conditions remain unchanged) in an unvarying balanceof life; and as this balanceadjusts itself to slowly changing external conditions, so, in the history of the individual mind, beliefs which sprang spontaneouslyfrom simple and single experiences , and from a naturally unlimited tendencyto generalization, ar~ held mutually in check, and in their harmony representthe properly balancedexperiencesand knowledgesof the mind, and by adaptivechangesare kept in accordance with changing external conditions , or with the varying total results in the

memory of specialexperiences .23

This extract revealsWright to be no master of lucid English prose, but the use of the very Darwinian ideas of variation , competition , and selection are

evidenthere. Insteadof a Darwinian competition among organisms, Wright

136

WithoutMiracles

describes

a

beliefs

and

better

- adapted

mental the

Already

ones

purposeful

psychological

conceptions in

neous

variation

which

the

selects lar The

an

conception

the

instability tion

of

. But

that

the

A problem be

by

there

. In

how

it

the other

random

it is a false

impossible on chance

.

to

By

of

a kind make

mind

not

imagine

bear

in

has

need

in

artificial

hosts , upon

of

only

of

and

social

] law one

formulation function

evolve

are

of

originally

outbirths

of

unstable

sponta

human

, preserves

or

variations

-

brain

,

destroys

due

to

the

is a spontaneous

brain

, and

itself

in

good

flashes

conceits

, are

mentioned

no

just

on

the

-

molecu

-

variation

other

that

in

, because

the

particular

and

an

the

exact

direc

bad

role

-

flashes

equality

central

. In

, we logical

of

a judicious

human

great much

patience

. It

can

in all

in

, re -

played

and

mind

number time is after

of

hours

the

by

value

nature

ideas at

the and

achieve

this

working smallest years

. Perhaps

our

at

the

utilized

meditation

in

entirely

to

time them

that

the

our

reasoning itself

on

.

to

us ,

persons

arrived

fumbling

same

at that

thought

. Those has

not

arrive

is

our

offers

mind

costs

will

themselves

a simple

discovery of

be

, but

dependence

perfect

knowledge by

it

to

us

except

; it

reduced

which

cross

which

end

it , perceives

present can

to

lead

begin

belatedly

are

and

results

to

before

very

which

will

must

substantially

ideas

conditions

ideas

way

direction

we

absurd

scholars

open

arrive

reasonable

could

even

will

the

of

thoughts is no

finds

addition

some

, and

our

conditions

. Of

marvelous

it

know

series

there

another it

these

have

, bizarre

how

takes

, perhaps

which

the

and

of

that that

. We

what

series

path

steps

know

the

first

idea

a solution

not

the

its

the

invent do

is evident

up

more

frivolous

must

how

those

idea

the

, and

refutes

upset

it

advance

considering

mind

or

and is that

case

selection and

that

which

the

, William

his and

excessively

of

absurd

we

know

sought

know

single

who

problem

the

note

every

what

takes

it more

, we

the

into

animals in

, accidental

the

out

1881

we

this

, retraces

at

tip

the

; but

, we . In

mind

route

leaving

insight

other

tendencies

scientific

flashes

in

which idea

words

immediately

and

important

ideas

, fancies of

notice

and

for

begin

. Our

fit

current

. . . .

to

Souriau

sought

should

other

:

is posed

met

less

and

confirms

as to

thing

invention

active

discovered

is such

Paul

in

both

development

morphological

. It

hypotheses origin . 25

an

Darwinian

images

sort

term

brain

Frenchman chance

with

those

the

simply

[ newly the

important

the triumphant spect of their

,

humans

activity

as it selects

of

having

of

, and

functional

the

sense

as

by

random

analogous

of

strictest

of

the

, just of

eliminate

concerning

environment

short

to

behavior influenced

shape in

accidents

8

, emotions

the

outer

, in

the

theories

. . . new

For

of

chapter

much

. 24

and

in

very

thought

produced

acting

beliefs

.

nature was

among

environment

described

J ames of

competition

, can , do

the

. Even sought

not same

genius - after

-


137

ideapresents itselfto theinventor . Hedoesnotsucceed withoutgoingastraymany times ; andif hethinkshimselfto havesucceeded withouteffort, it is onlybecause thejoy of havingsucceeded hasmadehimforgetall thefatigues , all of thefalse leads , all of theagonies , withwhichhehaspaidfor hissuccess .... If hismemory is strongenough to retainall of theamassed details , heevokes them in turnwithsuchrapiditythattheyseem to appear simultaneously ; hegroups them by chance in all thepossible ways ; hisideas , thusshaken up andagitated in his mind, formnumerous unstable aggregates whichdestroy themselves , andfinishup bystopping onthemostsimpleandsolidcombination .26 Americanpsychologist JamesMark Baldwin(1861- 1934) is of particular interestsincehespentconsiderable timein Franceandis believedto have had an influenceon the psychological theorizingof Piaget.Unlike Piaget, however , he believedthat Darwinianviews of cumulativevariation and selectioncould be appliedfruitfully not only to psychologybut also to ethics, logic, philosophy , religion, judgment, and logic. In a volumepublished in 1909 to mark the fiftieth anniversaryof the publication of Darwin's Origin, Baldwintreatedall thesesubjectsandexpressed thehope that hisbookwould alsostimulateDarwinianthinkingin otherfieldsof the humanitiessuchasanthropology , philology, political science , and literary criticism. Baldwinthoughttheselectionist theoryof adaptiveevolutionhad potentialapplicationto all thesedisciplines , but asa psychologisthe was particularlyinterestedin its applicationto humanthoughtandintelligence : Andhowfarthemethodof lawcalledbyDarwin"naturalselection " goes , whatits rangereallyis, wearenowbeginning to seein itsvariedapplications in thesciences of lifeandmind.It seems to be- unless futureinvestigations setpositive limitsto its application - auniversal principle ; for theintelligence itself,in itsprocedure of tentativeexperimentation , or "trial anderror," appears to operate in accordance with it.27 AustrianphysicistandphilosopherErnstMach, who livedfrom 1838to 1916, andwhosenameremainssynonymous with the speedof sound, was intriguedwith the function of thought, particularlyas it relatedto the advancement of science . Whenhe assumed the newprofessorship in "The HistoryandTheoryof InductiveSciences " createdfor him at theUniversity of Viennain 1895, heoffereda clearlyselectionist accountof scientificand artisticcreativity: Thedisclosure ofnewprovinces of factsbefore unknown canonlybebroughtabout byaccidental circumstances ....

138

After

Without Miracles

the

repeated

survey

advantageous

dominant

that

from

calls

,

as

,

,

poured

cumulative

them

or

the

it

it

say

that

,

per

resulted

the

simply

-

of

of

,

-

slow

outcome

thoughts

had

then

harmonizes

statements

that

they

all

;

has

were

the

the

imagination

which

if

or

background

which

that

as

they

and

the

highflown

light

is

of

word

impossible

and

explained

when

,

,

observer

as

on

he

of

,

One

,

in

a

Newton

,

melodies

,

retained

the

in

to

;

I

felt

combination

What

and

right

happens

then

that

they

never

know

useful

and

Perhaps

.

to

the

of

,

;

Epicurus

they

definitely

,

so

On

the

other

through

the

.

.

.

.

during

.

their

.

.

where

mutual

black

)

wrote

an

experiences

elements

and

to

coffee

and

interlocked

of

the

selec

criteria

could

,

so

of

not

to

-

har

sleep

speak

combinations

interest

,

blindly

and

upon

the

are

harmonious

ones

in

fruitful

without

esthetic

-

.

Ideas

making

a

formed

utility

sensibility

,

,

that

;

.

and

preliminary

unconscious

of

our

,

by

but

just

for

Consciousness

consequently

period

labor

.

combinations

complete

,

to

the

atoms

wall

and

of

the

wall

the

;

meeting

,

me

a

something

mind

so

of

Permit

as

repose

hooked

at

once

,

this

work

compari

like

these

the

atoms

complete

and

conscious

rough

are

rest

motion

may

consequently

-

hooked

-

be

without

in

any

.

a

from

space

Then

-

.

1912

creative

melo

thinker

.

them

,

the

act

-

according

numbers

effect

the

speak

detached

drank

without

certain

without

hand

are

own

of

pairs

explanations

During

to

I

great

are

elements

between

them

.

seek

prolonged

combination

the

all

.

are

,

until

without

future

his

mind

custom

only

.

,

to

1854

of

thoughts

creative

(

from

notion

.

precedes

Figure

atoms

also

insightful

recombination

collide

all

are

"

both

Poincare

unconscious

Among

them

ought

my

almost

.

always

son

,

beautiful

we

which

.

?

self

reason

will

.

fit

the

:

them .

subliminal

the

to

appear

and

derived

blind

is

leads

Henri

usefulness

creativity

may

sudden

the

and

of

gradually

mathematician

contrary

crowds

stable

the

inexplicably

products

,

explanation

nevertheless

creativity

beauty

evening

rose

an

that

emphasized

tpose

mony

' s

that

but

mathematical

which

tion

"

French

essay

Mach

process

harmonies

Renowned

in

in

selection

,

and

gnats

,

to

Then

be

to

free

appears

to

others

upon

significant

dies

of

in

,

are

and

a

arises

selection

interposition

the

appearance

which

.

the

with

relegated

future

design

suit

. 28

Particularly

less

or

that

gradually

form

,

for

traits

their

fancies

Thus

,

the

has

of

gradual

Wagner

had

and

mood

.

opportunity

all

particular

,

creation

Richard

harmonies ones

a

of

afforded

making

host

idea

of

act

Mozart

,

swelling

ruling

result

deliberate

has

,

that

the

the

vivid

suddenly

with

field

rendered

inappropriate

teeming

forth

a

has

more

are

the

fectly

a

,

thought

things

ly

of

accidents

period

the

of

they

impacts

are

apparent

put

rest

in

enclosed

may

,

produce

and

motion

unconscious

.

as

would

new

They

work

flash

,

for

combinations

in

exam

,

every

certain

direction

pIe

,

.

a

.

swarm

.

.

of

-

139


In the subliminal

self . . . reigns what I should call liberty , if we might give this name

to the simple absence of discipline order itself permits

unexpected

It is noteworthy divergent

that

fields

Poincare

, they

all were

perspective

useful

thoughts

a large

although

of inquiry

tionist

(beliefs , ideas ) can

" the inventions

of the scarlet

have

by accident

only

and abundance

of guesses

ery . " Wright memories

wrote

, experiences

tions , emotions

variation

in

brain

. " For

wrote

about

the

flown

imagination

great

numbers

almost

ly referred

in selecting ones . Thus

results

are something preserves social

appeared

thought which value

-

and

make

outer

it more

themselves

host

the

lead

at hand

thoughts

more

for

to natural

of fancies

which

. " Mach

(common

somehow

eliminating existence

remarked

that

confirms

that those

noted

a free and highflown

that

ones . be the

thoughts or refutes , and

sort . " Souriau

selection

" and

-

' which

morphological

in stating

substantially

to our mind , we note only in reasoning

the self ,

to useful

of an analogous

logical

" among

must

while

simply

, we can in addition

and

high -

subliminal

of thoughts

" ' struggle

environment

analogy

a free and how

in itself

' " ideas . James

accidents

clear

selection

and can be utilized

ing , swelling

) cannot

of

. " Mach

the

selects , in short , just as it selects

a very

of artificial

present

" the

due to molecular

to make

" by a kind

of the fittest

which

variety

the useful

wrote

human

at random

."

, the problem

and retaining

or destroys

variations

today

concep produced

of spontaneous

which

by

" of

unstable

described

out a great

evolution

except

utility

of discov -

" new

outbirths

formed

and without

churning

in the ' survival

Poincare

that , could

of repetition

are originally

of fancies

that

produces

requisites

excessively

to begin

blindly

interest

Wright

selec -

of imagination

how

evolve

the

host

. " And

to as brainstorming

useless

of

, swelling

For this , as in biological

to their

is convinced

power

explained

which

is no way

forth

simply

involved

the first

, expansive

activity

combinations

all are without

But of course

are among

for

all is the idea

" fertility

images , fancies , accidental

calls

respect

if the thinker

that

beliefs . James

" there

of

with

only

argued

tendencies

" the teeming

to mathematics

guesses . Thus , Bain

" unlimited

functional

Souriau

Bain

in quite

dye , of glass , of soap , or gunpowder

, and

in the shape of random

worked

to them

be found

blind

at truth

, and active

for

. Central

. " Jevons

of the

individuals

agreement

thought

and

eight

psychology

in remarkable

of varied

born of chance . Only this dis -

.29

these

, from

on human

number

come

and to the disorder

combination

perfect

" of all the which " from

that our ideas

have some the teem -

imagination

calls

140

Without Miracles

forth,suddenly thatparticular formarises to thelightwhichharmonizes perfectly withtherulingidea,mood,ordesign.AndPoincarØ foundthat

ofthemultitude ofideas which come tomind asonestruggles withaprob-

lem,onlycertain onesareharmonious, and,consequently, atonceuseful

and beautiful.

Butwhereas theseindividuals allshared whatwemightcalltheselec-

tionist insight concerning thought, theydidnotprovide particularly strong logical orphilosophical arguments, orcogent historical orpsychological

evidence to support theirtheories. Inaddition, theriseofbehaviorism in

theUnited States atthebeginning ofthetwentieth century made itvery

unpopular to baseanypsychological theory onsubjective andunobserv-

ablemental processes. Scientists such asKhler,Piaget, andChomsky who

swamagainstthebehaviorist tideandrejected trial-and-error theoriesof

overt behavior alsorejected trial-and-error theories ofthought. Itmaywell

beforthese reasons thatselectionist views ofthinking practically disap-

pearedduring thefirsthalfofthetwentieth century anddidnotmakea

comeback until the second half.

Thought as SubstituteTrial and Error:

Campbells Hierarchy ofKnowledge Processes

Itislargely duetothework ofAmerican psychologist Donald 1 Campbell (born in1916andnowprofessor emeritus atLehigh University inBethlehem,Pennsylvania) thata general selection theory ofthought isknown (andcontested) today. Campbell hasthreemajor accomplishments tohis

credit inthisregard. First,hedocumented thediscovery anduseofselectionist theories ofthought byphilosophers, psychologists, mathematicians,

andotherscientists since thetimeofDarwin. 30 Second, overa period of

morethan35 yearshe provided strongarguments thattheDarwinian process ofblindvariation andselective retention underlies allachievements

offit,including thefitofourperceptions totheworld theyrepresent, thefit ofourthoughts andmental processes tothereal-world problems weconfront andsuccessfully solve, andthefitofourscientific theories andpredictionsto theuniverse theydescribe. Finally, heprovided a hierarchy of

knowledge processes inanattempt to showhowtheevolution andonto-

geny ofallforms ofknowledge canbeaccounted forwithin ageneral selectionist framework.


Campbell his

1.

11

A

- level

blind

achievements to

environment

lays

out

four

basic

of

knowledge

hierarchy

- variation , to

- and all

genuine

- selective

premises

which

processes

- retention

increases

in

process knowledge

provide

the

141

foundation

for

. 31

is fundamental , to

all

increases

to

all

inductive

in

fit

of

system

.

Hereweseethecenterpiece of Campbell 'sepistemology in theboldclaim that blindvariationandselection is theonlynatural(that is, nonmiracu lolls) explanation for anyincrease in adapted complexity , for anyincrease in fit of onesystem withrespect to another . 2. In sucha process therearethreeessentials : (a) Mechanisms for introducing variation ; (b) Consistent selection processes ; and(c) Mechanisms for preserving and /or propagating theselected variations . Notethatin general thepreservation andgeneration mechanisms areinherently atodds , andeach mustbecompromised .

Whilethefirstpointmayappear to bea dogmatic andunsupportable claim , thesecond pointis quitedifferent in thatit positsmechanisms for variation , selection , andretention . Thus , Campbell makes thetestable predictionthatfor anysystem thatshows anincrease in adapted complexity overtime,careful examination will reveal mechanistic processes of varia tion, selection , andretention underlying thisadaptation . Inthisregard it is of interest to notethatDarwinbelieved thatthese threeprocesses were responsible forbiological evolution . Hecouldargue convincingly onlyfor mechanisms of selection ("survival of thefittest "), however , sincehehad noknowledge of themechanisms of eithergenetic variation or retention . If Campbell 's conjectures arecorrect , wewouldexpect thatresearch into natural andartificialadaptive systems wouldeventually yieldconcrete evidence oftheunderlying mechanisms of variation , selection , andretention . Thisindeed turned outto bethecase in thefieldsofimmunology andneurology , aswealready sawin chapters 4 and5. Chapters 14and15address howthesame appears trueincomputer science andmolecular design . 3. Themany processes whichshortcut a morefullblind -variation -and -selective retention process arethemselves inductive achievements , containing wisdom about theenvironment achieved originally byblindvariation andselective retention . 4. Inaddition , such shortcut processes contain intheirownoperation ablind -varia tion-and -selective -retention process atsome level , substituting forovert locomotor exploration orthelife-and -death winnowing oforganic evolution . Thisconcept ofshortcut orvicarious32 processes thatsubstitute forlonger, moretedious , andmorecostlyblindvariation andselective retention is

142

Without Miracles

central to Campbells hierarchy andsomerits further elaboration. Letus takea blind manasanexample. Ifplaced inanunfamiliar house, theblind mancanexitonlybymoving toandsearching thewallswithhishandsuntil

hefinds thedoorleading outside. Intheprocess hemaywelltriponfurnitureandother objects, ortumble down aflight ofstairs. Thisisobviously

verycostly intermsofenergy andtime,anda dangerous waytosolve his

problemof exitingthe house.

Butnowimagine thatthemanis givena longcanethathecanuseto

probe thespace around himmore efficiently andsafely. Note thatalthough

heisnotinanysense lessblind thanhewaswithout thecane, heisinmuch bettercircumstances sincebytapping hiscanehecansubstitute forthe

morecostlyanddangerous manualgroping hewouldhaveto dowithout

thecane. Thecane hasarather limited range ofeffectiveness, however, and itisstillnecessary forthemantomove about theroomashetaps. Letusnowsuppose thatthemanisequipped withaninfrared sensing device that,byemitting andtiming thearrival ofreflected infrared waves, cancompute thedistance ofanobject andpresent thisinformation bypro-

ducingan audible tonethatcovaries in pitchwiththisdistance so that closeobjects soundhigher thanmoredistant ones.Nowheisableto obtaininformation fromtheentireroomwithout having tomovearound it

atall.Slight dipsinfrequency would indicate thecorners oftheroom, with a dramatic dipindicating anopen window, doorway, orstairway. Notice thatsuch asystem begins toresemble vision initsoperation. Ahostage who istransported toa strange house while blindfolded andthenmanages to

remove hisblindfold isinmuchthesamesituation astheblindmanwiththe

infrared device. Initially, thehostage hasnoideawherea window ordoor

might be,andcanonly lookaround blindlyuntil hefinds something that

mightprovidea meansofescape.

Butcanitmake anysense torefer tovision asblind? Wecertainly dont justlookanywhere andeverywhere inusing vision tosolve problems. We normally dontsearch fordoors ontheceiling orfloor. Unlike thedrunkard, wedontlookforourlostkeys under astreetlight ifweknow wedropped them inadarkpartofthestreet. When faced withaproblem, wedontnormally justtryanything. Todosowould bevery inefficient andmake itquite unlikely thatwewould findthesolution toourproblem. Thatis,weuse

constraints (orrestraints) to solveproblems andmakenewdiscoveries.

TheDevelopment and Functioningof Thought

143

Undoubtedly suchconstraints canbeveryhelpful infinding solutions to

problems anddiscovering newknowledge, andtheymaybenecessary ifwe

wantto beabletofinda solution ina reasonable lengthoftime.Anexperiencedmechanic willnotcheckthemufflerof a carif themotorwillnot

start,andwillnotexamine thefuelinjection system ifthecarrunswellbut makesexcessive noise.Experience permitsthephysician to eliminate many

possible causes ofdisease basedonthepatients symptoms. Knowledge based onprevious similar experiences canconstrain thesearch space ofpossiblesolutionsthat makeuselessvariationslesslikelyandthereforepoten-

tiallyuseful onesmorelikely. Butwemustkeepinmindthreeimportant things concerning constraints. First, insofar astheyareuseful andtherefore fittheproblem athand,thisfitoftheconstraints themselves mustbeexplained. Thusifwewishto avoid providential orinstructionist explanationsforthefit,theexistence ofconstraints mustbeexplained astheresult of previousblindvariationandselection.

Second, no matterhowusefultheymaybe,constraints alonecannot accountfor solutionsto newproblems.AsarguedbyCampbell:

Intelligent variations require anexplanation forhowthese variations orhypotheses cameto bewise-in-advance. Thatmosthypotheses arewise,I havenodoubt.As such,theyreflect already achieved knowledge or,atveryleast,wiserestrictions on

thesearch space. Suchwisdom doesnot,however, explain further advances in knowledge. Thathypotheses, evenifnotwise, arefarfromrandom, I agree. But wiseorstupid, restraints onthesearch space donotexplain novel solutions. 33

Third, thereisnoguarantee thatconstraints thatproved useful inthepast willcontinue todosointhefuture.Indeed, inmanyrespects progress insciencecanbeseenasthesweeping awayofoldconstraints thatarenolonger considered valid.Theearlyconception oftheuniverse revolving aboutthe earthat itscenterconstrained astronomy inwaysthatwereoriginally use-

ful,butthisconstraint wascompletely rejected byCopernicus andlater astronomers. Theeveryday ideasthatlightis unaffected bygravityand massis unaffected by heatandvelocity wereconstraints important in

Newtons day,butwerediscarded byEinstein andlaterphysicists. Andthe constraints provided byreligious doctrine thatviewed thedesign of all creaturesas unchallengeable proofof the handiworkof a supremecreator were dramaticallyoverturnedby Darwin.

Nowthat we haveexamined the basicideasmotivating Campbells

thoroughly selectionist viewofknowledge, weareina betterposition to

144

WithoutMiracles

present

Table

9 . 1

Donald T . Campbell 11 C .... ; a "., r - ~ 11 . Science 10

. Cultural

9 . Language

8 . Socially

( 1974a

) hierarchy

of

selectionist

-

knowledge -- - -. - - - 0 -

processes r .a. ,-, - _ vv "" '-:J

-

cumulation

( overlapping

vicarious

7 . Mnemonically 6 . Visually

's

6

and

exploration supported

supported

7 )

; Observational

learning

and

imitation

thought

thought

5 . Habit 4 . Instinct

3 . Vicarious

locomotor

2 . Nonmnemonic 1 . Genetic -

devices

problem

solving

adaptation -

TheDevelopment andFunctioning of Thought 145

andavoiding danger, andcandothisincomplex, changing environments, it cannot profitfromexperience. Asitbegins totumble, ithasnomemory ofthedirection inwhich itwaslasttraveling andtherefore isjustaslikely to setoffinthissamedirection asanyother.Ofcourse, it ismuchmore

likely tochoose anewbearing, since there aremany more newdirections than the single old one.

Campbells third level isvicarious locomotor devices. Aswelearned with regard totheblindman,exploring onesenvironment witha remote sen-

sorysystem hasimportant advantages over locomotion anddirect physical contact. Particularly striking examples ofthisarethesophisticated echolocationsystems ofbats,porpoises, andcavebirds.Without sucha sensory

system, abatcould only make itswayoutofadarkcave byrepeatedly flyingintothewalls ofthecave until iteventually discovered theentrance. Its echolocation system makes suchcostly (interms oftimeandenergy) and

dangerous fumblings unnecessary, asitprovides a typeofsubstitute locomotordevice sothatthebatcanfinditswayintotheopenfaster, withless

energy, andmore safety thanbyflying intothecave walls. Campbell admits thatit isdifficult formostpeopleto conceptualize visionitselfasa vicarious locomotor devicethat operatesusingblindvariationandselection. He

maintains thatthisis,however, thecase, assuggested bytheblindmanwith the infrared distance detector.

Wenowmove toCampbells fourth andfifthlevels, habitandinstinct. As

described inchapter 3,instinct isgenerally understood asresulting from the evolutionary selection oforganisms withuseful behaviors, andhabitistypically considered toresult frombehavioral consequences oftheexperiences ofanindividual organism. Asshown inchapter 7,however, serious problemsariseifinstinct andhabitareconsidered asbeing linked tospecific behaviors, sincebehavior mustremain constantly adaptive (andtherefore variable) forittobeuseful inachieving thegoals ofanorganism living inan unpredictable anddisturbance-rich environment. Wecannonetheless keep theessential character ofCampbells thought ifwereconsider instinct tobe theresultoftheselection oforganisms withuseful (fit)perceptual control systems, andsubstitute forhabitthereorganization ofthese control systems that an organismaccomplishes to solveproblems.

AswemoveintothenextlevelofCampbells hierarchy wefinally meet

theprincipal concern ofthischapterthought. Thesixthlevel, visually supported thought, is:

the dom for of ins pr so in a , e . g a d b [wa ]sof Ko a req the su of a vi pr n W t e ron rep vic th se th is s a err of po loc in . Th " s " lo , S u ,to stit lev wit its se cr , ar p i o l n o i~ wh ap " , " p nt i n s u f furt ed in mo dir co w t en . 3 h To pro an ex of vis su h th im atte to rea the fu in yo liv ro to a an upr pia . In loo ov th ro a c fu , y cou rea im oth po ar . Y m t " T h be mo fro th ba w t un w o fr u spa for the , an tw ch cu u w to em co . " O c , th p n p o be the ac if yo re b ,.A '"i-+ 1 k.-"1 -1~C 146

WithoutMiracles

u""""\,.. ~~ LV lllC : UUlll -1lloooKcase , butother arrangements could easily be imagined asyou observe the room 'scurrent configuration and contents . Once adecision ismade ,however ,you could implement itdirectly ,without having totryout physically each ofthe arrangements you considered and then rejected inyour thinking . Asmentioned byCampbell above and asweconsidered earlier , this appears tobethe type ofthought process used byKohler 'schimpanzees tosolve various problems .However , the chimps didnotseem capable of Campbell 'sthought sev ,enth ofknowledge processes , that is,mnemonically supported : level Atthis level the environment being searched isvicariously represented inmemory or "knowledge ," rather than visually , the blindly emitted vicarious thought trials being selected byavicarious criterion substituting foranexternal state ofaffairs .The net result isthe "intelligent ," "creative ,"and "foresightful "product ofthought ,our admiration of which makes usextremely reluctant tosubsume itunder the blind variation -and -selective -retention model .36 Toextend our example ,imagine that you are now contemplating the pur chase ofapiano inthe dealer 'sshowroom and imagining how itcould be incorporated into your living room .Since you are not inyour living room and therefore cannot see it,you must rely onyour memory ofittodeter mine where the piano could be put .This memory represents the knowledge you must have tosolve the problem ofpiano placement .Inthis case ,knowl edge may well exist assome type ofmental image , although this ismost


147

The is no " d " infu or tra of kn ire or ha , ju a t is h no "by d acq of kno by ire ob in . A p o ,mak Bald wh ~ t the chi acq is a cri im , w h l to m e a trial and erro ma . He he tu fo ex a le that sou by a tria an err of vo he ch th mem of pat Re stu of th le o b s c f and elab sam mo 3 8 A par stri cha hu kn p is the role tha lan pla in tho an ac k . Alth Cam put lan at the nin le , he n th it o laps with the two pre lev ,sin itis ge ba to b mos imp too for bo mn su th sth

But even inthe case oflearning byimitation ,Campbell iscareful todis count aninstruction -based explanation instating that

lydependent knowledge processes .oflanguage He notes that through our use "the outcome ofexplora tions can be relayed from scout tofollower with neither the illustrative

148

WithoutMiracles

locomotion northeenvironment explored being present , noteven visually vicariously present ."39 Hefurther observes :

149

TheDevelopment andFunctioning of Thought

The

grand

beyond

breadth

the

sented

here

ogy

is

of

scope in

of

some

perhaps

cal

and

evolution

mechanisms

tion

for

such

or

human

fit

Campbell

thought

trials to

difficult

to

of

processes

with

radically

rejection

and

of at

selection

require

the

unfathomable

.

fit

,

is

is

pre

and

, from

the

to

the

assistance

a

of

-

of

point

only

of

mechanisms

biologi

progress

quick

provide

-

providential

levels

he

goes

It

epistemol

through

of

and

all all

thought

evidence

clearly

thought selectionist

acquisition

not

for

together

them

be

on as

the

.

years

. for

through

one

many

be

discussed

logical

the

But

he

his

out

explana

-

supernatural

passive

sensory

uses

to

IV

which

anecdotal selection

of

psychological that

it

thinking

and

problem

. This

immunology

) , become

and

admitted

' s hierarchy

theory in

, and

undertaken has of

Campbell

psychological

part

not and

studies

discoveries

in

historical variation

has

claims

why

practical

,

blind

scientific

reason

selectionist

to

, as

mainstream

the

( and

35

evidence

may

with

put

philosophical thinking

over

test

impact ,

offered

provide

. This

however

appreciated

does

considering

research

being

outright

sees

workings

has for

much

's

variation

that

the

Campbell

perception

of

knowledge

primarily

.

reasons

solving

its

of

dealing

knowledge

Campbell

the

instruction

in of

. Wherever

how

provider

since

unique accounts

hierarchy

chapter

detail

instructionist

science

this

a

has

may and

selection

more

not

very

had

well

very

change

,

neuroscience

theory

widely

is

,

is

known

now

and

10

_::.

-

Cultural Knowledge as the Evolution of Tradition , Technology, and Science

From the amcebato Einstein, the growth of knowledge is always the same: we try to solve our problems, and to obtain, by a process of elimination, something approachingadequacyin our tentative solutions. - Karl Popper The analogy that relates the evolution of organisms to the evolution of scientific ideascan easily be pushedtoo far. But with respectto the issuesof this closing section it is very nearly perfect. . . . Successive stagesin that developmentalprocessare marked by an increasein articulation and specialization. And the entire processmay have occurred, as we now supposebiological evolution did, without benefit ofa set goal, a permanentfixed scientific truth, of which eachstagein the developmentof scientific knowledge is a better exemplar. - Thomas Kuhn2

detail

new species.This is phylogeneticselectionamong organisms. We alsoconcluded, however, that n~tural selectionamong organismssuf"

fers from two notable drawbacks - it takes a considerable amount of time ,3

and it cannot be used by any individual organism to cope with an unpredictably changingenvironment or to learn from experience. An animal of a

152

Without Miracles

.species thatfor millionsof yearsdepended on a particularfoodwill have greatdifficultysurviving andreproducing if thefoodbecomes unavailable , unless it canlearnto findandusea substitute . Similarly , a mammal with a fixedrepertoire of antibodies will beunlikelyto surviveif it encounters a newvirulentstrainof virus. Butthroughontogenetic variationandselec tionoccurring withinananimal 'snervous andimmunesystems , it ispossi blefor theanimalto surviveandreproduce evenif itsenvironment isquite differentfrom that of its ancestors . Theadaptation resultingfrom such lifespan evolutioncanoccurquiterapidlyandpermitanorganism to cope withacomplex , changing environment , animportant partof whichismade of its felloworganisms with whichit mustcompete for food, shelter , and mates . Wenotedin previous chapters , however , thatsuchadaptation appearsto bebased onthesamebasicmechanisms of cumulative blindvariationandselection thatunderlie themuchslowerandlessresponsive process of phylogenetic evolution . Butthereisa thirdway,whichweconsidered onlybrieflyin chapter 9, in whichadapted complexity canarise . Althoughstrictlyspeaking it isa form of learning , it differsin someimportantrespects fromthetypeof learning wehavealreadyconsidered . Manyorganisms , particularlymammals and birds,areableto makegooduseof theexperiences of theirfellowcreatures . Eventhehoneybee is ableto profitfromtheexperiences of hivemates who havediscovered a newsourceof food. Thislearningfromthelearningof othersis nowheremoredeveloped thanin humanswhereit is generally referredto asculture . Its rolein boththeremarkable proliferation of our species to all cornersof theglobeandin ourcontinued survivaldemands thatweconsider itsadaptedness andexplainits originandevolution . Tradition Of RiceandReligion Riceishumankind '8singlemostimportantfood, providingnearlyathirdof all thecalories weconsume .4In tropicalclimates wherethreecropscanbe growneachyear , asmuchassixtonsof ricecanbeproduced annually from onehectare(2.47 acres ) of land. Andthiscanbeaccomplished fromthe same landyearafteryearfor centuries withouttheadditionof chemical fertilizersandpesticides .

Cultural Knowledge

153

Toachieveand maintaintheseimpressive yields,special,labor-intensive

cultivation practices arerequired. First,seedlings areobtained bysowing riceintospecially prepared seedbeds. Afterseveral weekstheseedlings are

transplanted byhandto a flooded fieldor terrace wheretheyareleftto growinSto20cm(2to8inches) ofwater forseveral months. Twotothree weeksbeforeharvest, thepaddyisdrained. Thericestalksarecutandtied intobundlesandallowed to drybeforethreshing separates thegrainfrom the rest of the plant.

Themoststriking aspect ofricecultivation istheuseofflooded fieldsor terraces, andthisis thekeyto obtaining high,recurring yields. Flooding

helpsto keepessential nutrients inthesoilthatwould otherwise belost through leaching andexposure totheatmosphere. Under flooded cultivationothernutrientscontaining iron,aluminum, manganese, andcalcium

release phosphorus forusebytheplants,instead ofit remaining chemically boundto thesemineralsas it is whendry.Morenitrogenis alsoavailable, sincetheoxidation ofammonia intonitratesisretarded,withtheresultthat

nitratesareproduced at aboutthesamerateas theplantscanassimilate

thenitrogen theycontain (much ofthis.nitrogen wouldotherwise belost intotheatmosphere). Inaddition, blue-green algaeanda smallwaterfern calledazollaestablish residence in thericepaddies.Theseplantshavethe

rareability totakeinnitrogen directly fromtheatmosphere, which isthen madeavailable to thericeplantasthealgaeandazolladieanddecompose.

Finally, theability oftheplants tothrive inflooded fields despite thelackof oxygen attheirwaterlogged rootsisduetotheirremarkably efficient systemofairpassages connecting theirleaves above watertotherootsbelow. TheIndonesian islandof Baliis in manywaysideallysuitedto theculti-

vatioiiofrice.Lying almostdirectly ontheequator, temperatures varylittle throughout theyearfromits26 C(790F)average. Raincanbeexpected to fallabout200 dayseachyear,andhumidityremainsat a fairlyconstant 70%to 80%.Themajority ofthepopulation inhabits thesouthsideofthe

islandontheslopes ofmountains whoselakesandriversprovide a reliable andcontrollable supplyofwaterforirrigation. It shouldtherefore comeas

nosurprise thatriceandwaterarekeyelements inBalinese traditional life. Every morning thewomen layouta fewgrainsofcooked riceonsquares madefrombananaleavesasofferings ofthanksto Wisnu,thewatergod, and DewiSri,the goddessof riceand fertility.

154

WithoutMiracles

Because wetland ricecultivation requires periodic flooding anddraining , anextensive network ofhillside terraces andirrigation systems exists onthe mountainsides . Butmorethanthisisrequired to ensure efficient riceproduction . Although theclimate issuchthatricecanbegrownatanytimeof theyear , thetimingandmovement of waterdowntheterrace slopes must becarefully planned . Ideally , planting should bearranged sothatthewater canbemoved downtheslope , providing forperiodic flooding anddraining thatcorresponds tothegrowing cycleofeach terrace . If decisions concern ingplanting andharvesting timeswereleftupto eachindividual farmer , therewouldbetheriskoftaxingthewatersupply beyond itslimitsduring onepartof theyearwhileseriously underusing theavailable waterduring otherperiods .

Cultural Knowledge

155

flood, when to plant, when to drain and so forth . All he hasto do is follow the calendar of Dewi Sri, the goddessof rice and fertility , and the crops are virtually guaranteed. Rice cultivation is the ultimate expressionof the Balinesereadinessto follow the edicts of some greater authority: the cult of the rice goddessnot only demonstratesthe integration of the secularand the spiritual worlds of Bali at the most fundamentallevel, it also provides an eloquent example of the functional significance of religion in human ecology.7

The fit between the cultural practices of the Balineseand the growing requirementsof rice is well evident. But what is the explanation of this fit ? If one were to ask a typical Balinese rice farmer , he would no doubt say that

rice cultivation practiceswere provided by the rice goddess, sinceit is only by observingher calendar that continued good crops are ensured. Indeed, the fact that rice cultivation is so successful would appear to be quite con-

vincing evidencethat Dewi Sri is both knowledgeable and kind, and deservesrespect, thanks, and praise, all of which shedoesin fact receive. But quite a different picture emerges if we try to understand the farmers '

knowledge of rice cultivation from a naturalistic perspective, that is, as somehow originating and developedwithout divine providence. This may first seemproblematic, sinceit is quite unlikely that they would know anything about the extraordinary

oxygen transport system that permits rice

plants to thrive in flooded terraces, or the intricate biochemistry of wet-rice cultivation that ensuresthe plants an adequatesupply of essentialnutrients without depleting the resourcesof the land. So it might first appear that their farming knowledgemust havebeenprovided to them from somewiser source. But if we recall the remarkably fit products that biological evolution has been able to produce through the processesof cumulative blind variation and natural selection, we can appreciatethat the rice-farming practices could have evolved similarly without the Balinesehaving explicit knowl edge of the underlying reasons for their effectiveness. In much the same way

birds have no formal understanding of physics, yet are able to use Bernoulli's principle quite successfullyto ly. If over a long period of time biological evolution has produced palm trees, lobsters, panda bears, and humans, it should not seemunlikely that cultural evolution could result in an efficient systemfor the Balineseto produce their staplefood. Indeed , the view that culture originates and changes over time in a man -

ner analogousto the cumulative trial and error of biological evolution has become increasingly popular among social scientists. As anthropologist John Readerexplains:

156

Without Miracles

CU[tUl'a/ Knowledge

157

areimportantadvantages for individualsto adoptthe agriculturaland othertraditionalpractices of themajorityof theircommunity .9 Sosincemanyaspects of traditionalricecultivationarenotindividually testable , weshouldnotbesurprised to findthats'omeof themarenotfunctionalor areevenmaladapted to therequirements of riceproduction . The difficultythat an individualwouldencounter in attemptingto analyze whichaspects areactuallywelladapted andwhicharenot, savefor thefact thattheymayplayimportantsocialfunctions , alsoargues for therational ity of accepting andobserving thetotalculturalpackage . HiddenAdapted Aspects of Tradition Ontheotherhand,manytraditionalcustoms , rituals,andtaboosthatmay at firstappear to lackanyadapted qualities or mayevenseem maladapted , may turnoutto haveinteresting adapted character .. oncloserexamination istics.A classicexample is the sacredcattleof India. With a population second onlyto that of China,Indiahasdifficultyfeedingherpeople . Yet in themidstof thisgreatneedfor food, millionsof cattleareallowedto roamthecountryside , tramplegardens , andsnarltraffic. According to traditionalHindubelief,thebodyof a cowishometo millionsof deities , and thenextreincarnation of thecow's soulwill takehumanform; killingthe cow sendsits soulbackto squareone, and86 morereincarnations are required to achieve cowstatusagain . ThusIndiansputupwith theinconvenience posedbythecattleandwouldneverconsider butchering andeatingone. Thiswouldappearto beanexample of howtraditionalcultural beliefsandpractices canbe maladapted with respect to basicbiological needs . "OrthodoxHinduopinionregards thekillingof cattlewith abhorrence , eventhoughtherefusal to kill thevastnumberof useless cattlewhich existin Indiatodayisdetrimental to thenation."lo If wetakea morecarefullook, however , wecanseethatgoodreasons existfor theapparent folly of India's sacred cattle.In a classic paperpublishedin 1966, Americananthropologist MarvinHarrisproposed that thesecattleareonthewholebeneficial to Indiansociety . Theyprovidesignificant of milkandmeat(thelattereatenbyMoslems , Christians , ----------amounts andlower-casteHindus) aswell ashidesfor India's leatherindustry , the world's largest . Morethan300milliontonsof manurearecollected each yearfor useascookingfuelin ruralhouseholds , 90% of whichhaveno

158

Without Miracles

other source offuel. Inaddition, thecattle areregularly pressed intoservice asoxen, allowing millions offarmers simultaneously andquickly toplow

theirparcels oflandtotakefulladvantage ofthesudden monsoon rainsfor

their crops.

Soweseethataspects oftraditional culture thatmayatfirstappear to beatoddswiththerequirements ofhuman lifeandreproduction may

actually turnouttobeinsome respects quitefunctional, even ifindividuals

areunaware ofthese adapted aspects andoffer other reasons fortheir prac-

tices,suchasconcerns aboutreincarnation. Thisdoesnotmeanthatthe

cattle situation inIndia could notbeimproved, onlythatifa striking although puzzling cultural practice haspersisted over along period oftime,

wemaywellexpect to findsomefunctional reasons foritsexistence ifwe

look long andhard enough. Buteven ifwedolook long andhard, wemay

stillfailtouncover functional explanations formany practices andbeliefs

foundthroughout theworldinbothso-called traditional andadvanced societies. Cana theory ofcultural evolution basedoncumulative variation andselection ofuseful beliefs andpractices account forthis?

Toattempt tomake sense ofthepuzzle ofapparently maladapted traditions, another problem willhave tobeconsidered. Many aspects oftraditional culture maybebeneficial tothesociety asa whole, butmaybe detrimental totheindividuals whopractice them. Forexample, inmost

societies people areexpected to respect theproperty ofothersandto do theirfairshareofwork.It iseasytoappreciate howthisisbeneficial for

thesociety asa whole, since ifeveryone simply stolewhattheyneeded

noonewouldproduce therequired goodsandservices. Yetto theindi-

vidual, itisadvantageous, atleast inabiological, evolutionary sense, tobe

selfish indoing theleastamount ofworkpossible, andtodevote asmuch of

onesenergies aspossible tostaying healthy, mating, reproducing, anden-

suringthesurvival andreproduction ofonesownchildren. Thecostinvolved totheindividual incooperating withsocietys conventions ismost

dramatically apparent inorganized warfare, where menandwomen are

expected tolaydown their lives forthegood oftheir community. Individual

costsareinvolved inallinstances ofaltruistic behavior, thatis,whenone individual paysthepriceofreduced reproductive success thatbenefits

anothers reproductive success. Thus wehave thedilemma ofexplaining howindividuals whoareinreproductive (genetic) competition nonetheless

quiteoftencooperate.

Cultural Knowledge

159

Onewayofunderstanding howsuchaltruistic andcooperative practices couldhavearisenis to considerthe geneticrelatednessof the potential

cooperators. Ifclosely related individuals shareahigher proportion oftheir genes thandolessclosely related ones,wecanunderstand whya mother wouldriskher lifefor herchildren,sincesheis alsolikelyprotectingthe

genes thatpredispose hertodoso.W.D.Hamilton madethefirstconvincingarguments forkinselection asanexplanation foraltruism intwopapers published in1964.These papers provided anexplanation forhowcooperative behaviorcouldevolvegenetically (that is, by biologicalevolution)

among genetically closerelatives. Andthereisconsiderable evidence that forhumans, cooperation ismuchmorelikelyto occur,anddangerous violenceis muchlesslikelyto occur,amongcloserelatives. 11

Butwealsofindcooperative behavior amongquiteunrelated members of

a species, andeveninvolving different species. Howcanthisbeexplained? Oneexplanation isthatit canbeadvantageous foraninteracting groupof organisms tocooperate instead ofcompete iftheriskofnoncooperation is high(forexample, deathincombat orlackoffood)andiftheindividuals involved arelikelyto meet(andrecognize eachother)again.RobertTrivers coinedthe termreciprocalaltruismto describethis typeof cooperation, 12

andpolitical scientist Robert Axeirod demonstrated using gametheory and computer simulations howcooperation amongunrelated individuals can evolve and remain stable. 3

Forsuchcooperation to evolve andpersist,however, it mustbepossible torestrict thereceipt ofaltruistic deedstothoseindividuals whowilllikely

reciprocate. Otherwise, cheating intheformofaccepting thebenefits of

others altruisticdeedswhileprovidingnothingor littlein returncould

become rampant andeventually undermine thecooperative natureofthe community (those attempting to cooperate become suckers ifrecipients do not reciprocate). It is herethat otherwiseapparentlyuselessor even

maladapted traditions maybeemployed to indicate membership in the samereciprocally altruistic community. Inthisrespect, easilyperceivable homogeneities indialect, dress,rituals, andscarification wouldbeparticularlyuseful. ThustheLuoofKenya knock outtwofrontteethoftheirmen, whiletheadjacent Kipsigis enlarge a holepierced intheirearstoa two-inch diameter. 4 Although in itselfa particular mannerofdressor speechmay seemto have no valueas an adaptation,it maywelltake on importanceas

160

Without Miracles

an easilyperceivable signof groupmembership andtherebyfacilitateingroupcooperation . In manyAmerican citiestoday , themannerin whicha youthwearshisbaseball capcanmeanlife or deathashepasses through areasin whichrivalgangs , distinguished in part by howtheysporttheir caps , compete for control. Fromthisperspective , a numberof otherwise puzzlingaspects of traditionalcultures become somewhat lesspuzzling . Forexample , somestriking similarities occurredamongthefirst city-statecivilizations to emerge in Africa, China,theMiddleEast , India, andtheAmericas , including : 1. A division of headed by a supreme , tyrannical ruler . laborandpoliticalcentralization 2. A. beliefin a supernatural cosmology that providedauthorityfor the ruler 3. A setof moralizing preachments that"preached thevalueof dutyto the politicalorganization anditscustoms . . . thedutyof self-sacrificial military heroismin defense of thestate. . . within-grouphonesty " and"preached against self-interested deviations fromduty(covetousness , jealousy , etc.)."15 4. A belief in an afterlife of "rewarding and punishing heavens , hells , and reincarnation ." 16 In markedcontrastto suchuniformity , thesecivilizations hadstrikingly divergent beliefsconcerning thespecifics of the supernatural forcesthat appear to haveplayedrolesin thedailylivesof theirinhabitants . Different godswerepraisedandplacated byvariousritualsandtaboos . Suchdiversity arguesfor the independent originof thesebeliefsandpractices , althoughthefunctionof eachseems quitethesame - to inhibitselfishbehaviorsthatcouldundermine thecooperative structure of thecommunity , andto encourage cooperative acts. To thisend, all of thesecosmologies (asdo almostall existingreligions ) appearto haveincludeda beliefin an afterlifeinvolvingrewarding heavens andpunishing hells.Wasteful royal funerals , in which"fully usefulhorses , soldiers , wives , weapons , jewels , andmoneywereinterred " 17area dramaticaffirmationof sucha belief. These funerals area strikingexample of apractice thatisdownright waste ful in squandering hard-earnedcommunityresources , but that may be adapted at thesociallevelasevidence of the realityof anafterlife . Thus theymightmotivateindividuals to foregoselfishness in thislife to obtain therewards of theirsociallyadaptive cooperation in thenext.

Cultural Knowledge

161

Soit doesappearpossible to accountforat leastsomeoftheapparently

adapted andmaladapted aspects oftraditional cultural beliefs andpractices usinga modelofcultural evolution thatworksatthelevelofthegroup.It alsodepends oncumulative variation, selection, andpropagation notof

genes, butratherofshared beliefs andpractices. Because suchaspects of traditionhaveprovedthemselves by the veryfactthat theyhavebeen aroundfora longtimeandthatmostpeopleofa community observe them, it makessensefor individualsalso to adopt them. But they are not con-

firmable asadaptive bytheindividual; indeed,evenexpressing doubtasto the truth of cherishedtraditionalbeliefsis oftenreasonenoughfor banishmentfromthe societyor worse.In addition,traditionsthat distinguish

members of onecommunity fromrivalonesdistinguish cooperative individualsfrom selfishones.Therefore,we can expectsomeof traditionto

appear useless andevenmaladapted tothebiological demands ofsurvival andreproduction, suchasexorbitantly wastefulroyalfunerals. 18 Technological Change

Thesecondaspectofculturethatwewillconsider istechnology. Although

todaywetendtothinkoftechnology asinvolving powerful machines and sophisticated electronic equipment, in its moregeneralsenseit refersto

anytoolsormethods thatallowaneasier ormoreefficient production of useful goodsandservices. Ahorse-drawn plowisa moreeffective andproductive wayofpreparing soilforcropsthanisa shovel orhoe;netsforfish-

inghaveadvaiitages overspears; andforcovering longdistances, airtravel issuperior inspeed, safety, andcomfort (ifperhaps notincuisine) to land andseatravel.Otherspeciesmayusetoolsto a limitedextent, 19 butnone

compares withusintermsofthevariety andsophistication ofourtechnological achievements. Eachgeneration inherits crucialknowledge in the formof toolsandprocedures for providing food,clothing,shelter,health

care,andprotection fromenemies. Theacquisition oftechnology byindividualsthrougheitherformalor informaleducationis essentialto a societys continuedsurvivaland prosperity.

Notallwouldaccepttheclaimthatall,or at leastmany,culturaltraditionsarewelladapted inanyobvious material sense, although itwasargued abovethat numeroustraditionalpracticesandbeliefsarewelladaptedin at

162

Without Miracles

leastsome respects. Theadapted nature oftechnology anditsprogress is

harderto doubt.Toincrease production or provide betterservices at less costandeffort,technology mustachieve an unlikely fitbetween itstools

andtechniques ononehandandthematerials withwhich itistowork, the constraints ofthephysical world, andthelimitations andpreferences of human operators andconsumers ontheother. Nonetheless, human history isreplete withexamples ofinnovations thatsuccessfully (andincreasingly) mettheformidable oddsimposed bythese demands. Stone tools, thewheel, thehorse collatheplow, iron,paper, thespinning wheel, theclock, thecottongin,theprinting press, steam andinternal combustion engines, barbed wire,theelectric motor, thevacuum tube,thetransisto ; andthecomputer microchip arejustafewexamples ofitems developed inChina, Europe, and theAmericas. Because every successful innovation provides ademonstrably better wayofproducing some good orproviding some service, eachrepre-

sents yet another puzzle of fit.

Thesimilarity between thebiological evolution oforganisms andthe

technological evolution of tools,machines, andinstruments hasnot es-

caped theattention ofthoseattempting todescribe, explain, andpredict technological progress. Asearlyas 1863English writerandcriticSamuel

Butler explored thetheory thatmachines develop ina waythatresembles

theevolution ofliving organisms. Indeed, Butler wassoimpressed bythe rapidevolution ofmachines during theIndustrial Revolution thathepredicted thattheywould eventually constitute anewclass ofliving things that

wouldsurpass thesophistication ofhumans andrelegate ustosecond-class status. 2

Although Butlersaccountoftheevolution ofmachines wasmoresatire

thanscience, therehavebeenserious attempts tounderstand thedevelopmentoftechnology asanevolutionary process. Three recent books provide

good examples.

The Evolution of Artifacts

In TheEvolution of Technology, 21 George Basalla continues Butlerscon-

sideration ofmade things (artifacts) asanalogous toliving organisms. First, heargues forthecontinuity oftechnological innovation: Anynewthing thatappears inthemadeworldisbasedonsomeobject already inexistence. 22 Newinventions donotemerge magically fromtheminds ofgreat

CulturalKnowledge 163 inventors 1793 was

, but

rather

Eli Whitney based

years

to

motor

remove

seeds

1831

engine . The

physicist

charka from

copied

, Brattain

Braun

points

out

nomic our

, and

proclivity

when

one . An example

designs

that

from

were

in the unsuccessful wood

- burning

In the same way there

are often

and be taken

that

From

the vast pool

's electric

in the

at Bell Laboratories to the work

steam

in 1947 of German

certain

of artifacts

that

are available

display

more

crystals

the fertility

to make

diversity

in psychological

is the more States

based

than

during

-

imagination

copying

an artifact

to than

, socioeco

of human

, and the imperfect

that

,

invari

-

on an already

1000

smokestack

the nineteenth

the escape of embers

cen -

and sparks

.23

organisms

element

Henry

of

that

to prevent

variations

up by the next

essential

thousands

1870s , found

in the United

attempt

more

third

including

locomotives

. In

cotton

.

of this diversity

patented

in use for

employed

much

cultures

attempts

artifacts

short - stapled

. Joseph

can be found

and fantasize

a person

existing

tury

certain

factors

to play

ably results

transistor owed

of this diversity

cultural

cotton

the diversity

that

been

mechanisms

who , in the

existing

seeds from had

one direction

society , recognizing

others . The source

of the

Shockley

in only

Second , Basalla

, which

of the first

, and

Ferdinand electricity

of previously

removed

long - stapled

many

development

by Bardeen

conduct

gin that

on the Indian

of

any

are modifications

's cotton

produce

more

of a given

generation

of Basalla

tool

offspring

of users . Thus

's view

can survive

than

selection

of technological

of human -designed variant

than

or machine

,

can survive becomes

evolution

the

.

artifacts , a few are selected to be -

come part of the material life of society . In nature it is the ability of the species to sur vive that counts - the fact that the organism , and especially its kind , can thrive and reproduce

in the world

in which

vive and pass on its form

it finds itself . The artifact

to subsequent

requires the intervention of human cation in workshop or factory .24

Technology

as Knowledge

Joel

's The

Mokyr

technological innovation occasional successful ductivity

Lever

advances as the

" free lunches innovation it makes

intermediaries

of Riches25 in Europe

major

motor

possible

. Like

who select the artifact

China

of economic

for

may also be said to sur -

of made things . This process

is an engaging and

" and many

is paid

generations

" cheap

historical

. Mokyr growth lunches

many

times

Basalla

, Mokyr

over

for repli -

account

of

sees technological , a way

of providing

" in that

the cost

by the increase

suggests

of a

in pro -

an evolutionary

,

164

WithoutMiracles

selectionist

account

primary man

role

for

The

I adopt

of species , and

genotype

, whereas

type

may of every

a role

as well

Invention emergence

' exists

's in

' outside

major

one

creation

of

that

model

strict

a compelling development Vincenti arriving refer

for

Campbell

the

biological case

the to

way

growth

's work

What

that

, the

somehow

more

concerned

- adjusted

cost

technologies

of

, he

technological Engineers

" as

present

detailed

case

the

success

of the

to garner

evidence

engineering

breaks

employed

from by

the

Basalla

a more

general

selectionist

the

advantages

of

is the

and

They

histories

How of

modern for

knowledge

away

achievements Know

five for

cases

analogy for

shape .

epistemological

is also quality

competing

take , but

exact

to speciation

and

important

airplane

a " variation ."

and

heavily

imposed Mokyr

account

by

and

of

. 28 In - selec -

Drawing

limits

the

makes

technological

.

successful actual determine

designs

building its

for

of a new degree

vicarious

technological

of

over

innovations

device

and

trying

it out

success

or

failure

. Such

direct

trials

. Direct for

its

direct

-

plays

Selection

of all

fly . In

these

emphasizes at

to

the

and

presents

paved

the

com -

pheno

.

that

, he uses

the

can

something

" lowest

. The

environment

its

,

are

of as the

in producing

equivalent

perspective to

among

character

a technique

" is not

word

way -

be thought

determine

psychological

marvelous

Vincenti

firm

forms

more

Variation

most

may

help , is thus

. " 27 His

selecting time

machines

chapter

rather

purpose

of the

It , Walter

innovations

tion

for

an evolutionary

technique

he gives

of the

, but

technology

's brains

as Vicarious

Certainly

last

that

sense

in a specific

by its genotype the

hu evo -

of a species . The

circumstances

a

of

technological

narrow

by the

giving

of the member

constrains

to

priority

at anyone

Technology

from

in the

have

of

growth

.

or service

a commodity

in part

idea

in

species

in the

utilized

is therefore

of people

criterion

in a society

Know

, the

insisting

economics

the

technique

adjustment

-

in them

instead the

analogues

a good

to produce

is determined

's approach

Basalla

with

changes

, but

emphasizes

living

techniques

, the emergence of a new of a new species .26

Mokyr

of

to produce

that

actual

. Similarly and

gives

of as the phenotype

organism

adaptability

than

the

and

is that

of how

be thought

, he

phenotype

of how

or conceptualization

modity

unit

and here

development

themselves

basic

genotype

, the knowledge

analogues idea

technological

artifacts

as the

the

approach

namely

of

the

knowledge

lution

his

to

in

trials

intended trials

of

Cultural Knowledge

165

overtvariationscharacterizethe attemptsof the Frenchbetween1901and

1908to develop flyingmachines. 29 DuringthistimetheFrenchbuilta wide varietyof aircraftandtestedoneaftertheother,usually withdisastrous results.Notbasingtheirtrialson anysystematic attemptto discover the basicprinciples of flight,therewaslittlein the failureof onedesignto informthe designof the next,and thecontinuous building(andsubse-

quentcrashing) ofworking prototypes consumed considerable timeand resources.

In contrast,vicarious trialsfordesigning newtechnologies permitboth expansion ofthevariations to beconsidered aswellasa muchmoreefficient(intermsoftimeandenergy) selection process. Vicarious trialscanbe eitherexperimental or analytical. Experimental onesconsistof the substitutionofpartialexperiments or complete simulation testsforprooftestor everyday use. 3 Anexample istheuseofwindtunneltestingin aeronauticalengineering. TheWrightbrothersextensive useofsuchtestsgavethem animportant advantage intheraceto buildthefirstairplane, sinceduring the time it wouldtake their Frenchcompetitorsto buildand launchyet

anothercomplete prototype, theAmericans built,tested,eliminated, and designed manydifferent scalemodels andtherefore mademorerapidprogress in their search for a successful design.

Analytical vicarious trialsareevenfurtherremoved fromdirecttrialsin that theyusemoreabstracttoolsto testthe worthiness of a particular designvariation. Theycanbeconsidered a kindof test runonpaper, 31 although sincecomputers areincreasingly usedforsuchteststoday,theyare more often test runs on computersilicon.As each hypotheticalarrange-

mentof partsis sketched eitherliterally or figuratively onthecalculation

padorcomputer screen, thecandidate structure mustbechecked byanalysis.Theanalysis consists of seriesof questions aboutthebehavior ofthe partsunderthe imagined conditions of useafterconstruction. 32 It is through suchteststhatscientific knowledge canbeexploited intechnological development. 33 The Wrightbrothersprofitedfrom an analytical

approach dueto theirknowledge of theprinciples of fixed-wing flight developed intheearly1800sbySirGeorgeCayley. Thisknowledge permitted the brothersto analyzethe problemof flightinto threesubproblems

consisting oflift,thrust,andcontrol. Byattacking eachoftheseproblems separately usingvicarious experiments, theymaderapidprogress indesigningthefirstmachinecapableofsustained mannedflight. 34

166

Without Miracles

Tnemphasizing theimportance of vicarious variation andselection, Vincenti ineffect hasdescribed atechnology oftechnology, thismeta-tech-

nology providing better andbetter tools fordeveloping newtechnology and

making possible theever-quickening paceofdevelopment. Asthesetools develop, theypermit moreefficient andreliable vicarious testing ofvaria-

tionsonexisting ideas. Thisalsopermits a broader scopeofvariations that

canbetested efficiently. Ineffect therangeofvariations canactually be increased without detriment andpossibly withgreatadvantage, since the newvicarious meansofvariation andselection aresorapidandefficient in

weeding outtheunfitvariations. AsVincenti hasremarked, engineers

havefreedom tobeincreasingly blindintheirtrialvariations astheirmeans

ofvicarious selection become more reliable. Oneseesengineers today, for example, using computer models toexplore amuch wider field ofpossibili-

tiesthantheywereabletoselect fromjusta decade ago. 35 The Developmentof Science

Thefinalaspect ofculture tobeconsidered hereisscientific knowledge. Technological knowledge canbeputtopractical usetosolve problems concerning theproduction ofgoods andservices. Scientific knowledge provides explanations forobserved phenomena interms ofunderlying mechanisms, thusproviding a basis forpredictions andperhaps ultimately controlling thephenomena under study. Technology may help toadvance science, especially inproviding newtoolsforexploration suchasorbiting telescopes,

electron microscopes, andpowerful computers. Andscience canaidtech-

nology, asparticle physics madepossible nuclear weapons andnuclear powerplants.Butthetwocanbeseenasdistinct. TheAustralian bushmans

ability toproduce boomerangs istechnological knowledge thatdoesnot depend onscientific knowledge ofphysics andaerodynamics. Andthe greatest scientific discovery ofthenineteenth century owedlittletotechnol-

ogy, other thanthatinvolved inbuilding andnavigating theBeagle, theship

thatallowed Darwin toexperience firsthand theworldsremarkable diversityoflife,inparticular thesubtledifferences in species foundscattered

amongthe GalapagosIslands. 36

Science provides whatmanyconsider tobethemoststriking andundeniable instances offitofonesystemhuman knowledgeto anotherthe

Cultural Knowledge

167

universein whichwe live.It is for thisreasonthat countlessphilosophers andscientists haveattemptedto understand howthisremarkably complex and fit aspectof our cultureis achieved. Bacons SystematicMethod of Induction

SirFrancisBacon(15611626)livedin Englandduringthe dawnof the modernWesternscientific andtechnological era.He wasgreatlyimpressed

bytheastronomical discoveries ofCopernicus andGalileo, which provided a radically newviewoftheuniverse, andthetechnological achievements represented bysuchinventions asgunpowder andtheprinting press.But Baconhadlittlepatience forthespeculative philosophers ofhistimewho, hebelieved, hadmadelittleif anyprogress in understanding naturesince thetimeoftheGreeks. Hefeltstrongly thatit wasonlythrougha carefully

implemented scientific methodology thatwewouldbeableto understand theworldaroundusandregainthemastery overnaturethatwaslostat the time of the fall of Adam and Eve.

ButBaconbelieved that morethanjustmasteryovernaturehad been lost.Thehumanmindwasalsocorruptedbyfalseimagesor idols,which

causedit to perceive theworldin unreliable andsubjective ways.These idolsincluded thetendency to seekoutandseeonlyconfirmations of al-

ready heldbeliefs, andtheproclivity tonotice themorestriking aspects of theworldwhilefailing toseethesubtlerones.Bacon heldthattheseimpedimentstoobjective andaccurate observation hadtobesweptaway, leaving themindas a blankslatesothattheworldcouldaccurately transmitand impressitstruenatureon the humanmind. Tothisendhe offereda systematic methodof inductionbasedon care-

fulobservation, comparison, andexperimentation. Forinstance, to understandthetruenatureof heat,onehasto findmanyinstances of heatand determine by observation whattheyhavein common. Butpositive instancesalonedo not suffice.Onemustalsofindnegativeinstances,in par-

ticular, instances thatareverysimilar tothepositive onesbutthatlackthe quality ofinterest. Itisonlybylisting andcomparing instances inwhich the phenomenon ofinterest ispresent withthoseinwhichitisabsent thatthe truenatureofthephenomenon canbeascertained. 37 Bymechanically fol-

lowingthismethod, Baconbelieved thatonewouldeventually arriveat validinductions concerning themakeup andbehavior ofthenaturalworld

168

WithoutMiracles

that

could

then

phenomena

.

pirical

the

's

of

' s

and

such

as

nent

,

Newton

both

empirical

,

.

But

this

death

,

referred

that

natural

to

today

careful

) ,

Newton

his

as

em

empirical

justified

-

ob

-

knowledge

laws

of

of

based

on

Darwin

of

,

acknowledged

societies

on

,

the

it

the

was

unbiased

observation

re

a

-

of

movements

only

be

-

success

govern

eventually

,

conti

scientific

remarkable

that

England

scientific

systematic

the

that

would

in

institutions

type

Indeed

nature

impact

various

similar

hinted

secrets

great

and

the

.

a

death

and

bodies

's

had

after

undertake

method

not

of

laws

to

and

of

least

partially

tance

he

power

( see

attempts

observation

And

therefore

him

hanging

escaped

of

matter

of

uncovered

time

using

and

the

objective

it

a

a

,

the

dreams

' s

emphasis

apparently

pulled

In

.

,

it

problem

,

,

as

he

as

Bacon

by

in

from

which

unprejudiced

that

indicated

shows

his

out

by

by

appear

the

after

induction

justified

would

and

century

rug

of

and

fact

Bacon

A

epistemological

method

discovered

.

by

influenced

the

instances

gods

,-

that

of

"

have

Aye

vows

omens

,

with

essential

did

answer

picture

would

their ,

the

a

and

concern

on

)

impending

good

temple

after ,

envisaged

he

foolproof

be

negative

was

shipwreck

of

a

world

the

manner

this

was

the

at

impor

account

of

-

the

:

in

drowned

6

at

could

of

on

prayer

the

whom

chapter

natural

aware

power

in

arrive

theories

the

placed

of

happen

scientist

Hume

all

general

to

empirical

David

under

it

belief

positive

Both

advocated

celestial

nature

was

breed

Bacon

the

( or

control

.

new

astrology

is

method

England

the

inductive

comparIson

were

to

and

of

its

.

in

discovering

all

and

.

shortly

had

terrestrial

before

on

and

Society

Bacon

in

laws

and

established

that

based

continent

,

Royal

were

search

predict

knowledge

certain

science

Bacon

the

is

with

and

European

to

scientific

it

us

on

the

debt

since

to

Induction

writings

on

their

of

content

Problem

deductively

view

provide

universe

Bacon

applied

a

,

can

The

But

Such

positivism

servation

the

be

asked

? "

And

say

he

by

again

the ,

"

but

the

like

instances

of

Bacon

not

are

all

in

to

showed

as

having

acknowledge

they

superstition

painted

,

that

whether

in

. 38

Popper

science

wonder

they

vows

now

anticipates

refutation

to

did

of

when

their

where

way

or

who

paid

he

,

is

one

had

whether

such

negative

occur

made

who

judgements

role

not

him

, "

divine

was

those

( to

,

as

it

be

discussed

did

later

's

similar

next

to

Hume

) .

,

Cultural Knowledge

169

thatif a negative instance(forexample, experiencing silentlightning) can leadoneto rejecta belief(thatalllightning produces thunder)previously

supported bycountless observations, howcouldonebesurethatforany general belief supported byobservation thata negative instance doesnot exist somewhere?Inasmuchas it is not possibleto provethat negative instancesdo not existsomewhere,it is not possibleto be absolutelycertain

ofthetruthofanygeneral belief, scientific orotherwise, nomatterhowwell the beliefhas beensupportedby past empiricalobservations.

Despite thisserious logical andmethodological difficulty, however, sciencedoesappearto makeprogressin waysin whichotherbeliefsystems,

suchasreligion, astrology, andpalmistry, donot.Ifscience isindeed ableto attainprogressively betterfitto theworldit describes, weareleftwiththe

puzzle ofhowthisfitisachieved, since itisevident thatempirically based induction ofthetypeadvocated byBaconandfoundwantingbyHumeis

simply notupto thetaskofproviding infallibly trueknowledge ofthe world.

Popper and Falsification

Sciencecontinuedto make importantbreakthroughsat an accelerating

pacefromtheseventeenth through thetwentieth centuries. It couldbe argued, however, thatnocomparable breakthroughs occurred in understanding howscience wasabletocontinue tomakeprogress, thatis,achieve betterandbetterfit withthe objectsandphenomena it described, untilSir

KarlPopper(19021994) confronted the problem. Poppergrewup in Vienna duringa timeofgreatintellectual andscientific activity inEurope in

general andintheAustrian capital inparticular. Asa student ofthephilosophy ofsience, hewasfascinated bytheability ofscience toachieve betterandbetterfitstotheworldit described, andconsequently hesethismind to determining whatit wasthatsetscience apartinthisrespect fromnonscientificdomains.What was it that allowedNewton and Einsteinto pro-

posetheories thatwereconvincingly betterthanthoseprovided bytheir predecessors, whereasthepolitical andeconomic theoryof communism offered byMarxandthepsychoanalytic theorydeveloped byFreudwere notdemonstrably betterthancompeting theories? Popperproposed a sim-

pieyetboldsolution tothisproblem ofthedemarcation ofscience from nonscience andin so doingoffereda solutionto thevexingproblemof induction raised by Hume.

170

Without Miracles

According to Popper, whatsetsscience apartfromnonscientific beliefs

andtheories isthatscientific theories arefalsifiable. Forinstance, oneof Newtonslawsof physicsis thatforceis equalto masstimesacceleration.

Thistheorycaninprinciple befalsified byexperimentation, sinceit makes specific testable predictions. Ifthetheoryiscorrect, it shouldtakethesame

force toimpart acertain acceleration toanobject asittakes toimpart twice

thisacceleration to anobjectwhichhashalfthemass.Ifthisisfoundnotto

bethecase,andnomethodological errors havebeenmade, thetheory must berejected anda betteroneformulated andtestedinitsplace. ThiscontrastswithFreudspsychoanalytic theory,whichis formulated in sucha

wayastomakeitimmune fromfalsification. Ifpsychoanalytic theory says

thatallmalesarejealousoftheirfatherandcovettheirmotherbuta certain

maledenies having these feelings, a confirmed Freudian would argue that

themalewasrepressing histruefeelings ofpaternal jealousy andmaternal desire. Similarly, fields suchasastrology andpalmistry arenotscientific, sinceifa prediction doesnotproveaccurate, reasons canalways befound

afterthe factwhythingsdid not turn out as foretold. Poppersdiscovery oftheimportanceoffalsification alsohada sidebene-

fitinthatit solved theproblem ofinduction. Asalready noted,a scientific theorythatproposes a general,universal lawof naturecanneverbera-

tionally justified, since byvirtue ofitsuniversality itmustgofarbeyond the limited observations ofmortal scientists. Sonomatter howmany times itis observed thatevent Aisfollowed byevent B(forexample, heating water to 1000C causesit to boil),it cannotbeprovedlogically thatallAarefol-

lowed byB.Butwhereas apparently confirming cases cannot justify a scientific theory, disconfirming cases doallow ustorefute it.Finding a clear instance ofAthatisnotfollowed byB(forexample, finding thatinDeath Valley waterat1000 Cdoesnotboil)means thatourtheory mustberevised or abandoned. Thena newtheorymustbeproposed thataccounts forall

thattheoldtheory accounted forasthewellasnewfindings thatitcould nothandle.According to Popper, thefitofscience isnotdueto observation

andinduction oftrue,justified (orjustifiable) theories, thatis,theaccurate,

instructive transmission of knowledge fromtheenvironment to thescien-

tist.Rather, science progresses through thecreation ofconjectures (guesses) andthesubsequent weeding outofinadequate hypotheses, leaving those thatarebetter thantheoneseliminated onlybecause theyhavenotyetbeen

eliminated themselves. 39

Cultural Knowledge

171

It didnotescape Poppersattention thathisviewoftheprocess ofscientificachievement hadmuchincommon withDarwins selectionist theoryof biological evolution:

Thegrowth ofourknowledge istheresult ofa process closely resembling what Darwincallednatural selection;thatis,thenaturalselection ofhypotheses: our

knowledge consists, atevery moment, ofthose hypotheses which have shown their

(comparative) fitness bysurviving sofarintheirstruggle forexistence; acompetitive struggle which eliminates thosehypotheses which areunfit. 4

Poppers viewof scientific progress as a cumulative selection process throws interesting newlightonscience anditsachievements. First,inthe samewaythatbiological evolution depends ontheexistence ofblindvariationinthestructure andbehavior oforganisms, science depends onsimilar blindvariation in hypotheses thatareproposed. Thisdoesnotmeanthat

thehypotheses arenotconstrained bytheknowledge already achieved. No respectable scientist isgoing topropose thatthecoreoftheearthismadeup ofstrawberry jamorthemoonssurface isSwiss cheese. Andnowhaleis

likely togive birthtoahorse. Inbothbiological evolution andscience, such constraints reflectthepastaccumulation ofknowledge bypriorblindvaria-

tionandselection andareessential innarrowing downthetypesoffuture variations thatappear. Buttheconstraints alonecannotaccount forthe

emergence ofnewandbetterfitsoforganism totheenvironment, andscientific theory to the universe.

Second, anevolutionary perspective accounts forthetentative nature of scientifictheories.Eachnow-extinct species,whichtogethermakeup a

muchlargernumber thanthosespecies stillextant,hadbefore itsextinction beensuccessful insurviving forquitesometime,insomecaseshundreds of millions ofyears. Butinnoneofthesecases wasthislongperiod ofsurvival

(which isclearevidence offittotheenvironment) abletoguarantee the

futuresuccess ofthespecies. Similarly, thelong-term, popular acceptance of a scientific theoryinnowayguarantees thatit willnotbecome extinctas bettertheories evolve tocompete withitandeventually replace it.Thephlo-

giston theory offire,thecaloric theory ofheat,theethertheory ofouter space, Newtons theory ofmechanics, andLamarcks theory ofevolution haveall beeneliminated in the strugglefor survivaldescribed byPopper,

despite thefactthateachwasthebestanddominant theory ofitsday.Soin

addition to explaining howscience canachieve progressively betterfitto theuniverse, theselectionist viewasproposed byPopper explains whywe

172

Without Miracles

cannever justify anyparticular theory asabsolutely andinfallibly true.In thesamewaythatpreviously successful species become extinct, scientific

theories areeliminated andreplaced bybetterones.

Finally, a selectionist viewemphasizes thecreative roleofthescientist.

Popper contended thatthescientists mindisnotanepistemological bucket thatisfilled withknowledge fromtheenvironment through theeyes, ears, andothersensory organs. Thisisnotunlikea Lamarckian viewofevolution

inwhich theenvironment somehow instructs thereproducing organism

to createnewadaptedforms.Instead,the scientist actively constructs

knowledge intheformofunjustified theoretical conjectures, which arethen testedandcompared tocompeting conjectures. Inthisway,thescientists experience oftheworlddoesnotprovide thetheories tobetested. Rathe ; observation isusedtoweed outtheunsatisfactory onesalready construct-

ed.Asit isnotpossible topredict thecourse ofbiological evolution, itis

similarly impossible to predictthefuturecourseof science. Andbecause

technological andsocial changes arestrongly influenced byscientific developments, itissimilarly impossible topredict thefuture course ofhistory. 41 Asfalsification isthekeyingredient toPoppers philosophy ofscience, it

should come asnosurprise thatPopper valued serious attempts tofalsify all

proposed scientific hypotheses because itisonlyinthiswaythatthebetter theories canbeselected andthepoorer oneseliminated. Forthisreason, Popper believed it is important thatscientific theories begivenan objectiveexistence in theformofspokenor (evenbetter)writtenwords andothersymbols thatcanbedisseminated widely to otherscientists for

scrutiny, forexample, intheformofconference presentations andpublica-

tions. 42 Thecritical worldwide attention anyimportant newscientific theorynowreceives makesit increasingly difficult forinadequate theories to survive forverylong,as shownbytheanimated flurryofresearch and

refutation thatfollowed theannouncement ofcoldnuclear fusion bytwo

University ofUtahphysicists in1989. Indeed, itwillbeargued laterthatthe

increasingly strong global selection pressure putontheories isanimportant

factor insciences rapidprogress, andthatconsiderably lessglobal selection

pressure hasbeenputontradition andtechnology.

It should notbesurprising thata philosophy ofscience asradically

different as Poppershasattracted considerable criticism. 43 Butwhereas somephilosophers continue theirsearch fora completely reliable scientific method basedonthefoolproof induction ofgeneral scientific lawsfrom

Cultural Knowledge

observation

and

has

joined

Popper

perspective

,

,

Genes

,

We

,

now

plexity

of

,

But

fortunately

are

,

begin

to

what

in

selection

merciless

what

.

survive

and

theory

and

how

their

variation

role

of

replicators

down

,

from

parent

ronment

in

the

fit

with

of

its

the

organism

inanimate

' s

and

.

be

the

In

this

interactors

will

sense

,

and

in

the

stage

groups

of

evolution

had

of

with

are

' s

individual

.

to

its

determines

and

envi

But

maturity

is

at

it

is

any

and

be

the

-

interaction

whether

can

it

the

handed

offspring

organisms

,

,

probing

play

that

to

of

,

genes

the

surviving

to

adequate

blind

respect

that

no

that

the

the

structure

themselves

,

successful

or

now

behavior

,

he

molecular

with

behavior

and

unable

the

interact

environment

be

their

know

of

that

and

organisms

on

We

copies

and

form

form

animate

organism

ducing

.

.

of

organisms

,

,

-

not

theory

organisms

However

Un

evolution

the

provide

make

,

determining

the

particular

they

offspring

biological

fit

genes

?

does

proposed

less

.

of

selectionist

evolves

among

the

operates

since

to

of

recombination

selection

,

from

evolution

he

-

and

away

culture

details

competition

com

beliefs

cultural

when

nothing

adapted

traditional

which

the

numbers

knew

the

turned

of

of

the

and

which

more

. 44

evolutionary

details

sufficient

and

mutation

on

of

toward

eliminating

in

,

form

position

see

for

by

know

nature

inheritance

much

Evolution

increasingly

,

same

could

of

reproduce

of

the

have

the

we

the

He

hammer

in

-

of

much

a

Adaptive

science

transmission

view

account

mechanisms

know

with

was

natural

be

actual

we

instead

of

to

theories

the

compare

Darwin

it

science

instructionist

what

,

evolutionary

Forms

attempts

or

empiricist

taking

often

of

Interactors

whether

,

and

.

and

and

and

as

current

technology

providential

ones

,

-

,

Science

that

culture

behaviors

and

an

,

tentative

philosophy

rejecting

induction

,

Replicators

seen

mainstream

as

justified

,

,

modern

insofar

fallible

Technology

Memes

have

least

on

probing

Tradition

,

at

based

cautious

to

experiment

173

repro

-

considered

level

of

inter

-

actorsthatselection takesplace . Thedistinction between replicator (genotype ) andinteractor (phenotype ) is importantasit helpsto clarifycertainlong-standing problems andcontroversiesconcerning the units of selectionin biologicalevolution .45 Similarly , thedistinction maybeof usein understanding culturalevolution aswell. Butthenwemustask, whatarethereplicators in culturalevolution?Forculturalevolutionwerequireanentitythat is analogous to the

174

Without Miracles

genein biological evolution . Thisreplicating entityisnowreferred to asthe meme .46 Thememe . As originallyintroduced by Dawkins : is a unit of culturalreplication

Cultural Knowledge

175

Tradition,Technology, andScience: Similarities andDifferences

Participation oftheenvironment intheselection ofmemes wouldappearto benecessary forallformsofcultural fit,butthereareimportant differences in howthisparticipation takesplace.Adapted traditional beliefs andbe-

haviors, asseeninriceplanting inBali, takeaverylongtimetodevelop and change. Thatis because theselection process doesnotdepend onindividualsasinteractors, butratheron a muchlargersocialgroup.Indeed,as

already mentioned, individuals oftendonotevenhavetheoption ofintentionally changing traditional beliefs andpractices, sinceto dosomight resultin banishment fromthegrouporperhapsevendeath.Sotheselection

process fortradition operates veryslowly, at leastwhencompared with other forms of cultural evolution.

Thesituation isquitedifferent fortechnological development. Inexperi-

menting withnewtechniques forproducing goods andservices, individual artisansandinventors areabletotesttheworthofvariousmemesandselect the onesthat are bestsuitedto the taskat hand.Sotechnological change basedonindividuals asinteractors canoccurmuchmorequicklythantra-

ditional change basedonsocieties asinteractors. Totheextentthattech-

nology evolves means ofvicarious variation andselection (asintheWright brothersuse of windtunnelsand scalemodelsto test airplanedesigns), adaptive technological changecanoccurat evenfasterrates. Thesameistrueforscience. Butscientific development differsfromtech-

nological development inanotherimportant way.Intechnological innovation,the individuals whocomeup withpromising innovations aremost

oftenthesamepeople whotestthem.It wastheWright brothers andthey alonewhotestedallthevariations in wingsandpropellers theyimagined

might besuccessful. Butinscience today, theindividuals involved intesting new theoriesare not usuallyrestrictedto the originatorsof theories.

Important newscientific theories aretypically published in widely read scientific journals, andsuchpublication requires stringent peerreview by fellowscientists.The moreabstractnatureof scienceand the easilycommunicated mathematical and natural language in which its memes are

expressed makesit possible forscientific memes to proliferate quickly to othersworldwide. Thisputsenormousselectionpressureon the memes

since large numbers ofscientists willbeeager tofalsify anynewtheory that receives widespread attention, andmanyofthemmaywellpropose theories of theirown.Of course,exciting newtechnological innovations arealso

176

Without Miracles

eventually tried andtested byindividuals notinvolved intheir creation, and

much scientific testing requires newtechnology aswell.Butwecannone-

theless expect anew scientific theory such asE=MG 3 published inaprestigious physics journal toundergo more rapid andmore widespread attempts at falsification thana newdesignforautomobile tires.

Although some important differences exist intheevolution oftradition, technology, andscience withrespect totheselection process, inanother

respect theyareverymuchalikeinthatthenewmemes thataretested arise

bytheblindvariation ofpreviously existing memes. When humans first

attempted tounderstand theinvisible causes ofvisible events such aslightning ordisease, they conjured upmyriad gods, spirits, andother mysterious entities andforces, many ofwhich remain with ustothisday. When aphysi-

cistorbiologist attempts tounderstand thesame phenomena, entities and

forces nolessmysterious appear intheformofelectromagnetic fields,

unimaginably small viruses, andstilltiniersubatomic particles. Atthis

pointthereisnofundamental difference between superstition andscience. Thedifference liesintheselection procedure andinthefactthatscientific memes mustbeinprinciple falsifiable insuchawaythatthephenomena to

which theyrefer participate intheselection. Wecannot falsify thebelief thatevilspirits cause aparticular illness, since thespirits arebydefinition undetectable andmake theirpresence known onlythrough thedisease. We

can,howeverejectthehypothesis thata particular virusor bacterium

causes acertain disease. Butalthough science cananddoes make significant progress inthecumulative introduction andwinnowing ofnewmemes, it can never lead us to the positive, infallible knowledge scientists continually seek. The Myth of CulturalTransmission

Although attempts tounderstand adaptive cultural change asa selection

process analogous inimportant ways toadaptive biological evolution have become increasingly common, it shouldnotbe overlooked thatthis approach hasmany critics. Adiscussion ofthecriticisms willbesaved for

chapter 15,butoneofthemost cited differences between biological and cultural evolution willbeaddressed heretheclaim thatbiological evolu-

tionisDarwinian andculturalevolution isLamarckian.

Bythisitismeant thatbiological evolution isaprocess ofblind variation

ofgenes andsubsequent selection oforganisms containing these genes by

Cultural Knowledge

177

theenvironment, whereas culturalevolution proceeds bythetransmission

ofacquired characteristics. Onthesurface, thismayindeed appear tobethe case.Wecurrently knowofnowaybywhichanyadaptive changes to an

organisms formorbehavior acquired during itslifetime canbeencoded in itsgenes andconsequently passed ontoitsoffspring. AsWeismann claimed (andmolecular biology hasnotyetbeenabletorefute), changes ingenes canresultin changes in thestructure andbehavior of organisms, but

acquired changes inanorganisms structure andbehavior resulting from use,disuse, orlearning cannot beencoded inthegenes. Inotherwords, information isnottransmitted fromenvironment to genome. Theroleof

theenvironment issolelyoneof selection, notinstruction or transmission.

Incontrast, it isclaimed thatcultural evolution isLamarckian inthe sensethatthecultural knowledge thatonepersonacquires duringa lifetimecanbetransmitted to another. IfI discover a betterwayofgrowing

potatoes, I canshow youhowit isdoneandthenyoucanusethenew methodas well.Mynewlyacquired potato-growing practices canbe inherited culturally bymychildren, aswellasbyotherindividuals with whomI comein contactor who learnaboutmy new,adaptedmemes

through books, wordofmouth, orotherindirect means, andthusthis process mayinitially seem Lamarckian innature. Infact,many whohave

madethemostvaluable contributions indeveloping evolutionary models of cultural change speak ofcultural transmission inthissense, 48 andDawkins

wasquoted earlier describing howmemes propagate themselves . . . by leapingfrom brain to brain.

Butto understandhowculturalchangemightbe Lamarckian, we must

consider carefully justwhatismeantbytransmission andwhatistransmitted.It shouldfirstbepointedoutthata selectionist theoryof adapted complexity doesnotruleoutallformsoftransmission. Indeed, onecan

certainly consider thereplication ofgenes thatisrequired forreproduction tobegenetic transmission. Thiscanbeseenmostclearly inasexual organismswhoseoffspring arealmost always genetically identical totheparent.

Thus, theparents genetic codeistransmitted totheoffspring. 49 Butwhat isessential tokeep inmind isthattransmission isnotresponsible forthefit ofa genome toitsenvironment. Thisfitcomes aboutthrough Darwinian selection, notLamarckian transmission orinstruction. Ifweareinterested inexplaining puzzles offit,weareprimarily interested inhowthefitcomes

178

Without Miracles

about, andonly secondarily interested inhow itisreplicated andpropagated once itisachieved. Sogenetic transmission plays anindispensable role inbiological evolution, butitsrole isasecondary one ofpreserving and

propagating thefitthathasalready been achieved through selection. Having madethedistinction between theachievement offitandthe propagation orreplication offit,wecanbetter assess therolethatLamarckian instructive transmission might playincultural evolution. With respect toachieving theinitial fitofcultural knowledge toenvironment, modern postpositivist philosophy such asPoppers, aswell asmuch current psychological theory, convincingly maintains thatthisfitcannot bethe result ofinstruction from theenvironment totheindividual. Ifthere isafit ofknowledge totheenvironment, theroleoftheenvironment isoneof selecting fromamong thevarious andsundry knowledge memes that

already were created bytheindividual. Solikebiological evolution, anevolutionary view ofadaptive cultural development depends onselection for theachievement offit,andnottransmission orinstruction fromworld to

mind byway ofthesenses ofthetype thatBacon andother empirical posi-

tivistsbelievedpossible.

Butwhataboutthepropagation ofknowledge tootherindividuals? Do nottransmission andinstruction playaroleherecomparable tothetransmission ofgenetic information from parent tooffspring? Alittle reflection

suggests thattheanswer isno.Theonly waythatJohncould possibly obtain knowledge from Mary isthrough hissenses. Mary cannot transmit

hergenes toJohn, norcanshegive him herbrain orparts ofit,orreplicate thepatterns ofitsstructure inJohns brain. Like therestofthephysical world, Mary issimply partofJohns environment, knowable tohim only through hissenses. This does notmean thatJohn cannot learn from Mary

andherexperiences. Butaswesawinchapter 7,learning appears tobe

nomorepossible through instruction thanistheLamarckian inheritance

ofacquired characteristics. Accordingly, wewill seeinthenext twochap-

tershowcurrent theories oflanguage andeducation alsoreject thenotion oftheinstructionist transmission ofknowledge fromoneindividual to

another (asfrom teacher tostudent), andinstead view linguistic communi-

cation andeducation asprocesses dependent oncumulative blind variation and selection.

11~ TheEvolution , Acquisition , andUse ofLanguage

Of

all

the

and

behaviors

yet

are

It

tant

to

to

not

is

tool

for

to

,

roadways

people

it

,

share

.

paper

,

,

science

and

development

its

theories

more

spoken

and

technology

public

of

and

their

peers

thus

.

,

that

it

subject

and

lived

appears

essential

allows

both

on

,

and

for

the

make

a

stone

possible

places

s ,

allows

of

on

~

language

others

first

it

skepticism

us

da

Language

with

,

-

Organized

members

to

the

is

impor

of

.

individuals

to

.

depend

makes

other

it

helps

alone

the

form

in

an

language

failures

,

,

humans

construction

all

disks

who

written

in

the

written

those

and

,

among

computer

Indeed

is

achieve

and

complex

.

important

,

shared

.

language

involved

permanent

on

language

vehicles

successes

be

so

conscious

which

individuals

is

language

more

could

,

transport

,

about

both

of

most

the

to

increasingly

something

language

and

experiences

in

now

Perhaps

of

or

in

individual

knowledge

Language

understand

,

none

to

two

ways

agriculture

activities

for

and

no

buildings

their

possible

munity

that

.

probably

listening

use

many

communal

the

to

ing

of

,

of

continuous

species

,

coordinate

the

things

warfare

and

gathering

think

human

-

,

canals

to

the

accomplish

hunting

a

in

difficult

engage

speaking

observe

involved

not

humans

as

unusual

who

which

commonplace

rather

of

in

-

mak

-

com

-

,

then

for

times

us

.

And

development

their

and

ideas

further

to

180

Without Miracles

Wemaynormally firstthinkoflanguage asa toolforcommunication among individuals, butit alsoappears to beusedsilently asa medium of

thought. Although itsspecific roleinfacilitating andshaping ourthought

processes remains thesubject oflively debate among scholars, thesubjec-

tiveexperience ofbeing human suggests thatlanguage plays animportant

rolein humancognitionandconsciousness. 2

Human language isalsoremarkable forhowitisputtogether, thatis,its grammar. Thestructural complexity ofall5000 to6000 extant languages is such thatdespite centuries ofanalysis, linguists have yettocome upwitha complete andaccurate grammatical description ofanyoneofthem.Lin-

guists have been successful indiscovering important regularities inmany languages, andhave alsoshown howdifferent languages thatmayappear quitedifferent onthesurface share underlying similarities. Nonetheless,

certain structural aspects ofalllanguages havesofareluded formal under-

standing. Since wearesointimately familiar with ourown native tongue we seldom consider thiscomplexity, yetitbecomes evident toanyone whohas

attempted to learna foreign language asanadult.

Thiscomplexity would bejusta curiosity, rather thana puzzle offit

demanding explanation, ifit didnotcontribute totheusefulness andex-

pressive power oflanguage. Weallexperience occasional difficulty in

putting ourthoughts intowords, butalllanguages, from thecautious, con-

servative French oftheacademicians totheever-changing inner-city English ofAfrican-Americans, provide forawide range ofexpressive possibilities,

andweareseldom unable todescribe orcomment onwhatweconsider to beanimportant aspect ofourphysical orsocial world. Wehavenouns to refertobothphysical objects suchascarsandhorses aswellastoabstract concepts suchasjustice andlove.Wehaveverbssuchaswalkto referto

actions, others suchasbelieve toreferto thestates oforganisms and objects, andstillothers such assitwhich canrefer toeither states oractions.

Wehave adjectives such asredandadverbs such asquickly tomodify nouns andadjectives. Wehave grammatical devices suchaswordendings and

wordorders tosignal therelationships among nouns andverbs. Wehave tenseandmoodsystems to specify further thetimeandmanner inwhich

events didtake, willtake, orwould takeplace. Wecanuselanguage decep-

tively todescribe events thatdidnotorwillnottakeplace, andwecaneven useitimaginatively andhypothetically todescribe things thatmight exist,

willexist,or that wewouldliketo exist.

TheEvolution , Acquisition , andUseof Language 181 We

also

have

make

language

social

environment

among that

all are

stream actions volved be

of in

walls hands

and free

the

it . The

of

( most

originating

in

tongue

, lips , and and

the

sounds

of

other

receiving

categorical

. Large

a rich

cords

, and

is well

quickly

through

air

and , unlike

, rather

than

gesturing

. Speaking tasks

and

is effective

matched

speech

in

to

both

the

the

is unique of

sounds

is modified the

brain

infants sounds

are

in -

appear

to

use

air

around

of

light

, can

manually of

by

. The

ocean

light

and

a continuous

) that

newborn of

tract variety

of

that

physical

producing

portions

perception

the

vocal

by

vocal

speech

well

language

travel

corners

teeth

the

characteristics so

human

production

is accomplished

" for

for

use

fits

. This

producing

turn

physiological

language

we

in , allowing

as a medium

us . Speech

which

human

how

speaking

sounds the

the

, and

in

for

" prewired

sound

consider

animals

used of

to

possible

penetrate , keeps

day

of

and

the dark

of night.3 Human languagethus appearsto be very well designedfor communicating with others and controlling our physical and social environments. Thus it provides another striking exampleof a puzzle of fit . We must considerat least three aspects of language. First we have its origin and evolution as a

characteristicof the our species. Secondis the puzzle of how it is possible that all normal children , who take so many years to develop fully their

cognitive, physical, and social skills, are able to learn the languageof their community with amazing speedand apparent ease. Finally, we must look into just how languageallows us to communicateour thoughts, questions, requests , and desires to others .

TheOriginandEvolutionof Language

Attempting toreconstruct theevolution ofhuman language isfraught with difficulties . Whereas thephysical remains of our ancestors mayendure manymillionsof years , unfortunately wehavenorecords of earlyspeech since"language leav ~snobones .,,4However , thishasnotprevented many scholars fromproposing accounts of theoriginof language andhowit pvolved oftongues thatarespoken today . "Forinstance , - .--.---intothethousands it hasbeen argued thatlanguage arose frommimicry ofanimal calls , imitationsofphysical sounds , or grunts of exertion - theinfamous 'bow-wow,' 'ding-dong ,' and'heave -ho' theories ."5Suchunfounded speculations became sorampant in nineteenth -century Europe thatin 1866theSociete de

182

Without Miracles

Linguistique deParis banned thetopic altogether from itsmeetings and

publications.

Butalthough evolutionary evidence ofapurely linguistic nature does not exist, recent research onhominid fossils andstudies ofthemodern human brain andvocal tractarebeginning toshed some light ontheevolution of

language thathasbegun tomove usbeyond therealm ofpure speculation. TheBiological Prerequisites ofHuman Speech

Since speaking requires theproduction ofacontinuously andfinely varied

stream ofsound, wecangainsome knowledge concerning theevolution of

language bystudying theevolution ofthevocal tract andcomparing itwith

corresponding systems inother mammals. Thehuman system isdifferent from thatofallother terrestrial mammals inonestriking way. Darwin himselfnoted the strange factthatevery particle offoodanddrink which we

swallow hastopass over theorifice ofthetrachea, with some risk offalling

intothelungs. 6 Sounlike mammals thatmaintain separate pathways for

breathing andfeeding, thusenabling them tobreathe anddrink atthesame

time, adulthumans areatamuch higher riskforhaving foodenter theirres-

piratory systems; indeed, many thousands dieeach year from choking. 7 In

addition, ourrelatively shortpalate andlower jawarelessefficient for chewing thanthose ofnonhuman primates andourearly human-like ances-

tors,andprovidelessspaceforteeth. 8

Butifthedesign ofthehuman throat andmouth isfarfrom optimal for eating andbreathing, itissuperbly suited forproducing speech sounds. All mammals produce oralsounds bypassing airfrom thelungs through the vocal cords, which arehoused inthelarynx (orAdams apple). Therisk ofchoking towhich weareexposed results from ourlarynx being located quite lowinthethroat. This lowposition permits ustousethelarge cavity above thelarynx formed bythethroat andmouth (supralaryngal tract) asa sound filter. Byvarying theposition ofthetongue andlips, wecanvary the

frequencies thatarefiltered andthusproduce different vowel sounds such asthe[i] ofseat,the[u}ofstupid, andthe[a] ofmama. 9 Wethusseean

interesting trade-off intheevolution ofthethroat andmouth, with safety andefficiency ineating andbreathing sacrificed toasignificant extent for thesake ofspeaking. This suggests thattheevolution oflanguage must have

provided advantages forsurvival andreproduction thatmore thanoffset

TheEvolution,Acquisition,and Useof Language

183

theseotherdisadvantages. Wewillsavefora bitlaterdiscussion of what these advantagesmight have been.

Theimportance oftheevolution ofthehumanvocaltracttofitthefunctionsof speech aresuggested bystudies ofNeanderthal man,wholived inEurope andtheMiddle Eastabout100,000 yearsago.Tofindoutwhat typesofspeech sounds Neanderthals mighthavebeenableto produce, Philip Lieberman, ofBrown University inRhode Island, andhisassociates reconstructed thevocaltractofthesehominidsbasedonfossilevidenceand

applied computer-modeling techniques to it.Theyconcluded thatNeanderthalshada relatively highlarynxandrelatively flattongue,andcould thereforehave,producedonlya limitednumberof nasalvowelsounds. Mostsignificant, theywouldhavebeenunableto produce thesounds[i},

[u},and[a], thethreequantal vowel sounds thataremosteasily distinguished byhuman listeners. ThusLieberman concludes thatevenif Neanderthalhominidshad had the fullperceptualabilityof modernhuman

beings, theirspeech communications, at minimum, wouldhavehadan error rate of 30% higher than ours. 1

Alsoexistingin Europeat thesametimeweretheCro-Magnons, a sepa-

ratehominid species whoappear tohavebeenslightly tallerbutwithlighter bonesandlesspowerful muscles thantheNeanderthals. Onecouldeasily imagine thattheNeanderthals superior strength would havebeenanadvantage forhunting andforanycompetitive encounters withtheirCroMagnon cousins. ButtheCro-Magnons appeared tohaveoneimportant advantage intheirfavora modern vocal tractcapable ofproducing allthe soundsofhumanspeech. It istherefore tempting to speculate thatboththe disappearance ofNeanderthals about35,000yearsagoandthesurvival andcontinued evolution ofCro-Magnons intowhatwearetodaywasdue at leastinpartto thesuperiorlinguistic abilityof ourancestors. 1

Thestudyoftheevolution oftheourvocaltractalsoprovides hintscon-

cerning theevolution ofourbrain. Obviously, thethroatandmouth would nothaveevolved thewaytheydidto facilitate language production and

comprehension while compromising eating andrespiration ifthebrainhad notbeencapableofproducingandcomprehending language.

Lieberman has proposedthreemajorstagesin the evolution of the neuralbasesfor language.Firstwaslateralization of the brain,meaning thateachhalfbecame specialized fordifferent functions. Formostofus,the

[find m an us too ] re th as u o t tw a , i a m h n ,Th ern ma thi as is sy . O h , le t s a ;wo ba ha the ot , rig , ac th m h i s c n .Alt pin for ex W o on ha is ne it is g th r .for us It se no fa to su c o f de in co wi a li d use the lef he fo un b th pre lim po . W a ge sy ( f la ) w e , it wo pr al be co p fr le h . If sp we ind a lat de , it w r to s t i e als un th di of th he a mo 1 2 bra lat ha pr o i b r of ha an la a n p 1 se co ev in th e o str re fo th vo , c s . we us ta an o l a ,pa gra it is int to co hu u o l w t c a mu sy of ot pa ob Ja G :.O 184

WithoutMiracles

lefthemisphere provides mostoftheneural circuitry required forlanguage production . Forabout 90%it alsocontrols thedominant (right)hand used fortasks involving finemotorcontrol , suggesting thatlateralization may haveoriginally evolved in response to selection forskilled hand -pressure

movements . As neurologistDoreenKimuranoted,

Figan 'sdifficulty inconcealing news ofthebananas from hisassociates contrasts sharply withtheease withwhich humans canuse language to deceive andmanipulate others , totalkofthepast , andtoplanforthe future . Lieberman attributes ourcontrol over language tocertain changes

TheEvolution,Acquisition,and Useof Language

185

in thebrain,including theevolution ofwhatisreferredto asBrocasarea, aswellastheenlargement oftheprefrontal cortex(thepartofthebrainjust behindtheforehead) anda rewiring ofconcentrations ofneuronsreferred to as the basal ganglia.

Thethirdcomponent intheevolution ofhumanlanguage involved the

ability toputsounds andwordsinspecific orders andto perceive these ordersasmeaningful. Inalllanguages, theorderinwhichwordsandparts ofwordsareproduced andperceived is crucialto themeaning. Thesen-

tence Mary sawJohnconveys adifferent meaning from JohnsawMary. Of allthecommunication systems usedbytheearthsanimals, it appears that

onlyhuman language derives itsexpressive power fromtherecombination ofa finite(though large)number ofwordsandwordpartsintoaninfinite numberof different sequential orders.Thewaysin whichwordsmaybe orderedand how thesedifferentordersrelateto meaningis syntax.

Lieberman suggests thattheevolution ofmotorcontrolforspeech itself

provided thebasisforthedevelopment ofsyntax. Thisisbecause articulatingevena simple wordsuchascatrequires theprecisely timedsequential coordinationof movements of the tongue,lips,and jaws.The orderof soundswithinwordsalsomakesa difference asto theirmeaning, withcat different fromtack.Lieberman thusconcludes that speechmotorcontrol

isthepreadaptive basis,thatis,thestarting point[for syntax]. Oncesyntax becamea factorin humancommunication, theselective advantages thatit confers. . . wouldhavesetthe stagefornaturalselection that specifically enhancedtheseabilitiesindependentlyof speechmotorcontrol. 15

Theimportance oflanguage andtheadvantages itprovides usincommu-

nicating, coordinating ouractivities, andthinking suggests that,inaddition tobeing a product ofourevolution, it alsoplayed a largepartinshaping ourevolution, particularly thatofourbrain.Lieberman hasconsequently propose [d} thatnatural selection toenhance fasterandmorereliable communi-

cationisresponsible forthesecond stageoftheevolution ofthesemechanismsthe evolution ofthemodernhumanbrain.Communication placestheheaviest functionalloadoncircuitryforbothelectronic devices andbrains. Thetransistors and solid-state devices thatmadedigital computers a useful toolwerefirstdeveloped for communication systems. Indeedonecanarguethatthedemands ofcommunication

preempt thehighest levels oftechnology andorganization ofa culture, whether couriers on horsesor lasersandfiber-optic bundlesarethemeansemployed. In short,evolution forefficient, rapidcommunication resulted ina brainthathas

186

Without Miracles

extremely efficient information-processing devices thatenhance ourability touse syntax. These brain mechanisms alsomaybethekeytohuman cognitive ability. As many scholars have noted, human language iscreative; itsrule-governed syntax and

morphology allowustoexpress new sentences thatdescribe novelsituations or convey novelthoughts. Thekeyto enhanced cognitive ability likewise seems to be

ourability toapply priorknowledge andrules orprinciples tonewproblems. 1 Given theprogress thathasbeenmadeinunderstanding theevolution of language, together withmodern biologys acceptance ofnaturalselection as theexplanation fortheappearance ofdesigninthestructure andbehavior

oforganisms, it mightcomeassomewhat ofa surprise to learnthatsome

scholars reject natural selection asanexplanation fortheappearance, struc-

ture,anduseof language. Whatis particularly noteworthy aboutthese critics ofnatural selection isthatsomeofthemarewidely recognized as leading thinkers andresearchers in theirrespective fields, among them

NoamChomskyand StephenJayGould.

Chomsky notonlyrejects natural selection asanexplanation forthe evolution ofhuman language, butalsorejects Darwinian explanations for

certainwell-understood biological phenomena. Hehasstatedthat evolu-

tionary theoryappears tohaveverylittletosayaboutspeciation, orabout anykindofinnovation. It canexplain howyougeta different distribution ofqualities thatarealready present, butit doesnotsaymuchabouthow

newqualities canemerge. 17 Thisisa curious statement, given thatDarwin proposed naturalselection to account forspeciation specifically andthat

theessence ofhistheory isstillaccepted today bymainstream biologists as

thesoleexplanation foradaptive evolution andspeciation. Asforevolution not say [ing] muchabouthownewqualities canemerge,thoseunconvinced bythebiological evidence willseeinpartIVhowartificial evolution is beingapplied increasingly to createuseful, innovative products from

drugs tocomputer programs. Anattempt tounderstand better Chomskys

rationale forrejecting aselectionist account oflanguage willbeoffered later

inthischapter, where wetakea lookathowchildren learn thelanguage of

their community.

Goulds reluctance toaccept a selectionist account ofhuman language stemsfromhismoregeneral concern thatadaptationist explanations of biological traitsareoftenmisapplied. Inanimportant paperwritten with Lewontin, 18 theyproposed thatcertain biological traitsmaynotbesolely

The Evolution,Acquisition,and Useof Language

187

dueto adaptivenaturalselection, but rathermayhavetheiroriginsas sideeffectsor by-products of otherevolutionary changes, whicharethen seized at a certaintimeby a newfunction.Wereferredto thisphenomenonof exaptation (originally calledpreadaptation by Darwin)in addressing theevolution of thehumanbrainin chapter5. Theexaptationist

perspective proposes thattheability to communicate bylanguage hadlittieor noroleinthechangesthatmadethebraincapableoflanguage. The

brain,inbecoming largeforwhatever adaptive reasons, acquired a plethoraofcooptable features. Whyshouldntthecapacity forlanguage beamong them?Whynotseizethispossibility as something discrete at somelater

pointinevolution, grafting uponit a rangeofconceptual capacities that achievedifferent expression in otherspecies(andin ourancestry) ?19

Perhaps wecanbetterunderstand theexaptationist viewoflanguage by leaving biology fora shorttimeandconsidering anexample oftechnologicalchange. 2 Theveryfirstcameras hadchemically treatedplatesofglass thatcapturedtheimageprojected bythelens.Inattemptsto makephotog-

raphylesscostlyandmoreconvenient, celluloid sheetswereintroduced, followed byrollsofcelluloid film.Butwhereas longrollsofflexible film wereintendedsolelyto facilitate stillphotography, theymadeit possible alsoto takemanypictures inquicksuccession, thusleading to thedevelopmentofthemotionpicturecamera.Wethereforecannotsaythatrollfilm evolvedto makethe motionpicturecamerapossible,sincethe ideaof a

motionpicture camera probably didnotevenexistwhenthefirstrollsof filmwereproduced. Instead, to useDarwinsterm,therollfilmofthestill camerawaspreadapted,althoughquiteaccidentally and unintentionally, foruseinthemotionpicturecamera. TouseGouldsmoreneutralandmore accurate terminology, thisfeatureofthestillcamerawasexaptedforusein

motion picture cameras. So,ineffect, Chomsky andGouldassertthatthe humanbrainis analogousto rollfilmin that it evolvedfor reasonsorigi-

nallyunrelated tolanguage concerns; butonceitreached a certain level of sizeand complexity,languagewas possible.

Exaptation is an important conceptual toolin understanding theevolutionof biological structuresand behaviors,but it alonecannotaccount forthecontinued evolutionofadaptedcomplexity. Althoughrollfilmmade

thefirstmotionpicturecameras possible, modernmoviecameras andthe

filmtheyusearemorecomplex andbetteradapted totheproduction of

188

WithoutMiracles

high-qualitymoviesthantheveryfirstmoviecameras . Andwhereas some of theseadditionaltechnological developments mayhavealsobeenexaptedfromotherfields,for example , developments in electronics andchem istrymakingpossible moreaccurate lightmetersandmoresensitive film, thewayin whichall thecomponent partsof a modernmoviecamera work together canbeunderstood onlyasresulting fromselection processes operatingatthelevelof theentirecamera , notjustitscomponent parts. Similarly , certainpreexisting structures andfunctions of thehumanbrain andvocaltractmayhavebeentakenover(or exapted ) for usein language . However , thiscannotbyitselfaccount for thewaysin whichthebrain,the vocaltract, andlanguage fit together to createa totalsystem thatis quite remarkably adapted to servethefunctions for whichlanguage is used . The factremains thattheprocess of cumulative blindvariationandselection is theonlyprocess currentlyunderstood thatcanaccountfor thenonprovi dentialappearance of theadaptive complexity thatis seenin thedesign of language , andthedesign of thehumanbrainandvocaltractfor language . AsPinkerandBloompointoutin theirimportantdiscussion of theroleof naturalselection in theevolutionof language , "language showssignsof complex design for thecommunication of propositional structures , andthe onlyexplanation for theoriginof organs withcomplex design istheprocess of naturalselection ."21 TheEvolutionof Language Butwhataboutthesounds , structures , andrulesthatmakeup language ? How did theyoriginateandevolveovertime, leadingto the languages spokenthroughout theworldtoday?As alreadynoted,we unfortunately haveno recordsof howlanguage wasusedby our prehistoric ancestors . Nonetheless , our currentknowledge of evolutionprovidesat leasta generalscenario of howit evolved . As PinkerandBloomobserved , for language to haveevolved bynaturalselection : Theremusthavebeengenetic variation among individuals in theirgrammatical competence . There musthavebeen aseries ofsteps leading fromnolanguage atall to language aswenowfindit, each stepsmallenough tohavebeen produced bya random mutation or recombination , andeachintermediate grammar useful to its possessor . Every detailofgrammatical competence thatwewishtoascribe toselec tionmusthave conferred areproductive advantage onitsspeakers , andthisadvan tagemustbelargeenough to havebecome fixedin theancestral population . And

The Evolution,Acquisition,and Useof Language

189

theremustbeenoughevolutionary timeandgenomic spaceseparating ourspecies from nonlinguistic primate ancestors. 22

Butsincethereareso manyconceivable waysin whichlanguagecould haveconferreda reproductiveadvantageon itsspeakersandsofewconclusivedataon thissubject,wecanonlyspeculate onwhichonesactually

wereimportant. Wealreadymentioned theuseoflanguage to coordinate humanactivity, anditisnotdifficult toimagine howtheability toplanand coordinate hunting,agricultural, andwarfareactivities wouldhaveconferredsurvival advantages to individuals andgroupswithlanguage skills. Also,as mentioned earlier,language makesit possible forindividuals to shareknowledge, thereby avoiding themistakes anderrorsthatothershave already made.

Wecannotgobackin timeto seehowlanguage wasusedbyearlyhumans,butwecanlearnfromthebehavior ofcommunities ofhunter-gathererswhostilllivetodayin muchthesamewayas allourancestors didas

recently as12,000 yearsagowhenthetotalhuman population oftheearth wasonlyabout10million. 23 Among suchgroupsarethe!Kung of the KalahariDesertin Namibiaand Botswanawho uselanguageto discuss

everything fromthelocation offoodsources tothebehavior ofpredators tothe movements ofmigratory game.Notonlystories,butgreatstoresofknowledge are

exchanged around thefireamong the!Kung andthedramatizationsperhaps best

of allbear knowledge criticalto survival. A wayof lifethatis difficult enough would,withoutsuchknowledge, becomesimplyimpossible. 24

Although sharingknowledge of thelocationof foodwouldcertainly seemto bea functionof languageprovidingimportantsurvivaladvantages, survivalin itselfcannotensurethat an individualsgenesand the language

abilities thatgowiththemwillbeinherited. Inheritance requires reproduction,andhumanreproduction, asdoesallsexualreproduction, requires a

partner oftheopposite sex.It should notbesurprising, therefore, tofind thatlanguage playsanimportant roleinsexualselection. Just asfemale birdsseemto havefavoredelaboratesongsbymales(notto mentionlong

andshiny feathers) whenchoosing amate,soprehuman females might have promoted a fancierformof language 25 by preferring menwithmore impressivelanguage skills.

Ofcourse,wedo notknow,andwillprobably neverknow,theactual eventsandselectionpressuresthat graduallytransformedthehoots,grunts,

190

and

Without Miracles

cries

of

tion

system

and

adaptive

sible

; and

our

ancestors

. But

into

two

tool

things

are

without

second

, at

that

natural

our

known

which

some

current

remarkable

: first

human

time

in

,

life

the

past

language

as it

vocal

is

we

did

know

not

a

it

exist

communica

highly

complex

would

be

. And

impos

-

regardless (...

1

of

- -- - -- - - -

arguments the

emergence

ian

evolution

for

the

that

is

the

may

's

language

, already

above

her

up ,

our

most responsible

to

the

available

be

of

child

the

fact

,

of

remains

and a

task

Darwin

-

explanation

useful

complete

~ -

explaining

that

nonmiraculous

remarkable for

the

ability

,

an

experience

ability

of

human

Eskimo

,

and

in

the

shoulders

, or

her

English

a

the

vocal

fit

concerned

by

is

expect

the

child

's

"

, to

this

be

learn

be

much

the

her

bio

language

it

-

she

Estonian

of

selection

and

.

found

apply

the

too

, natural

abilities

possible

to

whether

to

use

of

structures

" simply

,

evolution

language

biological

selection

language

the

make

the

community

not of

in

therefore

natural

and

should

about

that

and

parents

between

be

organs

' s task

centuries

good

to

community

child

. We

countless

provided

have

, provided

by

after

not

. The

endowment

Language

does

brain exist

spoken

task

quite

language

currently

Acquisition

, the

These

since

not

. 26

Fortunately

logical

of

only

of

believe

Child

hears

is

refinement

appearance

some

The

selection

and

consciousness

an

,

ordeal

should

,

have

requirements

of

the

.

Language

The

as

Learned

acquisition

from

the

an

of other

her

will

the

things

and

even

and

imitating

forced language

hears

first

-

by

television

actions

of be

set

they

satisfaction

acquisition

,

any

see

the different

drink

and

others

consequences ?

be

or

same

from

sounds

VCR

-

and . "

cups

,

of

, open

words

that

and ~ and

actions

typical

learn

and

close

from

and

their

hugs

Children

apparently

performing

that

to

encouragement

smiles " dada

different

spoken

considerable

the

very

unremarkable

language

the

in "

to

seem

of

provide

" mama

clothes

may

produce

evident

of

their

the

the

the

on

to

is

appear It

hours

also

as

utterance

put

operate

,

.

countless

Parents

speak

initially learn

begin

.

to

not

children

eventually

children

follow

many

who

structures

their

may

that

child

grammatical

for

language

things

American

around

ly

-

from

to

doors

do

,

observing

being

. Why

rein

should

-

The Evolution , Acquisition, and Useof Language

191

This is essentiallywhat was proposedby B. F. Skinner, introduced in our discussionof learning in chapter 7. It will be recalledthat Skinner's theory of learning attempted to explain the acquisition of any new behavior as a process of operant conditioning

by which new behaviors (such as a rat

pushing a bar) would be learnedto the extent that they were reinforced in some way by the environment (such as receiving food). He extended his behaviorist view of learning to human language. In all verbal behavior under stimulus control there are three important events to be taken into account : a stimulus , a response , and a reinforcement

. These are contin -

gent upon each other, as we have seen, in the following way: the stimulus, acting prior to the emissionof the response, setsthe occasionupon which the responseis likely to be reinforced. Under this contingency, through a processof operant discrimination, the stimulus becomesthe occasionupon which the responseis likely to be emitted .27

But he recognizedat the outset that the social use of languageis in one important respect quite unlike nonverbal behaviors operating on an inani mate

environment

.

When a man walks toward an object, he usually finds himself closer to it ; if he reachesfor it , physical contact is likely to follow ; and if he graspsand lifts it , or pushesor pulls it , the object frequently changesposition in appropriate directions. All this follows from simple geometrical and mechanicalprinciples. . . . However, when we use languageto act upon the world , as when we ask another for some water, the glassof water reachesthe speakeronly asthe result of a complex seriesof eventsincluding the behavior of the listener. . . . Indeed, it is characteristicof such [verbal] behavior that it is impotent againstthe physical world .28

Consequently, Skinner determinedthat although the links among stimulus , response , and reinforcement

may be less obvious

and more indirect

for verbal behavior, they nonethelessexist and can be used to explain languagelearning and use. For example, he noted that in a given language community certain verbal behaviors such as " Wait !" and " Sh-h!" are typi -

cally followed by certain consequences , such as someonewaiting or being quiet. Sucha result depends, of course, on the cooperation and behavior of the other person. But if the consequenceis achieved, that particular verbal responsewill be strengthenedand be more likely to occur in a similar instance in the future . Thus , like all other behaviors , language learning is

completely dependenton contingenciesof reinforcement. In essence , Skinner's analysisof verbal behavior is an attempt to show how languageis shapedby the environment in the sameway that a rat's

192

WithoutMiracles

lever pushing or pigeon's key pecking can be controlled by providing and withholding food. By giving reinforcementfor the sounds, words, and sentencesthe child producesthat approximate the adult form of the language, and by withholding such reinforcement when an utterance is in some way deviant, the child's verbal behavior is gradually shapedover time to approximate the languageof the community. Skinner arguedthat suchcontingenciesof reinforcementare not only responsiblefor the child's learning

language , but arethedeterminingfactorsfor all behavior , includingadults' languagebehavior. In insisting that reinforcementis the key to understandinglanguagebehavior , Skinner had to stretch the concept of reinforcement

to cover situa -

tions that are quite unlike those found in studies of animal learning. For instance, he claimed that many verbal behaviors are " automatically selfreinforcing," aswhen " the child is reinforced automatically when he duplicates the sounds of airplanes, streetcars, automobiles, vacuum cleaners, birds, dogs, cats, and so on." 29And " the young child alone in the nursery may automatically reinforce his own exploratory verbal behavior when he produces sounds which he has heard in the speech of others . " 30He stretched

the conceptof reinforcementto situations where the personproducing language is not even present when the reinforcement takes place, as when a

public speaker or writer is reinforced by "the fact that effects of verbal behavior may be multiplied by exposing many ears to the same sound waves

or many

eyes to the same page . " 31

Chomsky, in his influential and widely cited review32of Skinner's book, underscoredtheseand many other problems with an operant conditioning analysisof human behavior, including language. Concerningthe role of reinforcement, after having first cited the above and other examples, Chomsky concluded: From this sample, it can be seenthat the notion of reinforcement has totally lost whateverobjectivemeaningit may everhavehad. Running through theseexamples, we seethat a person can be reinforced though he emits no responseat all, and that the reinforcing stimulus neednot impinge on the reinforcedperson or neednot even exist (it is sufficient that it be imagined or hoped for ). When we read that a person plays what music he likes, sayswhat he likes, thinks what he likes, readswhat books he likes, etc. BECAUSEhe finds it reinforcing to do so, or that we write books or inform others of facts BECAUSEwe are reinforced by what we hope will be the ulti mate behavior of reader or listener, we can only conclude that the term reinforcement has a purely ritual function. The phrase " X is reinforced by Y" . . . is being

The Evolution , Acquisition , and Useof Language

193

used as a cover term for " X wants Y , " " X likes Y , " " X wishes that Y were the case , "

etc. Invoking the term reinforcement has no explanatory force, and any idea that this paraphraseintroduces any new clarity or objectivity into the description of wishing, liking , etc., is a seriousdelusion.33

It is interestingto note that Chomsky's analysisis not inconsistentwith perceptual control theory discussedin chapter 8, as he implied that people behaveto satisfy their internal " wishes," " likes," and " wants," and not because certain behaviors

were reinforced

in the past .

Languageas Innately Provided In his subsequentwritings on languageacquisition, Chomsky has attacked Skinner's and other learning theoriesby focusing on the syntactic structure of languageand the fact that all normal children show impressiveknowl edgeof this structuredespiteconsiderablevariation in their exposureto language. Chomsky's insightsand his developmentof what is calledgenerative grammar revolutionized our understanding of language. The particular generative grammars being developedto explain aspectsof various languagesare probably beyond the grasp of anyone who has not formally studied modern linguistics. The generalnotion of a generativegrammar is fortunately more accessible . Let us first considerwhat syntax is and why it is necessaryfor language. Due to the nature of the human vocal tract, we do not normally produce (or perceive) more than one speechsound at a time. This makes oral languagea serial medium, meaning that soundsare strung together one after another like beadson a string. As already noted, the order in which sounds are uttered to form words (compare pot with top ) and words are uttered to

form sentences(compare The dog ate the pig with The pig ate the dog) are related to the meaning of a sentence. All languagesdiffer in the degreeto which word order is crucial for understanding, and English is particularly choosy, with most orders being meaningless , or nearly so. As linguist Derek Bickerton

observed :

Try to rearrange any ordinary sentence consisting of ten words . There are, in prin -

ciple, exactly 3,628,800 ways in which you could do this, but for the first sentence of this paragraph only one of them gives a correct and meaningful result. That means3,628,799 of them are ungrammatical. How did we learn this? Certainly, no parent or teacherevertold us. The only way in which we can know it is by possess ing, as it were, somerecipefor how to construct sentences , a recipeso complex and

194

Without Miracles

exhaustive thatit automatically rulesoutall3,628 ,799wrong waysof putting together atenwordsentence andallows onlytherightone . Butsince such arecipe mustapply toallsentences , notjusttheexample given , thatrecipe will, forevery language , ruleoutmore ungrammatical sentences thanthere areatoms inthecos mos - andthere areatleast fivethousand different languages !34 Although manylanguages arenot asstrictasEnglish concerning word order , allofthemrequire atleast certain orders ofbasic sounds (called phonemes ) to makelipwords , evenif theyaremoreflexible in thepermissible orderings ofwordstoformsentences .35 Syntax

refers

words

in

a

cording

to

not

by

the

to

that

they

look

at

a

( 1 )

S

(2 )

NP

- 4

( 3 )

VP

- . t

+

) plus

a

( Det

) such

as

sible

verb

phrase

learns

these

permit

the

. Once

an city

form

a

the

city

the

city

earth

+

that

can

?

quake

, this ( VP

,

of

into .

sentence ( VP

,

and

most

of

then

of

learning

using

these

abstract

rules are

the

of

novel

speaker

syntax

-

sentences

by

which

by

rules

-

even

quite

before

generative

a

together

a

also

has

related

by

in

. Let

us

proposed

sentences ears

,

. In

sentences

voice

,

the

noun

. And

as

fit

which

such

as

a ) as

by

I

saw

The

that

The

Did

The

earthquake

city

pro

destroyed

-

destroyed trans

-

destroyed

earthquake

was

the

can

rules

the

-

that

throw

, one

earthquake

pos

. One

rules

John

classes

a

phrase

recursive

in

phrase

determiner

( 3 ) defines

, transformational

so

noun a

noun

including

into

(2

another

phrase

question

sentence

in

by

addition

of

defined

a

others

noun

words

into

passive

a

composed

is

) followed

some

which

of

be

followed (V

as

big

may

phrase

with

learns

transformed

the

that verb

sentence

dog

( 5 )

) . Noun

, or

) as

into

be

or

a

phrase

number

sentence

not

know

spoken

the

and

sentences

, we

nor

ordered

. Ac

NP

rules

That

up

sentences

heard

of

,

of

meaning

N

unlimited , or

make

generate

properly

sentences

. Instead

to

.

a , the

one

to

been

+

verb

use

that

examples

states

( NP

and

-

orderings

related

VP

Det

V

( 1 )

use

English

- . tNP

words

permissible

are

syntactic

sentences

we

simple

for

Rule

new

neither

few

of

the

orderings

produce

list

govern

these

patterns

create

have

that

how

we

a

that

Chomsky

duce

,

structural

frames language

the

principles

and

memorizing

the

ball

the

Chomsky

general

as

to

language

destroy

by

the


195

The magnitude of the child's achievementin learning languagecan be

appreciatedby consideringthe complexityof theserules. Take, for example, a simple question suchasDid you seemy toy?To produce sucha question the child must know that such a yes-no question is formed from the

correspondingdeclarative(nonquestion) sentenceby moving the first auxiliary verb beforethe subjector adding the auxiliary verb do to this position if no auxiliary already is present. This is a rather complex rule, and yet it is tacitly known by every English-speaking child who can ask a yes-no question! It is furthermore clear that such rules are not explicitly taught by parentsto their children, sincefew parentscould evenstatethem, and it is highly unlikely that a typical three- or four-year-old could comprehend them even if the parents did .

To make the acquisition of syntax evenmore remarkable, Chomsky says that the examplesof languagethat children hear are inadequatefor them to figure out the underlying generativerules on their own. Children hear only a relatively small subsetof sentences , and due to lapsesof attention or memory, false starts, slips of the tongue, interruptions, and other disturbances, many of the sentencesthey do hear are not well-formed. Consideringalsonoisefrom nearby machines, televisionsets, airplanes, and other sourcestogether with the frequent ear infections that many young children experience, and occasionswhen a child hears clearly articulated,

grammaticalsentences becomeevenlessfrequent.Chomskyhasreferredto these characteristicsas the " poverty of the stimulus," implying that languageas heard by a child is not sufficiently clear, accurate, and structured for the child to be able to deduceand learn its underlying generativerules. It is also clear that children go well beyond what they hear in coming up

with grammaticalrules. For example , manyif not all childrenliving in an English-speaking environment will use words such as breaked, drawed, balded, and cutted despitenever having heard adults use such words. But although they make certain types of errors, there are many other possible errors which they never seemto make. For example, no child has been heard to say " Kitchen the in is Daddy ?" to turn " Daddy is in the kitchen " into a question .

Chomsky, together with many other linguists and child language researchersinfluenced by his theories, have thus concluded that children do not and could not acquirelanguageby operant conditioning asproposed

196

Without Miracles

bySkinner, orindeed byanymethod oflearning aslearning isgenerally understood inpsychology today. Thefactthatallnormal children, regardlessofintelligence level, arequiteproficient speakers ofthelanguage to which theyareexposed bytheageoffouryearssuggests to Chomsky

thatmuchofthatknowledge ispartofthehuman biological endowment. Andalthough different languages certainly varyintermsoftheirsounds,

words,andgrammatical constructions, thesedifferences canbeunderstood as variationson a theme,the themebeingcharacteristics that are shared

byallnatural languages andthatlinguists calluniversal grammar. For Chomsky, knowledge ofuniversal grammar isasmuch apartofBaby Sues

biologicalinheritanceas is the nose on her face.

Ineffect, then,Chomsky believes thatsince language cannot beacquired byenvironmental instruction (dueprimarily tothepoverty ofthestimulus), to a verylargeextentit mustbe innate.To be sure,he andlike-minded

linguists recognize thatchildren mustbeexposed toa language forit to beacquired, butratherminimal andhaphazard exposure is allthatis

required totrigger itsacquisition. Inthisrespect spoken language isvery

different fromotherabilities suchasmathematics andreading skills, which normally require special experiences fortheirdevelopment, suchasformal schooling andprolonged practice. Eventhenmanychildren andadultshave difficultyacquiringtheseacademicskills.

Chomsky isnotaloneinhisinnatist viewofhuman cognitive abilites

related tolanguage. While hefocused onsyntactic knowledge, psycholin-

guistandphilosopher JerryFodorhasmaintained formanyyearsthatall human conceptual knowledge isinnate. Concepts suchasTRIANGLE,

DOG,andFREEDOM, andtherefore the meaning we attachto these

words, mustbeinnate since itisimpossible forsomeone tolearn aconcept

thatwasnotinsomeformalready knownbefore. It willbenotedthatthis argument bearsanuncanny resemblance to theonemadebyPlatoin the

dialogue withMeno discussed inchapter 6.ForFodor, whatappears tobe learning isactually fixation ofbelief, 36 using experience toselect among

a hostofinnateideas.Thismayinitially seemconsistent witha selectionist

(and therefore evolutionary andconstructive) theory oflearning andcogni-

tivedevelopment. ButFodorplacessomesevere limitations onwhathebe-

lieves such selection canachieve, asseen inhisstatement cited inchapter 9:


197

There literally isn't such a thing as the notion of learning a conceptual system richer than the one that one already has; we simply have no idea of what it would

belike to getfrom a conceptually impoverished to a conceptually richersystemby anything like a processof learning.37 But if evolution itself can be considered to be a form of learning in which

organismsover phylogenetictime acquire knowledge about their environment ,38it turns out that , despite Fodor , we do have a theory of learning ,

namely, natural selection, that explains how complex, adapted systems such as organismsand componentsof organismscan emergefrom simpler

or more"impoverished " ones. Indeed, the argumentthat knowledgein the form of new and richer conceptscannot be constructed, but must rather already exist and be innately provided, creates a serious problem for Fodor and others who use this can't-get-there- rom -here logic . Obviously , at some time (to be safe, let 's say one second after the Big Bang), concepts such as RED , MOTHER , CON SERVATION

OF ENERGY , and INTERROGATIVE

SENTENCE

did not

exist, yet today they do exist in the minds of humans. How did they originate? Certainly, biological evolution must havehad somethingto do with it sincetheseconceptsare clearly complex and functional characteristicsof our species.But if Fodor is right that there is no way " to get from a concep-

tually impoverished to a conceptuallyrichersystem ," he mustalsobelieve biological evolution to be impossible. Once it is recognizedthat the evolutionary process of cumulative blind variation and selection has in fact

resulted in the emergenceof more complex systemsfrom simpler ones, and that an evolutionary process involving the cumulative variation and

selectionof ideas, thoughts, and conceptscould alsobe an essentialand universal part of human cognitive development, Fodor's argument for the impossibility of learning appearsseriouslyflawed. Mark Bickhard, a cognitive scientistat Lehigh University, madejust such a critique, using in the following quotation the word representationto refer to Fodor's conceptualizationof knowledge: If representations cannotemerge , however , thentheycannotcomeinto beingat all. A narrowfocuson this point yieldsFodor's innatism:neitherlearningnor develop ment, ascurrentlyunderstood , canconstructemergent representation ; thereforethe basicrepresentational atomsmustbealreadypresentgenetically . Unfortunately , this conclusiondoesnot follow. If representation cannotemerge , thenit cannotemerge in evolutionanymorethanit canin development . Theproblemis logicalin nature,

198

WithoutMiracles

and is not specific to the individual . Conversely , if some way were posited in which evolution could yield emergent representation , then there is no a priori reason why that emergence would not be just as available in the development of the indi vidual . Fodor 's innatism , then , simply misses the basic issue. If representation cannot emerge, then it is impossible for it to exist , and evolution is in no better position in this respect than is individual development ; on the other hand , if representation can emerge, then there is something wrong with the models of learning and development that cannot account for that emergence. When those models are corrected , that emergence should be as available to the individual as to evolution . In either case, Fodor 's strong innatism does not follow .39 But isn 't it true that things can happen during evolution pen during the lifetime of an individual human ? Human evolved over a very long period of time through Yet no human

can " learn " to grow another

because such biological

structures

among -organism

knowledge

brain , and the brain is remarkably genome is not .41As explained

are determined

are dependent adaptive

in chapter

make adaptive changes through

a variation

abilities

underlying

on the structure

ontogenetically

of the

, whereas the

5 , the brain retains the ability

of language

knowledge

and Fodor

could

are wrong

and concepts

If the human

and linguistic

innatists , that this knowledge

genome could have acquired

among -organism at least a possibility

such knowledge

by way of the

that the brain could acquire similar knowledge selection of synapses . The implications

organism selection for such innatist views of human cognition sidered again at the end of chapter 15 . Language Acquisition

worthy

among linguists opposition

by way

of within will be con -

as Selection

Chomsky 's and Fodor 's views of language acquisition

chologists

of

must be innate .

selection of human evolution , then it must be considered

of the within -organism

popular

in

is innately

determined . But it does argue against their reasoning , and the reasoning other cognitive

to

and selection of synapses . Thus

None of this , of course , proves that Chomsky that our knowledge

a lifetime

by genes that do not

it is at least conceivable that new linguistic and conceptual emerge as a result of such within -organism selection .

their assertions

selection .

heart or arm during

change during one 's lifetime .4OIn contrast , the cognitive language and conceptual

that cannot hap hearts and arms

and cognitive

to their perspective

who take a less linguistic

are undeniably

very

scientists today . Nonetheless , note exists , much of it coming and more functional

from psy -

perspective

on

199


language

, its

viduals language

us leave

their

ter

of

not

the

by

pher

.42 He

. But

uses

used

initially

assumes

in

HOP

, or

time

of

some

way

mother word

not

to

any

( or

to the

ly has

, its

word

any

one

no

has have

-

some

state

( eat -

out

by

vocabulary

no

a

can -

and

says

the

would individual refers access

, for to some

to

those

of

ever

being

that

.43 Even

, the

current

possibilities it is not

matter

) . The

already

concepts

' and

child

to of

words

a

helpful

for

cat - plus

child

in her

- the - carpet

meanings

-

adult

meaning to

a word can

mind

that ) and

of scratching the

an of

know

animal

concerning

meaning

concept

our

two

if

behavior

possible the

in

of yet

evidence that

, a species

carpet

its

who

appearance much

is the

( actually the

as the

a person

certain

conundrum

, or

such

or

.

" cat , " how

transmit

the

how

fur -

concept

SHAPE

rabbit of

and

on

the

- LIKE

the

curse

matter

meaning

infinite that

to

no

to

it is making

realizes

name

itself

fur , its relationship sound

proper

same

animal

RABBIT

word

ears

. But to

-

whose the

long

rabbit refer

to

the

same

with

unrelated

philoso

hears

English

or

pointed

country

linguist

actually

FUR

was living

mammal

could

way

means

a strange

to the

that

the

clear

direct

and

or

pointed

learning

, the

an expletive

the

animal the

be

, or

reliably

cat

words a mat

hear

influential

, furry

or

argues

essentially

to

visit

quite

even

that other

even

visiting

a small

something

, it becomes

word

they

a word

' s most

his

HEAD

a rabbit

an

- on , the

information

that

to refers

its color

? When

of

learning

new

( on ) , action

what

gavagai

or

could

faces

cat

it - is - sitting

that

, he simply

points

the

of

realizes

languages

child

in

it is simply

word

that

is equivalent

pest . Quine

different

The

indi

view

to learn

that

is perhaps

a linguist

of

gavagai

even

collects

from

new

indicates

. During

presence

resembles

that

America of

MAMMAL

garden

appear

) , relation

determining

know

day . Gavagai

linguist

itself

the

perhaps

another

in

example not

he or

these

.

that

reflection

ANIMAL

of

is involved

appear

each

and

, arguably

the

does

first

( black

reflection

simple inherent

he

number

, a selectionist

what

children

between

perceive

further

that

gavagai

ther

can

W . V . Quine

language

. Since

, cat ) , quality

difficulty

out

use . A

implicitly

consider

, it might

association

child

adult

its

or

and

words

so quickly

an

be quite The

a while

of individual

example

ing ) that helpful

for

meanings

( for

, and

explicitly

.

forming

object

children

, either

syntax

meanings

~ nd

by

adopted

learning

Let the

acquisition

have

only , since

in the

adult

provide

to

a child suppose

she

sure -

's mind

,

200

Without Miracles

butshecanneverbeabsolutely surewhichit is.Thatchildren do make guesses andjumpto unjustified, tentative conclusions isclearwhena child

refers toa small black dogasa cat or,perhaps more humorously, when

sherefers tothevisiting parishpriestasDaddy. Itisinformative tocompare thechildinthissituation withthatofthescientisttestingtheories, suchasthatwatermustbeheatedto 1000Cforitto boil.Forthescientist, noamount ofevidence canbetakenasconclusive

proof ofthetheory. Onecanboilwater using heatfrom burning gas,elec-

tricalresistance, orsolarenergy andfindthatthesource ofheatmakes no

difference intheboiling point. Onecanboilwater invessels made ofsteel,

iron,aluminum, stone,orplastic, ordotheexperiment atdifferent timesof

dayandduring different phases ofthemoon andobtain whatappears tobe additional evidence forthetheory. These findings mayappear tolendsupporttothetheory, buttheycannot prove it.Indeed, awater-boiling experi-

mentconducted usingan accuratethermometer at an altitudeof 2000

meters abovesealevelwillshowthatthetheoryisinfactfalse,sincethe

boiling point ofwater depends onairpressure, which isreduced athigher

altitudes.

Thechilds situation withrespect towords andtheirmeanings appears

analogous. Ayoung childgrowing upwithlittleornocontact withnonsib-

lingchildren mayover aperiod ofmany years bepresented withabsolutely overwhelming evidence thatmommy refers totheoneparticular woman

whoisalmost always closebyandwhofeeds, clothes, bathes, andcuddles him.Imagine littleJohnnys surprise whenduringhisfirstvisitto kindergarten attheageoffivehehearsanother childusing thewordmommy to referto a woman whom hehasneverseenbefore! Johnny thenhasno option otherthantoreject hisinitially well-supported hypothesis about

themeaning ofmommy. Eventually hewillreplace itwiththetheory that mommy refers tonotoneparticular person, orevena class ofpersons, but ratherto a special kindofrelationship between onehuman being and another. Andofcourse, even thismeaning issubject torevision asyoung Johnny hears orreads about acow, dog,orcatwhoisthemommyofa calf,puppy,or kitten.

Butalthough learning wordmeanings mustnecessarily proceed through

a process oftheoryconstruction, rejection, andrevision, it isalsoclearthat

themeaning theories thatchildren entertain areeitheroftenaccurate or

The Evolution,Acquistion,and Useof Language

201

quiteclose totheadultmeaning, since itdoesntseem totakemanycycles oftrial-and-error elimination to arriveat theaccepted meaning. Ifthiswere

notthecase,children wouldbehardpressedto learnnewvocabulary as quicklyand easilyas theydo.

DonaldT.Campbell, whosevariation-and-selection perspective onlearn-

ingandthought wasdescribed inchapter 9,hastheorized thatchildren are aidedin theirguessesasto themeanings of newwordsbyan innateexpectationthat wordsreferto the moreeasilyperceivable, stableaspects(or

entities) oftheirenvironment, a characteristic hereferstoasentitativity. 44

Thus,sincea cupisperceived asa single entitythatcanbeseparated from andusedindependent oftherestofitsenvironment, a childwillexpect that thereis a wordthatrefersto a cup,andnot onethatrefersto thecombina-

tionofbothcupandsauceror to justthehandleandbottomofthecup.

Similarly, thechildwillexpect thatcatismorelikely torefertotheanimal sheseesmoving across therugratherthanto a combination ofthecatand rug,ortojustthecatsheadandtail.AsSteven Pinker concluded (referring to Quines gavagai example):

humans[arej innately constrained tomakeonlycertainkindsofguessesprob-

ablycorrect kindsabout howtheworld anditsoccupants work.Letssaythe

word-learning babyhasa brainthatcarves theworldintodiscrete, bounded, cohesiveobjects andintotheactions theyundergo, andthatthebabyforms mental categories thatlump together objects thatareofthesame kind.Letsalsosaythatbabies aredesigned toexpect a language tocontain words forkindsofobjects andwords forkindsof actionsnounsandverbs,moreor less.Thentheundetached rabbit

parts, rabbit-trod ground, intermittent rabbiting, andother accurate descriptions of thescenewill,fortunately, notoccurtothemaspossible meanings ofgavagai. 4

Suchstrategies, andlikelymanyothers, 46 areveryusefulin constraining

orbiasing thechildstheories ofwordmeaning, butthefactremains that

neitherthe childnor the adultcan everbe absolutely confidentthat his

meaning forawordisidentical tothatofanyotherperson. Even consulting a dictionary provides noabsolute assurance, sincea dictionary canonly definewordsthroughthe meaningsof otherwordswhosemeaningsare alsounverifiable. Butthemoreinteraction a personhaswithotherspeakers

ofthelanguage, themoreconfident (though nevercertain) hecanbethat meanings areshared,sincesuchinteraction provides forincreased opportunitiesfor the rejection,revision,and resultingfine-tuning of meanings. 47 Andasthechildsvocabulary increases, alreadylearnedwordscanbeused effectively to narrowdownthemeanings ofnewwords.

202

Without Miracles

It turnsoutthatacquiring themeaning of wordsalsohasimportantimplications foracquiring syntax. MITlinguist andcognitive scientist Steven Pinker, whose important article withPaulBloom onlanguage evo-

lution wasmentioned earlier inthischapte ; haspointed outthatmany puzzling exceptions tosome basic syntactic patterns inEnglish appear to render theirbeing learned bychildren verydifficult ifnotimpossible.

Consider thefollowing sentences (anasterisk precedes words andsentences

thatareun-grammatical inEnglish): (1) Beth sold the cookies to Eric. (2) Beth sold Eric the cookies.

(3) Beth pulled the cookies to Eric.

(4) *BethpulledEricthe cookies.

From thefirsttwosentences, itisclearthata speaker ofEnglish canuse oneoftwodifferent grammatical structures forsentences containing botha direct object (thecookies) andindirect object (Eric). Wecanputthedirect

object after theverb followed bytoandtheindirect object. Orwecandrop thetoandswitch thepositions ofthetwoobjects. Butnotice thatalthough

thesecond structure seems toworkfinefortheverbsold,itdoesnotsound

right tomost speakers ofEnglish fortheverb pulled asused in(4),despite

thefactthatbothverbsbehave similarly in(I)and(3). Letsconsider a fewmoresentences to showthatthisisnotanisolated

example.

(5) ChristopherkickedErin.

(6) ErinwaskickedbyChristopher. (7) ChristopherresembledErin.

(8) *Erinwasresembled byChristopher. Herewehaveexamples oftheactiveandpassive voices. In theactive

voiceconstruction ofsentence (5),thedoeroftheactionisbefore theverb

andtherecipient after theverb. Butinthepassive voice construction of(6),

therecipient isbefore theverb(towhich washasbeenadded) andthedoer

isnowaftertheverb(afterwhichbyhasbeenadded). Notethatcountless

verbs could besubstituted intosentences (5)and(6)andyield grammatical

sentences, suchasloved,heard,kissed, believed, andserved. Butforsome


203

reasonthe passiveconstruction using resembledin (8) is clearly not an acceptableEnglish sentence. N ow if children learnedlanguageby simply listening to and memorizing sentences , and if they never said a sentencethat they hadn't already heard, these inconsistencieswould pose no problem. But countless studies and observations reveal that children are not conservative in their language learning. We already noted their use of past tense forms they could not have heard from an adult , such as *cutted , *drawed , and *breaked . Some

additional examplesof children's creativity in attempting to figure out and apply rules of English grammar are: (9)

~4How was it shoelaced?

(10)

):-Jay said me no .

(11)

)"'I'm just gonna fall this on her.

(12 )

* ! ' m gonna pour it with water .48

So insteadof being linguistically conservative, children are quite creative speakersin venturing beyond the words and sentencesthey hear others produce. Now if it is true (asPinker believes) that children receiveno useful information from adults concerningwhich sentencesthey produce are un-

grammatical ,49then the fact that we do not producesuchungrammatical sentences as adults is one of the most interesting dilemmas of human

learning. If childrenobtain no informationconcerningthe grammatical ity of their sentences , how are they ableto eliminatethe ungrammatical ones? This is often referred to as the problem of the " learnability" of language.

Pinker attempted to provide a solution to this problem by showing that the syntactic exceptions of the types shown above are not arbitrary but dependon often quite subtle differencesin the meaningsof the verbs and the meaningsof syntactic constructions. For example, the so-called dative indirect object can be usedwhen the verb of the sentenceindicatesa change of possession , as in (2) Beth sold Eric the cookies, but not when only motion is implied as in (4) *Beth pulled Eric the cookies. Also, if the verb implies acting upon an object; then a passiveform is acceptable, asin Adam was hit by Anne and (6), but usually not otherwise as in ):4Moneyis lacked by Matilda and (8).

204

Without Miracles

SobyPinkersaccount, learning themeanings of wordsis essential to

producing syntactic sentences. Sohowdoes hesuggest thatthese meanings

arelearned? Toquote,witha bitofaddedcommentary inbrackets: What

weneed toshow isthatthechild iscapable ofentertaining asahypothesis anypossible verb meaning [thatis,consider anyofalarge number ofpossiblevariations], andthatheorsheiscapable ofeliminating anyincorrect

hypotheses [andconsequently selecting thebetterones]asa resultofob-

serving howtheverbisusedacrosssituations. 5 Pinkersoundsevenmore

Darwinian andselectionist when discussing thelearning ofmorphemes (the

meaningful entities thatmakeupwords) instating thatas thechildcontinuestoworkonthatmorpheme overa largesetofsentences, allincorrect

hypotheses willbediscarded atsome point oranother, anycorrect hypoth-

esiswillbehypothesized sooner orlater.. . andonlythecorrect oneswill survive in the limit. 51

Fortheunlearning(elimination) ofovergeneralized verbformssuchas

*drawed and*hitted, Pinker (aswellasseveral otherresearchers) invokes whatiscalled theuniqueness principle bywhich thechild expects thatthere cannotbetwowaysofexpressing exactly thesamemeaning. Sowhenthe childhearsanadultsayNicholas drewa nicepicture anditisclearthatthe adultisreferring to thepastactofdrawing, thechildwillunderstand that

drew hasthesame meaning as*drawed, andhewilleventually replace the

latter with the former. 52

Although Pinker would almost certainly notcharacterize histheory in

thisway,henonetheless hasproposed a procedure bywhich children are

abletogenerate guesses about thestructure ofthelanguage theyhearand theneliminate theincorrect oneswithout thebenefit ofhaving adults indicatewhich utterances areincorrect, thatis,without access tonegative evidence. Topullthisoff,children mustbesensitive to some verysubtle

semantic distinctions among verbs, whichisinitselfquiteremarkable. It

leads Pinker tospeculate thatsuchlexicosemantic concepts (which ap-

peartopertain onlytolanguage learning anduse)mustbeinnate andpart ofaseparate language component ofthemindhaving littletodowithother cognitive abilities. Buthisoverall theory oflanguage learning cannonethe-

lessbeunderstood asselectionist inthatinnate knowledge, arising from

natural selection among humans during biological evolution, interacts with selectionist cognitive processes ofhypothesis formation andelimination

(selection within humans) toarrive atadult language competence.

TheEvolution , Acquisition , andUseof Language 205 WhereasPinker's theory of languagelearning can be construedas implic~

itly selectionist , the one proposedby Americanpsycholinguists Elizabeth Bates and Brian MacWhinney, called the competition model, is explicitly so. According to the competition model, the child's learning of word meanings and grammar has three necessary stages: First, the child developsa function to express. We will call this functional acquisition . Then the child makes a first stab at a way of mapping the function into a form .

We will call this jumping in. Then a period of competition ensuesduring which the range of the form is narrowed or widened .53

Let us take a brief look at each of these three stages. Functional acquisi -

tion hasto do with the child's assignmentof meaningto the objects, actions, and relationships around her. Thesemeaning functions must be developed before shecan understandlanguagereferring to them, and before she can use language herself to express them . The child then has to associate the sounds of the language she hears to these meaning functions . Since at the early stages of language acquisition

shecan do no better than make a guessasto the meaningof a word, phrase, or sentence , this stageis referred to asjumping in. Suchinitial guessesmay be quite wide of the mark , but as the child learns more and more word

meanings, this knowledgecan be usedto help discoverthe meaningsof new words in much the sameway that the final words of a crosswordpuzzleare usually easierto identify than the first onesattempted. Thesejump-ins are, of course, nothing but preliminary guessesas to the meaning of a word (or

grammaticalform) andarethenecessary sourceof variationfor subsequent selection .

Selectionis accomplishedby a processof competition in which words

and grammaticalforms competefor meaningsbasedon the assumption that two different forms must have different meanings (as in Pinker 's

uniquenessprinciple), and if no difference can be found, one of them is wrong. For example, the child may hear the words plate and saucerreferring to what initially appearto be the sametype of object. Thesetwo words will then competefor thesetwo meaningsuntil , after severalpresentations and perhapssomecorrection (" That's not a plate, it 's a saucer" ), plate wins out for round, thin objectson which food is placed, and saucerwins out for round, thin objects on which cups are placed. MacWhinney likens this to the competition of two speciesfor the sameenvironment resulting in each speciesestablishingits niche in that part of the environment for which it is

206

Without Miracles

bestadapted .54In othersituations , oneof theformsmaybeeliminated entirely (become extinct ) aswhen*goedand*cutted areeventually replaced bywentandcut. Such competition isnotlimited towordmeanings . Different languages usedifferent ways ofexpressing grammatical relations , andthechildmust learnthese syntactic rules aswell.InEnglish , theorderofwords inasen tence istheprimary determinant in assigning therolesofsubject (almost always placed before theverb ) andobject (usually aftertheverb ), sothatin thesentence Thefoodatethedog , thefoodwould normally beunderstood astheentitydoingtheeating despite thefactthatthismakes littlereal worldsense (perhaps thedogfellintoavatofhighly acidic hotsauce ?). But thesame sentence structure incertain otherlanguages (such asSpanish in Lacomida cornia elperro ) wouldbeimmediately understood asthemore sensible "thedogatethefood ." Bates andMac Whinney andtheirassoci atesin theirextensive research onbothchildren learning theirfirstlanguage , andstudents andadults learning second languages , indicated that learners relyoncertain cues toresolve thevarious competing words , mean -

ings , andsyntactic forms .

It should beclearfromeventhisbriefdescription thatthecompetition modelis selectionist . Although Bates andMacWhinney donotusethat wordoritsvariants , theprocess ofcompetition theypropose isclearly one throughwhichselection (andelimination ) takesplace . Children make guesses concerning themeanings andformsof thelanguage theyhearand eventually fine-tunethese guesses bycumulative selection aswordsand formscompete forvarious meanings andfunctions . Thetheory isinformed bybothlinguistics andpsychology , andit drawssupport froma consider ablebodyof research . Andit doesnotassume thatthechildpossesses an extraordinary storeof detailed innatelinguistic knowledge . Theselection ismis clearwhenMacWhinney states that "theunderlying ideain the Competition Modelisthatmental processing iscompetitive intheclassical Darwinian sense ."55 A Near-CommonDenominator - Selection

Wehave nowconsidered several theories oflanguage learning thatdifferin anumber ofrespects . Oneuseful wayofcomparing andcontrasting them is toconsider theequation : Innate knowledge x experience x learning =language knowledge .


207

Thisequation states thatthechild'sabilitytouselanguage istheresultof theinteraction of thechild's innateknowledge , experience , andlearning . Innate knowledge isconsidered inthebroadsense , including thebrainand vocaltractstructures shaped by biological evolution , in addition to any moregeneral cognitive or specifically linguistic knowledge thatcouldbe considered to beapartofthechild'sbiological endowment . Notethatthe interaction among innateknowledge , experience , andlearning is consid eredto involve multiplication (notaddition ), sincethisrecognizes thatif anyoneof thethreefactors didnotexist(werezero ), thechild'slanguage knowledge wouldalsonotexist(wouldalsobezero ). Usingthisformula , it canbeseen thatthevarious theories of childlanguage acquisition discussed thusfardifferintheimportance theyascribe to thethreefactorsin a moreor lesscompensatory fashion . Skinner 's and otherbehaviorists ' accounts of language learning emphasize theroleof experience andlearning anddownplay innate knowledge . Chomsky andFodor , in contrast , minimize theroleof bothexperience (recallChomsky 's argument of thepoverty of thestimulus ) andlearning (Fodorbelieving thatanyformoflearning isimpossible )whileemphasizing theimportance ofinnate knowledge . Indeed , thestandard practice ofmost linguists todayisto putasmuchastheypossibly canin theinnate knowl edge factor , whichtheyrefertoas"universal grammar ," whilestillallowing children tobeableto learndifferent languages , thatis, whileproviding the minimum necessary roleforexperience , since it isclearthatchildren speak thesame language theyhear . It should alsobenotedthatlinguists ' concep tionoflanguage knowledge ismuchricherandsophisticated thanthatconceived ofbySkinner . Pinker appears to place moreemphasis onexperience thandoChomsky andFodor , buthenonetheless maintains oneaspect of thepoverty of the stimulus inhisbeliefthatnegative evidence , whichwouldallowthechildto reject easily mistaken hypotheses aboutthelanguage beingacquired , isnot available tothechild.Pinker alsoemphasizes theroleofinnate knowledge , since histheorydepends onthechildbeingableto classify verbsin quite subtle andsophisticated waysusingcategories andconcepts hebelieves to bequitespecific tothedomain oflanguage . Bates andMacWhinney 'scompetition modelof language learning contrastswith Chomsky 's andPinker 's viewsandbringsusbackcloserto

208

Without Miracles

The Evolution , Acquisition, and Useof Language

209

210 (1) (2) (3) (4) (5)

Without Miracles

Wheredidyouputthenewspaper ? Therewasaninteresting storyin thenewspaper yesterday Thenewspaper isgoingonstrike. Workers aredemonstrating outsidethenewspaper . Thenewspaper isexperiencing financialdifficulties .6O

I would venture toguess thatyoudidnothave difficulty understanding themeaning ofany oftheindividual words inthispassage .ButI would also venture toguess thatthepassage asawhole probably didnotmake much sense toyouwhen youfirstread it.ButifI now inform youthatthepassage has something todowithwashing clothes (and assuming thatyouhave had some experience inwashing clothes ), reading it again willlikely bequite a different experience asitwillnow elicit meanings it didnotbefore .This is because I helped youtoconstrain your hypothesis about what thepassage isabout .Butagain ,themeaning must actually becreated byyouand isnot transmitted bythewords ,phrases ,orsentences ofthepassage .

TheEvolution , Acquisition , andUseof Language 211

These examples aremeant todemonstrate thatlanguage comprehension isnotamatter ofreceiving meaning fromaspeaker orwriter , butrather is anactive process ofconstructing meaning asthelistener orreader attempts tomake thewords , intentions , andcontext ofthesituation fit.Aschild language researcher GordonWellsputit:

guageusehasimportant cation, to which we turn next.

12 -=-= Education

-

TheProvisionandTransmission of Truth, or theSelectionist Growthof FallibleKnowledge ?

But referring now to all things which we unde , we con , not the sp who utters words , but the guardi of truth withi the min itse bec we ha perhaps been reminde by words to do so . More , He who is con tea ; for He who is said to reside in the interio man is Chri , that , the un excellen of God and His everla wisdo , whic eve rati so do in deed consult . -Instructio .Instead S Augustin t of paper , we have pupils , pupils whos mind have to be imp wi th symbols knowled . Instea of type , we have class and the res of the ap ratus devised to facilitat the opera of teach . The ink is rep by vo of the masters , since it is this that conve inform to the min of the lis , while the press is school discipl , which keep the pup up to the wo an compels them to learn . -Selectio J Comeni ohn Let me first briefly summa the critica appr [ t edu o ] . It is ba on evolution epistem , which claim that we neve rece kno , bu rather create it ; we create it by modif the know we alre ha ; an w modify our existing knowle only when we unco inad in it th had not recogni hereto . Accep this as an expl of ho kn grows ,-H Iof have sugges that teach cons their role as fac of th growth their studen ' knowle . Perkins enry

Providence

Theimportant roles thattradition , technology , andscience playinthesurvivalandproliferation of ourspecies andthequality of ourlifemake it

214

Without Miracles

The purpose ofthischapter istoexplain thepuzzles offitthatarethe consequences ofeducation . Some ofthediscussion may overlap and draw

Education 215

on previous ones , but its focuswill beon formaleducation andpastand present viewsof howeducation resultsin newandimprovedknowledge andskills. EducationasProvidence

Perhaps themostwell-knowngeometry lessonof all timeis theonepresented by Socrates to Meno's slaveboy. Althoughweaddressed theMeno dilemma in chapter6, it will beusefulto startourdiscussion of education bytakinga closerlookatthatinteraction between teacher andstudent . TheproblemSocrates describes concerns theareaof a square . Afterestablishing thattheboyunderstands thata two-by-two square contains an areaof foursquare units,heaskstheboyhowlongeachsidewouldhaveto befor thesquareto containeightsquare units. Sinceeightsquare unitsis twiceasmuchasfour, theboyquickly(anderroneously ) concludes thata four-by-foursquare wouldcontainthedesired area. ButwhenSocrates sketches thisfour-by-foursquare , it becomes apparent thatsucha squarecontains not eightbut rathersixteensquare units. The boy, thenrealizingthatthedesired eight-square -unitfiguremustbelarger thantwo by two but smallerthanfourby four, guesses thattheansweris three.Butwhenaskedto reconsider thisanswer , theboyrealizes thatit is alsoin error,sincesucha square wouldhaveanareaof ninesquare units, noteight. At this pointthe boy admitsthat he doesnot knowthe answer , and Socrates pointsout to Meno, whois observing theinteraction with keen interest , that in hisadmittedignorance andconfusion theboyis actually betteroff thanhewasbeforewhenhebelieved heknewtheanswerbutin factdid not. With furthercarefulquestioning by Socrates , theboyfinally discovers thecorrectanswer , thatis, that a squaremadeof sidesequalin lengthto thediagonal of a two-by-two square(equalto thesquare rootof eight)hasanareaequalto thedesired eightsquare units. SinceMenostatesthattheboyhadneverbeeninstructed in geometry , Socrates concludes thattheknowledge theboyfinallydemonstrated in solvingtheproblemcouldnot havebeenlearnedearlierin hislife. Andsince Socrates didnotinstructhim, butonlyquestioned him, hedidnotgivethe boyhisnewknowledge . Socrates therefore argues thattheknowledge the boyusedto solvethe problemmusthavebeenpossessed by the boy all

216

WithoutMiracles

along andwastherefore simply recollected during theirencounter from knowledge provided byanimmortal soul . Thatis, inquiry istherecollec tionofknowledge thatwealready have - aprovidential view ofknowledge thatleads toaprovidential view ofeducation aswell . Such aview ofeducation didnotdiewithSocrates andPlato . St.Augus tine(354 -430 ), theintellectual father ofChristianity , fashioned aphiloso phyinwhich Godistheanswer toallquestions . Godmade theuniverse , was responsible foreverything thathappened init,andwas thesource ofall theknowledge thathumankind wasableto acquire about it. InhisDe Magistro ("concerning theteacher "), Augustine adopted thesame basic conclusion asSocrates concerning teaching , butinstead ofattributing all knowledge toimmortal personal souls ,heacknowledged God asthesource ofallknowledge , "theunchangeable excellence ofGod andHiseverlasting wisdom , which every rational soul does indeed consult ."4 Therefore according toAugustine , teachers cannot transmit knowledge totheirstudents byinstruction , since teachers canonlyutter words . If the words arealready familiar , students canlearn nothing fromthem . Andif thewords areunfamiliar , thestudents canstilllearn nothing fromthem : Foritisthetruest reasoning and most correctly said that when words are uttered we either know already what they signify orwedonotknow ; if weknow , then we remember rather than learn , butifto we do ,then wedonoteven remember , though perhaps we are prompted ask .5notknow Thenoninstructionist argument concerning words thatAugustine makes isremarkably similar totheoneattheendoftheprevious chapter thatlanguage usedoes notinvolve thetransmission ofmeaning orconcepts or information fromoneperson toanother . Tomake Augustine 'spointclear andrelevant toeducation today , imagine thatyouareaphysics teacher attempting toteach astudent theconcept ofacceleration . If thestudent already hasagood idea ofwhat velocity isandunderstands what it means foraquantity tochange overtime , youmayexplain acceleration asthe change invelocity over time , andthestudent may nowhave abasic idea of whatacceleration is. Thisisbecause shealready knew whatvelocity and change overtimemean , soputting these twofamiliar concepts together yielded anewone . Butif thestudent hasnoideaofwhatvelocity isor doesn 't understand what ismeant byachanging quantity , ateacher cannot domuch toteach theconcept . Nonetheless , bydemonstrating acceleration andallowing bothteacher andstudent toaskandanswer questions , the

Education

217

concept canbeacquired. Since it isnotclearhowsuchnewknowledge is acquired, Augustine canonlyimagine thatitisprovided byGod.Forhim, therefore, teachers canonlyhopeto guidestudents to knowledge thatis providedby God as divineillumination. Education

as Instruction

Buta quitedifferent viewofeducation wastoaccompany thenewphilosophybrought aboutbythebeginning ofmodern science insixteenth-centu-

ryEurope. Exemplified inthewritings ofSirFrancis Bacon, itrejected God, religious dogma, andtheclassic philosophical textsasthesolesources of knowledge. Instead oflooking withinoneself forGodsrevealed truth,or studying books andlistening tothewords oftheteacher, Bacon insisted that

naturebe studieddirectly, for it wasonlyin understanding natureand therebyadvancing science thatthehumancondition couldbeimproved.

According toBacon, ifyouwanttoknowhowmanyteethyourhorsehas, youmustlookintoitsmouth andcount them, notwaste timereading what

scholarsandphilosophers haveto sayaboutthematter.Thisnewphiloso-

phyalsostressed theimportance ofpractical knowledgeknowledge that couldbeputto useforgrowing crops,building factories, andconquering the seasand the unknownlandsthat laybeyondthe horizon.

Thischangein perspective concerning knowledge andscience wasto

havea significant impacton education. Whereas beforeeducation had beenconsidereda sort of initiationinto the cultureof the societyor a

process ofdiscovering thetruthwithin provided byGod,it nowbegan to beregarded asa process ofinstruction. Direct andcareful observation of theworldwouldallowknowledge to growbythetransmission ofinforma-

tionthrough thesenses tothemind. Where direct experience couldnotbe easily had,theknowledge already gained byteachers (whonowbecame instructors) wouldbetransmitted intotheminds ofstudents through their spokenandwrittenwords.Tofacilitate thelatter,textbooks cameto re

placeclassic texts.These textbooks weredesigned topresent knowledge to students inthemosteffective manner. Subject matterwasbrokendown

intosmall,easilydigested components andthencarefully organized and sequenced tofacilitate learning. Teachers andtextbooks thusbecame transmission relaystations whose jobwastoreflect knowledge oftheworldinto the minds of students.

218

Without Miracles

Thisinstructionist view ofeducation had , andstillhas , important conse quences forallaspects ofeducation . Firstofall, if education isthetrans mission ofknowledge fromtheteacher (ortextbook ) tothestudent , then theknowledge possessed bytheteacher (ortextbook ) must beaccurate . Instruction makes nosense if what istobetaught isnottrue . Teachers and textbook authors must therefore be(oratleast pretend tobe)unquestioned authorities onthesubjects being taught andwritten about . Second , aninstruction -oriented view ofeducation tends toputtheblame onthestudent forfailure tolearn , since if theinformation tobetransmitted ispresented clearly andaccurately , such failure must bedue toproblems on thereceiving end .Inattentivenss , laziness , andlackofmotivation arethere foreoften metwithpunishment andridicule inaneffort tomake students payattention anddotheirwork . Third,if education isthetransmission ofknowledge frominstructor and textbook tostudent , thentheusual testofitssuccess iswhether students can reproduce thetransmitted information in spoken orwrittenform.Thisencourages thememorization ofwhatis"known " to be"true" andtheuseof standardized , objective testsasindicators ofacademic success . In aword,atransmission -based , instructionist perspective oneducation isauthoritarian . AsNewYorkUniversity educational historian HenryPerkinson observed : . . . thetransmission metaphor [foreducation ] persisted through theeighteenth , nineteenth , andintothetwentieth century . It persists downtothepresent , where many educators stillremain caught initsspell . Believing thatweinductively learn fromexperience , theystrive valiantly to transmit knowledge to theirpupils . Inevitably , thisleads teachers toauthoritarianism ; theattempt tosecure greater con trolover theeducational process .Take forexample thesubject matter :teachers who seek totransmit knowledge attempt tocontrol thesubject matter bypackaging it intoatransmittable curriculum . . . . Take asanother example ofauthoritarianism , thetreatment of pupils : teachers whoseek to transmit knowledge to students attempt tocontrol them ; they"prepare " them , "stimulate " them , "motivate " them , "getthem topayattention ," "getthem moving ." Allthese tactics areattempts to control pupils canmore expeditiously andefficaciously transmit knowledge .6 sothatteachers Whereas theconsequences ofaninstructionist viewofeducation maybe problematic , itsgreatest problem isbeliefin thetransmission ofknowledge itself . DavidHumeidentified thisproblem in theeighteenth century with hiscritique of induction . Recall fromchapter 6 thatheconcluded thatwe

Education

219

can never justify knowledge by observation, sinceto be useful, our knowl edge(for example, objectsfall to the earth) must go beyond what we have

personallyexperienced (someobjectsfall to the earthat certaintimesand places ). But althoughHumeadmittedthe logicalimpossibilityof the environment being able to instruct the mind by way of the senses , he did not

rejecta psychological theoryof induction. That is, heconcludedthat learning from observation was logically unjustifiable; however, this is what we actually do, sincewe believethat repeatedobservationsof a phenomenon indicate a generallaw of nature. So it is logically invalid to assertthat the sun will rise tomorrow sinceit has risen everymorning up to and including today, but we still believethat we know that the sun will rise tomorrow basedon our prior repeatedobservationsof mornings and sunrises.7 If Hume is correct in asserting that no amount or type of experience can

provide us with certain knowledge, yet we believethat experienceleadsto knowledge, this puts educationin a very curious situation. For it meansthat although the knowledge possessedby teachersand contained in textbooks is almost certain to be mistaken , students will tend to believe (and will be

encouragedto believe) that it is unquestionably accurate. From this perspective, education is not a processby which studentsacquire and improve their knowledge , but is a type of indoctrination

in which students are en-

couraged and compelled to accept as true the inevitable errors of their teachers and textbooks .

But is it actually the casethat we acquirewhat must be uncertain knowl edge from our sensoryexperiencesof the world ? That is, does induction

work psychologically eventhoughit is a logicallyinvalid process ? To answer this question , we must consider again the work of Karl Popper who

was also discussedin chapter 6. The belief in the transmissionof knowledge from senseexperience to mind is part of what Popper referred to as both the

common-sensetheory of knowledgeand the bucket theory of mind; he convincingly argued that this view of learning is inaccuratefrom both logical

andpsychological perspectives . Humebelievedthat we areconditionedby habit to believethat repeatedobservationslead to useful generalizations. Popper noted that learning cannot proceed in this manner since to recognize that an observation is in some sensea repetition of a previous one requiresknowledge that cannot be obtained by observation alone. This is surely a difficult idea to understand, but it is essentialto grasp if we are to

220

WithoutMiracles

Education

221

transmission fromteacherandtextbookto student , andin thissense the teacher cannotliterallyinstructthestudent . Hume'sarguments against thelogicof inductionandPopper 's argument againstthepsychology of inductionwouldprobablybeconsidered irrelevant philosophical nonsense by mosteducators , particularlysinceour everyday experiences suggest that suchtransmission of knowledge from teacher to student takesplaceroutinely . Soalthough theseideashavehada considerable impactonphilosophy , particularly thephilosophy of science , weshouldnot expecteducators to bemuchinfluenced bythem,especially sincetheyareusuallymoreconcerned with thepracticaldifficulties of educationthanwiththephilosophical ones . Soaninstructionist conception of education remains verymuchaliveand influentialtoday.Nonetheless , the twentiethcenturyhasseenthe emer genceof educational theories thathaverejecte ~ theinstructionist viewof education . Thishascomeaboutasmoreandmorepsychologists andeducatorsdismissthe conception of students aspassive bucketsinto which knowledge ispouredbyteachers andtextbooks , andreplace it with aview of students asactivecreators of theirownknowledge . Education asDarwinianSelection PiagetandMontessori JeanPiaget'stheoryof cognitive development (touched onin chapter9) has hadperhaps themostimpactonthischange of perspective . Piagetshowed thatchildrenseeandunderstand importantaspects of theworldin ways verydifferentfromthoseof adultsdespite thefactthatbothareexposed to thesameworld. If knowledge resultsfromtakingin information , howis it thatchildrenperceive thingssodifferently ?Whydoesa two-year-oldcalla butterflya bird?Whydoesa four-year-old insistthat spreading out eight coinson a tableresultsin morecoinsthanwhenthe samecoinswere grouped closertogether ?Andwhydoesa five-year-old "explain " thatthe sundoesnotfalldownbecause it isyellow , or thatthesunpushes theclouds across thesky?Thefactthatchildrenhavehadfewerandlessvariedexperiences thanadultsmightexplainwhy theymaylackcertaintypesof knowledge or havelessknowledge . Butit cannotexplainwhytheirunder standings arequalitatively differentfrom thoseof adults , andhow they

222

Without Miracles

come upwithsuchextraordinary andcreative explanations thatcannot be traced to actual experiences . According to Piaget , these differences arisebecause children donotpassivelyreceive knowledge fromtheirenvironment butrathermakeit them selves . Fromthisperspective , knowledge resulting fromthetransmission of information fromenvironment (or teacher ) to student is replaced bythe concept ofgrowth . Children growin knowledge because theyconstruct it, oftenbyrecombining whattheyalready knowin newwaysandtesting it, andtheenvironment , whether it bethephysical environment of hard knocksor mommy 's verbalresponse , provides feedback concerning the adequacy oftheirconstructed knowledge . Thustheroleoftheenvironment for Piaget is notthatof a provider or transmitter or instructor of knowl edge , butratherastheselector oftheknowledge created bythechild. Piaget 'sownwords(translated fromFrench ) indicate thedegree towhich herejected a transmission viewof education andadvocated aconstructive one: I 'm not an educator ; I have no advice to give . Education

is an arena of its own and

educators must find the appropriate methods , but what I 've found in my research seems to speak in favor of an active methodology in teaching . Children should be able to do their own experimenting guide them

by providing

order for a child to understand reinvent it .1O

Renowned

Italian

and their own research . Teachers , of course , can

appropriate

materials , but the essential

something , he must construct

educator

get , had little time or inclination research . Instead she developed

Maria for

Montessori psychological

thing

is that in

it for himself ; he must

( 1870 - 1952 ), unlike

Pia -

the .ory or experimental

themostsuccessful method ofeducation inthetwentieth century , perhaps themost successful method in thehistoryof education . In Montessori schools , children of threeyears of agelearnto dust , to dry, to setthetable , to serve attable , to wash dishes . At thesame time,theylearnto takecareof themselves : theywash , bathe , dress andundress themselves , theyarrange theirclothes in theirlockeror in a drawer , tietheirshoes , combtheirhair,andsoOll.Byfourandahalfyears of age , theylearnhowtowriteandhowtoreadandcandobasic arithmetic calculations .11 Montessori did not seetheroleof theteacher asaninstructoror transmitterof knowledge . Instead , in a Montessori classroom theteacher tends to stayin thebackground , actingasa kindof valetwhocreates aninterest ing andchallenging environment . Thechild's ownnaturalcuriosityand

Education

223

desireto masterthe environment results in learning, learning that depends on attempting new skills , making errors , and learning from mistakes . As for

Piaget, the learning environment, including the teacher, actsnot as a source of knowledge but rather as a selectorof the knowledge constructedby the child .

Learning from Mistakes In his book that reviews twentieth -century educational theory , Perkinson

contendsthat the generalapproach of Piagetand Montessori to education (as well as aspects of those of B. F. Skinner, A . S. Neill , and Carl Rogers ) is

essentiallyDarwinian, and he contraststhis selectionistapproachto a transmission -instructionist conception with respect to the process of education ,

and the roles of the teacher, the subjectmatter, and the student. In contrast to a view of education

as a process of transmission

, selection sees education

as " a procedureof (Darwinian) growth; trial-and-error elimination; the continuous modification of existing knowledge ."

From a transmissionperspective" the teacherpreparesthe student, prepares the subject matter , and transmits (instructs , matches) the subject to the student in the form

of lessons that the student learns , " whereas from

a selectionistperspective" the teachercreatesan educativeenvironmentan environment that is free, responsive, and supportive- wherein the student can improve (modify) his presentknowledge through trial -and-error elimination

."

A transmissionview seessubject matter as "what is transmitted" but a selectionistperspectiveconsidersit " an agendathat specifieswhat aspects of the students' presentknowledge are to be improved" and as that which " evokes the students' present knowledge and tests it (reveals the inadequa cies in that present knowledge )."

Finally, transmissionconceivesof the studentas " a learner, a more or less passive receptor who needs to be controlled and motivated " while selection

seesthe studentas " a fallible, activecreator of knowledgewho seeksorder" so that "when he discernscontradictions (errors, mistakes, inadequacies)in his presentknowledge, he will modify that presentknowledge." 12 Perkinson does not pretend that Piaget, Montessori, Skinner, Neill , or Rogers would agree with his Darwinian , selectionist reinterpretation of their work . Indeed, we saw in chapter 9 how Piaget explicitly rejected

224

Without Miracles

Darwinian theory andinchapter 7howSkinner believed cultural evolution (which includes education ) todepend onLamarckian transmission . None theless , allof these individuals (withthepossible exception of Skinner ) considered education to bea process of creative growth thatresults in theindividual becoming betteradapted to theenvironment andthere forebetter abletocontrol aspects oftheenvironment . Since theyrejected bothprovidential andinstructionist explanations ofeducation , theirviews leadquitenaturally toaselectionist conception , even if theydidnot(again withtheexception of Skinner ) explicitly make Darwinian or selectionist arguments . Theselectionist orientation canperhaps bebest appreciated byassessing theroleoferror . If thisperspective views education asinvolving thefallible creation ofknowledge , theneducational theories andpractices mustcon sider errortobeanessential partofeducational growth .Thatisbecause the process creates fit notbytheclairvoyant , advance fashioning of adapted thoughts orskills , butrather bytheblindproduction ofvariations andthe subsequent hindsighted selection ofthethoughts andskills thatbetter fitthe needs andpurposes ofthelearner . It istherefore notsurprising thateduca tionaltheories thatreject transmission emphasize thepositive roleoferror , since aselectionist process inevitably results inmore errors (unfitvariations tobeeliminated ) thansuccesses (fit variations ). InPiaget 'stheory , thechildexperiences cognitive growth byrealizing the mistakes inherent in hisviewoftheworld . It isamistake to believe that there arenowmore coins onthetable simply because theyhave been spread outtotakeupmore space , andthechildwilldevelop more advanced modes ofperception andthinking when herealizes these mistakes . ForPiaget , cog nitivedevelopment proceeds asthechildcreates newandbetter waysof interacting withhisenvironment . Butsince these newways ofinteracting arenotdetermined orinstructed bytheenvironment , it isinevitable thatthe childwillmake many mistakes along theway . Thus , erroranditselimina tioncanbeconsidered tobethebasis ofcognitive growth andeducation . Montessori alsostresses thecentral roleoferrors ineducation : Supposing westudy thephenomenon oferror initself ; it becomes apparent that everyone makes mistakes .This isone oflife'srealities , andtoadmit it isalready to have taken agreat step forward .Ifweare totread thenarrow path oftruth and keep ourholdonreality , wehave toagree thatallofuscanerr;otherwise ,weshould all beperfect . So , it iswelltocultivate afriendly feeling toward error , totreat it asa

Education

225

companion inseparable fromourlives,assomething having a purpose, whichit

trulyhas... .Whichever waywelook,acertain Mr.Errorisalways present! Ifwe

seekperfection, wemustpayattention toourowndefects, forit isonlybycorrectingthesethatwecanimproveourselves. 13

Theessentialroleof error(thatis,unfitvariations)in a selectionist view

provides themoststriking contrast toa transmission orinstructionist view withrespectto theroleoftheenvironment. Sincefromaninstructionist per-

spective thepurpose ofeducation isto transmit knowledge to thestudent accurately andefficiently, anyerroronthepartofthestudent isanindicationthatsomething hasgonewrongin thetransmission process,usually with the studentconsideredat fault for inattentiveness or laziness.In con-

trast,forselectionism, theeducative environment reveals errorsofbehavior or thinkingto thestudentandis responsive to the studentsattemptsto revisehis behavioror thinkingfor the better.Thus, a teachersprimary

responsibility isnottotransmit knowledge, buttoassistthestudent indiscovering thewaysinwhichhiscurrentknowledge isinadequate. Butsince suchrevelations of inadequacy canbequitethreatening to a studentsselfesteem, theteachermustprovide anenvironment thatissupportive ofthe students attempt tobetterhisknowledge. Thisenvironment should alsobe

free,sothatthestudent willbenotbeprevented fromattempting boldnew solutions. 14 In short,studentsshouldbe eagerto encounter theirmistakes andwill,it is hoped,findthemselves in an environment that encourages themtorevisetheirthinking andactionstoarriveatbettersolutions totheir problems.

The Dangers of Instruction

Thiscentury hasseenthedevelopment andspreadofselectionist viewsof education;however,manyif not mosteducatorsretainan instructionist

approach. Thatisbecause education asusually practiced admittedly looks liketransmission. Andlearningobviouslydoestake placein classrooms whereteachersbelieveand act as if theyare in the businessof transmitting knowledge.

According to Perkinson, thelearning thattakesplacein suchenvironmentsisnot dueto thetransmission ofknowledge thatis attempted(which

is bothlogically andpsychologically impossible) butratherdespite such efforts. Even a transmission-orientedclassroomis free to some extent,

andeffective teachers usually doprovide a supportive environment fortheir

226

Without Miracles

students

{ although

Furthermore form

of

which

- and

reveal they

some

students

, there

the

coercion

not

improve

classifies

are

Second

those

those

skills

as the

are

out . These

created pears -

especially after the

The

third

the

Many

be

unable

or

skills

or

Selectionist

What

and

the

by

to

It

2 .

It

than

true

believers regard

, and

or

the

at

least of

a

to

reactions

from

be .

resentment

of

and

teachers

the

at

are

to

their

so

. The

be

-

do

teacher

who

textbook

or

game under

-

, knowledge , usually

have or

knowledge

the

skills

events

pupils

disap

-

undergone

the

into

experts

accord

in with

. IS

consider

and

game

- knowledge of

the

own

, however

revise

to

the ( or

first

a those

course

promulgate

realized

found

becomes

pseudo

teacher

modify

might be

reluctant were

modify

for

elimination

fabricating

create

. These

authorities

parents

then

, in

the

they

this

, that

last

such

improve

inadequate

have

get

it

shown

of

most

to

invite

is possible elimination

to

how

facilitate

present

them

, and

to

to

to

the

it

possibility

as

a student

his

or

her

a

would

knowledge

.

these

move ? It

educational

certainly

selectionist

is possible

students

is , and

practice

away

teachers

influ

-

sugges

-

- oriented

. Perkinson

' intellectual

psychologists

has growth

offered within

the

exist

-

:

subject

students

wherein

to

other

schools

accept

view

possible

orientation

, Montessori

educators

to

? Is

, transmission

is possible

- error

, which

. They

alternatives

Piaget

is possible

try It

teacher

must

least

arrangements

1.

who

the

they

educators

ing

, those

the

to

,

Teaching

are

enced

3 .

at

a coercive

tions

hipsters

. It

when

schooling

the

and

outcome

whom and

authorities

educators

for

wants

the

to

results

tend

possible

- error

concern

the

.

pupils

pronouncements

positive

of

them place

}.

in

test

classrooms

fear

- and

students

disadvantages

three

from

trial

understandings ones

teacher

for .

are

) as final

whatever

and

test

the

lead

serious

of

"

students , and

take

, these one

, either in

the

socialization

field

from

are

group

intellectual

and

those

withdraw

participate

stupid

what

standings

who

not

some

to

" good their

and does

mentioned leads

the to

, quizzes

learning out

already . This

pupils do

, there

finding

. As

to

feedback

' knowledge some

points

coercive

. They

them

least

limited ,al

, discussions

students

at

Perkinson

and

be

critic

sessions

the

approach

authoritarian

of

in

. Thus , but

may

provide

- answer

errors know

transmission

support

teachers

question

what

First

this

, the

matter to

rather

encounter

than

the

try

subject

to

transmit

matter

it .

critically

rather

it . critical knowledge

encounters grows

as .

a selection

procedure

of

tria

/-

Education

227

4. Regardless of institutionalconstraints , teachers canfacilitatethisgrowthby construingtheir role to bethat of creatinga classroomenvironmentthat is morefree, more responsive , and more supportive : a placewherestudentscan more readily learnfrom their mistakes . 5. Finally, it is possible , in the schoolsastheypresentlyare, for teachers to reconceptualize the aimof schoolingasan attemptto developconcerned criticswho can andwill facilitatethegrowthof our culture.16 Education as the Reorganizationof PerceptualControl Systems Although Perkinsonprovides convincing argumentsfor a selectionistview of education, he does not attempt to describethe specific mechanismsby which learning takes place, or examine teaching and learning as purposeful activities. To addressthese aspects, we will now consider educational growth from the perspectiveof perceptual control theory as describedin chapter 8. It will be recalledthat perceptualcontrol theory seesadapted (that is, fit ) behavior as allowing an organism to control some aspectof its environment. And sinceit is only through perception that an organism can know anything of its environment, adaptedbehavior is in effectthe control of perception. This view contrasts sharply with all other psychologicaltheories that consideran organism's perception to be in control of its behavior. For perceptual control theory, learning is the reorganization of an organism's control systemsthat allow it to control perceptual variables it could not previously control, a reorganization resulting from a Darwinian processof cumulative blind variation and selectiveretention. To apply the idea of reorganization to education, let us usethe example of a person learning to swim. In its most rudimentary form, being able to swim can be defined as staying alive in water that is deeperthan one is tall , that is, being able to tread water. One way to " teach" a nonswimmer to swim is to throw the person into a body of deepwater (we could call this the immersion method). This will likely createerror, sincethe student will havedifficulty keepingher head abovethe water.17This perceivederror in a crucial variable will trigger reorganization so that the student will immediately beginto move her arms and legsvigorously in random patternsto find some way to maintain her ability to breathe. If she finds a behavioral pattern (actually a perceptual-behavioral control loop) that allows her to

228

Without Miracles

breathe , if evenonlya fewgaspsbeforeshedisappears belowthesurface again , therandomness of themovements will declineuntil sheis ableto keepherheadabovewatercontinuously , atwhichpointwewouldsaythat shehaslearned to swim.In effect , thestudenthasnowgainedcontrolover a variablethatshecouldnot controlpreviously , andsoby our definition learning hastakenplace . Sincethestudentdidnot initiallyknowhowto swim, herinitialmove mentswereof necessity blindattempts to do so. Butalthoughshedid not knowhowto keephermouthabovewater , shecouldperceive howsuccess ful shewasin herattempts (gettinghereyesabovethewateris betterthan below , butnotquitegoodenough ). Thisthenprovides a criterionfor selec tion amongthevariousbehavioral patterns attempted , andallowsthestudentto learnfromhermistakes , eliminating patterns thatdidnotsucceed in gettingherheadabovethesurface andretaining thosethatdid. It is easyto imaginethat thelearnerwouldbeveryhighlymotivated , sincefailureto learnto swimwouldresultin deathfromdrowning . Accordingto perceptual controltheory , motivation simplyrefersto error(that is, a difference between a perception andthereference levelfor thatperception ) thatresultsin actionto eliminate theerror(seefigure8.1). Motivationis therefore considered to beinternalto thestudent , sincethereference variable isdetermined bythestudent , notbythe . levelof thecontrolled envlronment . Wemustpointout, however , thattheimmersion methodof swimming instruction maywellfail for anyparticularstudent , sincethereis noguaranteethat thestudentwill comeup with an effective controlsystemfor treading waterwithinthefewminutes available beforelackof oxygen leads to unconsciousness anddeath . Clearly , a lessdrasticapproach iscalledfor. Thismethodcouldbeimprovedin a numberof ways. First, wecould simplyallowmoretimefor learningto takeplace . Thiscouldbeaccom plishedby havingthestudentpracticeat theedgeof a swimming poolso thatshecouldreachout andholdontotheedgeof thepoolwhenshefelt herself goingunderwater . Or shecouldpractice in waterthatwasonlv . neck deepsothatshecouldsimplystandin thewaterat anytimeto breathe . Givenmoretimeto try out'newpatterns of movement andeliminate those thatareineffective , thelikelihoodof successful learning wouldincrease .

229

Education

Another

approach

with

verbal

your

sides

tion

, or

be

to

it

in

the

is

and

straints

on

different

to the

easier

of

swimming

could

the

and

onto . After

to

learning higher

since

learning than

gether

. 18 A

constantly upon student ganization

the that to

likely

a

tasks wants will

, first

the

that to take

are be

her

in

of

, if

how

when

she

someone

swimming

else

oneself

the

form

explicit

of

). con

-

con

-

pro

-

instructions

to

just

a

until

beyond

then

an

learning and

rising

water

sees

the

at

is

is

more to -

teacher

as

imposing

. Assuming

,

.

. Success

movements

continually

new

could

deeper

swimming

this

from

, she

subproblem

leg

abilities

control

student

problem in

of

upward

facilitates

each

and

abilities

achieves

in

water

, the

subproblems

a

. The

- depth

herself

to

less into

acquired

separately

arm

over

of

skill

feels

complex

these

series

this

feels

alone

control she

she

solution

both

be

masters

and

movements

gain

shoulder

more

view

a bit

in

easier

the

' s current

to

to

she

can

the

them

- depth

make

a

movements

into

offering

teacher

down

until

until

shoulder

to

and

suggesting break

she

leg

finding

arm

student

place

own

have

learn

arms

legs

and

a solution

able

to

. After

- reorganization of

her

will

in

stand

problem

learning

selectionist

by is

student

her

arm

effective in

. Even

she

, the

for

legs

kick

complex

make

skill

with her

the

finding

success

aware her

the

probability of

effective

( watching

to

opportunities

have

them

down

than

. One

make

and

combine

Breaking

way

a float

how student on

-

not

demonstrating

instruction

or

practicing

would

teacher

actually

time

verbal

goals

from

to

provide

more

and

weight

imita might

, she

information

cannot

student

movements

her

hold

they

knowledge

could

of

for

learn

the

from

.

models

horizontal

water

attempt

in

to

subskills

lifting

the

the

of

teacher

make

force

do

still of

herself that

a model

matter

perceptions

useful

water

instruction

teacher

must

the

no

process

the

example

. But

has

than

, but

learning in

" ) . Such

from

student

provide

to

the

; for

successfully

try

intermediate

number

and

giving

form

access

demanding

to what

this

hands

student

from

can

not

like

learned

experience

exactly

in

vide

different technique

her

be

, the

the

it

' s attempts

the

instruction

addition

to

accelerate horizontally

" ) , demonstrating

( " do

of

skill

very

what

concerning

two student

a model

to hands

again

perceptions

are

a very

Models

In

the

perform

straints

the the

provides

-

attempt your

movements

it . The

to

to

back

of

transmit

technique able

swim

and

vertical

not

imitate

the

front

constraining

teacher

to

is

the

attempt

could

be ( " move

a combination

useful

now

would

instructions

that

situation which

, time

the reor new

-

230

WithoutMiracles

demands areimposed(afterlearningto treadwater , the breaststroke is attempted ; afteradditionis learned , subtraction isintroduced ). Sucha view of learningis consistent with Russianpsychologist Lev Vygotsky 's (1896-1934)concept of the"zoneofproximaldevelopment " in whichthestudenttriesandeventually successfully masters newproblems that arebeyondherindependent capabilities butcanbelearnedwith the assistance ofa teacher .19 Notethattheteacher isnotatransmitter or instruc tor of informationor knowledge , butratheronewhoprovides supportto thestudent andarranges thelearning environment in suchawaythatsheis continuously challenged byproblems thatarejusta bit beyond hercurrent competence . In otherwords,theteacher arranges theenvironment sothat thestudent iscontinually encountering error,buterrorthatisnottoolarge, sothatanyreorganizing effortsaremorelikelyto besuccessful andsetthe stage for thenextintroduction of error.Thisviewisalsoconsistent withthe ideanowpopularin education that a successful teacher provideseduca tional"scaffolds " for students . These areplatforms thatprovidesupportin breaking downcomplexphysical andcognitive problems intomoreeasily mastered subproblems .2O All thisis applicable to thephysical skillof learningto swim, aswellasothermorecognitive skillssuchaslearningmathe matics , developing reading skills, andlearning to write.21 Buthow canwe accountfor knowledge that is acquired withoutany essential accompanying behavior , aswhena university studentis expected to learnbylisteningto lectures andstudying te:xtbooks ?Howcanpercep tualcontroltheoryhelpusto understand howthisispossible ? Herewewill haveto consider notjustonecontrolsystem asin figure8.1, butseethepersonasmadeupof acomplex hierarchy of controlsystems as described in chapter8 andillustrated in figure8.2. It will berecalled that thishierarchy hastwo principalfeatures . First, it is a hierarchy of percep tion. At the bottomlevel, perception is limitedto perceived intensityof stimulation of thesense organs - sounds canbeloudor soft, lightscanbe brightor dim. Butaswemoveupthe'hierarchy , moreandmorecomplex perceptions arepossible . Certaincombinations of intensities giveriseto particular sensations , for example , thetasteof orange juiceor thecolorred, andcertaincombinations of sensations resultin perceptions of configura tions,suchasthoseinvolved in seeing andrecognizing anappleor apencil. Perceived changes in configurations resultin theperception of transitions ,

232

Without Miracles

One way of explaining how we can plan and imaginecertain experiences is to suppose that " imagination connections " can exist in the control sys-

tem hierarchy betweenthe outputs at a given level and the input at the same level. When theseare activated, as indicated in figure 12.1, lower-level systems are in effect bypassedso that perceptionscan be madeto match refer.

encelevelswithout having to act on the environmentat all. Powersprovides an examplefrom playing chess: Supposeone is trying to find a good program- say a program for dealing with the next few moves in a chessgame. " If he moves his knight there, I'll move my rook here, but if he moveshis knight the other way, I'll take it with the pawn." Of course in chessone is not permitted to move the piecesfreely, nor would a good chess player (asI imagine him) babble to himself like this. Instead, he would simply imagine the relationships, not actually making the memory-derived imagesinto active referencesignals, but looking at them as if they had beenaccomplished. That is how a reference signal would look via the imagination connection - as if the lower -order

systemshad acted instantly and perfectly to make perception match the reference image .22

Purpose

Reference signal (r)

Com parator

. Sensor signal (s)

Error signal (e = r - s)

Imagination

Sensor

"

;If

Activator

SYSTEM ENVIRONM

/

ENT

~

Figure 12 .1 Imagination connection in a basic control system (after McClelland , 1991 ).

Education 233 Fromthis examplewe canrecallthat this hypothesized processof imagination is essentiallywhat we consideredin chapter 9 as thinking. But such thinking does not necessarilyyield the instantly perfect world to which Powersrefers. Indeed, the power of thinking liesin its ability to revealpro blems that would not have become apparent otherwise- realizing, for instance, that taking the knight with my pawn would exposemy queento dangerfrom my opponent's bishop, and that thereforeI had better comeup with another plan. Seenin this light, thinking becomesthe vicarious trial and-error-elimination processthat was presentedin chapter9, equivalentto Campbell's visually and mnemonically supportedthought. Thinking , then, becomesa way of discoveringproblems and attempting to solvethem without overt behavior. If the problem involves variablesfor which control systemsare already in place (for example, " what is 2 plus 2?" ), it can be solvedquickly and routinely. But if it doesnot permit the routine application of already existing control systems(for example, " what is the relationship betweenthe roots of two polynomials whose coefficients are reversed?" 23), reorganization as a form of blind variation and selective retention involving visually and/or mnemonically supported thought will be called on to solvethe problem (with , of course, no guaranteethat it will be solved).

We can thus conceptualizethe type of silent, covert learning that takes place in lecture and study halls as a type of internal problem solving that involves the reorganization of existing control systems. If I am taking a

courseon molecular evolution and want a good grade, I will haveto understand how DNA replicatesand how mutations can occur during the replication process. Right now, I perceivemyself asnot being ableto do this, and so there is error between a goal and a perception. Reorganization must therefore take place so that I can reducethis error. Attending lecturesand studying my textbook will facilitate this reorganization, not by transmitting to me the required information , but by enhancingthe production of new ideasconcerningDNA replication and mutation (blind variation) and helping me to eliminate my wrong ideas and retain the better ones {selective retention ).

Thus, a perceptualcontroltheoryview of thinking andthe reorganiza tion that must take place during the acquisition of new knowledge is consistentwith a selectionistview of knowledge processesin general and

234

Without Miracles

with education in particular. I should not give the impression, however, that such a view is widely known and acceptedin psychological and education circles. On the contrary, perceptual control theory is undoubtedly one of the best-kept secretsin psychology and education, since these fields continue to seebehavior and thinking as determined by a person's environment and have no adequateexplanation of how they are functions of the internal goals of the individual; that is, behavior and thinking are purposeful processes . Nonetheless, the selectionist perspectiveon learning offered by the theory is consistentwith both Campbell's insight into the ubiquity of cumulative blind variation and selectiveretention in all knowledge processes , as well as twentieth-century theories that emphasizeknowledge growth and reject the idea of education as the transmission of knowledge from teacherand textbook to student.

IV The Use of Selection

13

-

Evolutionary

Computing

SelectionWithin Silicon

When manwanted to fly, hefirstturned to a natural example - thebird- to develophisearlynotions ofhowtoaccomplish thisdifficulttask . Notable failures by Daedalus andnumerous bird-likecontraptions (ornithopters ) atfirstpointed inthe wrongdirection , buteventually , persistence andtheabstraction oftheappropriate knowledge (lift overanairfoil ) resulted in successful gliderandpowered flight.In contrast tothisexample , isn't it peculiar thatwhen manhastriedtobuildmachines to think,learn , andadapthehasignored andlargely continues to ignore oneof nature 'smostpowerful examples ofadaptation , genetics andnatural selection ? - DavidGoldberg1 Overthelastseveral decades thedigitalcomputer hasbecome animportant toolin agrowingnumberof humanactivities . Fromwritingschoolreports andmanaging household budgets to performing complex numerical analysesandsimulating astrophysical events , thecomputer quicklybecomes indispensable to anyone whotakesthetroubleto learnto useone.Its ability to store,manipulate , andanalyze largeamounts of dataandprovidestunningvisualdisplays thereofhavemadeit particularly usefulin all fieldsof scientific inquiry. Amongits manyscientific achievements , thecomputer hashelpedusto furtherour knowledge of biologicalevolutionandthecumulative blind variationandselection processes onwhichadaptive evolution depends . Perhapsevenmoresignificant is thatcomputers arenowbeingusedto solve extremely complexproblems in manyareasof science , engineering , and mathematics , generating solutions notimagined bythescientists whoposed theproblems andtranslated themintocomputer -readable form. In thischapterwe will surveysomeexcitingnewdevelopments at the intersection of computers andevolution . In doingso, wewill gaina better

238

Without Miracles

appreciation bothofthepower oftheevolutionary process andhowthis powercanbeharnessed tosolve complex problemsnotoverthecenturies

andmillennia ofgeological timeasinbiological evolution, butovermuch brieferperiodsof time.

Recreatingthe Processof Evolution

Darwinian evolution remains theaccepted scientific explanation fortheoriginanddesign ofallthelifeforms onourplanet; however, theslowness of

biological evolution makes itimpossible toobserve inthewaythatother

scientific phenomena canbestudied. Solareclipses canbeobserved from

startto finish; cannonballs canbedropped fromtowersandtheirdescents

timed; andtheentire lifecycles offruitflies, snakes, andrabbits canbestud-

ied.Butbiological evolution takes place oversuch longperiods oftimethat

itseffects cannotreadily beobserved during anyonehumanlifetime. 2 The

large timescale oforganic evolutionary change isnodoubt amajor reason foritsrejection andmisunderstanding bysomanyotherwise educated and

well-informed individuals.

Butcomputers havenowmade it possible to model theprocesses in-

volved inbiological evolution andtodothisquickly enough sothatthey

canbeobserved andstudied over periods oftime asshort ashours, minutes,

andevenseconds. Oneofthefirsttocreate suchsimulations wasOxford evolutionary zoologist Richard Dawkins whoreported hisresults inThe BlindWatchmaker, hispopular andcolorful introduction toevolution. 3

Dawkinss computer program ismade upoftwoparts. 4 Thefirstpro-

gram, called DEVELOPMENT produces onthecomputer screen a setof

treelike linedrawings called biomorphs, eachrepresenting anorganism. Thegeneration ofeachbiomorph isbased ona genome comprising nine genes, each ofwhich cantakeavalue from -9to+9.Forexample, onegene

controls width, sothata lowvalue forthisgenewould resultina narrow

biomorph andahigher value would produce awider one. Theremaining eight genes control other aspects ofthebiomorphs, such asheight andthe

number ofbranchings added tothecentral stem. Byspecifying values for each ofthegenes, DEVELOPMENT produces thecorresponding biomorph using thegene values asasortofgenetic recipe, notunlike thewayinwhich

realgenes direct thedevelopment ofliving organisms.

Evolutionary Computing 239

240

WithoutMiracles


necessary replication code from themore complex organisms .Butasthese parasites grew innumbers ,they began tocrowd outthehost organisms on which their own reproduction depended , and sothey began todieoffas well .This decline inthenumbers ofparasites ledinturntoanincrease in thepopulation ofhost creatures , causing anoscillatory cycle between the twotypes oforganisms inthesame way that cycles ofgrowth and decline between hosts and parasites orpredators and prey are observed innature . Butmore than justasimple moving pendulum between hosts and para sites was taking place .The hosts began toevolve characteristics that would make them resistant totheparasites , and theparasites found ways ofcircumventing these new defense systems . Anever -escalating evolutionary arms race was inprogress , thevery phenomenon thatisbelieved tohave provided thespringboard fortheincreasing adaptive complexity ofliving organisms over evolutionary time .8 Ray also found other types ofevolutionary processes occurring , such as forms ofcooperation among highly related organisms ,cheating ,and sexu alreproduction . Although hehad designed both the Tierran environment ched tothatofobserver once this and itsfirstinhabitant ,hisrole now swit Ancestor began toreproduce , mutate , andevolve . Butinstead ofobserving onlytheendproducts ofevolution ashehaddone during hismany years of fieldworkintherainforests ofCosta Rica , Raywasable toobserve the process ofevolution itself . Andthenewfieldofartificial lifewas finally on itswaytobecoming recognized bythescientific community asanimpor tantnewarena forresearch .9 TheworkofDawkins andRayonsimulated evolution arejusttwo examples ofthemany investigations ofartificial lifethathave been carried out.1o Their workwasamong thefirsttoshow howcomputers could be used to simulate thesame evolutionary processes responsible forallof earth 'slifeforms .Forthose skeptical thattheprocesses ofcumulative blind variation andselection canproduce thecomplex , adapted structures that make upliving organisms , research inartificial evolution reports striking evidence consistent withtheselectionist explanation proposed byDarwin fortheorigin andadaptive evolution ofspecies overa century ago . As McGill University biologist Graham Bellnoted , "Many people doubt that thetheory ofevolution islogically possible . . . .Now , one cansimply point totheoutput ofRay 'sprograms ; they aretheultimate demonstration ofthe

242

Without Miracles

logicalcoherence of evolution byselection ."11 Andalthough neither DawkinsnorRayinitiallydoubted thepowerof naturalselection , eventhey weresurprised to discover howeasy it wastomakeevolution happen ona computer oncethebasic processes ofvariation , selection , andreproduction weremodeled asprograms . Dawkins 'sandRay'sworkalsodemonstrated thecreative nature ofevolution.Even though theywereresponsible fordesigning thecomputer environment , algorithms , andoriginaldigitalorganisms , theorganisms that laterevolved andthecomplex interactions among themwerenotexplicitly programmed . Rather , theyemerged in bottom -upfashion asthecreative products oftheevolutionary process itself . Andif artificial evolution could beused tocreate complex linedrawings anddigitalcreatures abletosurvive andreproduce inachallenging silicon environment , it seemed reasonable to expect thatthesame process couldbeharnessed to solvepractical , realworldproblems , andperhaps tocreate intelligence itself. The ComputerCanKnow More thanthe Programmer

Attempts touseevolution -based techniques tofindsolutions tonovelproblemsandto develop artificialintelligence actually predate byquitesome timetheworkofDawkins andRay . In 1966Lawrence Fogel , AlvinOwens , andMichael Walsh published a slimvolume in whichtheydemonstrated howevolutionary processes implemented onacomputer couldbeusedto findsolutions toproblems involving prediction , diagnosis , pattern recogni tion, andcontrolsystem design . In theirprediction experiments , these researchers started outbycreating a single parent"machine " outof computer code . Thismachine wasmade upof a setof rulesrelating inputconditions to outputconditions , which wasusedto makethedesired prediction . Theprediction wouldbemade andevaluated , afterwhichtheparent machine wouldproduce anoffspring witha slight , random mutation . Theoffspring 'sprediction wouldthenbe evaluated andcompared withthatof theparent 's. If it wasbetter , theoffspring wouldbemutated to formthenextmachine ; if theparent 'spredic tionwasbetter , theoffspring wouldbediscarded andtheparent wouldbe permitted to reproduce again . Thisprocess wouldcontinue untilthepredictions werewithinadesired range of error , or untilsome predetermined

time or computinglimit was reached.


Thismethod canperhaps bebetterunderstood usingtheanalogy of preparing astew .Youstartwitharecipe , perhaps onefound inacookbook , orprovided byyourmother orother respected culinary artist .Aftermaking upapotofstew , youkeep it warmwhileyoumake upanother pot, randomly changing some partoftherecipe , perhaps adding more saltorcook ingit 20minutes longer .Youthencompare thetaste ofthetwopotsofstew anddiscard theone(andrecipe )youjudge inferior . If theoriginal stew was retained , another mutant stewwouldbeprepared , tasted , andevaluated , butif thesecond stew waspreferred it wouldbemutated again andcom pared withthenewresult .Byrepeating thisprocedure many times ,thestew should continue togetbetter untilanyfurther change wouldonlymake it worse (according toyourtaste , thatis). Of course , for someone withknowledge of foodsandcooking tech niques , theoutcome of arecipe willitselfsuggest certain changes - if the foodisundercooked , cookit longer ; if it istoospicy , cutbackonthechili peppers . Butin Fogel 'smethod , asin biological evolution , mutations are completely blindin thattheshortcomings of thecurrent computer code provide noinformation about howthecode should bechanged tomake it perform better . Soforthecooking analogy tohold , wehave toimagine a purenovice withabsolutely noknowledge ofcooking . Butthisprocedure isinonerespect quitedifferent frombiological evolu tion.Whereas natural selection typically operates onalarge population of organisms thatvaryin many different ways , thismethod compared only oneparent andasingle mutated offspring ateach step . Nonetheless , this approach toevolutionary computing used a(rather limited )source ofessen tialblindvariation that,whencoupled withcumulative selection , ledto foreseen or in anywayexplicitly pronew solutions thathadnotbeen grammed intothesimulation .Andalthough theirresults were quite modest anddidnothave amajor impact atthetimeonthefieldofartificial intelli gence , Fogel andhisassociates were among thefirsttodemonstrate anew , evolution -based wayof usingcomputers andto recognize itspotential almost 30years ago : entering

244

WithoutMiracles

Each

Evolutionary Computing

245

of shorthand fora computer program thatcanberunto provide atrial solution . These string -generated programs arethenevaluated , withthose mostsuccessful (forexample , thetop10%) retained andtheothers elimi nated . Theselected strings arethenallowed to pairoffand"mate " with each other , each mating resulting intwooffspring . Forexample , imagine tworelatively successful parent strings 10000 and01011 . If theposition between thesecond andthirdbitshadbeen randomly chosen asthecross overpoint , thetwoparent strings wouldbecutin thisposition , yielding 10-000and01-011.Bycombining thefirstpartofthefirstparent withthe second partofthesecond parent , oneoffspring would be10011 ; andcom bining thesecond partofthefirstparent withthefirstpartofthesecond parent wouldgenerate theotheroffspring 01000 . Occasionally , a bitor twoisrandomly flipped to simulate mutation ,13andeach stringof each newgeneration isagain evaluated .Thisprocess isrepeated again andagain , selecting thebeststrings andallowing them tomate andreproduce untila certain generation limitorerrorcriterion isreached .14 Holland wasbetting thattheprocesses ofcumulative blindvariation and selection thatovermany millions ofyears ledto remarkable instances of biological adaptation andknowledge couldbeexploited onthecomputer to evolve useful algorithms using thecomputer 'sspeed to compress hundreds ofgenerations intomere seconds .It turns outthathewonhisbet,ultimately finding genetic algorithms capable ofquickly andreliably evolving solutions toproblems ofdaunting complexity . Scientists andengineers the worldoverarenowusing these algorithms tosolve problems inmany different areas . General Electric applied them todesign gasandsteam turbines andto make thejetengines of thenewBoeing 777moreefficient . The Faceprints system developed atNewMexico State University allows awitness of acrime to evolve animage of a suspect 'sfaceusing a system not unlikeDawkins 's blindwatchmaker program described earlier . ThePre diction Company in Santa Fe,NewMexico , withthesupport of aU.S.based affiliate oftheSwiss BankCorporation , uses genetic algorithms to make predictions useful forcurrency trading . Genetic algorithms have also helped design fiberoptic telecommunication networks , detect enemy targets in infrared images , improve mining operations , andfacilitate geophysical surveys foroilexploration .ISEven personal computer spreadsheet users can nowfindsolutions tofinancial problems withgenetic algorithms .16

246

WithoutMiracles

But the evolution of evolutionary computation techniqueswas not to stop there. John Koza, a former student of Holland working at Stanford University, saw limitations in representingcomputer programs as unidimensional strings of zeros and ones. Sincemost complex computer programs are organized in a hierarchical fashion with higher-order routines making use of lower-order subroutines, he looked for a way of applying

cumulativeblind variation and selectiondirectly to hierarchicallystructured computer programs. He did this with a computer language called LISP, which is structured in such a way that it lends itself particularly well to the crossovertechniqueof sexualreproduction. Each LISP program is structured as a hierarchical tree composed of math -

ematical and logical operatorsand data. In the techniquehe namedgenetic programming , Koza would first generate a random population of such pro -

gram trees, including the operatorshe believedcould help solvea particular problem. Programs representedby the trees would then be run and their results evaluated. Most of theseprograms, sincethey were randomly generated, provided very poor solutions, but at least a few would always be better than others. As in the genetic algorithm technique, programs that achievedthe bestresultswould be mated, and offspring programswould be produced by swapping randomly chosenbranches- one eachfrom eachof the two parent trees- to produce two new program trees that were both

similar to and yet different from their parents (figure 13.1). Thesenew program trees would

be evaluated , and the process continued

until a program

providing a satisfactorysolution was found. Koza's 1992 book GeneticProgramming: On the Programming of Computers by Means of Natural Selectionand companion videotapeoffer a veritable tour de force of genetic programming applications .17These include , among others , the solution of algebraic equations , image compression ,

robot armcontrol, animalforagingbehavior , broombalancing , game-playing strategies, and backing up a truck to a loading dock. In applying ge-

neticprogrammingto theproblemof findingthe relationshipbetweenthe distanceof a planet from the sun and the time taken by the planet to complete one revolution around the sun, JohannesKepler's third law of planetary motion

was found , which relates the cube of the distance to the square

of the time. Particularly intriguing was that on its way to this conclusion,

thetechniquefounda lessaccuratesolutionto thisrelationshipthat wasthe

248

Without Miracles

sexually . Whereas such a technique would be hopelessly slow and ineffec tive if done manually , the recruitment of high -speed computers for this task has made it not only feasible but highly practical . Indeed , Koza showed that useful programs can be developed in this manner after as few as 19 generations . Among them are those that can generate crawling and walking behavior in simulated insects, perform aspects of natural language processing , make optimal bids and offers in commodity trading , perform finan cial analyses, control robots , generate art , and even produce bebop jazz melodies .19 Evolutionary computing techniques may turn out to be one of the most important

developments in computer science of the second half of the

twentieth century . The techniques are still quite young , but they are already beginning to find important commercial applications , and a number of regular scientific meetings, new journals , and electronic forums are avail able for disseminating the findings of research in this area.2o But the evolutionary approach to computing still encounters much resistance. The idea of using populations of randomly generated digital strings or LISP trees with trial -and-error elimination to write a computer program is so radically different from traditional approaches to software develop ment that such resistance is perhaps not surprising . As science writer Steven Levy recounts from an interview with Koza : Many traditional programmers, Koza explains, deplored the fact that a degree of chance was involved, both in the initial randomization and the probabilistic choicesof stringsthat survivedto the next generation. " Somepeoplejust broil at the idea of any algorithm working that way," saysKola . " And then you've got complaints that its sloppy. Becausethe first thing that comesto your mind when you hear about crossoveris, 'Oh, I can think of a casewhere crossoverwon't work .' Of course you can think of a case. The key to genetic algorithms [as well as natural selection] is the population, that thousandsof things are beingtried at once, most of which don't work out- but the idea is to find a few that do." 21

Indeed , many ofthecriticisms arenotunlike thearguments marshalled against biological evolution itself .Howcanacomplex , adapted system , be it aliving organism oracomputer program , possibly emerge fromchance andrandomness ?Howcansuch awasteful procedure , inwhich almost all oftheblindly generated entities intheinitialandsubsequent populations areeliminated , beeffective insolving acomplex problem , beit one imposed byaresearcher onacomputer program oroneimposed bynature ona

Evolutionary Computing 249 species? The power of nonrandom cumulative selection , of course, is the answer to these questions , and it works on the computer even better (or at

least a lot more quickly) than it doesin nature. And so the power of selection has begunto changethe role of the computer from that of a recipient of the human programmer's knowledge to that of a generator of new knowledge. Computers asWind Tunnels: Simulationsand Virtual Reality Captain Emerson was definitely ' on edge as he lined up the Boeing 747

jumbo jet on his final approach to runway 27. He had spent thousandsof hours in the air asan airline pilot , but this would be his first attempt to land the world 's largest commercial passengerplane. In many ways, the Boeing 747 handled much like the much smaller 737 he knew so well . But the view

of the runway from the cockpit situatedfour storiesabovethe landing gear was quite unlike that of any plane he had flown . Believingthat he still had severalmeters of altitude left when in reality he did not, the jumbo jet hit the runway hard , with a lot more force than even the monstrous landing

gear of the 747 could absorb. Not only ~ ould this be the captain's first landing of a 747 , it would also be his first crash landing .

Fortunately, however, Captain Emerson's error resultedin no injuries and no damageto the aircraft. For no actual aircraft or flying was involved. Rather, the captain had beentraining on a $16 million full -flight simulator, a pieceof equipment that in some respectsis actually more complex than the aircraft

whose behavior

it mimics . And so Emerson

was able to learn

from his error with no risk to life, limb, or equipment. He m~ght well make a few more rough landingson the simulator, but none would be asrough as the first , and most would

be better than the one before . When

his turn

would finally cometo pilot a real 747, his first landing would seemjust like

onemoreof themanyhehadalreadyperformedin simulation. The widespread use of flight simulators for pilot training is just one example of how computers are being applied to provide learning experi-

encesandresearchopportunitiesthat would otherwisebeprohibitivelyexpensive, time consuming, or risky. Although a certified flight simulator can cost many millions of dollars, many inexpensive simulation-based programs, including less sophisticatedversions of flight simulators, are now availablefor useon personalcomputers. For example, the popular SimCity


251

lem of powered flight relatively quickly, and the French went from one crash to another . As aeronautical

historian

Walter Vincenti

noted , " use of

vicarious trial , both experimental an~ analytical, was a strength of the Wright brothers in comparison with their French contemporaries.,'22In much the sameway, the computer now provides a meansby which homework assignments , hair styles, public transportation systems, and new supercomputer architectures (to name only a few examples) can be proposed, tested, and refined before producing a printed copy or real working model or prototype . This vicarious variation

and selection

has much in common

with

and

extendsthe use of the human brain to propose and test solutions to problems, as discussed in chapter 9. We saw there that important functions of

thinking are generating, testing, and selectingthought trials. Insteadof having repeatedly to rearrange all the furniture in his living room until an acceptablenew arrangementis found, the customerin the piano showroom can mentally generate, evaluate, reject, and finally selecta new furniture arrangementthat will accommodatehis new piano. But although the brain has quite good verbal, visual, and auditory pattern-generatingskills, it is lessadept at performing certain complex mathematical computations and logical reasoning. So by including thesemathematical and logical aspects in a simulated computer environment, a quite powerful combination of human intellectual and machine computational resourcesis formed for finding solutions to difficult problems. The ultimate developmentof this brain-computer collaboration can be seenin the creation of what is known as virtual reality. The goal of developers of virtual reality systemsgoes far beyond the small video displays and anemic loudspeakersfound on most computers by making it possible for users to experience a simulated environment in much the same way that they experience the real world . Instead of seeing a new design for a home or office building on blueprints , architect and client can instead don dis-

play goggles, step onto a steerabletreadmill, and walk through a virtual building. An ultrasound-generatedimage placed between physician and patient allows a surgeonto seeinside the patient to guide the initial critical moves of the scalpel. Molecular engineerscan now touch and manipulate atoms the size of tennis balls to create new molecular

structures

that may

have important scientific , engineering , and medical uses. Even virtual wind

252

Without Miracles

tunnels havebeendeveloped thatallowaerospace andautomotive engineersto entertheairstream virtually witha newvehicle designto examine

fromseveral perspectives howairflowspastit atdifferent speeds. 23 Dauntingtechnological challengesmust still be overcometo makevir-

tualreality a useful andaccessible tool, 24 butvirtual environments appear to betheultimate stepinfacilitating variation andselection processes for generating newknowledge to solvenewproblems. Avirtualreality-based architect caneliminate wallsandaddwindows witha sweep ofthehandto determine whether thechange should bekeptorundone. Thewings ofa newjetfightercanbemodifiedinstantlyto seeif somebettercombination

ofstability andmaneuverability canbeachieved. Andthesurgeon canrule out certaininterventions withouthavingfirstto subjectthepatientto exploratory surgery. Virtualrealityis stillin itsinfancy, butit seemsinevitable it willplayanincreasing rolein furthering ourknowledge and technology in manyfieldsdueto thewaythat suchvirtualenvironments

cangreatly facilitate andaccelerate thegeneration, variation, evaluation, andselection ofpotential solutions to complex problems. Neural Networks

Therelatively newfieldof artificial intelligence involves the creationof

computer hardware andsoftware designed tomimic andinsomewaysto surpasstheperceptual, thinking, andreasoning powersofthehumanbrain.

Although manyobservers of thisfieldhavebeendisappointed withthe resultsto date,therehasbeenconsiderable progress inworkthatinvolves computermodeling andsimulation oftheworkingsofthebrainandnerv-

oussystem. Based onourcurrentunderstanding ofthebrainasconsisting ofinterconnected networks ofbillions ofrelatively simple units(neurons),

developers of neuralnetworkscreatecomputersimulations of neurons

(sometimes referred toasneurodes) andconnect themupinvarious ways toformnetworks thatcanlearnandactintelligently insomeway. 25 Such networks areableto recognize human speech andwritten letters, play

gamessuchas backgammon, and determinewhethersonarechoeshave beenreflectedfrom undersearocks or mines. 26

Oneof the mostwidelyusedneuralnetworkarchitectures is knownas

thebackpropagation network (figure 13.2). Thisistypically a three-layer


253

Correct response pattern

Inputpattern

Backward error reduction

A:t!A!\OepJBMJO ;:l

~

~

Output response pattern

Figure 13.2 Backpropagation neural network (afterCaudill& Butler , 1992 ). networkwith a setof inputneurodes connected througha setof middle levelneurodes to a setof outputneurodes . It canproducecertaindesired responses (outputpatterns ) whenpresented withcertaininputpatterns . For example , theinputpatterns mightcorrespond to received sonarreflections of thetypeusedbysubmarine operators , andthedesired outputwould ~. be . a patternindicatingthepresence of eithera rockor a mine, provldmgill effectanautomatic andaccurate classification of theobjectbeingsensed . Fora neuralnetworkto behave in thisway, a particularpatternof connectionstrengths (analogous to thestrengths of synaptic connections betweenneurons ) mustfirstbefoundbetween eachneurode andtheneurodes in neighboring levels . With theconnection strengths amongtheneurodes

254

WithoutMiracles

initially

set

randomly

network

"

to

rock

"

the

dure

"

)

mine

used

.

"

to

a

into

the

-

an

the

weights

among

classify

correctly

new

ones

,

the

it

has

not

produce

is

of

particular

ence

selection

the

the

time

we

But

this

error

of

this

-

of

go

the

proce

-

discrepancy

synaptic

may

was

to

-

ac

to

the

network

it

cor

gives

the

learn

trained

,

classifying

to

but

also

sonar

instrument

sig

-

readings

characters

,

of

middle

and

,

new

to

design

is

.

input

know

actually

is

play

,

various

.

and

the

the

of

be

it

the

-

used

to

reduce

first

clear

from

part

of

an

the

com

,

-

not

network

in

.

the

And

such

pat

part

of

it

since

units

-

has

there

,

the

not

input

what

trained

on

-

will

generalize

of

num

important

the

to

the

Particularly

and

certain

on

differences

able

was

error

has

decide

number

and

is

network

.

few

optimum

trial

is

resulting

neural

important

not

which

differ

to

This

initially

network

too

but

the

the

output

that

blind

functioning

calculated

by

a

-

may

on

result

desired

change

must

in

subtle

patterns

and

.

training

neurodes

no

.

designer

system

beforehand

the

to

and

training

the

adaptive

network

error

during

encountered

of

neurodes

level

between

many

way

the

-

training

,

provided

the

and

manner

book

,

-

actual

Instead

and

instruction

First

-

the

trial

deterministic

approach

distinguish

too

to

no

the

medical

output

a

the

.

of

to

;

actual

this

trains

organization

able

of

changing

addition

no

.

pattern

the

instructionist

number

learned

tial

in

-

limitations

and

be

developer

seen

that

noteworthy

the

In

that

in

input

process

program

terns

' s

in

that

have

note

network

weights

next

example

puter

-

written

involved

network

synaptic

instructionist

is

are

neural

between

error

use

noting

which

from

recognize

to

backpropagation

modify

ber

,

interest

and

a

.

noise

(

made

. 27

variation

of

remove

speech

games

It

before

can

synthesized

board

is

experienced

networks

error

on

-

desired

weights

,

,

proce

strengths

This

and

either

random

are

iterations

,

the

patterns

at

the

connection

output

the

training

connection

.

many

output

reduce

set

a

with

the

the

input

,

error

After

desired

to

only

are

strengths

the

.

observing

and

neurodes

not

such

,

output

weights

compared

the

change

name

causes

as

However

changes

to

that

interpreted

connection

that

to

input

actual

the

level

be

.

is

to

its

will

wrong

pattern

network

providing

be

contributed

network

dure

nals

output

changes

most

input

connection

to

formula

that

the

that

initial

likely

and

to

pattern

the

mathematical

backpropagation

between

presented

output

each

,

neurodes

backward

is

most

whereby

pattern

cording

pattern

Since

are

output

those

a

an

outputs

is

rect

produce

or

initial

,

this

network

ini

-

255 Second, the correct correspondingoutput, that is, the proper classification for each encountered input pattern , must already be known to train

the network. A sonar-detecting backpropagation network cannot be used to distinguish rocks from mines if it is not already known which signals indicate

rocks and which

indicate mines . In other words , the instructionist

learning procedure used by backpropagation neural networks can work only if the instructor already knows all the right answersfor the training inputs. So this processcannot be used to discover new knowledge or develop new skills, but is a way of transferring knowledgefrom one sourceto another .

Third , backpropagation neural networks cannot always be trusted to find just the right combination of connectionweights to make correct classifications becauseof the possibility of being trapped in local minima. The instruction procedure is one in which error is continually reduced, like walking down a hill , but there is no guarantee that continuous walking downhill

will take you to the bottom

of the hill . Instead , you may find your -

self in a valley or small depressionpart way down the hill that requiresyou to climb up again before you can complete your descent. Similarly, if a backpropagationneural network finds itself in a local minimum of error, it may be unable to climb out and find an adequatesolution that minimizes the errors of its classifications

. In such cases, one may have to start over

again with a new set of initial random connection weights, making this a selectionistprocedureof blind variation and selection. One might also use other proceduressuch as adding " momentum" to the learning procedure, analogousto the way a skier can ski uphill for a while after having attained enough speedfirst going downhill , or adding noise to the procedure to escape such traps , again a form of variation and selection .

Finally, for those looking to researchon neural networks to further our understandingof how real neurons work in real brains, backpropagation neural networks are not biologically plausible. Unlike the neurodes and their interconnectionsin a backpropagation network that conduct signals in both directions (one way for responding, the other direction for learning), biological neural pathwaysconduct signalsin one direction only. Also, much if not all human learning occurs without the direct instruction and the

patient training of the type backpropagationnetworks require. Soalthough

256

WithoutMiracles

these

networks

their

can

and

Given

that

the

at

to

should

For

the

to

early

and

also

been

and

his

.

to

col

These

solve

re

learning

one

-

as

backpropagation

and

system

others

-

difficult

reinforcement

of

,

, 31

surprising

1980s

learn

as

example

cart

be

have

Barto

could

characteristic

movable

not

Andrew

in

referred

. 3O

a

by

networks

learning

hinged

it

adaptive

. 28

networks

Massachusetts

networks

pole

,

,

their

systems

neural

developed

process

supervised

applications

both

neural

designing

neural

selectionist

the

instructionist

a

problems

learned

solved

to

spatial

bal

-

learning

. 32

Another

ter

type

network

of

tive

,

Like

rode

is

to

the

each

the

effect

the

input

In

a

of

the

addition

that

is

ones

did

only

one

tions

of

their

so

previously

middle

a

specific

they

.

patterns

word

to

many

.

Teuvo

net

trials

will

An

exam

input

-

layer

of

layer

middle

But

-

output

pattern

is

- layer

each

middle

weights

that

con

-

neu

-

- layer

that

mediates

best

neurodes

matches

the

of

output

a

and

all

did

simulated

can

or

extensive

the

best

to

nearly

- fit

learn

than

to

that

all

work

-

adjust

closely

so

-

neural

with

more

categories

to

who

inhibit

allowed

even

network

useful

respond

this

,

layer

signal

such

,

all

arranged

while

neurodes

input

,

are

neurodes

result

and

Kohonen

the

middle

pattern

- layer

- layer

the

meaningful

neurode

middle

middle

on

the

task

the

.

competi

.

resulting

)

neighboring

The

match

After

into

- layer

having

the

signals

that

pattern

or

-

three

useful

the

every

random

the

.

a

,

of

instructor

such

fil

consists

make

an

in

it

competitive

middle

to

layer

input

excites

away

the

a

.

among

among

,

sends

neurode

the

,

of

For

input

( initially

Those

neurode

farther

.

the

active

connections

active

sending

input

.

help

and

each

in

by

most

competition

weights

Since

of

pattern

neurodes

activity

.

as

layer

the

words

extent

the

known

learns

word

pattern

are

highly

those

some

input

network

without

neurode

input

,

a

layer

different

pattern

but

the

The

detected

every

to

has

.

spoken

the

middle

and

activated

neurode

,

of

is

network

-

layer

recognize

sounds

on

to

network

backpropagation

patterns

to

the

activity

typical

output

input

learning

nected

neural

neurodes

the

to

is

the

and

responses

encodes

selectionist

interconnected

layer

ple

of

.

groups

they

of

interesting

on

biological

were

that

a

to

ance

rize

first

of

limitations

training

to

the

number

instructionst

University

by

opposed

ting

of

of

a

to

demonstrated

problems

their

applicability

Some

the

searchers

quite

impose

approaches

. 29

other

in

does

limitations

developed

ing

their

selectionist

leagues

useful

nature

flexibility

so

be

instructionist

catego

one

-

and

pronuncia

with

-

such

Evolutionary Computing 257 competitive filter networks, used such a systemto create one of the first voice typewriters .33 It is not difficult to see the evolutionary processes of variation , selection

basedon competition, and retention operating in such a network on the connection weights betweenneurodes. The initial distribution of weights for the middle-layer neurodesis usually determinedrandomly. Then competition occurs among theseneurodes, with the most active one winning and selectedto be modified to match the original input pattern evenbetter. But because this modification is directed toward a better matching of the just -received input pattern , there is no guarantee that this modification of

weights will be reta.ined, as more input patterns are received and further rounds of activation , competii :ion , selection, and modification

of connec-

tion weights take place. Another classof neural networks known asadaptiveresonancenetworks also use processes of variation , competition , and selection . For them , . . . the basic mode of operation is one of hypothesistesting. The input pattern is passedto the upper layer, which attempts to recognizeit . The upper layer makesa guess about the category this bottom -up pattern belongs in and sends it , in the guise

of the top-down pattern, to the lower layer. The result is then comparedto the original pattern ; if the guess is correct

. . . the bottom -up trial pattern

and the top -down

guess mutually reinforce each other and all is well . If the guess is incorrect . . . the

upper layer will make another guess. . . . thus, the upper layer forms a " hypothesis" of the correct categoryfor eachinput pattern; this hypothesisis then testedby sending it back down to the lower layer to seeif a correct match has beenmade. A good match results in a validated hypothesis; a poor match resultsin a new hypothesis.34

Gerald Edelman, whose neural Darwinism was discussedin chapter 5, has also been actively involved with his associatesin applying Darwinian selection to neural networks to simulate the adaptive functioning of the brain.35Using what he calls syntheticneural modeling, he and his associates havecreateda seriesof neural simulationsnamedafter Darwin that explore and demonstratethe principles of neuronal selection. Darwin III is a computer simulation of a sessile(seated) creature possessinga head with a movable seeingeyeand a jointed arm that has both touch and kinesthetic sensors.36Inhabiting a computer world of simulated stationary and moving objects of various shapes and textures , Darwin III learns to grasp certain

objects and repel others. Although certain initial biasesare built into Darwin III , for example, a preferenceto graspround, smooth objectsand reject

258

Without Miracles

square, bumpy ones, the actual perceptualcategoriesand behaviorsare not programmed. Rather, Darwin III learnsto fixate visually and track objects, distinguish different types of objects, and grasp somewhile rejecting others through Darwinian selectionof connection weights betweenthe neurodes, in much the same way it appears that the human

brain

is able to learn

through the variation and selectionof synaptic connectionsamong its neurons . With

the recent creation

of Darwin

IV , Edelman

and his associates

have moved beyond computer simulations and created a working robot that performs the tasks of its predecessors in a real-world environment of

physical objects.3? Beforeleaving our discussionof neural networks, let us return briefly to the instructionist networks describedat the beginningof this section. It will be recalled

that these systems of simulated

neurons

can indeed grow

in

adapted complexity through an instructionist processby which synaptic weights are modified in the proper directions without variation and selec-

tion . This being the case, however, we must wonder how it is that these network architecturesand training procedurescameto be, beingthemselves examplesof adapted complexity and thereby posing additional puzzlesof fit . Unlessthey were discoveredin some novel manner completely unlike that responsiblefor other scientific and technologicalinnovations, we must suspectthat they owe their designto repeatedcyclesof blind variation and selection, not in the working of the networks themselvesbut in repeated overt and vicarious (cognitive ) generation of the networks and their subse-

quent testing by scientists. And since, aswe saw earlier in this chapter, computers can now simulate

such selection

processes , we should

not be too

surprised to learn that recent work in genetic programming has shown that the computer itself can be used to evolve neural networks to solve problems and perform various tasks.38

It can be arguedthat the digital computer is the most important technological tool of the twentieth century, but it is quite often viewed as a sophisticated electronic combination

of file cabinet , calculator , and typewriter .

Most computersare indeed still usedin theseways. But we are now beginning to seeexciting new applications that go far beyond their standard office and scientific functions. Computers have made it possible to model the evolutionary processitself and thereby convincingly demonstrate the


cumulative tive

power

retention

allowing

, and

prove

their

techniques One

safely

of

today begin

approach

, only

and

years

software

logical power

are

even

processes increasingly

first

storage

discovered

within

them

that

from

that

silicon

among - based

time

can

over answer continue

ever - increasing

audiovisual carbon machines

- based

life

the

. Will

bred

machines will

ways

corn

are

they

run

eventually

designers

? Or

of computer near the

the

last

. But their

, the forms

fu -

in

, and

memory

capabilities

.

be

anything

achieved

themselves

acquiring , and

human

im -

using

lead

that

potential

achieving

only

the

, eco -

.

robots

and

,

, test , and

flowers

of their

has

invent problems

understanding

selec -

rapidly

ultimately

even

evolutionary

as computers

, thereby

, data

the

will

, ornamental

surpass to

solutions to

programs

and environments

programming this

perhaps

and

evolution

questions that

evolution , speed

? Will

organic

a certainty

all

faster

prevent

that

genetic

and

variation

complex

and

intelligence

trial

programmed

where

chickens

blind

real - world

many

exceedingly

that

limits

of

simulate

test

being

wonder

much

perhaps

? These

almost

are

programs

undiscovered

complexity

to

to

ways

evolve

and

certain ware

and

to

engineers

but

the

processes

used

algorithms

help

to

be

. They

genetic

to

combined

solutions

computers

bred on

and

own

cannot

analogous

the can

scientists

nomically

ture

of

. They

259

four it

billion seems techno

, processing evolutionary will

-

adapted

now own

will hard

be

used

-

14 The Artificial

Selection of

lSandMolecules Organism

It is wonderful what the principle of selectionby man, that is the picking out of individuals with any desiredquality, and breeding from them, and again picking out, can do. Even breedershave beenastoundedat their own results. . . . Man, by his power of accumulatingvariations, adaptsliving beingsto his wants- may besaid to make the wool of one sheepgood for carpets, of another for cloth, & c. - CharlesDarwin1 Ever since the appearance of the first life forms on our planet , organisms have influenced each other in their evolution and resulting adaptations . They compete for sunlight , food , shelter, and mates, with the most successful passing down the accumulated knowledge of their genomes to the next generation . Speciesinvolved in parasite -host or predator -prey relationships may evolve sophisticated offensive and defensive equipment and behaviors in a continuing evolutionary arms race. Other species have come to depend on each other for survival as symbiotic relationships evolved , such as when flowering plants offer sweet nectar to insects in return for the insects' dissemination of the plants ' pollen . Although relative latecomers to this scene, humans driven by their con tinuing and increasing need for food , shelter, clothing , fuel , beauty, and amusement , have had a particularly

dramatic effect on the evolution of

some species. In this chapter we will see how that influence predates by many thousands of years our knowledge of evolution , and how our current understanding of natural selection coupled with advances in biotechnology have provided us with unprecedented power to create new and useful organisms and molecules .

262

WithoutMiracles

TheHumanSelection of PlantsandAnimals Formostof its existence , our species livedmuchlike othermammals , roaming its habitat and obtainingfood by hunting, scavenging , and gathering edibleleaves , nuts, seeds , fruits, and tubers. But this nomadicway of life beganto changeashumansdiscovered theadvantages of takingmoredirect control over the growth and breedingof plants and animals. The first domesticationof food plants probably took placebetween10,000 and 13,000 B.C. in southeastAsia wherecropssuchas rice and beanswere plantedand harvested . Two othersiteswhereagricultureappearsto have developed independently weretheFertileCrescent , whichincludespartsof present -dayIraq, Iran, Turkey, Lebanon , Israel, andJordan, andsouthcentral Mexico. In theFertileCrescent animalswerealsodomesticated , suchas thehorse, donkey,camel,andsheep .2 This changein occupationfrom hunting and gatheringto cultivating food cropsand raisinglivestockhad importantconsequences for human cultural evolution. In contrastto moving from one temporarycampto another , agriculturepermittedtheestablishment of permanentvillagesand ultimatelycities, city-states , andempires . Increasing agriculturalproductivity freeda significantportionof thepopulationfrom thedemandsof growing food and allowedthemto take up scientific, technological , religious, andartisticendeavors . But in additionto thesesweepingandcomparative ly rapid culturalchanges , anothersloweronewastaking place- we were beginning to directtheevolutionof increasing numbersof plantandanimal . speCIes . No understanding of evolution or genetics was necessary to beginthis transformation of useful organisms ; neither was a conscious and purpose -

ful selectionof plants and animals. That is becausethe practice of agriculture itself necessarilyimposesnew environmental conditions on plants and animals. Crops are planted at certain times of the year using specificmethods of cultivation . As in natural selection, plants that by chanceare better suited to these human -made conditions grow well and produce more seeds

for the next planting. For example, most wild plants produce seedsthat fali to the ground and are dispersedat maturity, but a few hold on to their seeds,

. facilitatingtheir gathering. By selectingand then SOWIng more seedsfrom

the latter, a selectionpressurewas created toward the evolution of plants with so-called non-shattering seedsthat were easierto harvest.

The Artificial Selectionof Organisms and Molecules

263

But although such early artificial selectionmay have well been accomplished unintentionally, even these prehistoric peoples no doubt noticed that, among living plants and animals, like begetslike. Horses give birth to horses, not ducks. From the seedsof a fig tree more fig trees sprout, not palm trees. And human couples reproduce children who bear an obvious resemblanceto their parents. So as these people chose seedsto plant the next season's crop from the largest and most productive sterns of wheat, barley, corn, or rice, and allowed the largest, strongest, besttasting, or gentlest horses, cattle, or camelsto mate, they were effectively exploiting the mechanismsof cumulative blind variation and selectionto produce plants and livestock that were better and better adaptedto human req uiremen

ts .

Indeed, it was the observation that domesticatedplants and animals changedunder the selectionpressuresimposedby humansthat led Charles Darwin to his theory of natural selection. The first chapter of the Origin is entitled

" Variation

Under Domestication

, " and here Darwin

presents many

examplesof how domesticatedplants and animals changed over time to become better adapted to human needs. He explains these changes as " man 's power of accumulative selection : nature gives successivevariations ; man adds them up in certain directions be said to make for himself

useful to him . In this sense he may

useful breeds . " 3 When

he realized

that such

selection pressures also exist in nature without meddling by human agents, his theory was born .

The theory made explicit the principles of evolution that agriculturists the world over had beenunwittingly using for millennia. For someobscure reason nature servesup variations, and only certain ones are selectedby humans for breeding the next generation. These ancient farmers had no way to control the amount or direction of these naturally occurring varia tions , but by eliminating the undesirable and breeding from the desirable , they could control just about any observable or measurable characteristic

of plant or animal. The amazingsuccessof early plant breederswas demonstratedby Native Americans who over the course of 4000 years transformed a stingy grass

into one of the world 's most productive food crops. When Europeansfirst set foot on the shores of the New World

264

Without Miracles

Adapted maize [corn ] cultivars extended fromthesouthern partof South America tothenorthshore oftheSt. Lawrence River ; fromsealeveltoelevations of3,355m (11,000ft.). Types included flint, flour,pod,andpopcorn aswellasred,blue , black , yellow , white , andvariegated kernels . There isnodoubtaboutthecompetence in plantbreeding of theAmerican Indians , asit tookthedevelopment of F1hybrid maize ofmodern genetics toexceed theperformance capability ofIndian maize . It is understandable thatEuropean settlers grewtorespect themaize cropanditsdevel opers , formaize issaidtobethegreatest giftfromtheIndians .4 But althoughNativeAmericansand otherearlyagriculturistsweresuccessfulin breedingplantsandanimalssuitedto their conditionsandneeds , it wasnot until the insightsof Darwin into evolutionandthoseof Gregor Mendelconcerningthe geneticbasisof inheritancethat plant and animal breedingcould beginto beput on a firm scientificfoundation. This foundation permittedrapid advances in breedingtechniquesbeginningin the 1900s. Sinceplant and animalbreedingis a form of adaptivebiologicalevolution, it dependson thethreecomponents of variation, selection , andreproduction. Advancesthereforeinvolveoneor moreof thesecomponents . By increasingvariation, theprobabilityof findinga variantwith desiredcharacteristicsis increased . Breedersdo this with techniquessuchas raising largenumbersof plants or animals, cross-breedingdifferentvarietiesto producenew hybrids, and usingirradiationand chemicalsto increasethe occurrence of mutations. Other techniquesthat improvethe accuracyand efficiencyof selection includeproceduresinvolvingphysicaland chemicalcomparisons coupled with statisticaland screeningmethodsto eliminateunsuitablespecimens quickly. For example , to developa newvarietyof wheatthat is resistantto highlevelsof salinity, highlevelsof saltareappliedto thesoil. Thiswill kill almostall of theaffectedwheatplants, but almostcertainlya fewplantswill survive. Thesesurvivorscan then be usedto breeda line of salt-tolerant plants. Screening (thebreeder 's termfor selecting certainindividualsfrom a largepopulationof specimens ) is now possiblein the testtubewhereit is referredto as in vitro selection . Individualcellsfrom plantsmay be subjectedto certainchemicaltoxins. Thosethat areableto grow despite - the presenceof a toxin are mutantsthat are naturally resistant . Using this method,cornwith 10to 100timesmoreresistance to certainherbicides has beendeveloped .s

265

TheArtificialSelection of Organisms andMolecules Advances Once

in molecular

a particular

characteristic

to make

inserted

gene

cutting

in half olant

the

netic

Other

that

sexual

mammal

clones

bull

desirable

inseminate vitro bull birth

have

using

fertilization and their allow

animals

kernels

such as high of only

with

asexual

and

animals

whether

this

tech -

oil content

with

yield

, thus

a newly

. With

other

have

varieties

or standability plants

that

. Ge -

retain

the

little

cannot

bovine

of eggs from transplantation

into

particularly siblings

cloning

many

surrogate

desirable

many

gen , no

cows

than

mother

. And

the sperm cows

to produce

permits he could

techniques

with

couple

over

useful

(at this writing

insemination

love -making

for

sexual -

it permits

variation

more

cow

reali -

to reproduce

of normally , and

be avoided

), artificial

a desired

. The

led to attempts

or no genetic

produced

reproduction

all the genes necessary

propagation

is called

reproduction yet been

contains

organism

animals

traditional

, all of them

or animal

the genes (or

high

those

to enhance

or animal

genes to impregnate

a single

with

cross -bred

.

a desired

plant

characteristic

being

selection

cells . This and

to be preserved . Where

with

the

plants

breeding

.

single plants

until

to determine

genetic

have been developed

using

with

required

of a new , identical

ly reproducing

on

is associated

of immature

characteristics

oil content

that

an impact

can do it by examining

is now

the rapid

had

to wait

to produce

corn

each cell of a plant

organisms

erations

this

allows

the production

varieties

normally

cause corn

methods

. They

has the desired

desirable

gene for high

zation

time

, and

other

also

have

) of the cells

or animal

screening

longer

selection

markers

identified

having

no

their

A nique ., genes that been

have

gene has been identified

, breeders

matures

hred

biology

a

ever the

in

of a desired

for gestation a large

herd

and of

.

GeneticEngineering

These twentieth -centuryadvances thatpermitmorecontroloverthevariation, selection , andreproduction of livingorganisms greatlyfacilitated the breeding of domesticated plantsandanimals . However , noneofthese methodsmakesit possible to producethedesired variations directly . Variation canbe increased ; moreaccurate andquickerselection methodscanbe employed ; andreproduction of desired varieties canbeincreased . Butthe breeder is still limitedto selecting amongthosechance variations thatthe organisms themselves provide .

266

WithoutMiracles

To gain more direct control over desired characteristics , breeders had to

await developmentsof the secondhalf of this century in the relatively new field of molecular biology. After Watson and Crick's discoveryof the structure of the DNA molecule in 1953 and the subsequentbreaking of the geneticcodeby which sequences ofnucleotides orchestratethe construction of proteins, it was only a matter of time before scientistsdevelopedmethods to reach deepinto the center of living cells and manipulate their genes. Many different techniques are now employed in genetic engineering, which is also known as genesplicing and recombinant DNA technology.6 They all involve manipulating an organism's genes. One techniqueinvolves introducing a genefrom a plant or animal cell into a bacterium. Many bacteria have tiny rings of DNA , known as plasmids, that are more accessible and easierto manipulate than the more tightly packed genesin the chromosomes. When a gene that produces a desired protein is removed from a

plant or animal cell and spliced into a bacterium's plasmid, the genew~ll continue to produce its designatedprotein product in its new bacterial setting. Provided with a nutrient-rich environment, the bacterium quickly grows and divides, producing more cells that also contain the foreign gene and that therefore also produce the desired protein ~which can then he h,t;lr -

vestedfrom thesebacterial chemicalfactories. With this technique, bacteria have been genetically engineeredto produce insulin, human growth hormone, and the anticancer drug interferon, among thousandsof other important

substances .

But what about introducing foreign genes into the chromosomes of plants and animals to alter their characteristicsand to create brand new organisms, such as the fire-breathing chimera of Greek mythology, which combined the head of a lion , the body of a goat, and the tail of a snake? Unfortunately, manipulating chromosomal DNA is a more difficult affair. For certain plants , the aid of Agrobacterium tumefaciens is enlisted . This

soil bacterium has beendoing geneticengineeringof its own for millions of years by inserting genesfrom its plasmids into the chromosomesof plant cells. The foreign DNA causesthe plant cells to form tumors that produce unusual compounds (opines ) that serve as food for the bacteria . Thus ,

this bacterium is a naturally occurring genetic engineerthat has evolved the ability to alter plant cellsto provide food for itself. This ability has been exploited by human geneticengineerswho can now splicedesiredgenes(for


267

example , onethat providesresistance to a pathogenic virus) into Agrobacterium 'splasmid , whichinsertsit intoaplantcell'schromosomal DNA. Thetargeted plantcellis thencoaxedinto developing intoa complete fertile plantthatwill passon theengineered DNA andaccompanying viral resistance to itsprogeny . Formanipulating thegenes of animalsandthoseplantsfor whichthis technique , rice, and ... will not work, suchasthemajorfoodcropsof wheat com, a messenger virusis used , or thedesired geneis directlyinjected into thenucleus of a plantcellor into a one-cellanimalembryo . In thecaseof animals , thegenetically alteredembryos areimplanted intotheuteriof surrogatemothers , andtheresulting transgenic animals canbeusedto breed moreprogenyby traditionalmethods , with eachoffspringcontaining the alteredgene . Thefirst transgenic animalwasproduced in thismannerin 1981whena rabbitgenewasinserted intoa mouse embryo . Althoughgeneticengineering is a veryyoungfield, it hasalreadyhad someimpacton foodproduction andpromises to havemuchmorein the future.Abouthalfthecheese produced in theUnitedStates usesanenzyme created by bacteriacontaining a cowgene . Similarly , recombinant bovine somatotropin (rBST)is grownin bacteria , isolated , andinjected intodairy cows,increasing milkproduction upto 20%. TheFlavrSavrtomato,genet icallyengineered to stayfreshlongerbeforespoiling , madeits appearance in American supermarkets in May 1994. Andextensive research is under waybyfirmssuchasCalgene , Monsanto , andDuPontto genetically engineercropsresistant to herbicides , harmfulinsects , viruses , bacteria , and fungi. Whereas genetic engineering nowprovides means for directlymanipulat inggenes , thesetechniques havenoteliminated theuseof cumulative trialand-errorresearch fromthedesign of newandvaluable organisms . Firstof all, to knowwhichgeneto insertintoanorganism , thegene 'sfunctionmust beknown. Onewayto determine thefunctionof a particulargeneis to expose a largenumberof organisms to amutagen , select thosethatdifferin someinteresting wayfromnormalorganisms , andthenestablish through DNA sequencing whichgenewasmutated . If, for example , a mutatedbac teriumis unableto replicate itsDNA, andthegenes thatwerechanged by mutationcanbedetermined , it will beknownthatthese genes mustin some waybeinvolvedin DNA replication . Thismethodallowsno controlover

268

Without Miracles

what mutations willresult (thus they remain blind variations), andorgan-

ismsareconsequently selected onthebasisofsome particular characteris-

tics.Thistechnique canbeeffective forlarge populations ofsingle-cell or-

ganisms whosegenomes arerelatively small(F.colicontains about3000

genes initsgenome), butit islesssoforlarger organisms withlarger

genomes (thecommon fruitflyhas20,000genesandthemouseabout

200,000). Forthese organisms, researchers traditionally relied onthestudy

ofnaturally occurring mutations, buta newtechnique allows themtoma-

nipulatethe genesof their choice.

Currently applied tomice, with which weshare more than90percent of

ourgenes, targeted genereplacement (alsocalled homologous recombination),developed byItalian-born Mario Capecchi attheUniversity ofUtah

School ofMedicine, makes itpossible tocreate mice with amutation inany gene ofinterest. 7 Butthisrelatively precise genetic control isstilllargely a hit-and-miss process since itissuccessful inonly avery small percentage of

treated cells. Itismuch more likely thattheintroduced gene willeither not beincorporated intothecellschromosomes, orthatitwillbe,butatthe

wrong location. Byincorporating twoadditional marker genes intotheforeigngene,onethatprovides resistance toa certain drugandtheotherthat provides sensitivity toanother drug,itispossible toscreen thetreated cells

using thetwodrugs andeasily findtheonecellinamillion whose genome hadbeen altered inthedesired way. Only these cells areselected andinjectedintoamouse embryo, eventually leading totransgenic mice with thepre-

cisedesired mutation. Themiceareexamined forphysical orbehavioral

abnormalities thatprovide clues tothefunction ofthealtered gene. Soonce again weseethataselectionist screening procedure isnecessary toseparate successfully genetically engineered cells from theothers. Similar screening methods employing antibody orenzyme detection areapplied inmany of

thegenetic engineering techniques described earlier. 8

Buteventheidentification ofthefunction ofindividual genes doesnot

permitoneeasily to createneworganisms withdesired characteristics for

thesimple reason thatmany ofthemost important traits arecontrolled by

combinations ofhundreds orperhaps even thousands ofgenes. Itisthere-

forenecessary totestvarious combinations ofgenes todetermine which one

provides thedesired characteristic, such asearly maturity, high oilcontent,

height, ordrought resistance, while maintaining otherbeneficial traits.Itis

269


for thisreason thatdespite therapidprogress in genetic engineering , much moreresearch mustbedonebeforeplantscanbegrownthatcombine the advantages of bothcornandsoybeans , or animals arecreated thatbothgive milklikecowsandgrowwoollikesheep . TheEvolutionof DrugsandMethodsof DrugDesign Regardless

of

plethora tion

of does

stand

its

not

fashion

certain

vastly

outnumbered

saries

using

generated

discovery

portant

times

pain

, the

liver

to

do

and

the

, refinement

, and

a

of

rates

fight

, we

back

of

a

evolu

-

now

under

-

and

various

organisms

human as

of

much

our

and

have

hosts

,

and

larger

popula

. But

although

mutation

against

knowledge

us

these

infecting

us , such

high

microbial

-

adver

technology

iodine

agents

methods

could .

malady

recovered

another

person

animals

. Such as

willow approach functioning

) to

have

Perhaps after

certain

quinine

to

trees

to

to

that

ingesting

treat

the

it

has

drug treatments .

to

and and ,

coupled

, inevitably

the

fever

by

an

individual

aspirin

substance they

with

to of

- like

. However

resulted

or . Or

discovery

,

ignorance in

many

consuming ( which

early

users

be

trial a

, and

tried

have

in

of

the

ineffective

it

on

from useful

of

-

certain

observed

suffering

certain

lack

- and

with

may

of

, these

crudest

substance the

relieve

physiology

the

-:

since

to

sponges these

only

im

used

Chinese

and

appeared

the

most

been

since

food illness

they

led

malaria

research body

ingesting

that

same

when

pain

, the

. But

our

opium

chemistry

a certain the

is

have

Arabs

developed noted

from

Drugs

worms

goiter of

been

probably

treat

drug of

treat

healers

plants

methods

for

drugs

prescribing

- century

understanding

suffering

eating

health

no

.

Romans oil

twelfth

content

of

battle and

castor

, and

virtually

administration

continuing Greeks

taking

anemia

high

other

this the

Egyptians

had

error

in , with

treat

drugs

such

and

intelligence

weapon

ancient

and

. Indeed

. Many

over

presence biological

between

for

, and

can

that

.

The

have

viruses

advantages

, we

our

continued

organisms

methods

reproduction

the

reminder

competition

, and

effective

, prolific

,

constant

Darwinian

, bacteria

important

sizes

a

adapted

as

remarkably

have

is

perfectly

diseases parasites

evolved

achievements

illnesses

infectious

pathogenic

tion

remarkable

human

the

a

ill

drugs bark

of

systematic

structure and

and harmful

270

WithoutMiracles

TheBeginning ofScientific Drug Development Important advances indrug development began about 1800 , although they toooftenresulted fromserendipity rather thansystematic scientific research . Inthe1790s English physician Edward Jenner heard thatdairy maids whocontracted therelatively mildcowpox disease seldom suffered fromthemuch more serious andoften fatalsmallpox . Consequently , he suspected thatcowpox somehow provided immunity against smallpox , and developed asmallpox vaccine fromthepusofcowpox sores . Inthe1870s French chemist Louis Pasteur observed thatchickens inadvertently injected withweakened cholera bacteria developed immunity to thisdisease ; he later developed vaccines against anthrax andrabies . Andin1928 , Scottish bacteriologist SirAlexander Fleming observed thatamoldthatacciden tallycontaminated aculture ofStaphylococcus appeared tostopthebac terium 'sgrowth , andthereby discovered penicillin .9 Thedevelopment ofthese andother modern drugs wasgreatly facili tated byknowledge ofthecauses ofdisease . Itwas onlythrough hisaware ness thatinfectious diseases were caused bymicroorganisms thatFleming was able torecognize theimportance oftheantibiotic produced bythepeni cillium mold . Recent discoveries ofthekeyroleofenzymes andvarious receptors incellular activity permitted important advances inthedevelop ment ofawide range ofnewdrugs . Finding Drugs byLarge -Scale Random Screening Drugs exerttheireffect byproviding molecules thatareableto fit and attach themselves toother molecules inthebody , notunlike thewaythatan antibody fitsanantigen , asdiscussed in chapter 4. Forexample , angio tensin -converting enzyme (ACE ) actsasacatalyst toconvert angiotensin I toangiotensin II, thelatterbeing avasoconstrictor , which reduces thediam eterofblood vessels , thereby increasing blood pressure . Thedruglisinopril hasadistinctive molecular shape thattightly fitsandeffectively plugs upthe active siteofACE , thereby inhibiting itsactivity inconverting angiotensin I tothevasoconstricting angiotensin II. Theneteffect ofthisdrugistherefore to reduce levels of angiotensin II, resulting in lowered bloodpressure in hypertensive patients . From themolecular perspective , finding aneffective drugcanbelikened tofinding akeyforalock.Inwhatisnowreferred toastheclassic method

TheArtificialSelectionof Organismsand Molecules 271

ofmodern drugdesign, asmanyastensofthousands ofsynthetic andnaturalsubstances mayberandomly testedingroupsof50to 100aspotential

keystoadesired locksuchasACE. Ifitisfoundthatoneofthesesubstances bindswiththetargetsite,furtherscreening is doneto isolatetheparticular

keylike compound, whichisreferred toastheleadcompound ormolecule. Oncea leadcompound isfound,muchmoreworkremainsto bedonein the formof additionalmodifications andmoretests.Variationsof the lead molecule arethensynthesized andtested.Someof thesevariationsmaybe lesseffectivethan the originalmolecule,and othersmay be muchmore

potent.Anexample ofthelatterisetorphine, a variantofmorphine thatis 1000timesmorepotentthanthemorphine. Butthisis onlythebeginning. Theabilityof themolecule anditsvariations to reachthetargetsite(its

bioavailability) mustbetestedusinglivingcellsandtissuecultures. Then thetoxicity ofthedruganditssideeffects mustbedetermined intrialswith animalsand humans.The objectof allthesemodificationsand testingis to

findthecompound thatismosteffective whilehaving theleasttoxicity and fewestunwantedside effects.Modern drug developmentinvolvesmany

sophisticated techniques forzeroing inonsuchcompounds, 1 andtheentire processisclearlyoneofa ratherlengthycumulative-variation-and-selection

procedure. Because ofthelargenumber ofcompounds tobescreened, followedbyrequired rigorous testingoncells,tissues, animals, andfinally humans, it maytakemorethan10yearsof intensive research effortand manymillions ofdollars todevelop a newdrugandbringit toyourmedicinecabinet.Thisprovidessomeexplanation for the highpriceof many drugs.

Nonetheless, theinitialinvitrochemical screening oflargenumbers of moleculesis muchmoreefficient(andsafer)than havingfirstto test all

thesecompounds onanimalsor humans,particularly sincepharmaceutical firmsemploytechnology that largelyautomatesthe randomscreening

process. Indeed, large-scale molecular screening canbeseenasatypeofvicariousblindvariationand selectionin whichmanytypesof moleculesare

tested,andonlythosethat arefoundto havean affinityforthe target enzyme orreceptor areretained foradditional testinganddevelopment. Butthismethodof molecularselectionhas an importantlimitationthat becomes obviouswhencompared withthecumulative variationandselec-

tionof biological evolution (aswellastheevolutionary computing tech-

niques considered in thepreceding chapter). Namely, onlynaturaland

272

Without Miracles

synthetic compounds thatarealreadyonhandcanbeinitiallyscreened , as thereisnowayto createspontaneously newvariations of successful moleculesin thewaythatmutations andsexualgenetic recombination provide newvariationsin biologicalevolution . In otherwords, the desireddrug molecule mustalreadybeprovidedby theresearchers , andno stepin the selection process canautomatically fine-tunethosemolecules that show somefit to thetargetmolecule . Anysubsequent finetuningmustbecarried out painstakingly by chemists who attemptto determine thereasons for the selected molecule 's success , andapplytheir knowledge of chemistry to improve it further through repeated rounds of variation and selective . screemng . Structure -Based RationalDrugDesign It shouldalsobenotedthatthisclassic methodof drugdevelopment does not requireinitial knowledge of the actualstructureof the interacting molecules . Butsincemoderntechnologies involvingX-raycrystallography andnuclearmagnetic resonance spectroscopy canprovidedetailedinformationabouttheatomicstructure of manymolecules , thisinformationis now beingusedin whatis calledrationaldrugdesign . If theclassicapproachto drugdesign canbelikenedto findinga keyamongtensof thousandsof keysthatwill fit a particularlock, structure -basedrationaldrug design is analogous to havinginformation abouttheshape of thetumblers insidethe lock. Sincesuchknowledge wouldfacilitatethemakingof a workingkey, informationon thestructureof a targetsitefor a drugcan facilitatethefindingof aneffective drugthatwill bindto thetarget. Butevenwithdetailed structural information , designing aneffective drug isnotasstraightforward asonemightexpect . A drugmolecule cannotsimply begroundintoshape in thewaythata keycanbemadefroma blank. Instead it mustbeassembled fromconstituent atomsthatwill themselves fit togetheronly in certaincombinations andarrangements . To complicate mattersfurther , a molecule 'sconfiguration maychange dramatically when broughtintoproximitywith anothermolecule , andatomiccharges (which arenot easilymodeled ) canaffectthebindingof onemolecule to another . Sowhereas knowledge of thetarget 'sstructure usefullyconstrains thenumberof candidate drugmolecules , aconsiderable amountof trialanderroris

TheArtificial Selection of Organisms andMolecules 273

required tofinda goodfit.Thissearch isnowfacilitated bythree-dimensionalcomputer displays ofthetargetmolecule thatallowresearchers to

design molecules atom byatom tofitthesimulated target site,inmuch the samewaythatwindtunnels andothercomputer-simulated environments facilitate thedesignof newproducts asnotedpreviously. Butsincecom-

puter models arenotperfect, promising compounds stillhave tobetriedin

a testtubeandthenoncells,tissues, liveanimals, andfinally humans in clinical trials.Eachstageinthislongprocess represents yetanother selective

filter, withonlythose molecules thatpassevery testfinally finding theirway

ontopharmacists shelves. Thecumulative variation andselection involved

inthisprocess isapparent inthedescription provided bythreepioneers of structure-based drugdesign oftheirattempts tocomeupwitha compound thatwould enhance theeffect ofcertain anticancer andantiviral agents and behelpfulin treatingautoimmune disorders: Thisiterative strategyincluding repeated modeling, synthesis andstructural

analysesled ustoahandful ofhighly potent compounds thattested wellinwhole

cellsandinanimals. Hada compound encountered difficulty inthecellular oranimaltests(such astrouble passing through cellmembranes), wewould haverevisit-

edthecomputer tocorrect thedeficiency. Then wewould have cycled a modified drugthroughthe circuitagain. 11

Recognizing bothclassic andstructure-based drugdesign astypes ofevolutionary processes involving cumulative variation andselection helps usto

understand thepotential advantages ofthelatterapproach overtheformer. Intheclassic method, thousands ofcompounds arefirstrandomly tested, andthesuccess ofonecompound provides littleifanyinformation onthe

potential usefulness ofothers stilltobescreened. Variation andselection occur tobesure,butthereisnocumulative variation andselection inthe initialscreening procedure, sothatevery oneofthousands ofcompounds mustbetested.Intherational, structure-based approach, information con-

cerning thestructure ofthetarget siteisusedtoconstrain thepossible can-

didates. Soinsteadofscreening thousands ofcompounds, theresearchers

quoted above hadtoprepare onlyabout60substances tofindapromising

drug,thereby saving considerable time,effort, andmoney. Inaddition, using thecomputer asatypeofvirtual testtubefortesting drugs allows a vicarious meansofvariation andselection thatcanbemoreefficient and cost-effective thanconductingallinitialscreenings chemically.

274

Without Miracles

Directed Molecular Evolution : Selection inaTest Tube Butisit reallynecessary to knowthestructure of thetarget siteto producedrugs efficiently ?Biological evolution hascome upwithstunning achievements ofdesign without anysuch knowledge atall. Ofcourse , evo lutionhashadoverthree andahalfbillionyears . Butif molecules them selves couldsomehow bebredusingverylargepopulations andrapid iterations of cumulative blindvariation andselection , it should thenbe possible todirect theevolution ofuseful molecules fordrugs andother purposes inmuch thesame waythatbreeders traditionally directed theevolu tionofdomesticated plants andanimals . Thefirstdemonstration thatanevolutionary approach to molecular design wasindeed possible wasprovided in theearly1960s by Sol Spiegelman of theUniversity of IllinoisatUrbana -Champaign .12Using a technique thatcauses strands of RNA(themolecular messenger of the genetic information archived inDNA)toreplicate witharather higherror rate , Spiegelman began to breed a huge number of variations of a particular RNAmolecule . Forhisselection criterion , hechose arather simple one - speed ofreplication . Since heprovided progressively shorter periods oftimeforreplication ,molecules thatwere thequickest atmaking copies of themselves were more likelytobeselected asparents forthenextgeneration of molecules . After74generations of this"serialtransfer " experiment , Spiegelman hadbredanRNAmolecule thatwas83%different fromthe original ancestor molecule andreplicated itself15times faster .Theartificial evolution ofmolecules inatesttubehadbeen achieved . Variations ofthistechnique , referred toasdirected molecular evolution , arebeginning to beused to design molecules fordrugs andotheruses . Instead ofselecting self -replicating molecules thatreplicate mostquickly ,a typeofmolecular obstacle course issetupthatinvolves binding toatarget molecule . A varied population of manymillions of molecules is passed through a'filtration column thatcontains thetarget molecule . Since thisinitialpopulation isarandom collection ofmolecules , onlyaverysmall percentage arelikelytobindtothetarget molecule . Butsince thepopulation is extremely large andvaried , even anexceedingly small percentage of"hits" isvirtually guaranteed toproduce atleast some molecules withanaffinity forthetarget , andthevastmajority thatdonotstickaresimply washed down thedrain(awatery version ofDarwin 'shammer ). Onlytherelatively

TheArtificialSelection of Organisms andMolecules

275

276

Without Miracles

Ittherefore seems quite likely thatbefore long anewgeneration ofpow-

erfuldrugs andothersubstances willbeavailable thatevolved inthelaboratoryundertheselective guidance ofhuman researchers. Butincontrast

totherational design ofmolecules, theresearcher will(atleast initially)

havenorealunderstanding ofwhya particular molecule issosuccessful at

doing whatit does, inthesame waythattraditional plantandanimal

breeders know nothing ofthegenes underlying thedesirable characteristics thattheyselect. It isofcourse likely thatsuccessfully evolved molecules

will beanalyzed andperhaps even improved byrational methods ofdesign. Butthepotential power ofcumulative blind variation andselection using

large,heterogeneous populations ofmolecules issuchthatrational fine-

tuning maynotbenecessary. Obviously impressed, Nobel prize-winning biochemist Manfred Eigen callsdirected molecular evolution the future of

biotechnology.

Three different technological revolutions have nowtaken place. Inthenine-

teenth century theindustrial revolution involved theexploitation ofhuge

amounts ofenergy formanufacturing, agriculture, andtransportation. The information revolution ofthemid-twentieth century provided telecommu-

nications andcomputer hardware andsoftware to generate, analyze,

manipulate, andtransmit vastamounts ofinformation. Themostrecent revolution, thatinvolving biotechnology, nowprovides themeans to

manipulate thevery core ofliving cells anddirect thecourse ofbiological evolution formany species from bacteria tohumans. Ithasalso given usthe ability todevelop newdrugs, vaccines, andgene therapies tofight disease

andimprove thelengthandquality ofhumanlife.

Itwasargued inchapter 10thatalltechnological development isdepen-

dent oncumulative blind variation andselection. Butwhat isparticularly

intriguing aboutrecent developments intheinformation andbiotechnologydomains isthatthese fields arenowmaking explicit useofartificial evo-

lution. Thesequence ofevents leading tothisdevelopment isnoteworthy:

evolution ofthehuman brainbycumulative natural selection; theuseof thisbrains evolved capacity forthought (itself a formofvicarious cumulativevariation andselection) tofashion anunderstanding ofthenature and power ofnatural selection; andtheapplication ofthisknowledge tosolve problems byexploiting forms ofartificial selection to direct evolution in agricultural plots,barns,computers, andtesttubes.

TheArtificialSelection of Organisms andMolecules

277

.

ultimatelypOlson su per -

V The Universality of Selection

15 From Providence

Through Instruction to Selection A Well-Traveled Road

Looking backintothehistory ofbiology, itappears thatwherever a phenomenon resembles learning, aninstructive theorywasfirstproposed toaccount fortheunderlying mechanisms. Inevery case,thiswaslaterreplaced bya selective theory. Thusthespecies werethought tohavedeveloped bylearning orbyadaptation of individuals to theenvironment, untilDarwinshowedthisto havebeena selective

process. Resistance ofbacteria toantibacterial agents wasthought tobeacquired by

adaptation, untilLuriaandDelbruck showed themechanism to bea selective one. Adaptive enzymes wereshownbyMonod andhisschoolto beinducible enzymes

arising through theselection ofpre-existing genes. Finally, antibody formation that wasthought tobebased oninstruction bytheantigen isnowfoundtoresult from theselection ofalready existing patterns. Itthusremains tobeaskediflearning by

thecentral nervous system mightnotalsobea selective process; i.e.,perhaps learningis not learningeither.

Niels

Jerne 1

It isnevertheless worthnotingthatinthehistoryofideasinstructivehypotheses

havemostoftenpreceded selective hypotheses. When Jean-Baptiste Lamarck tried tofoundhistheory ofdescendence ona plausible biological mechanism, heproposedtheheredityofacquired characteristics, a tenetthatadvances ingenetics would eventually destroy. Onehadtowaitalmost halfa century before theideaof selection wasproposed byCharles Darwin andAlfred Wallace andvalidated the principle, ifnotallthedetails ofitsapplication. Inthesamewaythefirsttheories abouttheproduction of antibodies wereoriginally basedoninstructive models

beforeselective mechanisms replaced them.It couldconceivably bethesamefor theories of learning. Tounderstand thereasonsforthistemporal succession, we mustobviously examine thefunctioning of thescientists brain.Aninstructive

concept consists ofonlyonestep.It isthesimplest possible approach. Moreover, whetherwelikeit or not,it contains an egocentriccomponent. Naturedirects

formsmuch asthesculptor models clayintoastatue....Theconcept ofselection, ontheotherhand,implies further reflection. Itinvolves twosteps, anditsatisfies the

quest foramaterial mechanism totally devoid ofintentional aspects. Itisnatural thatthismorecomplicated procedure, moredifficult to execute, shouldhavesys-

tematically appeared insecond placethroughout thehistory ofscientific thought. Jean-Pierre

Changeux 2

282

Without Miracles

. . . chance Pure

alone

chance

evolution

: this

possible

or

esis , the

only

tific

one , in

now to

such

one

of

we

in

reflection tive

of in

how

the

biosphere

one

among

hypoth

. . . . There

the

of

other

conceivable

is no

anthropocentrism

from

.

edifice

than

intensely

-

scien

-

this

teleonomic

are , an

that of

the the

of theory

the

adapted

it

has

fields

are

evident

adapted

. Certainly the

been of

finally

, insofar

provided

now

these

is

by

another

growth

, if

we

the

the

was

find

also in

~ ..-

basic

advanced

in

adapted

is

com

quite

the

-

strik

-

growth

and

' s knowledge , are provided complexity

an

skill

extension

that

of

its

, and

by

to

of power be

the

a supernatural

. Although

modern

everyday

also

. From

of

liv

-

had

a crea

-

organisms

is

a

this

products entity

of , Paley

design

these

that

function

. Similarly

forms

the

eloquently assume

functional

life

intelligence

our so

and

creator

the

under for

naturally design

these

considered

of Paley

we

of in

of

explanation

intricate

complexity

, skill

progression

Reverend

a watch the

and

this

reasoned

. As

and

conclusion design

of

first

works

intelligence

watches

knowledge of

of

examples

organisms world

of

creator

, as

and

many

to

have

evolution

, it origins

characterized

system

origin

2 , the

a maker

adapted

, like

knowledge their

be

one

obvious

the

clearest

living

the

to

of

theories

of

chapter

examination leads

to

fit

across

the

had

to

of

concerning

quite

for

Theology

due

fields

Fit

the

have

, organisms

recipients dential

one saw

Natural

organisms and

longer

fact

of

in

stupendous

.

of

must

of

patterns

on

of

reasoned

,

the

sole

tested

protest

considered

be

similarity

provides

complexity

tor

no the

destructive

selectionist

account

Perspectives

adapted

ing

and

creation

of

today

different

already

itself

understanding

watch

is

is

conclusions are

similar

. As

many

that

to

standing

the

biology

instinctive

puzzles

-

another

Changing

reasoned

so

book

knowledge

Biology

root

. It

, more

an

, of all

very

observed

sciences

general

should

this

. But

watch

the

knowledge

domains

human

the

modern

with

constitute

remarkably

the

of

some

and

these

-

squares

surveyed

forms

of

plexity

of

innovation , at

hypotheses

so arouses are .

conclusion

all

ing

any

every

blind

concept

that

entertain

complexity

theme

of

but

Monod3

Having

as

source free

conceivable

no other that we

Jacques

time

the

central

even

concept

one , and creatures -

is at

, absolutely

perspec

and in

biological

this

-

passive provi science

-

From Providence Through Instruction to Selection

283

in its quest for naturalistic explanations no longer considersthis to be a

viabletheory, it is probablystill the mostpopularlyheldbeliefconcerning the origin of the biosphereamong nonscientists, even in a technologically advanced country

such as the United States .

But it eventuallybecameclear to somethat living organismswere not the finished, final product of an omniscient and omnipotent creator. Instead, the accumulating fossil evidenceindicated that the biosphere has undergone- and is still undergoing- dramatic changes, with somespeciesdying out, new ones appearing, and others undergoing extensivemodifications. Lamarck's explanation for these changesdid not explicitly reject God's graces, but addedanother mechanismby which organismswould be modified advantageouslyby their interactionswith the environment in ways that were passeddown directly to their descendants . Lamarck 's theory of evolution had some important

advantages over

Paley's purely providential account of the designof life forms. It took into account the fact that speciesare not immutable but rather change over time .

It also attempted to provide a naturalistic theory of evolution that did not rely exclusively on the omniscienceand omnipotence of a supernatural being . But Lamarck 's instructionist theory was inadequate to the task of

explaining biological evolution, as it provided no explanation for either why the modifications undergoneby organismsin their interaction with the environment should be adaptive (why should exercise make a muscle stronger or cold weather make a mammal 's fur grow thicker ?), or how these

modifications could be transmitted to the next generation. The theory was an important step in moving from a supernatural to a naturalistic understanding of adaptive evolutionary change, but it did not go far enough in that it still required processesthat were " smart," and the origin of this smartness was itself unaccounted for .

Natural selectionprovided an alternative explanation that removed all smartnessfrom the process. Organisms varied, but why, Darwin did not pretend to know. By absolutely blind and ignorant luck, certain variations were better adapted to their environment, giving them a competitive edge over other organismsand allowing them to leave behind more offspring. Theseoffspring, being more like their parents than not (again, Darwin did not know why), would share the traits and reproductive successof their parents, but with continued variations in their own offspring.

284

WithoutMiracles

Darwin's accountrequiredno divineprovidence . Neitherdid it have magical instructions fromtheenvironment tellingorganisms howto adapt andhowto passonthisinformation to theirprogeny . Darwinhit onatheory that, althoughincomplete because of his understandable ignorance of genetics andmolecularbiology , providedthe first explanation for how adapted complexity couldemerge onitsownwithnooutside guidinghand or mysterious environmental instructions . Butdespitehis insightthat so stunned theworldof science , hewasunwillingto rejectcompletely either theprovidentialism of Paleyor theinstructionism of Lamarck . Ashestated in thefirsteditionof theOriginandrepeatedly emphasized in latereditions , "I amconvinced thatnaturalselection hasbeenthemainbutnottheexclusivemeans of modification ."4Wewill seelaterthathealsodidnotrejecta providential account of life'sinitialemergence . Andsoit wasleft to theyoungerandmoreradicalultraselectionists , in particularAugustWeismann , to asserttowardtheendof thenineteenth centurythatnaturalselection wasthesoleprocess bywhichspecies grewin adapted complexity . Andnowmorethan100yearslater,thispurelyselec tionistviewof theemergence of designhassofar withstoodall challenges (includingsomequiterecentonesto beconsidered in thenextchapter ) and continues to bethefoundation for modernbiology . Asrecounted in chapter 4, a remarkably similarsequence of theoryevolution, fromprovidential throughinstructionist to selectionist , canbeseen in thefieldof immunology . Ehrlich's originalside-chaintheoryof antibody production wasprovidential notin thesense of callingonGodasanactive creatorof antibodies , butratherbecause it assumed thattheinformation necessary to produceall possible antibodies wasalreadycontained in the genome .5Butthistheorywascastinto doubtwhenit wasrealized thatthe immunesystem couldproduce antigens thatwereableto matchandthere byrecognize asforeigncompletely novelmicroscopic invaders . As in biology , thefirst nonprovidential theoryproposed for thefunctionaldesignof antibodies wasan instructionist onewith a definiteLamarckianflavor. Thetemplate theoryproposed that antibodies obtained information for theirproduction fromtheirenvironment , thatis, fromthe antigens with whichtheycarnein contact . But afterabout20 yearsit became apparent thatthetemplate theorycouldnotaccount for certainkey findings , suchasthatlater-produced antigens wereusuallymoreeffective in bindingwith anantigenthanearlierones .

From Providence Through Instruction

to Selection

285

To explain these findings, Jerne proposed a selectionisttheory of antibody production that in its basic conceptualization is still acceptedtoday. The building plans for antibodies are not all specifiedin the genome. Nor are antibodies created from instructions or templates obligingly provided by the invading antigens. Instead, new antibodies having remarkable fit to a given antigen are produced through the random generation of a large population of varied antibodies, with those having the best fit selectedfor additional

rounds of blind variation

and selection .

So as in evolutionary biology, immunology proceededthrough the same stagesto explain its puzzleof fit . The formulation and acceptanceof a selectionist theory of antibody production was of particular significancesinceit provided the first clear demonstrationthat the samebasicprocessof cumulative blind variation and selectionthat occurs over eons of phylogenetic time among organisms also occurs within organisms during their much shorter life spans .

Although selectionist theories of the achievementof fit have become mainstays of evolutionary biology and immunology , they have not been

well acceptedinto other fields seekingexplanationsfor the puzzlesof fit in animal and human learning, human thought, scientific progress, and cultural adaptation. In philosophy, the selectionistperspectiveis well representedby the work of Popper. But Popper's emphasison human fallibilism and his evolutionary epistemology (which seesall knowledge growth as never-ending cycles of conjectures and refutations) is not very popular today. A minority of philosophers seehis work as an important advance, but mainstream philosophy seemsmore attached to providential and instructionist theories of the origins of knowledge. An example of current providentialism is the linguistically influenced innatism of Chomsky and Fodor , and instructionism survives in the continuing efforts to wrestle with

the problem of induction in attempts to explain how our limited sensory experiencescan instruct us with justified, certainly true beliefs about the wor ld .

Much the same state of affairs can be seen in the cognitive and social

sciences . Despite the work of Campbell and a handful of others, the view that perception and cognition constitute forms of substitute trial -and-error -

elimination is not widely held amongpsychologistsand cognitive scientists. The same appears generally true of sociology, anthropology, and other fields concernedwith understanding the dynamics of human culture and

286

Without Miracles

socialinstitutions. Soselectionist accounts ofcognition andculture certainlyexist,butmainstream currents inthesefieldsstillseemto vacillate between theprovidentialism ofinnatist theories (asinE.0. Wilsons socio-

biology 6) andtheinstructionism ofmuchlearning theory, which seesthe

environment in control of behavior.

Selection theory isdoing somewhat better intheneurosciences, partly

duetotheinfluence ofChangeux 7 andtoEdelmans neural Darwinism. But therecontinues tobemuchresistance toselectionist theories ofbraindevelopment andfunctioning, perhaps at leastpartlyattributable tothedifficultyofEdelmans writing 8 andpossibly resentment against anoutsider from thefieldofimmunology attracting somuchattention intheneurosciences. 9

Thelargest successes ofselection theory outside biology, immunology, andagriculture havebeeninthedesign ofcomputer programs andmolecules.Butevenhere,particularly in computer science, muchresistance remains. Formany, resorting tothecreative powers ofignorant, cumulative variation andselection asimplemented ingenetic algorithms andprogramming isanadmission ofprogramming incompetence, forcing onetorelyon

thecomputation muscle offast,parallel processors to finda brute-force

solution intheformofa hacked program thatmaydothejob,butisinelegantandperhaps even incomprehensible totheprogrammers eye. The Rejectionof Selection

Whyareselectionist theories ofdesign solittleknown andusedinthesci-

entific community outside ofevolutionary biology andimmunology toexplainpuzzles offitandthegrowth ofknowledge? There appear tobea

number ofreasons, some ofwhich werementioned inthediscussion ofbiological evolution inchapter 2.Butnowhere hasmoredebate arisen thanin

theattempt toapply ittothegrowth ofhuman knowledge infields suchas thephilosophy andhistory ofscience, psychology, andcultural change. And soitisbacktothese fields wewillturntoconsider thearguments against a selectionist account oftheontogenetic growth ofknowledge. TheWastefulness andImprobability ofSelection

Someofthearguments against a within-organism selectionist account of

human knowledge growth aremuch thesame asarguments thatwere (and

FromProvidenceThroughInstructionto Selection

287

stillare)putforward against naturalselection asanaccount oftheadapted products oforganic evolution. Some ofthemareprimarily basedonreligiousoraesthetic considerations. Wesawinchapter 9 thatPiaget rejected bothDarwinsselectionist accountof evolution and anyapplication of selection theoryto thecognitive domain, citingthealarmingwaste and

the fruitlesstrials acceptingsucha theorywouldentail.Indeed,Piagets

concerns werewarranted, evenif hisconclusions werenot.Theprocessof cumulativeblind variationand selectionis exceedinglywastefulsince

almostallvariations, ignorantas blindprobings mustbe,aredeadends

(literally soinorganic evolution). Butthiswastefulness canbeunderstood astheunavoidable pricethatmustbepaidfora processfromwhichnew

adaptations andknowledge canemerge without miraculous outside providentialistinsightsor mysteriousinstructionist guidance. That we tend to forgetor ignorethe manyfailuresin our attemptsto

betterunderstandand controlour surroundings and remember onlythe successes makesa selectiontheoryepistemology appearunnecessary and

unappealing. Thisis muchthesameas considering onlythelivingand adaptedendresultsof biological evolution andignoring thecountless unadaptive variations. Butunlikebiological evolution wherethefailed organisms andextinctspecies areusually wellhiddenfromview,wecan take noticeof the fruitlesstrialsof our experimentsand conjectures. An informativecase from the historyof technological innovationis

ThomasAlvaEdison two-yearattemptto findan appropriate substance

forthefilament in thefirstelectric lightbulb.Aftertryingoutdozensof substances, including redhairfromamansbeard,Edison finally foundsuccessin December1879usingcarbonizedsewingthread.His oft-quoted

statement thatgeniusis1%inspiration and99%perspiration reflects the

longhours andcountless failures thataccompanied thisandhismany other inventions and technological advances.

Butwhenwebuya newproduct,whetherit bea newvideocameraor a

moreeffective laundrydetergent, wetakeno noticeof the manyfailed

attempts thatpreceded itsdevelopment. Wealsoknownothing ofthe countlesswould-beinventorsand scientistswho do not producenote-

worthy breakthroughs. Weusually consider onlythesuccesses andnotthe failures, making itappear asifscience andtechnology progress through the sheerbrilliance andinsightofscientists andinventors ratherthanthrough

288

WithoutMiracles

painstaking trial and error and only occasionalexhilaration of trial and success. Walter Vincenti , whose account of advances in aeronautical engineer-

ing was introduced in chapter 10, remarks: From

outside

or in retrospect , the entire process tends to seem more ordered

and

intentional- lessblind- than it usually is. It is difficult to learn what goeson in even the consciousminds of others, and we all prefer to rememberour rational achievements and forget the fumblings and ideasthat didn't work out.1o

Another argumentagainstselectiontheory stressesthe improbability of a blind, unintelligent processcoming up with useful solutions to complex problems. This sameargument was used repeatedlyagainst natural selection in the evolution of species. It is of course true that any single blind change to a working system is almost certainly not going to make that sys-

tem function better. A random wiring change inside the computer I am using to record thesewords is quite unlikely to make it perform better and much more likely to lead to a trip to the repair shop. However, iterative cyclesof blind variation and selectionbasedon large populations of such variations is quite another matter . The recent impressive successesof selec-

tionist approachesto the designof both computer programs and molecules may convince some of the more open-minded skepticsof selectiontheory that such a processworking cumulatively on populations can effectively tame the improbability of a blind processgeneratinga fitter solution. Examining Variation, Selection, and Transmission But not all who argue against selectionist accounts of the growth of knowl -

edge are offended by the wastefulnessof variation and selection, nor do they all doubt its power, at least not in biological evolution. Rather, some insist that there are important differencesbetweenthe biological evolution of adaptedcomplexity and the growth of knowledge as manifestedin both the progress of scienceand the cognitive development of the individual human brain . These arguments focus on the three principal themes of varia tion , selection , and transmission

.

Probably the most frequently raised objection against the selectionist view of knowledge growth takes the form of an argument against the view that the variations are generatedblindly as to their ultimate success . Other equally unflattering adjectivessuch as " unjustified," " unforesighted," "non-prescient, " " undirected," " haphazard," "random," "groping,"


289

" stupid," and " dumb" have also been used to describe these necessary variations, but the word blind perhaps makes the essentialpoint most clearly. The argument against blindnessin the variations of thought and theory leading to advancesin human knowledge is that, although in biological evolution the generation of genetic variations may indeed be blind, the growth of human knowledgeis consciouslyand purposely directedtoward finding solutions to specificproblems. As University of Waterloo (Canada) philosopher of scienceand cognitive scientistPaul Thagard proposed: Whereasgeneticvariationin organisms is not inducedby the environmental conditionsin whichthe individualis strugglingto survive, scientificinnovationsaredesignedby their creatorsto solverecognized problems ; theythereforearecorrelated with solutionsto problems. . . Scientists alsocommonlyseeknewhypotheses that will correcterrorin their previoustrials. . .11 But, we must ask, how doesthe fact that scientistshave purposes(which few would doubt) provide emancipation from the necessityof blind variations in pushing back the frontiers of knowledge?The fact that a young scientist may be spending almost all of her waking hours in pursuit of room-temperature superconductivity does not, unfortunately, provide her with any clairvoyance as to the desiredsolution, if one does exist. Stating that thesevariations are " correlated with solutions to problems" begsthe questionasto how suchprior guiding knowledgemight havebeenachieved in the first place. Our scientist, unlike the processof organic evolution, most certainly does have a definite goal, and she generatesmethodological and theoretical variations in an attempt to accomplish this goal. But to the extent that new discoveriesare made for which prior knowledge did not exist, this growth of scientific or technological knowledge is possibleonly by producing and testing new experimentalvariations whose outcomesare unknown until tested.12As Campbell put it, " rather than foresightedvariation, hindsighted selectionis the secretof rational innovation." 13 But we must be careful to make clear what is meant by blind in this context. First, blindnessdoesnot imply that all variations are equally probable. For this reason, the word random is probably not a suitabledescriptor since to some it may carry that connotation. Second, blindness does not mean that the processof producing variations of ideas, theories, and experiments for testing is necessarilyunconstrained. Our superconductivity-seekingscientist is not likely to throw just anything into her concoction of chemicals,

290

WithoutMiracles

suchas someof last night's leftover soup. Instead, shewill rationally tryout those substances in those proportions

and under those conditions that ,

based on her knowledge of previous research and current theory, she believes have the greatest chance of success .

Soit cannot be deniedthat previously achievedknowledgehas an important role to play in constraining the variations to be investigated. Nonetheless, the new concoction

is still a blind

variation

in the sense that

the

scientist doesnot know, and cannot know, if the resulting material will be an improvement over previous ones. It is in this important sensethat the variation, although far from random and unconstrained, remains blind. The manner in which you grope about in a dark room to find the light

switch changessignificantly after making contact with the wall on which the switch is located . What were three-dimensional gropings now become

two -dimensional ones. And as you encounter the molding along which you know the switch is located , your gropings become further constrained

to just one dimension. But although they may becomeprogressivelyand usefully constrained over time, an unavoidable blind component exists in your gropings until you actually find the switch. The same could be argued- although it is a much harder sell- about our use of vision to find objectsand help us navigatearound our environment (recall the discussion of the blind man in chapter 9). To the extent, however, that constraintsare effectivein advancingknowl edge(for example, whatever it is that preventsthe scientistfrom adding the soup to her would-be superconductingmaterial), they must be seenasadditional puzzlesof fit requiring explanation. And unlesswe are to return to providential or instructionist explanationsfor the existenceof theseadapted constraints, they can be explained only as the products of prior blind variation and selection.14As such, they may be quite well suited to guiding researchinto new, unexplored areas. But their fallibility must also be recognized sincetheir usein finding answersto new problems may on occasion actually hinder progressrather than facilitate it . So, " it is not only the case that there is no prescience about which variations will lead to success, there

is also no prescienceabout what part of the wisdom already achievedmust be abandonedin order to go beyondit . In exploring new regionsthe cognitive constraint system is itself up for grabs. " 15 It has also been argued that the process of selection in the advancement

of human knowledge is very different from natural selectionin biological


291

evolution. Again, Thagardcontrastshumansasintentionalagentsin their role asselectors of theoriesin thegrowthof scientificknowledgewith purposeless naturalselection : Thedifferences between epistemological andbiological selection arisefromthefact thattheoryselection is performed by intentional agents workingwith a setof criteria, whereas naturalselection istheresultof differentsurvivalratesof theorgan ismbearing adaptive genes .16 Thiscertainlyis a noteworthydifference , but wemustagainaskhow it in anyway invalidatesa selectionist explanationof scientificachievement . In biologicalevolution, organisms that by the luck of their genomearebetter suitedto their environmentleavebehindmoreprogenyandthereforemore copiesof their genesthanthoselesswell adapted . It is thiswinnowingaway of theless-fit organisms , not anyforesightedness or clairvoyance onthepart of the geneticvariations,that is responsible for the fit of organismto environment. And differentenvironmentsresult in the selectionof different adaptations , suchaswingsandlungsfor air, andfins andgillsfor water. Similarly,science progresses by theselectionof theoriesthat betterfit the criteriausedby scientists , suchasexplanatorypowerand parsimony . This is not to denythat certainpracticesandcriteriausedby scientists andcommunitiesof scientists maybe irrelevantor evendetrimentalto the progress of science , suchasthetendencyto fund or follow a line a research duesolely to thepower, prestige , or popularityof its leadingproponent.But insofar as sciencebecomesprogressively betterat describingand explainingthe objects, forces, and processes in the universe , it must be because the universesomehowinteractswith the experimentsand thoughtsof scientists andtherebyplaysa role in determiningwhichtheoriesandhuncheswill be retained. It hasalso,beennotedthat biologicalevolutionshowsdivergence leading to a greatdiversityof life forms, whereassciencein markedcontrastultimatelyleadsto convergence . The biosphereis rich in manystrikinglydifferenttypesof life forms, but physicists theworld overusethesametheories of relativity and quantumphysicsto accountfor and predictmechanical events . Thisdifferencehasbeentakenby someasevidence that organicevolution and conceptualdevelopmentmust be fundamentallydifferent. Thagardpointsout this differenceby statingthat

292

Without Miracles

survivalof theoriesis theresultof satisfactionof globalcriteria, criteriathat apply overthewholerangeof science . But survivalof genesis theresultof satisfactionof local criteria, generatedby a particularenvironment . Scientificcommunitiesare unlikenaturalenvironments in their ability to applygeneralstandards .I7 But just how are scientificcommunitiesable to apply these" generalstandards" ?Is it not becausethe local criteria of modern scientistsare much the sameno matter where on the globe they may be located?The most obvious explanation for why scientifictheoriestend to convergeis that they all share a very similar local environment. Light behavesvery much the samein Sri Lanka as it does in Switzerland. So do falling bodies, chemical reactions, and cell division. Indeed, much of the technology and effort of scientific researchis directed toward ensuringthat experimentsare conductedunder highly controlled conditions that can be duplicatedelsewherewith the same results. Thus a successfulexperimentthat revealsa new regularity of nature should be replicable by other scientistswith similar equipment anywherein the world . In addition, the goals of scientistseverywhereare much the same in their searchfor powerful yet simple theorieswith high explanatory and predictive powers. American biologist and philosopher David Hull addressesthis issuein pointing out: Conceptual evolution,especially in science , is bothlocallyandgloballyprogressive , not because scientistsareconsciousagents , not because they arestrivingto reach bothlocalandglobalgoals, but because thesegoalsdo exist. Eternalandimmutable regularitiesexistout therein nature. If scientists did not striveto formulatelawsof nature, they would discoverthem only by happy accident , but if theseeternal, immutableregularitiesdid not exist, anybeliefthat a scientistmighthavethat heor shehaddiscovered onewould beillusory.18 So regardlessof the differences between biological evolution and the work of scientists, one can argue that scientific theories, like organisms, develop as they are edited by the selectionpressuresof their environments, which, although necessarilylocal, reflect both universal (asfar aswe know ) regularitiesof nature and the sharedpractices, beliefs, and goalsof modern earth-bound researchers . The unavoidable local nature of theseapparently global criteria may somedaybe made quite clear when life is discoveredon another planet that doesnot conform to terrestrial theoriesof life, or when it is revealedthat the laws of physicsin the vicinity of black holes have lit tle resemblanceto those that were formulated to account for phenomena closerto home.


293

It should also be noted that biological evolution, like science, showsconvergencewhen similar problems are confronted by quite different organisms. The case of flight is perhaps the most striking example, with the asymmetricallycurved wing having evolved independentlyin insects, reptiles (the extinct pterosaurs), flying fish, birds, and mammals (bats). The similar shapesof fish and marine mammals such as dolphins and whales is another example of convergentevolution. Indeed, the phenomenonoften makesit difficult for biologiststo disentanglethe phylogeneticrelationships among organisms basedon physical appearancealone. Just becausetwo organismssharea common feature, this doesnot necessarilymeanthat they are closeto eachother on the phylogenetictree. Similarly, just becausetwo scientistsmay comeup with the sametheory to solvesomeproblem, it does not necessarilyindicate that one of the scientiststook the idea from the other. The independentdiscoveryof natural selectionto explain the origin of speciesby both Darwin and Wallaceis a casein point . So certain aspectsof selectionmay at first appear different in biological evolution and scientific development, but its basic function of eliminating the less fit and retaining the fitter is actually very much the same. As Campbell put it so simply, " rather than foresightedvariation, hindsighted selectionis the secretof rational innovation." 19And this " hindsightedselection " is asmuch a featureof scientific discoveryasit is of organic evolution. Differencesin the transmissionof accumulatedknowledge betweenbiological (genetic) and scientific systemshave also beenemphasizedby opponents of selectiontheory epistemology. It has beensuggested , evenby such widely recognizedexpertsin evolutionary theory asGould2 and Dawkins,21 that the cultural transmissionof ideas, including scientific ones, is a type of Lamarckian instructionist process, sincethe knowledge discoveredby one individual can be promptly passedon to another. In this sensethere appears to be a type of inheritance (or at leasttransmission) of acquiredcharacteristics that simply doesnot seempossiblein biological evolution. However, such a view of the origin and progressof scientific and cultural knowledge encounters difficulties no less severe than Lamarck's account of organic evolution. Certainly, a strict Lamarckian interpretation is untenablesince" in order for sociocultural evolution to be Lamarckian in a literal sense, the ideas that we acquire by interacting with our environment must somehow becomeprogrammed into our genesand then transmitted to subsequentgenerations." 22

294

WithoutMiracles

But even if a less literal interpretation of Lamarckism were applied to human knowledge, imposing problems remain. It was statedin chapters11 and 12 that knowledge cannot simply be transmitted from one individual to another, either by language or any other means currently known . Instead, the processof understandingthe ideasof another, whether expressed in oral or written languageor other signs or gestures, requires the active generation of a variety of candidate ideas on the part of the " receiver," and

the subsequentselectionof the best ones. Virtually all modern theories of learning, education, and knowledgeacquisition emphasizethe activerole of the learner in the construction of meaning, even if they do not explicitly embrace a selectionist

account of communication

.

So although scientific and cultural knowledge may appear to be transmitted from individual to individual and from generation to generation

through a Lamarckian process, this turns out to be as unlikely as the inheritance of acquired characteristicsduring biological evolution. For such transmissionto take place, it would have to be possible somehow for the knowledgecontainedin my brain to betransferredto yours in the sameway that I can copy the computer file containing this chapter from my disk to yours. It is possiblethat such a techniquecould be developedin the future, perhapsby reading the pattern of synapsesin one brain and then rewiring part of another

brain to match this pattern . Until

that time , however , we

must tolerate the rather slow (although still many orders of magnitude faster than biological evolution) and inefficient necessityof recreating the knowledge of others using languageand educational settingsas facilitators of this constructive

, selectionist

process .

Other Criticisms

Many other criticisms have beenmade of the application of selectiontheory to the growth of human knowledge. It would be too tedious to consider all of them here , but many if not most of them share one or more of three characteristics .

The first is taking the processof biological evolution asthe gold standard of selectionismand consequentlytreating any differencesin the mechanics of variation , selection, and reproduction

between that process and the

growth of human knowledge as reason enough to discount a selectionist view of the latter. But asthe previous chaptersshould havemadeclear, bio-

From Providence Through Instruction

logical

evolution

and

is just

selection

though

scientific

different

ing

, these

on

the

and

of

reason

that

by

insofar

as

process

. But

Second

not all

to

the

accepted

rather

critics

part

of

research

, and

that

sequences genome

. As

one

out

is

new

and

efforts

what

cognitive

to

apply

DNA be

is nec -

complexity

this

may

acquired

knowledge

in

techniques

for

being of

dictum

variation

,

. We

, having

example

' s selectionist blind

theory

as is now

an

and

. In

acquired

or show . The

a

deci -

done

this

as

type

(which

and

to

how

know

capabilities

how

of

sees

selective

all

reten

-

only

that

not

an appro

knowledge

search

to those

evolution

abilities

that to

seems

-

is at

to occur

instances

argument

it

need

itself make

it

wasteful

has

in

when

we

possible

that

development not

and

endowed

blind

contends

phylogenetic , but

previous

.24

resorting

, this

of the fact and not the

. It is certainly

blind

applies

cognizant

of knowledge

much

infrequent

suggest

the

sufficiently

knowledge

Dictum

mental

essence

been

to note

without

applied

not

and the advancement

's Dictum

of critics

sensory

selection

,

organic

knowledge

of current

of

from

typically

we already

knowledge

appropriately

sory

activity

, a number with

and

] have of already

mundane

beyond

adaptive

a scientist

use

may

Campbell

the origins

of Campbell

in one 's daily

"

selectionist

adapted

of old

how

strands

resulting

is for

evolution a

selection

fit , new

, is able

.23 The

in

through . It

:

or extension

of our

for

field

-

" selectionist

, is

mimic

that

new

application

project of

of fit

concerns

most

tion

defender

of Campbell

work

acquire

work

human

criticism

species

his

genetic

utilization

Finally

of his

bear

. Biological

to

account of

routine

the

[critics

are going

to

of increase

into

explanation

fashion

the Dictum

priate

to take

little

epistemology

complexity have

-

progress

" and

very repli

.

emergence the

knowledge

) pointed

They

not

details

the

instances tion

fail the

chapter

in

al -

, and

selection

" evolutionary

adapted

processes

biological

make

. So

exist

in themselves

epistemology

1974

in

may

hindsighted

original

seminal

selectionist

its

a selectionist

routine

phering

all

his

the

variation

another

, selection

science

theory

to

in

have

and

to

genes

of variation

and

" selection

increases

explain

knowledge require

thought

blind

system

, and

modes

variation

preferable

of

, many

only

not

both

to

one

differences

blind

terms

leads

in

essary

that

of cumulative

of

practices

distinct

superficial

Campbell

it

evolution

, cultural

other

the

instances

adaptation

employ

" are

proposed

of many

the

cumulative

epistemology

new

and

argument

a process this

to

theories

forms

cation

one

leading

295

to Selection

should

for

mental selection of not

human be

our us varia

to -

theory sen applied

296

WithoutMiracles

to the ontogenetic acquisition of knowledge and skills developedduring the life span of individuals .2sIn the same way that a girl does not have to

figure out by trial and error how to grow breastsat puberty, this having been already figured out during the course of phylogenetic development of the human speciesand encodedin her genes, during the course of human evolution we acquired sensory organs and brain structures that allow

us to acquirenew knowledgewithout the needfor further variationand selection .

But this argument that the growth of knowledge is able to bypassvariation and selection in its basic mode of operation has at least three flaws .

First, the achievementsof biological evolution are due to what was selected in the past, and can provide no guarantee that these achievements will con-

tinue to be of use in the future. Fortunately, many aspectsof our physical environment have remained relatively stable over long periods of time so that many of evolution's achievements(such as lungs, gills, wings, and feet) are still of usetoday. But the fact that extinction, not perpetual survival, is the usual fate of a speciesatteststo the tentative nature of the knowledge achievedby a selectionistprocess. It therefore appearsunlikely that natural selectioncould have provided us with innate abilities for acquiring all new knowledge, particularly in fields where our usual intuitions concerning space, time, and causality are violated, as in relativity theory and quantum mechanics .

Second, aspresentedin chapters4 and 5, it has becomeincreasinglyclear that at leastsome(and perhapsall) aspectsof ontogeneticdevelopmentare dependenton selectionistprocesses . In the mammalian immune system . it i.~ . widely acceptedthat the fit of antibody to antigen is not solely due to past achievements of evolution , but requires an accompanying process of blind

variation and selectionduring the lifetime of the animal. Selectionismis not quite asfirmly establishedin the brain sciences , but thanks to increasingevidence, there is growing agreement that the fit of the brain to the needs of its

owner is largely due to a continuous processof blind variation and selection of neurons

and neuronal

connections . Therefore , the argument

that selec -

tionist processesunderlie only phylogeneticchangeresulting from amongorganismselectionand not ontogeneticdevelopment, which is now known to require within -organism selection, is no longer tenable.


297

Third , some critics have pointed to instructionist proceduresby which new knowledgeis attained ascounterexamplesto the conjecturebeing consideredhere that all knowledge has its roots in cumulative blind variation and selection. One exampleis Pavlovian conditioning as discussedin chapter 7. It will be recalled that in Pavlovian conditioning an animal learns to

make an old responseto a new stimulus, aswhen a dog begansalivating on seeinga dark-colored liquid , after having had such a liquid containing acid placed its mouth. Such learning appearsinstructionist, since no obvious trial and error of responses is involved , and by pairing a neutral stimulus with an unconditional one that elicits a built -in response, the environment does appear to be instructing the organism to use the neutral stimulus to anticipate the unconditional

one. It cannot be denied that such learning

makesan animal's behavior better adaptedto its environment, as it is now able to anticipate and therefore avoid dangeroussituations (such as a bear fleeing at the sight of hunters ) and seek out favorable ones (as in using the

squeakof a mouseto locate food).26 But whereas such instructionist adaptation may occur and therefore pose a challenge to universal selection theory , it must be realized that such

mechanismsof knowledge acquisition are quite limited in what they can accomplish. As already noted, Pavlovian conditioning does not account for the learning of new responsesand the developmentof new perceptual abilities, but instead can account only for the associationof old responses to new stimuli. And thesenew stimuli must be onesthat immediately precedethe unconditional stimuli. Stimuli that either follow or are presented long before the unconditional stimulus do not readily becomeconditional stimuli , at least not in nonhuman

So Pavlovian conditioning

animals .

may be an instructionist

mechanism for

attaching new meaningsto old stimuli, but it cannot by itself provide innovative solutions to problemsthe way that natural selectionand other selectionist processeshave done and can do. Similarly, simulated nonstochastic neural networks may be capable of certain forms of learning without making use of cumulative blind variation and selection, but they are there -

by condemnedto sticking to their " innate" architecture and do not have the flexibility to reorganize themselvesas demanded by novel problems in the way that evolutionary algorithms can. (Of course, it is the human researcher who changes them , using cumulative blind variation and selection , when it is clear that such networks

have to be fine -tuned or more

298

Without Miracles

drastically overhauled.) Instructionist processes mayplayanimportant part incertain contexts foracquiring newknowledge, buttheyclearly cannot be thewholeanswer.Andif theyareeffective, thisin itselfconstitutes a

puzzle offitthatmustbeultimately accounted for,mostlikelybya selectionist explanation.

Finally, selection theory isnotintended toaccount forchanges thatare notcharacterized byincreases inadapted complexity. It ispossible thata species maychange overtimeinways thatareessentially neutral orperhaps

evenmaladaptive. A decrease in theozonelayermayresultin increased

ratesofradiation-induced mutations sothataspecies maynolonger beable topreserve andpassdowntheaccumulated wisdom ofitsgenome. Forvari-

oussocial, geographical, andeconomic reasons a society mayloseadapted cultural andscientific knowledge; bookburning andpersecution ofscientistsoccurred many times during periods ofsocial conflict andupheaval. It is,ofcourse, interesting toconsider justhowmuchevolutionary, cultural, scientific, andcognitive change isadaptive, anda comprehensive studyof anyofthesefields willcertainly havetogobeyond selection theory. Butto theextentthatadaptive change hasoccurred, a selectionist perspective

wouldseethatcumulative blindvariation andselection mustbeinvolved.

Totheextent thatnochange occurs orthatchange isnotadaptive, other

mechanisms (or factorsinterfering withcumulative blindvariationand selection) mustbeinvolved. Selection theoryisneitherablenorintended to explain stasisorneutral ormaladaptive developments, andtocriticize iton thiscountindicates a misunderstanding ofitsintent. Butalthough it is neitherablenorintended to explainneutralor mal-

adaptive change, it canprovide clues forunderstanding suchprocesses, especially whenadaptive change would bedesired ornormally expected. Forexample, despite continued attempts bybreeders, nocattlehaveyet beendeveloped thatreliably produce significantly morefemale thanmale

offspring. 27 Since thesebreeders havebeensuccessful inselecting forother traitssuchasincreased milkproduction, selection theory suggests thatthere isnoheritable genetic variation forthesexratioofoffspring. Aswesawin chapter 10,cultures mayfailtocontinue toadapttochanging environmentalconditions. Selection theorywouldleadusto placetheblameinsuch casesonlackofvariation incultural practices (perhaps noinnovations are permitted by the society) or lackof information withwhichvariations

FromProvidence Through Instruction toSelection 299 could fective

be

compared

in

suspected ciently responsive

fostering of

free

and the

not

selected growth

providing

to

allow

in

giving

. 28 And

students sufficient

of an

students

educative to

generate feedback

a

school

or

classroom

' knowledge environment and for

tryout students

and that their to

that skills was

is would

both

variations select

inef

be

suffi

-

, and the

better

ones. To return to what selectiontheory is intendedto explain, it was observed that: None of Campbell 's criticshaveproposedrival modelsof how knowledgecould havearisenout of ignorance , or how stupidprocesses couldleadto intelligentadaptation. In rejectingCampbell 's Dictum the criticshaveall notedthat advances in knowledgeareall baseduponprior knowledge , but . . . this is largelyan irrelevant criticism. In any situationin which the advancesin knowledgeare not wholly explicablein termsof previouslyattainedknowledge , a BVSR[blind variationand selective retention] process mustbeat work. Unlessof course , we reallydo live in a world in whichprayeror meditationor passiveinductioncanleaddirectlyto new knowledgewithout anyneedfor blind trials.29 So despitethe many attacks on selectiontheory, no one has yet demonstrated how a process that completely circumvents blind variation and selectioncan generatenew knowledge. This, of course, doesnot mean that such a demonstration may not be forthcoming. But until such time, selection theory appearsto be the only ballgamein town . The Innatist Misconstrual of Selection

Unfortunately , a ratherseriousproblemhasarisenfrom the useof withinorganismselectionistexplanationsof fit. Somequite prominentscholars have used selectionismto advanceinnatist conclusionsconcerningthe ontogeneticachievement of puzzlesof fit that areinconsistentwith selec tionistprocesses andproducts.This misconstrualinvolvesemphasizing the processof selectionwhile ignoring or deemphasizing the generationof novelvariantsamongwhich selectiontakesplace. The work of cognitivescientistMassimoPiattelli-Palmariniis onesuch example .3OPiattelli-Palmarini, who haswritten on conceptand language acquisition , makesmanystrongarguments againstinstructionist(whichhe calls"instructive") accountsof conceptandlanguageacquisitionandfor a selectionist(his "selective " ) one. He is alsoimpressedby the selectionist

300

Without Miracles

functioning ofthemammalian immune system andusesthis(asI didin

chapter 4)todemonstrate boththatsomatic, ontogenetic selection proc-

esses dotakeplace, andhowtheories ofimmunology haveprogressed from

instructionist onesto a selectionist process.

ButPiattelli-Palmarinis perspective onwithin-organism selection isquite different fromtheoneadvanced hereinthathisrequires allvariations tobe innately provided before selection cantakeplace. Forexample, inhis1989 discussion oftheselectionist functioning oftheimmune system hementions onlyinpassing thesomatic (ontogenetic) genetic reshuffling andmutations thatprovide thenecessary variation inantibodies. Anddespite thefactthat these antibodies arenotspecified inthegenome butrather emerge fromthe

countless possibilities thatthegenome allows,herefersto thisas the innaterepertoire ofantibodies. 31 Hethusleadsthereaderto surmisethatall

possible antibodies areinnately specified andtherefore provided before any selection takesplace. Insodoing, heessentially resurrects thelong-dead, genetically providential germ-line theory ofantibody production proposed byEhrlich in1900(seechapter 4).Piattelli-Palmarini makes thesamebasic

argument forconcept andlanguage acquisition instating thatAnypattern

isalready intheactualrepertoire oftheorganism. 32

Thisinnatist (andtherefore genetically providential) perspective onselectionism ignores thefactthatin theevolution of species andin the production of antibodies unpredictable novelty emerges fromtheblind recombination andmutation ofDNA sequences inthegenes regardless of

thefactthatit isalways thesameoldfourgenetic building blocks thatare reshuffled. Indeed, immunologists nowmakean important distinction between whattheycallinnate immunity andadaptive immunity, withthe

latter clearly notinnate. 33 Similarly, a selectionist account ofhuman cogni-

tionshould leadtoanexpectation ofnovelty intheemergence ofhuman concepts, ideas,andproblemsolving, dependent ontheblindvariation and

selection ofneuronal connections asdiscussed inchapter 5.ButforPiattelli-

Palmarini (andalsoforFodor, asobserved inchapter 11),allsuch products ofhuman cognition mustbeinnately specified before theycanbeselected,

despite thefactthatsucha conclusion is inconsistent withwhatis now

known oftheimmune system, thatis,allantibodies arenotinnately specifiedinthegenome butrather existonlyaspotentialities, themajority of

whichremain unrealized. It isalsoinconsistent withwhatisknownabout biologicalevolution.


301

Compare atypicaladulthuman withatypicaladultmouse . I willtakeit asuncontroversial thatthehuman possesses knowledge oftheworld,concepts , andlanguage thatthemouse does not. Andyetremarkably "anestimated99percent or moreof thegenes in miceandhumans arethesame andserve thesame purpose ."34SOmouse genes areverymuchlikehuman genes , differing onlyinthesequence ofrelatively fewnucleotide base pairs andhowthese areorganized intochromosomes , withameager 1%ofgenes thathumans donotshare withmice(andperhaps alsodifferences in how thegenes areorganized ) responsible for human cognition andlanguage . NowbyPiattelli -Palmarini 's (andFodor 's35 ) reasoning , wewouldhaveto consider themouse ashavinginnateknowledge of human concepts and language (thegenetic building blocks areallthere ; theyjustneed to berearranged a bit). Sucha conclusion wouldof course beabsurd . Theconfusionhereappears to stemfromnotappreciating thattherecombinatory shuffling ofgenes , aswellastherecombination ofsynapses , computer algo rithms , andpartsof molecules , canresultin theemergence of completely novelandunpredictable possibilities foradapted complexity . Herearetwomoreexamples .36Shakespeare 'splayMacbeth isessentially a long,ordered stringof about35typographical characters , including the 26letters of theEnglish alphabet (ignoring thedifference between upper andlower -caseletters ), a fewpunctuation marks , andspaces between words . Knowledge of these writtencharacters , however , doesnotin itself provide anyknowledge whatsoever abouttheplay.Nordoes knowledge of allthesounds ofRussian provide anyclues astowhatBorisYeltsin saidyes terday . It should beclearthatpossessing theelements thatmakeupacomplexstructure isnotthesame asknowing thecomplex structure itself. Piattelli -Palmarini is nottheonlywell-knownscholar to misconstrue selectionism asrequiring innatelyprovided variations amongwhichto choose . In arecent selectionist attempt to understand theworkings of the human mindandbrain,neurobiologist Michael Gazzaniga made thesame innatisterrorin hisbookNature "'s Mind. AlthoughGazzaniga , unlike Piattelli -Palmarini , doesrecognize thatthegeneration of antibodies involves thesomatic mutation oflymphocyte cells , thecreative andvariation generating aspect of selection theoryisnotincluded in hisviewof human behavior andbrainfunction . Thisoversight is madeclearwhenhestates "fortheselectionist , theabsolute truthisthatallwedoin lifeisdiscover

302

Without Miracles

whatisalready builtintoourbrains 37 andselection theory ishardonthe

nature/nurture issue inarguing thatallwearedoing inlifeiscatching up withwhatourbrain already knows. Wearediscovering built-in capacities. 38 Asonereviewer remarked, Gazzanigas perspective onknowledge is notunlike Socrates providentialist doctrine ofrecollection (oranamnesis; seechapter 6)inwhich all knowledge isalready inthemind, waiting to beremembered appropriately. Michael Gazzanigas doctrine isremarkably

similar, although itsays,selectedratherthanremembered. 39

Aquick comparison ofmouse andhuman (oramceba andmouse) makes itevident thattheamong-organism selection ofbiological evolution results intheemergence ofnewpossibilities thatwere notpresent before, andthat natural selection istherefore a creative process thatconstantly fashions

innovative variations among which toselect. Weshould therefore expect

within-organism ontogenetic selection todothesame. Indeed, aswesawin

chapter 5,thebrainnotonlyselects frompreexisting neural circuits, but actually engineers newcircuits byfirstadding newsynaptic connections between neurons andthenselecting some while eliminating others. And

although themechanisms forsuchwithin-organism variation andselection

may beconsidered innate, itsproducts cannot beregarded assuch, since by thesame reasoning, onewould beobliged toconclude thatthegenetic information forthehuman body, brain, andcognitive abilities wasalready contained inthevery firstorganism thatused DNA foritsgenes. Gazzaniga andPiattelli-Palmarini arerightincontending thatselection

theory hasimportant implications forthenature-nurture debate inpsy-

chology. Buttheyerrintheirconclusions thattheimplications arenecessari-

lyinnatist. Ifthey were right, selection could operate only among already

achieved variations, andthatwould meanthatwecould nothaveevolved

to bewhatwe are today.

AsImmanuel Kant noted: TheCreation isnever over. Ithadabeginning butit hasnoending. Creation isalways busymaking newscenes, new things, andnewWorlds. 4 Inthesame waythatnatural selection among organisms provides anexplanation forthecontinuing adapted complexity

andadaptive creativity oforganic evolution, cognitive selection within the nervous system provides anexplanation fortheadapted complexity and

continuingadaptivecreativityofthehumanmind.

16 -=

~

-

Universal

The

Selection

Second

. . . the bling

what

Darwinian

growth Darwin

of

our called

Theory

Revolution

knowledge ffnatural

is

the

selection

result " ; that

of

a

process

is , the

natural

closely

resem -

selection

of

hypotheses . . . - Karl Popper1

In theprovocative essay "Universal Darwinism " (referred to attheendof chapter 2), Richard Dawkins maintains thatif lifewereto befoundelse where in theuniverse , wewouldhaveverygoodreasons to suspect thatit hadevolved asit didonEarth , thatis, bynatural selection . Letusrecallthat Dawkins 'sconclusion isbased ontheargument thattheprocess ofcumula tiveblindvariation andselection istheonlycurrently available scientific explanation thatisin principle capable of explaining theemergence ofthe adapted complexity required forlife.2 ButDawkins 'sargument isnotlimitedto adaptive biological evolution . Wesawinthepreceding chapters howselectionist explanations forpuzzles of fit themselves haveevolved in manyotherdisciplines thatarealsoconcerned withunderstanding various formsof knowledge growthinvolving thespontaneous emergence offit between twoormoreinteracting systems . Scottish philosopher andpsychologist Alexander Bainmayhavebeenthe firstto applya Darwinian perspective to human thought(seechapter 9), butit wasnotuntil100years afterthepublication ofDarwin 'sOriginthat DonaldT. Campbell envisioned a comprehensive , all-embracing rolefor selection theory .3Hisselectionism emphasizes thepsychological , scientific , andculturalgrowthof knowledge , buttherecent selectionist discoveries in immunology andneurophysiology , andtheapplications of selection theoryto theengineering of molecules andcomputer software finallyhave

304

Without Miracles

attracted theattention ofthescientific community totheimportance and potentially universal applicability ofDarwins selectionist insight. Indeed, under certain conditions, theevolution offitbywayofcumulativeblind variation andselection appears inevitable. Consider apopulation ofself-replicating entities thatvaryinways relevant totheirreproductive

success, andthatinhabit anenvironment oflimited spaceandresources that

doesnotundergo largefluctuations fromonegeneration ofentities tothe next.Iftheseentities produce quite(butnotalways perfectly) accurate copies ofthemselves, aftera fewgenerations thewinnowing effect ofselec-

tionwillbenoticed asthepopulation inexorably shifts toward apreponderance ofnewentities thatbetter fittheir environment. This, inanutshell,

iswhattheprocess ofcumulative blindvariation andselection isallabout. Howeveaswehavenowseen,theseentities donothavetoberestricted

toliving organisms andthegenes theycontain. Theycanbemolecules, antibodies, neural synapses, behaviors, scientific theories, technological products, cultural beliefs, words, or computer programs. Andselection doesnothavetoberestricted tothenatural andpurposeless selection of Mother Nature, butmayinvolve purposeful humans selecting forplants growing bigger tomatoes, cows giving more milk, scientific theories providingbetter predictions, automobile engines yielding greater efficiency, or

molecules providing morepowerful drugs. Therobustness oftheselection process wasdramatically demonstrated bythefindings of artificial life

researchers such asThomas Ray(see chapter 13)whowere amazed atjust howeasy itistogetadaptive evolution happening ontheir computers. As long asthebasic conditions ofsome (butnottoomuch) variability, accurate (butnottooaccurate) replication, andafairly stable environment prevail, themechanistic, unforesighted evolution offitappears inescapable. Theselectionist explanation fortheemergence ofadapted complexity andnewknowledge mayalsoappear inevitable onlogical grounds. As Campbell putit,increasing knowledge oradaptation ofnecessity involves exploring theunknown, going beyond existing knowledge andadaptive recipes. Thisofnecessity involves unknowing, non-preadapted fumbling in

thedark. 4 Ofcourse to thiswemustaddtheselection andretention of

those occasional results ofthisfumbling thatarefound byhindsight topro-

videa bettersolution to theproblem at hand.

Universal Selection Theory 305

Whenstatedthis way, particularlyusingthe phrase"of necessity ," a selection theoryepistemology maybeinevitable by tautology . Thatis, it appears to betruesimplybydefinition(asisthestatement "abachelor isan unmarried male") with nopossibility of beingdisproved andconsequently replaced with a bettertheoryof knowledge growth.sA selectionist episte mologymaybedescribed tautologically ; however , thisdoesnotmeaneither thatit isfalseorthatit cannotbestatedin amorefalsifiable way.Forexam ple, Campbell alsomadetheboldclaimthat A blind-variation -and-selective -retention process is fundamental to all inductive achievements , toallgenuine increases inknowledge , toallincreases inthefit ofsys temto environment .6 And it wasobserved that

Thus , it shouldbepossible to testtheboldselectionist hypothesis of Campbell anddiscover - if theyexist - nonselectionist processes underly ingtheemergence ofadapted complexity andknowledge growth . Thismay notbeeasyin disciplines suchasthehistoryof science andcognitive psychology , butthenewselectionist explanations forantibody production and braindevelopment arebased onempirical research findings , notonempty tautologies . Campbell wouldthushaveusbelieve thatall knowledge , all problem solving , all skills , alladaptive physiological andneural changes , alluseful cultural beliefs andpractices , andallscientific andcultural progress haveat theirrootscumulative blindvariation andselection - phylogenetic (among organisms ), ontogenetic (withinorganisms ), or both- of thesame general typeproposed byDarwinto account fortheevolution ofspecies . Thefindings , theories , andrationale for selectionist explanations of puzzles of fit reviewed in thisbooksuggest thatsuchauniversal selection theoryshould nowbetakenseriously .

Universal Selection Theory 307

308

WithoutMiracles

Universal SelectionTheory

computer

programs . The basic components

cruise control

maintain

systems , such as the

device you may have in your car , are made up of rather sim -

ple input -output chapter

of control

309

devices . Connecting

8 ) gives the resulting

them together

control

a desired goal despite unpredictable

This phenomenon

of emergence

provides the complexity the detailed properties

without

in a special way (see

system the lifelike

is clearly

" willpower

" to

disturbances . very powerful

stuff , and it

which life surely could not exist . But since

of emergent variations - whether

they be molecules ,

organelles , cells , organs , organ systems , organisms , societies , or social in stitutions - cannot nent building

initially

be predicted

from

knowledge

of their compo -

blocks , the only way that newly emerged entities can lead to

adapted complexity

is through

a process of blind variation

and hindsighted

selection . But we can make selection more powerful

still . Not being content with a

single -step selection process , we can instead take the best of the variations , vary them , and then select the best of the new generation , repeating the nrocess over and over again . This , of course , is constructive cumulative .sLelection

(figure

sional accidently

16 .1 ). This process of selecting and fine -tuning useful emergent

we should not be surprised tion , antibody

production

the occa -

system turns out to be so powerful

that the adaptive

processes of biological

that evolu -

,11learning , culture , and science all employ it , and

KnowledgeProcesses

Figure16.1 Typesof knowledge processes : instruction andthreetypesof selection .

310

Without Miracles

311


y

0

0

0

0

Figure16.2 Four problemspaces(or fitnesslandscapes ).

ond

time

since

.

.

thief

has

not

this

the

not

only

way

a

form

makes

likely

card

in

that

be

The

found

the

is

a

canceled

find

and

through

a

and

the

,

)

to

large

(

.

Unfortunately

a

solution

if

the

although

,

using

multidimensional

series

.

of

blind

guesses

character

at

,

provide

Nonetheless

noncumulative

tedious

to

succeed

universe

very

essentially

best

.

and

are

possible

one

eventually

stolen

left

can

at

should

reported

time

.

he

therefore

previous

selection

,

spaces

solution

is

the

noncumulative

is

enough

knowledge

than

patience

problem

this

less

of

and

the

a

of

or

selection

problem

process

is

be

for

hindsighted

more

time

method

and

no

before

might

this

worst

an

of

impossibly

the

long

.

In

in

So

accumulation

is

enough

perhaps

there

gradual

trial

solution

So

No

each

the

the

second

neighborhood

panel

we

of

notice

the

a

peak

broad

provide

peak

to

better

the

results

solution

,

than

so

those

that

trials

farther

312

WithoutMiracles

away . Anexample ofsuch aproblem wouldbefinding theamount offertilizerthatmaximizes acertain cropyield , where toolittlefertilizer leads to diminished yields , asdoes toomuch . Nowit isnotnecessary totryallpos siblequantities offertilizer , since if it isfound thatmore fertilizer isbetter , onewillcontinue totrystillmore untiltheyieldbegins todropagain . Not muchblindvariation isinvolved in thisproblem , except perhaps forthe guess oftheinitialquantity . Such single -peak problems canoftenbesolved quitereadily using whatarecalled hill-climbing techniques or algebraic methods based ondifferential calculus ,thatis, solving forthex value whose associated yvalue hasaslope ofzero if thefunction relating ytox isknown . Indeed , wediscussed inchapter 13andintheprevious section thatback propagation neural networks achieve theirfit byatypeofinstructionist hillclimbing technique . A variation andselection procedure canalsowork quite well,andmuch more quickly thanintheprevious problem , aslongas newguesses about thevalue ofx takeadvantage oftheknowledge ofthe partial successes ofprevious trials . Things become bothmoreinteresting anddifficult when weconsider a two-peak problem space asshown inthethirdpanel . Nowwehave alocal maximum (thelowerpeak ), inaddition totheglobal maximum represent ingtheoverall best solution , soweruntheri.skofgetting stuck onthelocal maximum andnever finding abetter (orthebest ) solution . Anexample of such aproblem wouldbefinding theoptimum dosage of a drugto treat hypertension , where alowdosage hassome effect inlowering blood pres sure , ahighdosage hasthegreatest effect , andintermediate dosages have littleeffect . Theonlywaytoescape thetrapofalocalpeak istotry . some new , random values to seeif theyleadto a higher peak . Since it isnot known where thatpeak maybelocated , onecanonlytakeablindjumpoff thelocalpeak andhope thatit leads toabetter solution . Thecumulative variation -and -selection approach tothisproblem wouldbetostartwitha wide -ranging population of values of x andallowthemto recombine ,. reproduce , andmutate sothatit isunlikely thattheglobal peak wouldbe missed . Buteven thistwo-peak example ismuch simpler thanmany if notallof thereal -worldproblems thatweandotherspecies encounter . Thefourth panel represents a verycomplex problem space withmany jagged peaks andvalleys . Onfirstconsideration , finding thevalue ofx thatprovides the


313

bestsolutionmightseem asdifficultasthefirstproblemconsidered above . Regardless of thecomplexity of thelandscape , constructive andcumulative blindvariationandselection maybesuccessful in findinga solutionto this andsimilarproblems if knowledge concerning thebestsolutionof previous trialsisappliedto construct newvariations . Thisiswhatisdonein thebiologicalworldwith sexualreproduction , andit is alsousedin genetic algorithmsandgeneticprogramming asdescribed in chapter13. By taking valuesthat on previous trialsprovidedthebestsolutionsandusingthem to breednewvalues , patterns thatmaybetoocomplexfor thehumaneye to perceive canbeexploited anda solutionreached . Byrepeatedly breeding (with mutationor sexualrecombination to ensurea continued sourceof newvariations ) fromthebestsolutions foundsofar, knowledge obtained time from pasttrialsis preserved , andat the same newblindvariations are introduced in each generation that continue to gropefor still better solutions .13

Representingproblems as fitness landscapescan be useful for understanding how a selectionistprocedurefinds solutions. It can be somewhat misleading, however, giving the impressionthat the solutionsto be searched all exist beforethe searchtakesplace- in other words, that nonconstructive selectionis involved. Instead, thesesolutionsexist only aspotentialities, and each one must first be constructed before it can be evaluated . This is obvi -

ous in biological evolution where pairs of sexually reproducing organisms whose offspring, no matter how many, representonly a minute fraction of the possiblegeneticrecombinationsof the parents' genomes. And when to this is added various possiblemutations and other errors such as extra or missing chromosomes, the possibilities that can be constructed and evaluated by selectionbecomeinfinite . The importance of the construction and consequent emergence of new forms to test and select led Henry Plotkin , an

important advocateof universal selectiontheory, to usethe phrase " generate -test -regenerate " 14to describe what cumulative

is referred

to here as constructive

selection .

Despite the very powerful generation-and-searchproceduresthat constructive cumulative selectionprovides, it does have its limits . Notably, it

cannotbeguaranteed to find theoptimumsolutionfor everyproblem. But it has now been well demonstratedthat many problems exist for which constructive

cumulative

selection

can find , if not the best , at least useful


315

attemptingto extendthe selectionist perspective to otherfields. If cumulative blind variationand selectionis found to be lackingasan explanation for theemergence of designin organicevolution, an extensionof selection ist principlesto other achievements of adaptedcomplexitywould be suspect. We will thereforenow confront five suchwould-be challengers to natural selection : punctuatedequilibrium, directedmutation, exaptation , symbiosis , andself-organization . Punctuated Equilibrium Accordingto classicDarwinian selection , biologicalevolutionproceeds throughthe accumulationof verysmallchanges overlongperiodsof time. Gradualchangeis essential , sinceit is the only way that blind variationis likely to comeup with improvements for selection . Whereasit is always possiblethat a largegeneticchange(or macromutation ) mayresultin a fitter organism , for example , the transformationof an organIsmcompletely insensitive to light to onewith a functioningeyein onegeneration , thelaws of probabilityarealmostcertainto makelargerandomchangeslessadaptive ratherthanmore. As Dawkinsexplains: To "tame" chance means to breakdowntheveryimprobable intolessimprobable smallcomponents arranged in series . No matterhowimprobable it is thatanX couldhavearisenfromY in asinglestep , it isalways possible to conceive of a series of infinitesimal graded intermediates between them.However improbable a large scale change maybe, smaller changes arelessimprobable . Andprovided wepostu latea sufficiently largeseries of sufficiently finelygraded intermediates , weshallbe ableto deriveanything fromanything else , withoutastronomical improbabilities .16 The problemis, however , that the fossilrecorddoesnot provideclear evidencefor the gradualchangeof evenonespeciesinto another . Darwin recognized the incompleteness of the fossilrecord, but believedthat it was only a matterof time beforetheseintermediate"missinglinks" would be found to providehard evidencefor the gradualemergence of new species overtime. That thesefossilgapsremaindespitemanynewfossilfindshas beentakenby someasan indicationthat Darwin's emphasis on the gradualismof evolutionwasmistaken , andthat evolutionproceeds not by slow, gradualchanges but ratherby largeanddramaticjumps, or saltations . Truesaltationistsarenot easyto find amongmodernevolutionarybiologists, sinceit is generallyrecognizedthat large, blind macromutations from parentto offspringare almostcertainto be maladaptive .1? But a

316

Directed

Without Miracles

Mutation

Similar compatibility is not the case, however, for another view of evolution that has attracted

considerable

interest and led to much recent controversy .

In 1988 John Cairns, a well-respectedmolecular biologist and cancer researcher, published with two associates a paper in the prestigious British journal Nature that threatened to undermine the basic tenets of Darwinian evolution

.2O

Cairns and his colleaguesclaimedto havefound evidencethat E. coli was able somehow to direct its mutations to achieveadaptive changeswhen placed in a new, challenging environment. This researchinvolved placing bacteria that could only metabolizeglucosein an environment where only a foreign sugar (lactose) was available. Here the stressedbacteriacontinued to duplicate and, aswould be expected, someof the descendantscontained mutations that permitted them to metabolize the new sugar. This in itself is

not surprising, since the genetic changenecessaryto transform an E. coli


317

from a glucose- to a lactose-eating bacterium is quite small, and in a large colony it would be expectedthat at least some of the naturally occurring mutants would have stumbled on it by sheer blind chance. But these scien-

tists reachedthe highly unorthodox conclusion that instead of being produced randomly, the bacteria were somehow able to produce the adaptive mutations at a much higher frequencythan other, nonadaptive mutations. In other words, they believedthat their studiesprovided evidencethat " bacteria can choosewhich mutations they should produce" which would " provide a mechanismfor the inheritanceof acquired characteristics." 21 As would be expected, these statementsimmediately elicited both considerable interest and controversy , since the central dogma of biology was

being challenged, that is, that changesin the environment cannot direct (instruct) changesin the genome. Someresearchersrejectedthis conclusion out of hand, but others were impressedenoughto attempt to find possible mechanisms by which the environment could somehow instruct the genome

to produce just the right mutations to allow it to digest the new sugar. Cairns himself proposed that environmental changescould affect changes in proteins that could consequentlyinstruct the DNA to make certain adaptive changesin the genes, in flagrant violation of the central dogma. However, it may well be that this and other explanationsfor directed or " instructed " mutation are not necessary after all . Australian microbiologist

Donald MacPheeand his colleaguesprovided evidencethat, when placedin a medium of lactose, the mutations produced by glucose-metabolizing E. coli are indeedproduced blindly.22What seemsto happenunder the stressed condition of a glucose-poor environment is not a specificincreasein the rate of adaptivemutations, but rather a generalincreasein the overall mutation rate due to inhibition of the mechanismthat usually checksand repairs the genetic errors that arise during the normal functioning of the bacterium . So

while mutations continue to be produced blindly, the higher rate of genetic changeallows the bacteriato stumble on the adaptivegeneticchangemore quickly than they would if left in their normal glucose-rich environment. But let us continue to imagine for a moment that a bacterium was able to

changejust thosegenesregulating metabolismin just the right way to allow for the digestion of a foreign sugar. If this were the case, it would be yet another example of a puzzle of fit demonstrating that the bacterium had somehow acquired the ability to sensea new sugar in its environment and

318

Without Miracles

alter its genometo digest it . But then we would be led to ponder how this adapted complexity could have originated in the first place, with cumulative blind variation and selectionas a prime candidateto explain the source of this remarkableability that somehowpermitted the bacteriumto instruct its genometo make the required changesto digest the new, strange food that was being served. Although no convincing evidenceexists that adaptive changesin genes can be directed by the environment in a Lamarckian manner, the findings of Cairns and MacPheeand their respectivecolleaguesare important . If organisms are able to increasetheir mutation rate in the presenceof new environmental stressesbut keep mutations in checkwhen thesestressesare absent, it would enableorganismsto exert a certain degreeof control over classicneo-Darwinian perspective. Instead onstant

evolution

of

that

producing

,

varied

from

may

just extant

the

at

organisms

offspring species

absent

mutations

conditions

the

is

when

a

c

rate

produce

such

more

innovative

regardless

mutations

variation

of

and

is

environmental

therefore

necessary

more

to

keep

. 23

This view ascribesto the evolutionary processdecidedly more " intelli gence" than does the neo-Darwinian perspective. It nonethelesspreserves the required blindnessof geneticvariations. What is altered is only the rate of production of thesevariations. This sensitivity of mutation rate to environmental stresscould simply be the result of a stress-related breakdown of geneticrepair mechanisms.Or it could be the result of a more sophisticated activemechanismthat itself had evolvedby natural selection, sinceindividuals that by chanceproduced more geneticvariability under difficult environmental conditions would have beenmore likely to leavebetter adapted progeny than those insensitiveto environmental stress. The work of Cairns and MacPheeconcernedthe metabolism of different types of food. It is not difficult to imagine how other types of biological functions could also be involved, suchasthermoregulation. For example, as temperaturesdropped at the onset of an ice age, mammalswould undergo stressas did Cairns's bacteria when placed in an environment where no useful food was available. This would lead to an increasein the mutation rate during reproduction, resulting in a secondgeneration of animals with greater variation in the length and texture of their coats. Those particular descendantshaving, by chance, longer and thus warmer coatswould suffer


319

lessfrom the cold environment, resulting in lower mutation rates and consequentlylessvariation in the coats of their third generation, extra-hairy offspring. But those second-generation animals with short coats would maintain a higher rate of mutation, so that at least someof their offspring would likely havewarmer coats then their parents. This hypothesishas someinteresting consequences . As in the ice-ageexample, by varying the mutation rate, a specieswould adapt more quickly to changingenvironmental conditions. It is also of interestto realizethat such stress-dependentmutation rateswould result in occasionalshort periods of relatively rapid (although still gradual) evolutionary changeseparatedby longer periods of little or no changeduring periods of environmental stability. And this is exactly what Gould, Eldredge, and their associatesrefer to as punctuatedequilibrium. Exaptation Another challengeto natural selectionis posedby those advocatingexaptation asa major mechanismof biological evolution. Although we briefly discussedexaptation in chapter 11, some additional remarks are appropriate here, as this perspectivehas receivedconsiderableattention as a potential rival to the selectionistaccount of adaptiveevolution. It will be recalled from chapters 5 and 11 that exaptation refers to the emergenceof somefeature of an organism that fits a current function, but did not originally evolve for this use. We consideredhow bird and insect wings, now used for flight , appear to have have evolved originally to aid in cooling and heating. Darwin 's own example was how the lungs of terrestrial animals evolved from the swim bladder of fishes.24Exaptation also refers to the emergenceof a current adaptedly complex characteristic that arosefor no functional reason, probably as a nonselectedcorrelate of someother adaptedcharacteristic, but turns out later to be useful for some function. The concept of exaptation is particularly valuable in understandinghow the necessarygradualism of natural selectioncan account for the evolution of complex adaptations that would appear to be functionally uselessin their incipient forms. Let us consideragain the bird 's wing. If wings evolved gradually over a long period of time, the first protowings would have been nothing but stubby protuberancesfrom the backsof protobirds. Sincethey

320

WithoutMiracles

would have beenill suited for flight , the question arisesas to why and how they would have begunto evolvein the first place. Exaptation, by disentangling current use from the original selection pressures, makes this understandable . The protowings may not have been

of much use for flight , but could have eas.ily aided in eliminating excess body heat when stretched out in a shady breeze, and increasingwarmth when extended to catch the rays of the morning sun. Selectionfor larger and larger wings for more effectivethermoregulation would have also laid the groundwork for flying, although the first aeronautically useful wings probably mainly provided someprotection from falling from treesor other high places. Once past a certain size, the switch in the primary role of wings

from thermal equipment to flying equipment would have beenmade, and further refinementswould have increasedtheir fitnessfor flying. But although someseeexaptation as a challengeto natural selectionand as a part of a " new theory of evolution," 25it actually provides no competition at all as an explanation of adapted complexity. For wings to have evolved as useful instruments for flight , they first had to evolve as useful instrumentsof thermoregulation. The only explanation for how this could havehappened, asidefrom the astronomicalimprobability of a singlelucky mutation, is by natural selection. In addition, further refinements were undoubtedly necessaryto fashion wings for efficient flight . Once again, the gradual and cumulative selectionof blind variations in wing shapesis the only processcurrently understood that can account for the evolution of such design. Therefore, to identify a feature as an exaptation doesnot mean that feature " is not the product of natural selection" - only that it did not always serveits presentfunction. Indeed, most human behaviors serving biological functions that enhance fitness should probably be regarded as exaptations, modified through natural selectionfor subsequent specialization. The manual dexterity enabling specializedhuman hand movements now used in food gathering, eating, caretaking, tool use, and communicational expressionsderivesfrom the ability of prosimian ancestorsto grasp branches . Human emotional expressions evolved from ritualized facial and vocal

displays in other ancestralspecies, display derived from behaviors that originally had no signaling function. . . . The use of stereotypical vocal patterns in human mothers' speechto infants is also undoubtedly an exaptation, with evolutionary origins in ancestral non -human primate vocalizations used for very different purposes .

The claim that such human behaviorsare exaptations rather than adaptations is a claim about origins, which does not require rejection of an adaptationist explanation for the current fit betweenthese behaviors and the biological functions they serve . 26


321

Sowhereasexaptation permits new, unanticipatedtwists and turns in the evolution of a speciesas new roles and consequentlynew selection pressuresare found for old features, it relies on natural selectionfor all initial adaptationsand consequentrefinementsof theseadaptationsfor new functions. Exaptation may make it easierfor natural selectionto do its stuff, but it is most certainly not an alternative explanation for the adaptedcomplexity of naturally occurring design. Symbiosis Other individuals interestedin accountingfor the products of evolution but who are not particularly enamoredof Darwinian natural selectionemphasizethe role of cooperation among organisms. One particularly outspoken advocateof this view is American biologist Lynn Margulis. Shehas offered strong criticism of the Darwinian selectionistaccount of evolution, stating, " It 's totally wrong. It 's wrong like infectious medicine was wrong before Pasteur. It 's wrong like phrenology is wrong. Every major tenet of it is wrong." 27 Margulis doesnot doubt that natural selectionoccurs, but objects both to the gradualnessof evolution and its emphasison competition. In 1965 she proposed that eukaryotes, the nucleatedcells that make up all organisms except bacteria, originated from the fusion of two different types of bacteria. Although her proposal was initially greeted with considerable skepticism, the theory of the symbiotic evolution of eukaryotes is now widely acceptedamong biologists. But if symbiosisdid occur, how doesthis make Darwinian theory " totally wrong" ?To be sure, Darwin knew nothing of the geneticbasisof life and so he could hardly have imagined that new organismscould emergefrom the incorporation of the geneticmaterial of one genomeinto the genomeof another. Also, the cooperation betweendifferent speciesthat characterizes evolutionary symbiosis does constrast with his very competitive view of evolution. But it must be kept in mind that the basic core of Darwinian theory is its reliance on hindsighted selectionfrom a population of blindly constructed variations, and in this respectMargulis's symbiotic theory is Darwinian at its core. The variations available, however, are no longer limited to genetic mutations and recombinations within the genepool of one species , but include geneticcombinations acrossspecies.And theseinitial interspecific (between-species) genetic transfers must have occurred

322

Without Miracles

blindly, with probably almost all of them harmful to one or both organisms (as occurs when the geneticmaterial of an influenza virus invadeshuman cells) or of no consequence . So by pure chanceit must have happenedone lucky day that the accidentalpooling of the geneticmaterial or organellesof two types of unicellular organismsprovided a new cell with survival and reproductive advantagesover both of its constituents. But if one is to suggest that anything other than chancewas involved in forming such symbiotic relationships, then one will be in the unenviableposition of having to explain the origin of the knowledgerequired to circumvent chance. Indeed, if suchsymbiotic evolutionary processeshavein fact taken place, it would make Darwinian selectionisma more powerful, rather than a less powerful, explanation for adaptive evolution. Evolution would no longer be limited to fiddling with the genesof separatespeciesby mutation and sexualrecombination. Instead, a whole new world of possibilities becomes available in the mixing and matching of genes, organelles, and cells across species. A specieswould not have to evolve photosynthesisfor itself if it could incorporate the photosynthetic know-how of another species , which is now believedto have occurred. But there still is no way that a primitive one-celled organism could possibly " know " which other type of cell or organism it should incorporate or borrow from to improve its chanceof survival and reproduction. It could only await the blind forces of nature to bring a promising partner within reach, or await the invasion of an adventuresomevirus bearinggeneticmaterial from a previously infected cell. Al though the potential benefitsof certain symbiotic unions would be great, it would not be possibleto avoid the very sameblind variation and selection that characterizesthe natural selectionthat Margulis attacks so vigorously- at leastnot without the guidanceof an intelligent and beneficentmatchmaker working behind the scenes . Self-Organization But there may well be somethingmore to adaptive evolution than natural selectionafter all. The secondlaw of thermodynamics is the well-known law of increasingentropy, which statesthat in an isolated system(that is, a systemthat can neither gain nor lose energy or matter) we can expect order to decrease , energyto becomelessavailable, and a stable(and lifeless) equilibrium to be reached. B.ut in an open systemable to draw on sourcesof


323

outsideenergy , thesituationcanbedramaticallydifferent. Theevolutionof life itself is the moststriking exampleof a naturallyoccurringincreasein complexity . But inanimateobjectsandsystems canalsodemonstrate naturally emergingcomplexityin certainsituations.Anyonewho hasmarveled at the intricatesymmetricalbeautyof a snowflake , observedthe coordinatedballetof grainsof rice in a simmeringpot of water, or encountered the organizedfury of a tornadohasnoticedthat complexitycanalsoarise spontaneously in theinanimateworld. And this spontaneous emergence of complexity , or self-organizationas it is now usuallycalled, has recently attractedthe attentionof a wide rangeof scientists , from physicistsand biologiststo cognitivescientists andeconomists . That organizedcomplexitycanemergespontaneously in inanimatesystemsmayhavefar-reachingimplicationsfor understanding theoriginof life and its continuingevolution. One of the major difficultiesin comingup with a convincingnonmiraculousaccountis explaininghow inanimate matter could haveorganizeditself into the very first self-replicatinglife forms. The degreeof complexityrequiredfor this first stephasseemed to manybiologiststo be just too unlikelyto be dueto the randomforcesof nature. Darwin himselfwas reluctantto advancea nonmiraculousargument, andin the lastparagraphof latereditionsof the Origin refersto the power of life "havingbeenoriginallybreathedby the Creatorinto a few formsor into one." Soif the blind lawsof natureoperatingon inanimateentitiescouldwith high probability leadto the emergence of complex, self-orga,nizedmoleculesandnetworksof molecules , thenthe origin of life itself, aswell asits continuedevolution, becomes somewhatlessof a mystery . This is themessagethat AmericanbiochemistandbiophysicistStuartKauffmanhasbeen deliveringfor thepastdozenor soyears,andprovidesin detailin his influentialbook, TheOriginsof Order.28 It is now recognized that the lawsof physicsactingon nonlivingentities canleadto spontaneous complexity , but nothingin theselawscanguarantee adaptedcomplexityof the type seenin living organisms , that is, the ubiquitousbiologicalpuzzlesof fit. Of all the complexsystemsandstructuresthat may self-organizedueto the forcesof nature, therecan be no assurance that all or any of them will be of usefor the survivaland reproductionof living organisms . Selection , therefore , must chooseamong

324

Without Miracles

thesevariouscomplexsystems theoneswith characteristics bettersuitedto survivaland reproduction , and eliminateothers. As Kauffmanremarked , "evolutionis not just 'chancecaughton thewing.' It is not just a tinkering of theadhoc, of bricolage , or contraption.It is emergent orderhonoredand honedby selection ."29 The study of self-organizing systemsis amongthe newestand most ambitious scientific ventures of the late twentieth century, and its discoveries may ultimately havea major impact on evolutionary theory and our understanding of the emergenceof life itself. But from our presentviewpoint it is difficult to see how self-organization could ever replace, as opposed to complement, natural selection. It may help to jump-start natural selection by blindly offering up a variety of already complex systemsfrom which to choose. But it is only after-the-fact selection that can eliminate the nonviable complex systemsand retain the viable ones. To return

one final

has become puzzles

and

directed

and instructionist been found

or without

different

of many

movement be considered

from

evolution

of adaptive , selection

account

of puzzles

of humankind providential

suggestion

different

Darwinian

book , selection

disciplines

change theory

previous

and knowledge provides

of fit , whether

a truly

growth

have

naturalistic with

in many

to selection

of selectionism of inquiry

pro -

providential

of theories

instructionist

revolution

theory

to explain

this fit occurred

. The evolution through

fields

either

them , as in genetic

. Where

of the superiority

in so many a second

of this

system , or to create

molecular

theories

the assistance disciplines

ist is a provocative

may

theme

part

to be inadequate

and nonmiraculous

recent

to the major

of fit , as in the immune

gramming

different

time

an important

. And

constitutes

-

this what

A

In a numberof respects this revolutionis unlikethe first. Thefirst made a dramaticdebut on 24 November1859 with Darwin's The Origin of Species sellingout its entirefirst editionof 1250copiesthe first dayit was offeredfor sale. It involveda singlediscipline , biology, and a singlequestion, the evolutionof species . And Darwin himself(if not all of his fellow Darwinians)wasinsistentthat processes otherthannaturalselection ----such as the instructionist Lamarckianprocessesof use and disuse - were involvedin theemergence of adaptedorganicstructuresandbehaviors .


325

In contrast, this secondDarwinian revolution cannot be limited to any one significant event. Its roots lie even deeperthan the theory of organic evolution. It involves parallel and simultaneousdevelopmentsin many different disciplines, in much the sameway that different speciesevolvesimultaneously in conformance with the demands of their particular environments. The secondrevolution also appearsto be moving in an exclusively selectionistdirection in accountingfor the emergenceof truly novel knowl edge. And although selectionismcurrently rules in evolutionary biology and immunology, it remainsvery much a minority viewpoint in the other fields surveyedin the precedingchapters. But if the history of sciencehas anything at all to tell us, it is that our current theorieseventuallybecomelessthan satisfactory. Theoriesthat seemto be well founded and clear improvementsover previous onesare eventually seenasinadequateand are replacedby newer, more encompassingperspectives in the way that Newtonian physics gave way to Einstein's relativity and Bohr's quantum mechanics. This continued development in science dependson the relentlesscriticism of both currently acceptedand newly proposedtheoriesin the form of continuing efforts to discoverhow they are inadequateand how they can be improved. From this historical perspective, it would appearhighly unlikely for selectionism to be the final explanation for the emergenceof all puzzlesof fit . Although it has already endured for a remarkably long period of time in evolutionary biology, we would normally expect it eventually to be surpassed. Newer and better theoriesmight then explain everything in a given field that selectiontheory can, plus other phenomenathat it cannot, even if at this moment it is exceedingly difficult to imagine what such a replacementwould look like. That such a theory would have to be selected among competing ones (including selectiontheory itself) posessomewhat of a paradox to thosewho want to overturn selectionistaccountsof knowl edgegrowth.3O But since universal selectiontheory is itself a blind variation, although one that has so far resistedthe arguments that have been fatal to many providential and instructionist accountsof adaptedcomplexity, I must welcome the criticism and selectionpressurethat this book and its thesiswill undoubtedly provoke. Such scrutiny is absolutely necessaryto lead us to understandingsthat may go beyond selectionism, if indeedthat is possible.

326

Without Miracles

I do hope, however, that such criticism will not be basedon faulty understandingsof universalselectiontheory and the claims beingmadefor it , and consequentlyresult in a return to providential and instructionist theories that have been shown to be inadequate to the job of accounting for the emergenceof adaptedcomplexity. In the meantime a strong casecan be made that universal selectiontheory providesthe bestexplanation for both naturally and artificially produced puzzlesof fit . It relieson patient, iterative cyclesof blind variation and selection that over the course of time can result in biological adaptations and new species , functional human cultures, technological breakthroughs, and scientific revolutions. And what is perhapsmost appealing from the naturalist perspectiveof modern science, it provides this explanation without miracles- except for the illusory miracle of how such an inherently blind, stupid, wasteful, and sluggishprocesscan be found at the very foundation of life, its marvelousdesign, and all the subsequentknowledgethat life in its human form has generated.

.a..a. Appendix TheTroubleWith Miracles

Throughout this book miraculous accountsof the origin of adaptive complexity were discounted and natural, scientific ones advocated in their stead. But what exactly is the problem with miracles?Why should miracles and miraculous explanations be rejectedin sciencein favor of nonmiraculous ones? First, we have to define what we mean by miracle. In everyday usage, there seemto be two meanings. The first refers to an event that is considered to have a very low probability of occurrence, but that happens nonetheless . A man and his wife both winning the grand prize of a statelottery on successiveweeksmight be consideredsuchan event. But, of course, there is really nothing miraculous about such happenings, since they are possible and their rarity is consistentwith their low probability. We have also seenhow the processof cumulative variation and selectioncan dramatically increase the probability of adaptive evolutionary change that would be highly unlikely to emergein a singlestep. The secondmeaningof miracle hasto do with eventsthat violate the laws of nature. As defined by the Oxford English Dictionary a miracle is a marvelouseventoccurringwithin humanexperience , which cannothavebeen broughtaboutby humanpoweror by theoperationof anynaturalagency , andmust thereforebeascribedto thespecialinterventionof theDeityor of somesupernatur al being; chiefly, an act (e.g. of healing) exhibitingcontroloverthe lawsof nature, and servingas evidencethat the agentis eitherdivineor is speciallyfavouredby God. Abraham Lincoln coming back to life and walking the streetsof Springfield, Illinois , would be such an event. So would the creation of all the earth's plants and animals in a single day. Sucheventsare thought to be the work

328 Without Miracles of God or some other powerful supernatural circumvent the usual laws of nature . The

ever - skeptical

induction

was presented

rather , belief standing

in

. The

occurred

only

that

The plain

in

first

philosopher

in chapter

miracles

published

arguments

dence

Scottish

, in

if the evidence

consequence

tual destruction

that

is " That

no testimony

them . Hume would

of miracles

, or

Human

subject that

is sufficient

Under -

involved

two

a miracle

had

than

to establish

the evi -

a miracle ,

only gives us an assurance

violation

Lincoln

coming

be rational

suitable

the inferior ." l

she saw the sixteenth

friend

president

. Now , this would

from of the appear

of the laws of nature as we understand

advise you to consider

be more of a miracle

which

would

be more miracu -

back to life , or your friend

being either

believe the less miraculous

to believe in Lincoln 's resurrection

. In this

only if it would

that your friend could have either been mistaken

or

her report .

The second part of Hume 's argument

consists of four

actual practice there could never be compelling though

of

that an old and trustworthy

or deceitful - and accordingly

in making

this

critique

was stronger

the streets of his hometown

to be an event in flagrant

deceitful

a critique

to believe

remains , after deducting

told you that yesterday

United States walking

case , it would

, whose

to

it :

of arguments , and the superior

lous - Abraham

on

it had occurred

To give an example , imagine

mistaken

Hume

has the power

Concerning

reasoning

put

that

be of such a kind , that its falsehood would be more miraculous , it endeavours to establish ; and even in that case there is a mu -

to that degree of force , which

Springfield

Enquiries

it was rational

it had not . As Hume

unless the testimony than the fact , which

David

6 , also offered

his

1748 . His

was that

entity

evidence for a miracle , even

evidence for one could exist in principle

dence against it . Only

one of these reasons will

having to do with the unreliability

reasons why in

that outweighed be mentioned

of human reports of miraculous

the evi here , that events :

. . . there is not to be found , in all history , any miracle attested by a sufficient num ber of men , of such unquestioned good sense, education , and learning , as to secure us against all delusion in themselves; of such undoubted integrity , as to place them beyond all suspicion of any design to deceive others ; of such credit and reputation in the eyes of mankind , as to have a great deal to lose in case of their being detected in any falsehood ; and at the same time , attesting facts performed in such a public manner and in so celebrated a part of the world , as to render the detection unavoid able: All which circumstances are requisite to give us full assurance in the testimony of men .2

The Trouble With Miracles

329

In other words, Hume insiststhat you simply cannot trust the testimony of your friend from Springfield, at leastnot when reporting a miracle, sinceit is always more likely that her testimony is unreliable. And such is the case not only for your friend but for any individual or group of individuals pro viding an account of any miraculous event.

It should be madeclear that Hume doesnot discount the possibility that miracles may have in fact occurred, but only that it is never rational to believethat one has occurred. Neither doesanything in his argument suggestthat we might not be happier if we did believein miracles, such asthat the world and everythingin it were made by a benevolentcreator who continues to provide for our welfare. But a stubborn belief in miraculous accounts of events for which we have nonmiraculous

accounts is inconsistent

with the scientific enterprise that involves the continual search for the sim-

plest and most parsimoniousexplanationsof the goings-on of the universe. It is of interest that CharlesDarwin , certainly no philosopher, expressed very much the sameconclusion as Hume in his autobiography where he recountshis gradual loss of Christian faith : By further reflecting that the clearestevidencewould be requisite to make any sane man believein the miracles by which Christianity is supported,- that the more we know

of the fixed laws of nature the more incredible

do miracles

become ,- that the

men at that time were ignorant and credulousto a degreealmost incomprehensible by us,- that the Gospelscannot be proved to havebeenwritten simultaneouslywith the events,- that they differ in many important

details , far too important

as it

seemedto me to be admitted as the usual inaccuraciesof eyewitnesses ;- by such reflectionsasthese, which I give not the leastnovelty or value, but asthey influenced me, I gradually came to disbelievein Christianity as a divine revelation. The fact that many false religions have spreadover large portions of the earth like wild -fire had someweight with me. Beautiful as is the morality of the New Testament, it can hardly be deniedthat its perfection dependsin part on the interpretation which we now put on metaphorsand allegories.3

To the remote desertdweller witnessingthe effectsof refrigeration for the first time , seeing water turn into a solid may indeed appear to be a miracle ,

a remarkable violation of a very basic law of nature (that water is always fluid ), which could only have beenaccomplishedby a supernaturalpower. But science has a different

view , since scientific

progress would

cease if it

proposed and retained theories that could at any time be violated by miracles, instead of proposing theories and then rejecting them (and replacing them with better ones) when their predictions are in error, or when more powerful and parsimonioustheoriesare generated.

330

WithoutMiracles

This is why, in our scientific endeavorto make senseof the emergence of adapted complexity, we have no choice but to be rational and seek nonmiraculous explanations of the type that universal selection theory provides.

Without Miracles: Universal Selection Theory and the Second Darwinian Revolution

Darwinian Populations and Natural Selection

Evolutionary game theory, natural selection, and Darwinian dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

The Second Information Revolution

Universal Algebra, Second Edition

Universal Algebra and Lattice Theory

Universal Algebra and Lattice Theory

The Theory of Cultural and Social Selection

Darwinian Psychiatry

Second Language Acquisition and Universal Grammar

Miracles

Don't Start the Revolution Without Me!

Don't Start the Revolution Without Me!

Philosophy and the Darwinian Legacy

Miracles

Miracles

Miracles

Don't Start the Revolution Without Me!

Miracles

Miracles

Darwinian heresies

Miracles

Miracles

Miracles

Miracles

Miracles

Fractals and Universal Spaces in Dimension Theory

Without Miracles: Universal Selection Theory and the Second Darwinian Revolution

Darwinian Populations and Natural Selection

Evolutionary game theory, natural selection, and Darwinian dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

Evolutionary Game Theory, Natural Selection, and Darwinian Dynamics

The Second Information Revolution

Universal Algebra, Second Edition

Universal Algebra and Lattice Theory

Universal Algebra and Lattice Theory

The Theory of Cultural and Social Selection

Darwinian Psychiatry

Second Language Acquisition and Universal Grammar

Miracles

Don't Start the Revolution Without Me!

Don't Start the Revolution Without Me!

Philosophy and the Darwinian Legacy

Miracles

Miracles

Miracles

Don't Start the Revolution Without Me!

Miracles

Miracles

Darwinian heresies

Miracles

Miracles

Miracles

Miracles

Miracles

Fractals and Universal Spaces in Dimension Theory

Recommend Documents