You are on page 1of 39

St yl e, Funct i on, and Cul t ur al Evol ut i onar y Pr ocesses

Rober t Bet t i nger


Ant hr opol ogy Depar t ment
Uni ver si t y of Cal i f or ni a- - Davi s

Rober t Boyd
Depar t ment of Ant hr opol ogy
Uni ver si t y of Cal i f or ni a
Los Angel es, CA 90024

Pet er Ri cher son
Di vi si on of Envi r onment al St udi es
Uni ver si t y of Cal i f or ni a
Davi s, CA 95616




Ver si on 4. 4 Oct ober , 1994

Introduction

When expl ai ni ng human behavi or , ant hr opol ogi st s
f r equent l y di st i ngui sh t he t hi ngs t hat peopl e do of t hei r own
f r ee wi l l f r omt he t hi ngs t hey do because t hey have t o. I n
much of ant hr opol ogy, and most Amer i can ar chaeol ogy, t hi s i s
t he di f f er ence bet ween style and function. Funct i onal
behavi or s ar e t he t hi ngs peopl e ar e const r ai ned t o do;
st yl i st i c behavi or s ar e t he t hi ngs peopl e do when
unconst r ai ned. Wher e necessi t y st ops and f r ee choi ce begi ns
i s, of cour se, a cl assi c pr obl emof soci al sci ence t heor y, but
wher ever t he boundar y i s pl aced, i t i s gener al l y i mpl i ed t hat
t he domai ns t hus di vi ded ar e not of equal i mpor t ance
( Bet t i nger 1991: 49- 50) . Few st r addl e t hi s f ence: Mat er i al i st s
emphasi ze f unct i on and downpl ay st yl e; st r uct ur al i st s and
post moder ni st s do t he opposi t e. Recent at t empt s t o appl y neo-
Dar wi ni an concept s t o t he ar chaeol ogi cal r ecor d pr edi ct abl y
si de wi t h mat er i al i st t r adi t i on, r epeat i ng t he pr emi se t hat i t
i s most i mpor t ant t o expl ai n f unct i onal behavi or ; st yl i st i c
behavi or i s i nt er est i ng onl y f or l ocal i zi ng soci al uni t s i n
t i me and space.

Any at t empt t o cr eat e a r i gi d boundar y bet ween st yl e and
f unct i on wi l l f ai l . For exampl e, t he at t empt t o use f r ee wi l l
as a di st i nct i on f ounder s on t he f act t hat conf or mi ng t o
st yl i st i c convent i ons of speech, dr ess and bel i ef i s
f r equent l y compul sor y and al most al ways sanct i oned. We may
of t en have mor e oppor t uni t i es f or f r ee choi ce of mundane
ut i l i at r i an obj ect s. St yl e of t en has f unct i ons, and t he most
basi c f unct i ons- - eat i ng, def acat i ng, and havi ng sex- - ar e
usual l y done i n st yl e.

I n t hi s paper we ar gue t hat mat er i al i st s f or f ei t t oo much
when t hey di smi ss t he i mpor t ance of st yl e f or humans. Human
st yl i st i c behavi or over t he l ast f or t y t housand year s i s
ext r aor di nar i l y ext ensi ve and el abor at e. Thi s i s cr i t i cal
because f or neo- Dar wi ni ans compl ex, r i chl y- st r uct ur ed f or ms
al ways si gnal t he oper at i on of nat ur al sel ect i on or related
evolutionary forces. Ther e ar e si mpl y no known mat er i al
pr ocesses except nat ur al sel ect i on, and anal ogous evol ut i onar y
f or ces i n t he cul t ur al r eal m, capabl e of account i ng f or
phenomena t hat appear t o be " desi gned" ( see Dawki ns [ 1986] f or
an excel l ent i nt r oduct or y summar y of t he adapt at i oni st f or mof
t hi s ar gument ) . Fr omt hi s vi ew, ar t obj ect s, l anguages, and
super nat ur al i deol ogi es seemas much t o be t he pr oduct of
evol ut i onar y pr ocesses as subsi st ence t echnol ogy and cannot be
i gnor ed.

Wi t h t he advent of unambi guous st yl i st i c f eat ur es i n t he
ar cheol ogi cal r ecor d at t he Upper Pal eol i t hi c Tr ansi t i on,
subsi st ence st r at egi es al so i mpr oved and popul at i ons of humans
j umped (
and Gambl e, 1993) . We ar gue t hat t hi s coi nci dence i s not
acci dent al . St yl e has f unct i ons. The st yl e- f unct i on di chot omy
embr aced by mat er i al i st s and non- mat er i al i st s al i ke obscur es
under st andi ng of t he f undament al pr ocesses t hat gener at e human
behavi or .

Tr eat i ng st yl e and f unct i on as a di chot omy ar i ses f r oman
over si mpl i f i ed pi ct ur e of evol ut i onar y pr ocesses. I n ani mal s,
st yl e ar i ses due t o sexual sel ect i on and per haps mor e
gener al l y t o soci al sel ect i on ( West - Eber har t , 1983) .
Evol ut i onar y bi ol ogi st s si nce Dar wi n have engaged i n a compl ex
debat e about t he f unct i onal i t y of pl ant and ani mal st yl e.
However , t hi s debat e has har dl y ever had t he char act er
di vi di ng sexual l y sel ect ed t r ai t s of f i nconsequent i al . The
moder n debat e has f ocussed on whet her st yl e count er - f unct i onal
( t he r unaway hypot hesi s) or whet her i t i s an i ndex of t he
over al l f i t ness of a pot ent i al mat e ( t he handi cap hypot hesi s) .
Too much t i me and ef f or t go i nt o st yl e f or i t t o be neut r al !

Human cul t ur al st yl es cannot be expl ai ned wi t hout
under st andi ng t he cul t ur al anal ogs of t he sexual sel ect i on
mechani sm. The convent i onal st yl e f unct i on di chot omy i s a
r esul t of not t aki ng account of t he var i et y of t hese f or ces i n
cul t ur al syst ems, and how t he r el at e t o t he act i on of nat ur al
sel ect i on of adapt at i ons. Human cul t ur e i s i nf l uenced by a
compl ex of evol ut i onar y f or ces t hat ul t i mat el y der i ve f r omt he
oper at i on of nat ur al sel ect i on, but whi ch have pr oxi mal
pr oper t i es t hat di f f er subst ant i al l y f r omi t ( Campbel l , 1965;
Boyd and Ri cher son, 1985) . Sever al of t hese i nvol ve t he
choi ces of cul t ur al mat es and par ent s and ar e l i ke mat e choi ce
sexual sel ect i on. Ot her s ar e per haps mor e di r ect st and- i ns f or
nat ur al sel ect i on t hr ough psychol ogi cal pr edi sposi t i ons
( Cosmi des and Tooby, 1992) . When such " r el at ed evol ut i onar y
f or ces" ar e t aken i nt o account t he r i gi d di st i nct i on bet ween
st yl e and f unct i on di ssol ves. St yl i st i c var i at i on r esponds t o
a compl ex of r andomand di r ect i onal evol ut i onar y f or ces and
can ser ve i mpor t ant f unct i ons pr eci sel y because i t i s
ar bi t r ar y and symbol i c. Expl i ci t evol ut i onar y model s of
st yl i st i c var i at i on cl ar i f y our under st andi ng of st yl e and i t s
r el at i on t o f unct i on and cul t ur e hi st or y, and i mpr oves our
under st andi ng of t he pat t er ns st yl e mi ght l eave i n t he
ar chaeol ogi cal r ecor d.

I n t he f i r st par t of t hi s paper , we r evi ew r ecent debat es
on t he i mpl i cat i ons of st yl i st i c behavi or f or ar chaeol ogy.
Then we anal yze t he l i mi t at i ons i nher ent i n assumi ng t hat
evol ut i onar y pr ocesses can be col l apsed i nt o sel ect i on act i ng
on f unct i onal at t r i but es and r andomef f ect s act i ng on
st yl i st i c var i at i on. Fi nal l y, we ar gue t hat r ecent advances i n
t he t heor y of cul t ur al evol ut i on pr ovi de a r easonabl e account
of t he pr ocesses t hat af f ect t he compl ex of st yl i st i c and
f unct i onal of cul t ur e.


Style and Function in Neo-Evolutionary Perspective


Style versus function

" Nat ur al sel ect i on i s t he pr i mar y expl anat or y mechani sm
i n {Bob, t ypo soci al ?} sci ent i f i c evol ut i on. . . . St yl e and
f unct i on ar e def i ned i n t er ms of nat ur al sel ect i on. Because of
t he di st r i but i onal ent ai l ment s of nat ur al sel ect i on, each has
a di st i nct i ve, whol l y pr edi ct abl e di st r i but i on i n t he
ar chaeol ogi cal r ecor d. " Dunnel l 1980: 49, 88.

" St yl e denot es t hose f or ms t hat do not have det ect abl e
sel ect i ve val ues. Funct i on i s mani f est as t hose f or ms t hat
di r ect l y af f ect t he Dar wi ni an f i t ness of t he popul at i ons i n
whi ch t hey occur . . . The di chot omy i s mut ual l y excl usi ve and
exhaust i ve. " Dunnel l 1978: 199.

" St yl i st i c var i at i on i s sel ect i vel y neut r al . Hence
st yl i st i c t r ai t s ar e sor t ed st ochast i cal l y: by dr i f t . " Nei man
1993: 1.

The posi t i on of Dunnel l ( 1978, 1980) , a pr omi nent
ar chaeol ogi cal exponent of t he r i gi d st yl e- f unct i on di cht omy,
ser ves as an excel l ent st ar t i ng poi nt f or our count er -
ar gument . He ar gues t hat t he di st i nct i on bet ween st yl e and
f unct i on i n ant hr opol ogy i s essent i al l y bet ween behavi or s t hat
ar e subj ect t o pr ocessual expl anat i on and behavi or s t hat ar e
not . Because Dunnel l advocat es a neo- Dar wi ni an vi ew of
pr ocess, f or hi mfunctional r ef er s t o t hi ngs expl i cabl e as
adapt at i ons due t o nat ur al sel ect i on and style means,
ef f ect i vel y, af unct i onal or neut r al - - t hi ngs wi t hout di r ect
posi t i ve or negat i ve sel ect i ve val ue.


Dunnel l ar gues t hat because t hey ar e f r ee of sel ect i ve
const r ai nt , st yl i st i c t r ai t s wi l l var y st ochast i cal l y, much
l i ke adapt i vel y neut r al t r ai t s i n bi ol ogy. He not es t hat t hese
pr oper t i es make st yl e especi al l y appr opr i at e as measur es of
t i me ( e. g. , i n ser i at i on) , soci al i nt er act i on, and cul t ur e
hi st or y ( cf . Nei man 1993) . As wi t h ot her t echni ques t hat
empl oy pr esumabl y neut r al t r ai t s t o measur e descent
r el at i onshi ps ( e. g. , non- codi ng DNA r esembl ances,
l exi cost at i st i cs) , shar ed f eat ur es of st yl e ar e t aken t o be
homol ogous si mi l ar i t i es r ef l ect i ng common cul t ur al her i t age.
Dunnel l obser ves one compl i cat i on i n t hi s si mpl e st yl e-
f unct i on di st i nct i on: i f a f unct i onal r equi r ement admi t s
al t er nat i ve sol ut i ons, t he same t r ai t can be st yl i st i c
( neut r al ) and f unct i onal , dependi ng on t he l evel ( scal e) at
whi ch i t i s def i ned. Var i ant s of a f unct i onal t r ai t ( e. g. , Z-
t wi st and S- t wi st cor dage) may be neut r al wi t h r espect t o
each ot her , even t hough al l ar e ut i l i t ar i an.

Dunnel l ' s def i ni t i on of st yl e and f unct i on i s wi del y
accept ed by t hose i nt er est ed i n appl yi ng evol ut i onar y
pr i nci pl es t o t he ar chaeol ogi cal r ecor d. Most mat er i al i st
ar chaeol ogi st s agr ee t hat f unct i onal f eat ur es wi l l be non-
r andoml y pat t er ned as a r esul t of sel ect i on, and f eat ur es of
st yl e wi l l be mer el y st ochast i c. Thi s i s f r equent l y r ead as
meani ng t hat t he f or mer ar e subj ect t o pr ocessual expl anat i on,
t he l at t er ar e not ( e. g. Ki r ch 1980, Leonar d and J ones 1987,
O' Br i en and Hol l and 1992; Nei man 1993) .

I n por t r ayi ng ar t and st yl e - - t he t hi ngs ant hr opol ogi st s
have hi st or i cal l y i dent i f i ed as di st i nct i vel y " cul t ur al " - - as
beyond t he r each of neo- Dar wi ni an expl anat i on, Dunnel l
ar t i cul at es a t r adi t i onal t enet of mat er i al i st ant hr opol ogi cal
i nqui r y, expr essed f i r st i n moder n f or mby St ewar d ( 1938) , and
subsequent l y i n i ncr easi ngl y ext r eme f or m, by ear l y cul t ur al
mat er i al i st s ( Har r i s 1968) , neof unct i onal i st s ( Vayda and
Rappopor t 1967) , New Ar chaeol ogi st s ( e. g. , Bi nf or d 1962; cf .
Whi t e 1959) , and, most r ecent l y, by human evol ut i onar y
ecol ogi st s ( e. g. O' Connel l , J ones, and Si mms 1983) . Dunnel l
si mpl y oper at i onal i zes t he t r adi t i onal ar gument t hr ough t he
pr edi ct i on t hat ar t and st yl e al ways pat t er n r andoml y.


Style as Style

" Ever y st yl e i s necessar i l y pr el i mi t ed. . . The r ange of i t s
channel ed ski l l s wi l l ext end so f ar ; beyond t hey f ai l . Then we
say t hat t he st yl e has exhaust ed i t sel f , i t s char act er i st i c
pat t er n has br oken down. . . I t i s commonpl ace t hat al l aest het i c
st yl es, r i se and f al l and per i sh. " Kr oeber ( 1948: 329- 330)

Many ant hr opol ogi st s, of cour se, di sput e t he mat er i al i st
account of st yl e and t he nat ur e of st yl i st i c change. Kr oeber
( 1948) was one of many f or whomst yl i st i c change was di r ect ed
r at her t han r andom. Kr oeber vi ewed st yl es as basi c t hemes
( anal agous t o st yl es i n ar t or musi c) upon whi ch cul t ur es
el abor at ed. Because i t seemed i nconcei vabl e t o hi mt hat such
el abor at i on coul d cont i nue beyond a cl i max i n whi ch t he
possi bi l i t i es i nher ent i n t he st yl e wer e exhaust ed, Kr oeber
bel i eved t hat st yl i st i c change f ol l owed a non- r andom
hi st or i cal t r aj ect or y, a posi t i on t hat Dunnel and ot her s
expl i ci t l y r ej ect . Cl ai ms of t hi s ki nd ( Sahl i ns, 1976) ar e
of t en pr esent ed as al t er nat i ves t o t he f unct i onal account
st yl e, but t hi s i s not necessar y. Even among Kr ober ' s l engt hy
r umi nat i ons, one can f i nd t he ker nel of an i dea r econci l i ng
t he vi ew t hat st yl e shapes and const r ai ns cul t ur al change wi t h
t he f unct i onal i st vi ew t hat cul t ur e i s adapt i ve.

" . . . For t hi ngs t o be done wel l t hey must be done def i ni t el y,
and def i ni t e r esul t s can be achi eved onl y t hr ough some
speci f i c met hod, t echni que, manner , or pl an of oper at i ons.
Such a par t i cul ar met hod or manner i s cal l ed a st yl e i n al l
t he ar t s. . . . A st yl e. . . may be sai d t o be a way of achi evi ng
def i ni t eness and ef f ect i veness i n human r el at i ons by choosi ng
or evol vi ng one l i ne of pr ocedur e out of sever al possi bl e ones
and st i cki ng t o i t " ( Kr oeber 1939: 329)

Sacket t ( 1982, 1985) has mor e r ecent l y l abel ed such
behavi or isochrestic: pat t er ned behavi or r ef l ect i ng
essent i al l y ar bi t r ar y choi ces bet ween essent i al l y f unct i onal l y
equi val ent ways of doi ng t hi ngs. Thi s i s Dunnel l ' s scal e
ef f ect - - speci f i c var i abi l i t y i n t r ai t s t hat equal l y sat i sf y
t he same adapt i ve f unct i on. Dunnel l , Sacket t , and Kr oeber ,
t hen, al l seemt o agr ee t hat j ust how one ski ns a cat can be
f unct i onal l y l ess i mpor t ant t han t he f act t hat one ski ns i t at
al l . They di sagr ee f undament al l y, however , i n what t hi s
i mpl i es about t he mechani sms dr i vi ng t he hi st or i cal
t r aj ect or i es of al t er nat e var i ant s of cat - ski nni ng.

For Kr oeber ( as exempl i f i ed by t he l at t er quot e) and
Sacket t i sochr est i c var i at i on i s mai nl y a pr oduct of t he
f or mal i zat i on and r out i ni zat i on of t echni que, whi ch makes t he
t r ansmi ssi on of t he knowl edge about how t o make a compl ex
obj ect easy t o i mi t at e, r emember , and execut e. These benef i t s
evi dent l y r esul t ed when si mpl er f or ms of i ndi vi dual and soci al
l ear ni ng wer e r epl aced by ar bi t r ar y convent i ons t hat
st r eaml i ned acqui si t i on by cul t ur al t r ansmi ssi on and
coor di nat ed compl i cat ed cul t ur al behavi or s. Once cr af t smen
become ski l l ed at maki ng and usi ng a t ool one way, t hey may
r at i onal l y r esi st change because of l ear ni ng cost s. Thi s i s
consi st ent wi t h f or mal model s suggest i ng i t pays t o r et ai n a
subopt i mal t ool when sear chi ng f or t he opt i mal al t er nat i ve i s
cost l y or er r or pr one ( Si mon 1959, Hei ner 1983, Boyd and
Ri cher son 1992a) . At t he same t i me, as Kr oeber and
Ri char dson' s ( 1940) cl assi c paper on dr ess st yl e shows, t her e
i s not hi ng i n t he concept of i sochr est i c var i at i on t hat deni es
t he possi bi l i t y of non- r andomhi st or i cal t r aj ect or i es of
change.

Ot her s wor ki ng wi t h st yl e f i nd i sochr est i c var i at i on
met hodol ogi cal l y pr obl emat i c and f avor st yl i st i c i nqui r i es
t hat emphasi ze i coni c and symbol i c var i at i on. Bi nf or d
( 1989: 52- 53) sees t he st yl e- f unct i on di chot omy as an
opposi t i on bet ween consci ous, expl i ci t l y- r at i onal , pr obl em-
sol vi ng behavi or , on t he one hand, and unconsci ous, r ot e-
l ear ned mot or habi t s and soci al l y or symbol i cal l y- mot i vat ed
behavi or , on t he ot her . Wi t hi n t he l at t er , he evi dent l y now
f ol l ows Wei ssner ( 1985: 162) i n equat i ng i sochr est i c var i at i on
wi t h t he unconsci ous or r ot e- l ear ned mot or habi t s ( Bi nf or d
1989: 56, 58) . Because i t i s al ways possi bl e t hat what appear t o
be i sochr est i c var i ant s connect ed wi t h i ndi vi dual s or et hni c
uni t s act ual l y have f unct i onal si gni f i cance, Bi nf or d bel i eves
t hat when def i ni ng act or s or act or - gr oups i n t he
ar chaeol ogi cal r ecor d i t i s saf est t o f ocus on t he most
obvi ous sor t s non- f unct i onal var i at i on r el at ed t o soci al or
i deol ogi cal behavi or . The cl oser a st yl e i s t o pur el y
symbol i c, t he l ess l i kel y i t i s t o be f unct i onal , i . e. ,
pat t er ned by r at i onal choi ce. Backed and sel f bows ar e at some
l evel f unct i onal l y equi val ent but t hi s equi val ence i s not
guar ant eed so gener al l y as woul d be t he f unct i onal equi val ence
of al t er nat i ve geomet r i c desi gns pai nt ed on t hem.

Bi nf or d and Wi essner have di scussed st r at egi es f or
wor ki ng wi t h f or mal var i at i on i n mat er i al cul t ur e t hat
consci ousl y t r ansmi t s i nf or mat i on about soci al or per sonal
i dent i t y. The emphasi s her e i s on t he use of t he var i at i on as
embl ems or i cons of soci al or pol i t i cal gr oups ( embl emat i c and
i conol ogi cal st yl e) or expr essi ons of i ndi vi dual i dent i t y
wi t hi n such gr oups ( asser t i ve st yl e) . I n t hese cases,
var i at i on i s sai d t o be pur el y symbol i c but ser ves a f unct i on
( communi cat i on) and i s sur el y not r andom. Wei ssner ( 1985: 162;
and evi dent l y Bi nf or d 1989: 54- 55) ar gue t hat because soci al or
i deol ogi cal st yl i st i c var i at i on i s mani pul at ed t o sui t
changi ng soci al and i ndi vi dual cont ext s, i t shoul d var y
subst ant i al l y t hr ough t i me and space, i n cont r ast t o
i sochr est i c var i ant s r esul t i ng f r omst r eaml i ned cul t ur al
deci si on- maki ng, whi ch ar e st abl e once est abl i shed. As j ust
not ed, however , t he i sochr est i c concept does not r equi r e t hi s
and, as Dunnel l ar gues, i sochr est i c choi ces made by
i ndi vi dual s can cer t ai nl y gi ve r i se t o behavi or al change at
bot h t he i ndi vi dual and popul at i on l evel s.

I n t he mai n, Dunnel l di smi sses t he r el evance of t hese
di st i nct i ons f or t he ar chaeol ogi cal r ecor d. I n cont r ast t o
Sacket t , f or whomcul t ur al t r ansmi ssi on st r eaml i nes ( hence
const r ai ns) deci si on- maki ng, f or Dunnel l cul t ur al t r ansmi ssi on
i s adapt i ve f or t he opposi t e r eason: i t br oadens access t o
behavi or al al t er nat i ves. I t i ncr eases t he amount of
f unct i onal l y si gni f i cant var i at i on f r omwhi ch i ndi vi dual s can
choose and, t hus, upon whi ch sel ect i on can act . Thi s i ncr eases
t he speed and r ange of adapt i ve r esponses r el at i ve t o si mpl e
genet i c t r ansmi ssi on, wher e popul at i on var i at i on i s mor e
f i ni t el y const r ai ned by such t hi ngs as gener at i on l engt h,
mut at i on r at es, and exi st i ng genet i c var i at i on ( Dunnel l 1978:
198) . Sel ect i on sor t s ( hence pat t er ns) t he f unct i onal t r ai t s,
l eavi ng st yl i st i c t r ai t s, symbol i c and i sochr est i c, t o dr i f t
r andoml y. I f t hi s i s so, t he met hodol ogi cal and ont ol ogi cal
compl exi t i es of non- symbol i c i sochr est i c var i at i on t hat char ge
t he t heor et i cal debat e bet ween Sacket t , Bi nf or d, and Wei ssner
ar e empi r i cal l y uni mpor t ant . St yl e and f unct i on ar e mor e
cl eanl y di st i nct under t hi s assumpt i on, and non- st yl i st i c
var i at i on i s al ways f unct i onal l y si gni f i cant .

To summar i ze, i f our r evi ew of t he l i t er at ur e above i s
cor r ect , ar chaeol ogi st s i dent i f y t hr ee t ypes of ar t i f act
var i at i on: ( 1) f unct i onal var i at i on uncompl i cat ed by st yl i st i c
f eat ur es, ( 2) i sochr est i c var i at i on i n f unct i onal t r ai t s i n
whi ch var i ant s ar e qual i t at i vel y di f f er ent but near l y equal i n
f unct i on, and ( 3) i sochr est i c var i at i on i n i coni c and symbol i c
t r ai t s der i ved f r omar bi t r ar y, f unct i onl ess decor at i ve
el ement s. These t hr ee t ypes seemt o be vi ewed as poi nt s i n a
t wo di mensi onal space wi t h one di mensi on r epr esent i ng
f unct i on, r angi ng f r omcompl et el y f unct i onal t o compl et el y
i sochr est i c, and t he ot her di mensi on r epr esent i ng
communi cat i on, r angi ng f r om var i at i on t hat i s hi ghl y sal i ent
as expr essi ve or embl emat i c communi cat i on t o var i at i on t hat
communi cat es not hi ng and i s soci al l y i r r el evant . Not e t hat
al t hough t her e mi ght be some t endency f or speci f i c t r ai t s t o
l i e on a di agonal l i ne i n f i gur e 1, not hi ng pr event s t he
exi st ence of of f - di agonal cases, as Wi essner , Bi nf or d, and
Sacket t r epeat edl y not e. The upper r i ght of f i gur e 1, f or
exampl e, woul d accomodat e t he many known cases i n whi ch
st yl i st i c di spl ay i s cost l y ( e. g. , Cohen, 1974) or i n whi ch
f unct i onal di f f er ences ar e meani ngf ul as expr essi ve or
embl emat i c symbol s, as when past or al i st s t ake pr i de i n owni ng
cat t l e and despi se t hei r l i vest ock- poor f ar mi ng nei ghbor s. By
cont r ast , t he l ower l ef t of f i gur e 1 woul d i ncl ude cases wher e
f unct i onal l y neut r al var i at i on i s compl et el y i gnor ed f or
communi cat i on, as Sacket t supposes f or var i at i on i n San
pr oj ect i l e poi nt s.


hi gh 3 f unct i onal var i at i on
used communi cat i vel y,
cost l y symbol syst ems

Communi cat i ve
i mpor t ance 2
of var i at i on

f unct i onal l y neut r al
t r ai t s not used
communi cat i vel y 1
l ow +- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
l ow hi gh

Funct i onal i mpor t ance of var i at i on

Fi gur e 1. The t hr ee " t ypes" descr i bed i n t he t ext may r esul t
f r oma t endency f or act ual var i at i on t o cl ust er al ong t he
di agonal as i ndi cat ed by t he number s cor r espondi ng t o t he
number ed i t ems i n t he t ext . However , i t i s an empi r i cal mat t er
how t hi ckl y cases ar e scat t er ed of f t he di agonal .
Ther e i s l i t t l e agr eement about what Fi gur e 1 means f or
t he pat t er ns one mi ght expect t o f i nd i n t he ar chaeol ogi cal
r ecor d. Many comment at or s, r epr esent ed by Dunnel l , woul d
appar ent l y be comf or t abl e ar gui ng t hat var i at i on pr oj ect i ng on
t he f unct i on di mensi on wi l l be cont r ol l ed by sel ect i on, whi l e
t hat pr oj ect i ng on t he communi cat i ve di mensi on by r andom
pr ocesses. St r uct ur al i st s, r epr esent ed i n our br i ef r evi ew by
Kr oeber , i magi ne non- r andompr ocesses, but emphat i cal l y not
sel ect i on, t o be act i ng on var i at i on wi t h hi gh communi cat i ve
f unct i on, per haps l eavi ng onl y var i at i on near t he bot t oml ef t
of f i gur e 1 t o r andompr ocesses. I n cont r ast , Wi essner
( 1985: 162) and Bi nf or d ( 1989: 54- 55) ar gue var i ant s i n bot t om
l ef t of f i gur e 1 shoul d be hi ghl y st abl e and not subj ect t o
r andompr ocesses, a pat t er n Bi nf or d ( 1989: 54) ext ends t o many
hi ghl y- adapt i ve, non- symbol i c f unct i onal char act er s ( i . e. ,
var i at i on i n t he l ower l ef t of f i g. 1) . Sacket t ( 1986: 630- 631)
seemi ngl y r ej ect s t he not i on t hat i sochr est i c var i at i on wi l l
consi st ent l y conf or mt o any speci f i c pat t er n.


A Critique of the The Selectionist Program in Modern
Archaeology

We bel i eve t hat one cannot oper at i onal i ze t he model s of
st yl e pr esent ed by Dunnel l , Kr oeber , Sacket t , Bi nf or d, and
Wei ssner , much l ess di f f er ent i at e t hemwi t h r espect t o
pat t er n, because nowher e i n t hei r wr i t i ngs can one f i nd an
expl i ci t model of t he cul t ur al t r ansmi ssi on and " sel ect i on"
pr ocesses t hat gi ve r i se t o st yl i st i c var i at i on. Opl er ( 1964)
t ook Kr oeber t o t ask sever el y on t hi s count but t he ot her s
ment i oned her e ar e equal l y cul pabl e. Indeed, given the
importance contemporary archaeology places on model-building
it is remarkable these individuals have not been more severely
criticized on this point by processualists ( post -
pr ocessual i st s, of cour se, have not over l ooked t he pr obl em,
but t hei r cr i t i que i s beyond t he scope of t hi s di scussi on) .
The absence of cr i t i ci smi s sympt omat i c of a t endency of
cont empor ar y mat er i al i smt o r educe cul t ur al pr ocess t o
sel ect i on/ adapt at i on, wi t h t he i mpl i cat i on t hese ar e a
cl ear l y- under st ood, st r ai ght - f or war d pr ocesses. When t hi s
si mpl e sel ect i oni st account f ai l s, mat er i al i st s ar e much t oo
r eady t o abandon t he i nqui r y ( e. g. , Bi nf or d and Wei ssner ) or
r esor t t o expl anat i on i n t er ms of r andomf act or s ( e. g. ,
Dunnel l ) .

The pr obl em, i n shor t , i s t hat t he bel i ef t hat cul t ur al
var i at i on i s adapt i ve i s not suppl ement ed by a concer n wi t h
t he det ai l s of t he pr ocesses t hr ough whi ch adapt at i ons
act ual l y ar i se. Despi t e r ef er ences t o t he i mpor t ance of
cul t ur al t r ansmi ssi on ( e. g. , Dunnel l 1978: 198) , t he
adapt at i oni st s ar e suspi ci ous of model s of t hose pr ocesses.
The gr oundwor k was f or t hi s vi ew was l ai d by Whi t e and
subsequent l y expl i cat ed i n det ai l by Bi nf or d and ot her s i n t he
New Ar chaeol ogy movement . They exhor t ed ar chaeol ogi st s t o
abandon model s of cul t ur e as a syst emof i nher i t ance and
r epl ace t hemwi t h a model s of cul t ur e as adapt at i on.

" A nor mat i ve t heor i st i s one who sees as hi s f i el d of
st udy t he i deat i onal basi s f or var yi ng ways of human
l i f e. . . For adher ent s of t he nor mat i ve school , t he assumpt i ons
about uni t s or nat ur al " packages" i n whi ch cul t ur e occur s ar e
dependent upon assumpt i ons about t he dynami cs of i deat i onal
t r ansmi ssi on. . . The nor mat i ve vi ew l eaves t he ar chaeol ogi st i n
t he posi t i on of consi der i ng hi msel f a cul t ur e hi st or i an and/ or
pal eopsychol ogi st ( f or whi ch most ar chaeol ogi st s ar e poor l y
t r ai ned) . . . [ t her ef or e] a new syst emat i cs, one based on a based
on a di f f er ent concept of cul t ur e, i s needed t o deal
adequat el y wi t h t he expl anat i on of cul t ur e pr ocess. . . [ t hat
bei ng] . . . cul t ur e as man' s ext r asomat i c means of adapt at i on
[ Whi t e 1959: 8] . " Bi nf or d 1965: 203, 204, 205.

Dunnel l , Sacket t , and Bi nf or d ( and of cour se many ot her s)
evi dent l y bel i eve t hat i gnor i ng t he det ai l s of cul t ur al
t r ansmi ssi on i s j ust i f i ed because sel ect i on f avor s f ai t hf ul
cul t ur al r epr oduct i on.

" I f we acknowl edge t hat a cul t ur al syst emi s a syst emof
ext r asomat i c t r ansmi ssi on f or behavi or al l y r el evant
i nf or mat i on f r omone gener at i on t o t he next , t hen a cul t ur al
t r adi t i on i n i t s r epr oduct i ve mode woul d be most ef f ect i ve i f
t he t r ansmi ssi on of i nf or mat i on f r omone gener at i on t o t he
next i s exact and unchanged i n t he pr ocess. " Bi nf or d 1983:
222.

" I t hi nk i t shoul d be cl ear t hat di scussi ng a cul t ur al
syst emi n t er ms of . . . t he dynami cs of cul t ur al
r epr oduct i on. . . i s not l i kel y t o hel p us under st and t he
dynami cs of descent wi t h modi f i cat i on. " Bi nf or d 1983: 222 ( but
compar e Bi nf or d 1983: 221) .

For Bi nf or d ( and most ot her adapt at i oni st s) , Whi t e' s
def i ni t i on of cul t ur e as an ext r asomat i c means of adapt at i on
sel ect i vel y f avor s exact or near - exact cul t ur al t r ansmi ssi on:
t he i ni t i al phase of each new cul t ur al gener at i on i s a near -
dupl i cat e of t he t er mi nal adul t phase of t he pr ecedi ng
gener at i on, di f f er i ng onl y i n mi nor and r andomways, as
Dunnel l assumes. These aut hor s bel i eve t hat such f ai t hf ul
t r ansmi ssi on r ender s t r ansmi ssi on i nconsequent i al . To pr edi ct
human behavi or i n a par t i cul ar envi r onment , one needs onl y
det er mi ne what behavi or s ar e adapt i ve i n t hat envi r onment - - -
sel ect i on wi l l sor t t hi ngs out so t hat expl i ci t at t ent i on t o
t he pr ocess of adapt at i on i s unecessar y.

Exper i ence i n evol ut i onar y bi ol ogy suggest s t hat t hi s
vi ew i s al most cer t ai nl y wr ong. Adapt at i oni st t hi nki ng has
been ext r emel y usef ul evol ut i onar y bi ol ogy, and adapt at i oni st s
have of f er ed many usef ul i nsi ght s about t he behavi or of humans
and ot her speci es. However , compl et e dependence on adapt i oni st
t hi nki ng, par t i cul ar l y t he si mpl e ver si on used by Dunnel l ,
Bi nf or d and ot her s, f or f ei t s t he most power f ul el ement s of
Dar wi ni an t hi nki ng because, cont r ar y t o t he vi ew wi despr ead i n
ant hr opol ogy, t her e i s mor e t o Darwinism than natural
selection. Sel ect i on i s j ust one of sever al Dar wi ni an
pr ocesses, and " nat ur al sel ect i on" i t sel f i s a het er ogenous
compl ex of pr ocesses, some whi ch do not pr oduce adapat i ons i n
any i nt ui t i ve sense.

Mor e f undament al l y, Dar wi ni smi s not a l i st of pr ocesses
or der ed by r el at i ve i mpor t ance; i t i s a met hodol ogy gui ded by
t he cent r al assumpt i on t hat t he key t o under st andi ng evol ut i on
i s good bookkeepi ng. Even t he si mpl est evol ut i onar y f or ces
i nt er act at sever al l evel s i n compl ex ways, and t o under st and
t hese compl exi t i es r equi r es concer t ed at t ent i on t o account i ng
f or how some i nher i t ed var i at i on r esponds at t he popul at i on
l evel t o f or ces l i ke sel ect i on. I t i s t he i nt er act i on of t he
f or ces, at var i ous l evel s, t hat i s of i nt er est .

Sexual sel ect i on r ef er r ed t o al r eady i l l ust r at es t he
i mpor t ance of keepi ng car ef ul account s of t he i nt er act i on of
mul t i pl e pr ocesses. Dar wi n bel i eved t hat exagger at ed,
pr esumabl y mal adapt i ve, mal e char act er s l i ke peacock t ai l s
ar ose because f emal es pr ef er r ed mal es wi t h such char act er s.
Dar wi n coul d not expl ai n, however , why f emal e pr ef er such
mal es. Thi s quest i on i s now bei ng hot l y debat ed by
evol ut i onar y bi ol ogi st s ( Bar t on and Tur el l i , 1991; Eber har d
1986, Maynar d Smi t h 1991, Ki r kpat r i ck 199x [ ARES r evi ew] ,
Pomi ankowski 1988) . The r un- away hypot hesi s, one of t he
compet i ng posi t i ons i n t hi s debat e, wi l l i l l ust r at e our poi nt .
I n t hi s vi ew f emal e pr ef er ence f or showy mal es i s a
pat hol ogi cal consequence of f emal e choi ce ( e. g. Lande, 1981) .
Suppose t her e ar e cr ypt i c mal es wi t h pr act i cal camof l age t ai l s
and showy mal es wi t h br i ght pr edat or - at t r act i ng t ai l s. I f
f emal es who pr ef er mal es wi t h showy t ai l s ar e suf f i ci ent l y
common, t hei r choi ces can i ncr ease t he f r equency of " showy-
t ai l genes" even t hough such t ai l s ar e ot her wi se
di sadvant ageous. Such choi ces wi l l al so cause t he genes t hat
gener at e a pr ef er ence f or showy t ai l s t o co- occur wi t h t he
genes t hat cause showy t ai l s. As a r esul t an i ncr ease i n t he
f r equency of " showy- t ai l " genes wi l l al so cause i ncr ease i n
t he genes t hat cause f emal es t o pr ef er showy mal es. Dur i ng t he
next gener at i on, sexual sel ect i on wi l l f avor showy mal es even
mor e st r ongl y, whi ch i n t ur n wi l l f ur t her i ncr ease t he
f r equency of f emal es who pr ef er showy mal es, f ur t her
i ncr easi ng t he st r engt h of sexual sel ect i on, and so on, unt i l
mal es become spect acul ar l y el abor at ed. A key pr obl emf or t he
r unaway hypot hesi s i s t he magni t ude of t he associ at i on bet ween
genes f or t ai l s and genes f or f emal e pr ef er ences, whi ch i n
t ur n depends on t he i nt er act i on bet ween nat ur al sel ect i on
act i ng on bot h set s of genes, sexual sel ect i on, and t he
mechani cs of l i nkage and r ecombi nat i on. I t i s si mpl y
i mpossi bl e t o under st and t hi s pl ausi bl e evol ut i onar y mechani sm
wi t hout det ai l ed model s t hat car ef ul l y t r ack t he net ef f ect s
of t hi s compl ex of i nt er act i ng pr ocesses ( Bar t on and Tur el l i ,
1991) . Nor i s t hi s a si ngul ar case, si mi l ar compl exi t i es ar e
conf r ont ed i n t he sexual sel ect i on debat e ( e. g. Hami l t on and
Zuk, 1982; Ryan, et al . 1990) , model s of speci at i on, model s of
t he evol ut i on of sex, r ecombi nat i on, mat i ng syst ems, t he
shi f t i ng bal ance t heor y, and a number of ot her cur r ent
pr obl ems i n evol ut i onar y bi ol ogy.

The l esson i s cl ear . The evol ut i onar y i nt er pr et at i on of
human behavi or , cont empor ar y and ext i nct , r equi r es
ant hr opol ogi st s t o const r uct expl i ci t model s of cul t ur al
pr ocesses and cal cul at e t he i mpl i cat i ons of t hose model s.
Bi nf or d has r el ent l essl y exposed t he f al l aci es of
i nt er pr et at i on t hat r esul t when we t r y t o i nt ui t t he meani ng
of ar chaeo- f aunas wi t hout f or mal model s t hat f or ce us t o keep
t r ack of var i ous f or mat i on pr ocesses act i ng at var i ous l evel s.
Dar wi n' s met hods of " popul at i on t hi nki ng" encour age a si mi l ar
at t ent i on t o t he det ai l s of how par t i cul ar var i ant s i ncr ease
or f ai l t o do so under t he i mpact of speci f i c envi r onment al
and soci al ef f ect s. I t i s qui t e cl ear t hat an evol ut i onar y
per spect i ve of cul t ur e pr ocess r equi r es model s of cul t ur al
t r ansmi ssi on t hat ar e anal yt i cal l y separ at e f r ommodel s of
ot her pr oceses t hat act on t hat f or mof var i at i on, such as
sel ect i on.

I n t hi s r egar d, cont empor ar y adapt at i oni st s i n
ar chaeol ogy t end t o f ol l ow t he i nt er pr et i ve t r adi t i on of
ant hr opol ogy, whi ch emphasi zes gener al i zat i ons about
consequences, r at her t han t he mor e pr ocess- or i ent ed t r adi t i on
of evol ut i onar y bi ol ogy ( Bet t i nger 1991) . The dat a of a
par t i cul ar ar chaeol ogi cal case ar e " expl ai ned" by means of
empi r i cal gener al i zat i ons about t he ar chaeol ogi cal r ecor d and
by ar gument s about t he l ar ger " meani ng" of t hose r ecor ds. For
cont empor ar y adapt at i oni st s, t he ar chaeol ogi cal r ecor d i mpl i es
t he over whel mi ng i mpor t ance of nat ur al sel ect i on. Thus
adapt at i on, l i ke pr ogr ess f or ear l i er schol ar s, i s used as t he
i nt er pr et i ve t ool t o di ssect and expl ai n a case at hand.

I n cont r ast , evol ut i onar y bi ol ogy devot es much of i t s
ef f or t t o st udyi ng t he act ual pr ocesses of evol ut i on.
Genet i cs, popul at i on genet i cs, and popul at i on ecol ogy ar e
most l y about t he pr ocessual i nner mechani cs of t he i nher i t ance
of var i at i on and i t s modi f i cat i on by t he popul at i on- l evel
i mpact of envi r onment . The adapt i ve i nt er pr et at i on of t he
st r uct ur e and behavi or of par t i cul ar or gani sms depends on t he
knowl edge we have about t hese pr ocesses, gai ned f r ommany
ki nds of st udi es of many ki nds of speci es. Somet i mes adapt i ve
i nt er pr et at i ons ar e f ai r l y obvi ous and don' t depend cr uci al l y
on a cl ose knowl edge of pr ocess, but t he opposi t e i s qui t e
of t en t r ue. I n t he case of sexual sel ect i on, f or exampl e, t he
debat e i s t i ght l y f ocused on t he det ai l s of model s of t he
sexual sel ect i on pr ocess and upon t he i nt er pr et at i on of dat a
( l ar ge scal e sur veys of bi r d col or at i on by Hami l t on and Zuk,
and of i nsect i nt r omi t ent or gans by Eber har dt ) . Rat her t han
use t he t heor y t o i nt er pr et cases, cases ar e used t o deci de
how t he t heor et i cal model s appl y. Onl y i f t hi s sear ch l eads t o
gener al concl usi ons do we obt ai n some war r ant f or a mor e
gener al i nt er pr et i ve st r at egy.


Processes of Stylistic Evolution

The t r adi t i onal def i ni t i on of st yl e r equi r es t hat
behavi or al var i ant s not be subj ect t o nat ur al sel ect i on.
St yl i st i c var i ant s must be neut r al wi t h r espect t o nat ur al
sel ect i on ( and sel ect i on- der i ved, adapt at i on gener at i ng,
deci si on- maki ng ef f ect s such as Boyd and Ri cher son [ 1985]
di scuss under t he headi ngs of bi as and gui ded var i at i on) . Thi s
def i ni t i on f ai l s t o do j ust i ce t o st yl i st i c var i at i on i n
t hr ee maj or ways. Fi r st , many i sochr est i c var i ant s of
ut i l i t ar i an ar t i f act s may be subj ect t o f r equency dependent
ef f ect s. When Qwer t y keyboar ds ar e common, i t i s sensi bl e t o
adopt t hem, even t hough r ar e keyboar ds ( l i ke Dvor ak) ar e
act ual l y bet t er . Sel ect i on i t sel f can mai nt ai n syl i st i c
het er ogenei t y. Second, pur el y symbol i c char act er s wi l l come t o
have f i t ness ef f ect s i f t hey become t he obj ect of choi ce, as
i n sexual sel ect i on. Thi r d, st yl i st i c var i at i on may be
cont r ol l ed by evol ut i onar y f or ces t hat gener at e non- r andom
pat t er ns, even i n t he neut r al case. ( I t i s per haps al so wor t h
ment i oni ng t hat nat ur al sel ect i on i n a r andoml y var yi ng
envi r onment wi l l t end t o i mpose t hat r andomness on f unct i onal
var i at i on under i t s cont r ol and t hat chaot i c dynami cs mi ght
mi mi c r andomvar i at i on. )

We sket ch bel ow a t axonomy of t he pr ocesses t hat mi ght
af f ect t he evol ut i on of st yl i st i c f eat ur es. Thi s di scussi on
l eads t o t wo concl usi ons: Fi r st , wel l - def i ned cul t ur al
evol ut i onar y pr ocesses can r esul t i n det ect abl e, non- r andom
pat t er ns i n adapt i vel y- neut r al st yl i st i c var i at i on t hat wi l l
of t en be di f f i cul t t o di st i ngui sh f r omt he ki nds of pat t er ns
t hat r esul t f r omnat ur al sel ect i on act i ng t o pr oduce
adapt at i ons. Second, some of t he r easons f or pat t er n i n st yl e
have t o do wi t h indirectly f unct i onal f eat ur es. The ar gument
t hat t her e shoul d be a si mpl e di st i nct i on bet ween r andom
st yl i st i c and adapt i ve f unct i onal pat t er ni ng i s suppor t ed
nei t her met hodol ogi cal l y nor onot ol ogi cal l y.


Pattern generated by non-selective random processes

I f i ndi vi dual s acqui r e st yl i st i c t r ai t s by f ai t hf ul l y
copyi ng ot her s, and t hen make i nnovat i ons t hat ar e r andomwi t h
r espect t o adapt at i on, t he r esul t i ng pat t er ns may be r andomi n
t he sense t hat t her e i s no cor r el at i on bet ween st yl i st i c
f eat ur es and envi r onment al var i abl es af f ect i ng f i t ness. Onl y
cul t ur al var i ant s i n t he bot t oml ef t ext r eme of of f i gur e 1
( much of Sacket t ' s i sochr est i c var i at i on) wi l l have such
si mpl e dynami cs, but t hi s case i s of consi der abl e i nt er est
her e because t he t r adi t i onal st yl e- f unct i on di chot omy hol ds
t hat such dynami cs shoul d pr oduce " r andom" pat t er ns t hat ar e
di st i nct i vel y di f f er ent f r omt hose char act er i zi ng var i ant s at
t he bot t omr i ght of f i gur e 1. Even t hi s si mpl e compar i son
cont ai ns enough compl exi t i es t o suppor t t he ar gument t hat
ar chaeol ogi st s must pay cl oser at t ent i on t o t he det ai l s of
pr ocess.

I magi ne a ver y l ar ge, wel l mi xed popul at i on wi t h a
st yl i st i c r eper t oi r e of n di scr et e el ement s, a t r ansmi ssi on
r ul e i n whi ch each i ndi vi dual acqui r es one of t hese var i ant s
at r andom, and a r ul e f or i nnovat i on i n whi ch i ndi vi dual s
( wi t h some pr obabi l i t y) swi t ch t o anot her var i ant wi t h equal
pr obabi l i t y. A popul at i on usi ng such r ul e wi l l mor e or l ess
r api dl y conver ge t o a st at e i n whi ch each var i ant i s pr esent
i n t he popul at i on wi t h equal f r equency ( 1/ n) , no mat t er what
t he st ar t i ng poi nt .

Evol ut i onar y syst ems wi t h pr oper t i es f or mal l y ver y
si mi l ar t o t hi s ki nd of st yl i st i c cul t ur al var i at i on have been
ext ensi vel y st udi ed by popul at i on genet i ci st s i nt er est ed i n
what i s cal l ed t he neutralism controversy. The debat e i s
br i ef l y r evi ewed her e because i t cont ai ns i mpor t ant l essons
f or t hose i nt er est ed i n t he evol ut i on of st yl i st i c cul t ur al
var i at i on ( Ri dl ey, 1993; Ki mur a 1983; and Gi l l espi e 1987,
1991, f or r evi ews of t hi s subj ect ; see al so Caval l i - Sf or za and
Fel dman 1981 f or t heor et i cal appl i cat i ons t o t he speci al case
of cul t ur al var i at i on) . I n t he 1960s advances i n mol ecul ar
genet i cs demonst r at ed t he exi st ence of a huge amount of
genet i c var i at i on i n popul at i ons ( Sel ander 1976) . I ndi vi dual s
ar e het er ozygous at ca 5- 15%of l oci and at t he popul at i on
l evel 15- 60%of l oci ar e det ect abl y pol ymor phi c ( at l east one
r ar e al l el e wi t h a f r equency gr eat er t han 1%) . Ki mur a ( 1968)
ar gued t hat nat ur al sel ect i on coul d not possi bl y mai nt ai n so
many pol ymor phi c l oci because r ecombi nat i on woul d ensur e t hat
each i ndi vi dual had a subopt i mal genot ype at many l oci . Even a
smal l amount of sel ect i on agai nst each subopt i mal l ocus woul d
cumul at i vel y ensur e a huge sel ect i ve l oad on t he popul at i on.
Ki mur a ar gued t hat such a l ar ge amount of var i at i on coul d be
mai nt ai ned onl y i f most al l el es wer e neut r al wi t h r espect t o
nat ur al sel ect i on. The ensui ng debat e i s of i nt er est si nce
i sochr est i c cul t ur al var i ant s ar e so si mi l ar i n concept t o
neut r al al l el es i n genet i cs. The pr obl emi n bot h cases i s t o
di st i ngui sh t r ai t s under nat ur al sel ect i on f r omt r ai t s t hat
ar e not .

Test s pi t t i ng Ki mur a' s neut r al i st cl ai magai nst t he
al t er nat i ve t hat sel ect i on pl ays a r ol e i n mai nt ai ni ng
var i at i on pr oceeded by t he const r uct i on of model s t o deduce
t he uni que pr edi ct i ons of t he neut r al i st and sel ect i oni st
hypot heses. The f i r st compl i cat i on her e i s t hat even i n t he
compl et el y neut r al ( i . e. , " st yl i st i c" ) case, one must t ake
i nt o account t hat popul at i ons ar e not i nf i ni t e. Pat t er ns i n
t i me and space wi l l ar i se i n f i ni t e popul at i ons i f one
i ncl udes t he ef f ect s of genet i c dr i f t ( r andomef f ect s at t he
popul at i on l evel ) . New genes wi l l be i nt r oduced i nt o t he
popul at i on by mut at i on ( r andomef f ect s at t he i ndi vi dual
l evel ) , and t he chances of " sampl i ng" dur i ng r epr oduct i on i n
f i ni t e popul at i ons wi l l cause some genes i ncr ease and ot her s
t o decr ease by chance ( r andomef f ect s at t he popul at i on l evel -
- genet i c dr i f t ) . For t he mut at i on r at es and popul at i on si zes
t hought t o char act er i ze ani mal popul at i ons, t he t heor y
pr edi ct s t hat many genet i c l oci shoul d be monomor phi c, but a
f ai r l y l ar ge pr opor t i on shoul d have var yi ng degr ees of
pol ymor phi sm. The dat a f i t t hi s pr edi ct i on appr oxi mat el y,
al t hough t her e i s consi der abl e debat e r egar di ng t he par amet er
val ues t hat must be assumed f or mut at i on r at es and popul at i on
si zes. For exampl e, f or Drosophila popul at i on si zes have t o be
r at her smal l t o account f or t he l ow l evel s of var i at i on
obser ved. Oht a ( 1976) ar gued t hat t he f i t i s bet t er i f one
assumes most al l el es ar e subj ect t o sl i ght negat i ve sel ect i on.
Gi l l espi e ( 1987, 1991) concur s t hat t he Oht a ver si on i s t he
most empi r i cal l y r easonabl e ver si on of t he neut r al t heor y
( al bei t al so f l awed) . At l east aspect s of t he neut r al t heor y
can be r escued wi t h ot her assumpt i ons, f or exampl e t hat
popul at i ons wer e on aver age smal l er i n t he Pl ei st ocene.
Regar dl ess of t he si t uat i on f or genes, t he t heor y may be qui t e
appr opr i at e f or some ki nds of cul t ur al t r ai t s, especi al l y
i sochr est i c var i ant s near t he l ower l ef t of f i gur e 1.

I f we sur vey a popul at i on over t i me, t he neut r al
hypot hesi s pr edi ct s t hat t her e wi l l be a mor e or l ess r api d
t ur nover of genes as dr i f t " sel ect s" f or at f i r st one and t hen
anot her genet i c var i ant by chance. Super f i ci al l y si mi l ar
r epl acement s wi l l occur , however , i f t he l ocus concer ned i s
r espondi ng t o sel ect i on due t o envi r onment al f l uct uat i ons or
any ot her t i me- st r uct ur ed envi r onment al f act or . Ther e i s
si mpl y no war r ant at t hi s l evel f or t he ar chaeol ogi cal
assumpt i on t hat r andomand sel ect i ve ef f ect s wi l l have
qual i t at i vel y di f f er ent pat t er ns i n t i me. Gi ven er r or i n
sampl i ng or r andomvar i at i on i n t he di r ect i on of sel ect i on, or
some dr i f t super i mposed upon a t r aj ect or y of sel ect i on, t he
gr oss t i me t r end of neut r al and sel ect i ve evol ut i on can be
ver y si mi l ar . The pr ocesses of sel ect i on and r andomevol ut i on
by mut at i on and dr i f t ar e suf f i ci ent l y compl ex t hat model s of
bot h cont ai n enough " t unabl e" par amet er s t o mi mi c each ot her
cl osel y. Thi s i s one r eason why a seemi ngl y t r i vi al debat e
coul d vex popul at i on genet i cs f or a gener at i on.

For t he anal ogous case of human st yl i st i c var i ant s, t he
whol e debat e over par amet er val ues woul d have t o be conduct ed
anew, but per haps some guesses wi l l gi ve an i dea of how r andom
evol ut i on pr oceeds. Let us st ar t wi t h a popul at i on i n whi ch
one var i ant of a st yl i st i c t r ai t i s over whel mi ngl y common.
I ndi vi dual s acqui r e t hei r var i ant by copyi ng someone of t he
par ent al gener at i on at r andom, but al so, r at her r ar el y,
cer t ai n i ndi vi dual s at r andomi nnovat e one of t he many ot her
st yl i st i c var i ant s t hat ar e possi bl e. Suppose some i ndi vi dual
i n a popul at i on of N i ndi vi dual s i nvent s a par t i cul ar new
st yl i st i c var i ant . Assumi ng, f or si mpl i ci t y, t hat each
i ndi vi dual uses onl y one var i ant what i s t he chance p t hat t he
new var i ant wi l l event ual l y become i n t ur n t he over whel mi ngl y
domi nant var i ant i n t he popul at i on? I t i s si mpl y,

p
N
=
1
( 1)

whi ch i s easi est t o see i f we not i ce t hat by chance dr i f t
( r andomvar i at i on i n t he r ol e of speci f i c i ndi vi dual s i n
t r ansmi ssi on each gener at i on) i n t he l ong r un, some one of t he
cur r ent st yl i st i c al t er nat i ves wi l l become t he onl y one used
( supposi ng no mor e i nnovat i on) . Si nce ever y exi st i ng per son' s
st yl e has an equal chance of bei ng t he one t hat " dr i f t s t o
f i xat i on" as t he popul at i on genet i ci st says, any gi ven new
i nnovat i on has a chance of bei ng t hat l ucky var i ant equal t o
i t s f r equency at t he poi nt i t f i r st appear s, whi ch i s ( 1) i n
t he absence of si mul t aneous i nnovat i on. Of cour se, on aver age
i t wi l l t ake a f ai r l engt h of t i me f or some gi ven var i ant t o
be r epl aced by anot her by chance, and i t much mor e l i kel y t hat
any gi ven i nnovat i on wi l l be l ost due t o chance non- i mi t at i on
of i t s or i gi nat or or successor s ( 1- p) . I f we suppose t hat
t her e i s a cer t ai n per - i ndi vi dual r at e of i nnovat i on, u, we
can ask what t he r at e of st yl i st i c t ur nover i n t he popul at i on
mi ght be. I n l ar ge popul at i ons t her e wi l l be mor e i nnovat i ons
each gener at i on, Nu, but accor di ng t o ( 1) t he r at e at whi ch
t hey wi l l become f i xed i s an i nver se f unct i on of N. I n t hi s
si mpl e model t he t wo exact l y cancel , so t hat t he t i me f or one
st yl i st i c var i ant t o r epl ace anot her , k, i s expect ed t o be

1 1
k
Nu
N
u = = ( 2) .

That i s, t he t ur nover r at e i s j ust t he r eci pr ocal of t he
i nnovat i on r at e.

I f we i magi ne t hat soci et i es ar e f ai r l y conser vat i ve as
r egar ds st yl i st i c i nnovat i on, say an i nnovat i on r at e of a f ew
t ent hs of a per cent t o a f ew per cent per i ndi vi dual per
gener at i on, t hen t he t i me t o r epl ace one st yl e wi t h anot her i s
a f ew hundr ed t o a f ew t ens of gener at i ons, i ndependent of
popul at i on si ze. That i s, i n a popul at i on i n whi ch a f ew
t ent hs t o a f ew per cent of peopl e i nnovat e each gener at i on,
t he t ur nover of st yl i st i c f eat ur es wi l l occur on an
ar chaeol ogi cal l y i nt er est i ng t i me scal e. A st yl i st i c f eat ur e
wi l l dr i f t i n and out of a popul at i on over t he cour se of
hundr eds or t housands of year s, j ust l i ke t he st andar d
bat t l eshi p cur ves of st yl i st i c ser i at i on, as Nei man ( 1993)
i l l ust r at es.

The spat i al pat t er ns gener at ed by r andomst yl i st i c choi ce
wi l l be gover ned by sub- popul at i on si ze, i nnovat i on r at es, and
di f f usi on r at es ( see Nei man 199?) . I f popul at i on si zes ar e
l ar ge and mi gr at i on i s hi gh r el at i ve t o i nnovat i on r at es,
chance ef f ect s al one wi l l not be suf f i ci ent t o cause
popul at i ons t o di ver ge. Of r el evance her e i s t he cont r over sy
i n evol ut i onar y bi ol ogy as t o whet her genet i c dr i f t mi ght be
r esponsi bl e f or popul at i on di f f er ent i at i on and, i n combi nat i on
wi t h gr oup sel ect i on, pl ay a r ol e i n movi ng popul at i ons acr oss
sub- opt i mal t r oughs i n t he adapt i ve l andscape, a f amous
hypot hesi s of Sewal l Wr i ght . The condi t i ons ar e f ai r l y
r est r i ct i ve i n t he bi ol ogi cal case, but t hen mut at i on r at es of
genes ar e assumed t o be ver y smal l , on t he or der of 10
- 6
per
l ocus per gener at i on.

I f t he cor r espondi ng i nnovat i on r at es ar e somet hi ng l i ke
10
- 2
i n t he case of cul t ur e, and st yl i st i c di f f usi on r at es ar e
not t oo hi gh, chance st yl i st i c di f f er ent i at i on of l ocal
popul at i ons i s easy t o i magi ne. I n gener al , i f i nnovat i on
r at es ar e gr eat er t han di f f usi on r at es, we woul d expect chance
di f f er ent i at i on t o be i mpor t ant wher e sel ect i on i s negl i gi bl e
( Sel ect i on compl i cat es t he si t uat i on by r et ar di ng di ver gence
bet ween popul at i ons i n whi ch t he same var i ant s ar e f avor ed,
enhanci ng di ver gence bet ween popul at i ons i n whi ch di f f er ent
var i ant s ar e f avor ed) . The r at e of di f f er ent i at i on wi l l al so
depend upon popul at i on si ze ( Nei man 199?) . Raw popul at i on
si ze, however , wi l l be l ess i mpor t ant t han t he por t i on of t he
popul at i on t hat i s act i ve i n t r ansmi ssi on of genes or cul t ur e,
t he " ef f ect i ve" popul at i on si ze i n t he j ar gon of evol ut i onar y
bi ol ogy.

Cul t ur al t r ansmi ssi on i s l i kel y t o be sensi t i ve t o
ef f ect i ve popul at i on si ze because i t of t en t akes t he f or mof
" one- t o- many" t r ansmi ssi on, i n whi ch some t r ai t s ar e
t r ansmi t t ed by r el at i vel y f ew " t eacher s" t o l ar ge number s of
ot her s. I n t hi s case, t he ef f ect i ve popul at i on si ze i s much
smal l er t han a si mpl e headcount woul d i ndi cat e, whi ch, ceteris
paribus, st r engt hens t he ef f ect of dr i f t ( Caval l i - Sf or za and
Fel dman, 1981) .

Bar t h ( 1987) gi ves t he exampl e f r omNew Gui nea of t he
Mount ai n Ok, among whomt he t r ansmi ssi on of r i t ual knowl edge
i s cont r ol l ed by t he handf ul of ol der mal es i n each communi t y
who have succeeded i n passi ng t hr ough a l ong ser i es of r i t ual
i ni t i at i ons. Because t he t r ansmi ssi on of t hi s r i t ual knowl edge
i s i nf r equent and subj ect t o er r or s of memor y, i nnovat i on
r at es ar e much hi gher t han t hey woul d be i n genet i c
t r ansmi ssi on. As a consequence, t he esot er i c l or e of t he
semi - i sol at ed Ok r i t ual communi t i es di ver ges ver y r api dl y. Ok
shamen i ndi vi dual l y at t empt t o r emai n f ai t hf ul t o t r adi t i on,
and, when t hey occassi onal l y vi si t i ni t i at i on r i t es i n ot her
communi t i es, ar e shocked by t he al ar mi ngl y l ar ge devi at i ons
f r omwhat t hey t ake t o be anci ent , i mmut abl e Ok t r ut hs.

Sci ent i f i c di sci pl i nes ar e a mor e f ami l i ar case. Most
moder n di sci pl i nes count t hei r pr act i t i oner s and t eacher s i n
t he t housands but ar e shar pl y st r at i f i ed wi t h r espect t o
i nf l uence so t hat t ext book wr i t er s, successf ul i nnovat or s, and
i ndi vi dual s wi t h many st udent s have di spr opor t i onat e wei ght .
I n t he r el at i vel y nar r ow subdi sci pl i nes wher e most change i s
gener at ed, t he " ef f ect i ve" number of i nf l uent i al i nvest i gat or s
i n any one gener at i on can be ver y smal l , per haps l ess t han
t en, so chance ef f ect s i n t he evol ut i on of sci ence ar e per haps
l i kewi se al ar mi ngl y l i kel y. ( We ar e i ndebt ed t o J . R. Gr i esemer
f or t hi s l ast exampl e. )

Ret ur ni ng t o t he st yl e- f unct i on di chot omy, t he t r oubl e i s
t hat t her e ar e no si mpl e qual i t at i ve r ul es t o di st i ngui sh
t hese dr i f t - i nduced pat t er ns f r omt hose pr oduced by si mpl e
adapt i ve pr ocesses l i ke sel ect i on, by ot her adapt at i on-
pr oduci ng and non- adapt i ve cul t ur al pr ocesses ( out l i ned
bel ow) , or by t he i nt er act i on of sever al of t hese pr ocesses.
To t ake a si mpl e exampl e, f avor abl e t echni cal i nnovat i ons t end
t o occur at i r r egul ar i nt er val s, and each sweeps t hr ough t he
popul at i on once di scover ed. The hi st or y of i mpr ovement of a
t echnol ogy t hus t ends t o be char act er i zed by a successi on of
i mpr oved f or ms i n t i me. For t hi s r eason par t i cul ar t echni cal
f or ms can of t en be expect ed conf or mt o bat t l eshi p cur ves t hat
ar e i ndi st i ngui shabl e f r omt hose t hat pr ovi de t he basi s f or
st yl i st i c ser i at i on ( e. g. , Phi l l i ps, For d, and Gr i f f i n 1951:
Fi g. 11. 3) . We def y t he r eader t o di st i ngui sh wi t h r espect t o
pat t er n Mangel sdor f ' s ser i at i on of changi ng cor n f r equenci es
i n t he Tehuacan Val l ey, Mexi co ( Mangel sdor f 1974: Fi g. 15. 23) ,
whi ch i s pr esumabl y di r ect ed by sel ect i on, f r omDeet z and
Det hl ef sen' s ( 1967: Fi g. 1) ser i at i on of changi ng New Engl and
gr avest one desi gns, whi ch i s pr esumabl y not di r ect ed by an
adapt i ve pr ocess at al l . I n det ai l , t hese pr ocesses make qui t e
di f f er ent pr edi ct i ons about behavi or . For exampl e, a gi ven
t echni cal i mpr ovement can sweep r api dl y t hr ough even ver y
l ar ge popul at i ons, wher eas f i xat i on by dr i f t i s a sl ow pr ocess
i n l ar ge popul at i ons. However , unt i l we make r easonabl e
est i mat es of t he mai n par amet er s of t he pr ocesses, such as
i nnovat i on r at es, magni t udes of sel ect i ve di f f er ences, and
ef f ect i ve si zes of popul at i ons, we cannot t ake advant age of
t he knowl edge t hat sel ect i on and dr i f t wi l l pr oduce di f f er ent
ef f ect s.


Pattern resulting ordinary adaptive forces

The ar gument t hat st yl i st i c var i ant s must al l have equal
f i t ness, r est s on t he i mpl i ci t assumpt i on t hat al l adapt i ve
pr obl ems have unique sol ut i ons. St yl i st i c var i at i on must be
neut r al , t he ar gument r uns, because onl y neut r al var i at i on can
per si st . I f t wo st yl i st i c var i ant s di f f er r ed si gni f i cant l y i n
f unct i on sel ect i on woul d r api dl y el i mi nat e t he i nf er i or
var i ant . Per si st ent di f f er ences bet ween gr oups i n f unct i onal
t r ai t s must t hen be t he r esul t of an envi r onment al di f f er ence.
Thi s r easoni ng f ai l s, however , i f t her e ar e t wo or mor e
locally st abl e t r ai t s. Nat ur al sel ect i on i s onl y a myopi c
opt i mi zer - - - i t causes a popul at i on t o cl i mb up t he adapt i ve
t opogr aphy, event ual l y comi ng t o r est at a l ocal opt i mum. Most
model s of adapt i ve pr ocesses i n cul t ur al evol ut i on suggest
t hat t hey ar e si mi l ar l y myopi c ( Caval l i - Sf or za and Fel dman
1981, Nel son and Wi nt er ( 1982) , Boyd and Ri cher son 1985,
Dur ham1991) . I f t her e i s mor e t han one l ocal opt i mum,
popul at i ons t hat begi n f r omdi f f er ent posi t i ons may r each and
mai nt ai n di f f er ent equi l i br i a even when some equi l i br i a ar e
bet t er sol ut i ons t han ot her s. Cl ear l y, envi r onment al one
cannot account f or such di f f er ences.

Ther e i s subst ant i al evi dence t hat adapt i ve pr obl ems
t ypi cal l y have many l ocal opt i ma. Engi neer s have shown t hat
many desi gn pr obl ems have t hi s pr oper t y. For exampl e,
Ki r kpat r i ck et al . ( 1983) r epor t t hat wher e t he pr obl emwas t o
mi ni mi ze t he number of sl ow connect i ons bet ween chi ps i n t he
I BM 370, t her e ar e about 10
1503
possi bl e ar r angement s, many of
whi ch ar e l ocal l y opt i mal . Local t r i al - and- er r or sear ch cannot
i mpr ove t hese l ocal opt i ma even t hough t hey t end t o have about
4 t i mes as many connect i ons as t he best ar r angement t he
engi neer s di scover ed. Among t he many l ocal opt i ma ar e a
subst ant i al number ( 70) of desi gns t hat ar e qual i t at i vel y
di f f er ent but essent i al l y i dent i cal i n f unct i on t o t he best
ar r angement f ound. Opt i mi zat i on t ext s ( e. g. Wi l de, 1978)
suggest t hat vi r t ual l y al l r eal wor l d desi gn pr obl ems " f r om
dams t o r ef i ger at or s" ( p. xx??) have many equi l i br i a. We see
no r eason t o suppose t hat t he desi gn pr obl ems f aci ng peopl e i n
subsi st ence economi es ar e any di f f er ent ( Bet t i nger 1980) .

Economi st s bel i eve t hat i ncr easi ng r et ur ns t o scal e,
par t i cul ar l y t hose r esul t i ng f r omwhat t hey cal l net wor k
ext er nal i t i es, of t en gener at e mul t i pl e evol ut i onar y
equi l i br i a i n moder n economi es ( Ar t hur 1990) . Net wor k
ext er nal i t i es ar i se when mor e wi del y avai l abe goods have an
adavt age mer el y because t hey ar e wi del y avai l abl e. I f you use
a common make of comput er you have access t o mor e sof t war e and
mor e add- on har dwar e, mor e of your f r i ends ar e abl e t o hel p
you l ear n t o use i t , and i t i s easi er t o col l abor at e wi t h
ot her s. As a r esul t , comput er t echnol ogi es wi t h an i ni t i al
numer i cal advant age may come t o pr edomi nat e even t hough t hey
ar e i nf er i or t o al t er nat i ve t echnol ogi es. Si mi l ar phenomena
l i kel y occur i n subsi st ence economi es. I t coul d be t hat t he
smal l er ! Kung ar r ow poi nt s r eal l y ar e bet t er t han t he l ar ger
poi nt s made by t he ! Xo, but t hat ! Xo who adopt ed smal l er
poi nt s woul d be wor se of f because t hey woul d be unf ami l i ar t o
exchange par t ner s, har der t o l ear n t o make and use and so on
( See Wi esner 1983) .

Model s suggest t hat many t ypes of soci al i nt er act i ons
al so l ead t o mul t i pl e evol ut i onar y equi l i br i a. The si mpl est
exampl es ar e coor di nat i on games i n whi ch f i t ness i s f r equency-
dependent but t her e i s no conf l i ct of i nt er est among
i ndi vi dual s Sugden, 1986) . Dr i vi ng on t he l ef t ver sus r i ght
si de of t he r oad i s an exampl e. I t does not mat t er whi ch si de
we use, but i t i s cr i t i cal t hat we agr ee on one si de or t he
ot her . Reci pr oci t y pr ovi des a good exampl e. Such model s ( Boyd
1992b) suggest t hat t her e ar e a l ar ge number of di f f er ent
st r at egi es t hat can capt ur e at l east some of t he pot ent i al
benef i t s of l ong r un cooper at i on. I n or der t o per si st when
common, r eci pr ocat i ng st r at egi es must r et al i at e agai nst
i ndi vi dual s who do not cooper at e when cooper at i on i s
appr opr i at e. When such a st r at egy i s r ar e, i t wi l l i nt er act
most l y wi t h ot her st r at egi es whi ch cooper at e and expect
cooper at i on i n a di f f er ent set of ci r cumst ances. I nevi t abl y, a
r ar e st r at egy wi l l r et al i at e or suf f er r et al i at i on and
cooper at i on wi l l col l apse. Thus, a common r eci pr ocat i ng
st r at egy has an advant age r el at i ve t o r ar e r eci pr ocat i ng
st r at egi es, even i f t he r ar e st r at egy woul d l ead t o gr eat er
l ong r un benef i t wer e i t t o become common. I nt er act i ons of
t hi s ki nd ar e omni pr esent i n soci al l i f e. Di f f er ent soci al
syst ems may of t en l ead t o var i at i on i n ar t i f act s avai l abl e i n
t he ar chaeol ogi cal r ecor d ( e. g. Bet t i nger and Baumhof f , 1982
and bel ow) . Syst ems wi t h coni cal cl an pol i t i cal or gani zat i on
wi l l t end t o have a mi nor i t y of gr aves wi t h r i ch f ur ni shi ngs
wher eas syst ems wi t h segment ar y l i neages wi l l t end t owar d a
mor e egal i t ar i an di st r i but i on of gr ave goods.


Pattern resulting from novel adaptive forces of cultural
evolution

Cul t ur al l y t r ansmi t t ed det er mi nant s of behavi or ar e
pot ent i al l y subj ect t o a number of evol ut i onar y pr ocesses t hat
Campbel l ( 1965) t er ms " vi car i ous f or ces. " These r esul t f r om
nat ur al sel ect i on act i ng i n t he l ong r un t o pr oduce deci si on
r ul es t hat i n t ur n vi car eousl y sel ect cul t ur al var i ant s. That
i s, i ndi vi dual choi ces about what t r ai t s t o adopt and i nnovat e
wi l l gui de cul t ur al evol ut i on r at her t han sel ect i on act i ng
di r ect l y on cul t ur al var i at i on, al t hough t he di r ect ef f ect s of
sel ect i on ar e not necessar i l y negl i gi bl e. Ther e i s not space
her e t o gi ve even a cur sor y r evi ew of t he compl exi t i es t hat
t hese f or ces engender ( see Dur ham1991 and Boyd and Ri cher son
1985) . I n pr i nci pl e, however , t hese f or ces have ef f ect s t hat
ar e di st i nct i vel y di f f er ent f r omeach ot her , f r omdi r ect
sel ect i on, and f r omt hose i n syst ems af f ect ed onl y by r andom
var i at i on and dr i f t .

Consi der as an exampl e t he f or ce Boyd and Ri cher son cal l
" conf or mi st t r ansmi ssi on. " Thi s i s a ver si on f r equency-
dependent bi ased cul t ur al t r ansmi ssi on ( Boyd and Ri cher son,
1985: Ch. 7; Lumsden and Wi l son, 1980) . " When i s Rome, do as
t he Romans do" i s a f ami l i ar exampl e of a conf or mi st or
" posi t i ve" f r equency dependent r ul e. Conf or mi st t r ansmi ssi on
causes peopl e t o di scr i mi nat e agai nst r ar e t ypes, and i s a
pot ent suppr essor of var i at i on wi t hi n soci et i es. Thi s can
qui t e adapt i ve i n a spat i al l y het er ogeneous envi r onment
because i t causes peopl e t o di scr i mi nat e agai nst mi gr ant s, who
mor e t han l ocal s, ar e pr one t o car r y t r ai t s bet t er adapt ed t o
ot her envi r onment s. On t he ot her hand, i t t hwar t s i nt r oduct i on
of new var i at i on, and so may i mpede adapt i ve t r acki ng of
envi r onment al change over t i me. Such a si mpl e deci si on r ul e
may be most adapt i ve when i t i s appl i ed wi t hout much
j udgement , as a ki nd of r ul e- of - t humb, t o save on deci si on-
maki ng cost s. Accor di ngl y, conf or mi st r ul es coul d be be
appl i ed t o whol l y neut r al t r ai t s as a bi pr oduct of t hei r
advant ages wi t h r egar d t o adapt i ve t r ai t s, causi ng neut r al and
adapt i ve t r ai t s t o pat t er n si mi l ar l y.

The compl exi t i es i nt r oduced by such pr ocesses can be
gl i mpsed i n at t empt s t o expl ai n t he obser ved spat i al
di st r i but i on of house f or ms i n Af r i ca. Ther e, t he gr ound pl an
of houses ( r ect angul ar , r ound, el l i pt i cal , et c. ) i s hi ghl y
var i abl e f r ompl ace t o pl ace but r el at i vel y uni f or mwi t hi n
i ndi vi dual soci et i es. Ther e i s consi der abl e spat i al
aut ocor r el at i on so t hat soci et i es wi t h si mi l ar house f or mt end
t o co- occur geogr aphi cal l y. Caval l i - Sf or za and Fel dman
( 1981: 209f f ) ar gue t hi s pat t er n i s per f ect l y consi st ent wi t h
dr i f t - l i ke ef f ect s i n whi ch: 1) t he l ow var i at i on wi t hi n
soci et i es i s due t o t he one- t o- many dr i f t - enhancement ef f ect ;
2) t he spat i al aut ocor r el at i on i s due t o mi gr at i on bet ween
cl osel y adj acent gr oups; and 3) t he di f f er ent i at i on of di st ant
soci et i es i s due t o i sol at i on.

I t i s easy, however , t o pr oduce a count er - scenar i o t hat
coupl es a di f f er ent f or mof t r ansmi ssi on wi t h
adapt at i on/ sel ect i on. I magi ne t hat house bui l der s use a
" bi ased sampl i ng" r ul e f or acqui r i ng cul t ur al t r ai t s i n whi ch
t hey sur vey a number of cul t ur al model s and i mi t at e t he house
f or m most common among t hose model s. As Af r i cans adopt ed
agr i cul t ur e and began t o bui l d houses a f ew t housand year s
ago, subt l y di f f er ent house t ypes may have been advant ageous
i n di f f er ent pl aces, per haps because of di f f er ences i n r aw
mat er i al s f r ompl ace t o pl ace. Such ear l y acci dent s coul d have
been f r ozen by conf or mi st t r ansmi ssi on, and di st r i but ed about
t he l andscape by mi gr at i on or by t he t endency of non- house
bui l der s t o acqui r e houses f r omnear est nei ghbor s.
Al t er nat i vel y, adapt i ve consi der at i ons such mi cr o- cl i mat e or
avai l abi l i t y of bui l di ng mat er i al s may have t ended t o
det er mi ne t he st andar d house f or mi n par t i cul ar l ocat i ons and
t her eaf t er conf or mi t y act ed t o suppr ess var i at i on ar ound t hat
st andar d. Bot h hypot heses di f f er f r omt hat of Caval l i - Sf or za
and Fel dman i n t hat house f or mi s det er mi ned i ni t i al l y by
adapt at i on and subsequent l y by conf or mi st t r ansmi ssi on, i . e. ,
by a mi xt ur e of adapt at i on and t r ansmi ssi on.

Non- conf or mi st t r ansmi ssi on r ul es and mor e compl ex f or ms
of " t r end- wat chi ng" ar e qui t e concei vabl e ( Lumsden and Wi l son,
1980) . The non- conf or mi st ver si on of f r equency dependent
t r ansmi ssi on wi l l pr ot ect var i ant s f r oml oss wi t hi n soci et i es
by dr i f t , t endi ng t o pr eser ve var i at i on ar i si ng by i ndi vi dual
i nvent i on. Thi s pr ocess wi l l mi mi c a si t uat i on wi t h hi gh
i nnovat i on and mi gr at i on r at es pl us dr i f t . Agai n and agai n t he
poi nt emer ges t hat empi r i cal pat t er ns, even t he most r i gi dl y-
st r uct ur ed ones, ar e of t en consi st ent wi t h a var i et y of
di f f er ent pr ocessual hypot heses. I t i s unr eal i st i c t o expect
t o be abl e at a gl ance t o segr egat e t hemunambi guousl y as
r esul t i ng f r omei t her sel ect i on or neut r al t r ansmi ssi on.


Patterning as the result of correlations among characters

One possi bl e way t o di st i ngui sh adapt i ve f r omst yl i st i c-
neut r al pat t er ns of var i at i on i s by t he pr esence of pl ausi bl e
sel ect i ve f act or s capabl e of expl ai ni ng t he obser ved pat t er n.
I n t he genet i c case, pat t er ns r el at ed t o sel ect i ve f act or s
( e. g. , cl i mat e) ar e of t en f ound ( e. g. Cl egg and Al l ar d, 1972;
Wat t , 1977) . However , as pr oponent s of t he neut r al t heor y
count er ed, linkage of adapt i ve and neut r al genes coul d easi l y
gi ve r i se t o pat t er ns of neut r al al l el es t hat
i ndi st i ngui shabl e f r omt hose of adapt i ve var i ant s. A neut r al
al l el e at one l ocus can " hi t ch- hi ke" t o hi gh f r equency i f i t
i s st at i st i cal l y associ at ed ( l i nked) t o an adapt i ve var i ant at
anot her l ocus. I n t he case of genes, t he st at i st i cal
associ at i on i s gener al l y assumed t o r esul t f r omphysi ci al
pr oxi mi t y on t he chr omosome, so t hat i f a gene f or hai r col or
i s l ocat ed on t he same chr omosome and ver y near t he gene f or
col d- t ol er ance, t he pat t er n of hai r col or ( neut r al ) mi ght end
up bei ng cl osel y associ at ed wi t h pat t er ns of cl i mat e due t o
sel ect i on on a l i nked gene i nf l uenci ng l i mb l engt h or some
ot her di r ect adapt at i on t o col d.

I mpor t ant t echni cal i nnovat i ons t hat pr oduce waves of
popul at i on expansi on coul d easi l y dr ag a host of neut r al or
near neut r al genet i c and cul t ur al t r ai t s t o hi gh f r equency
because of a chance hi gh f r equency i n t he t he popul at i on whi ch
f i r st acqui r es t he adapt i ve t r ai t . Physi cal l i nkage anal ogous
t o genet i c l i nkage may be i nvol ved. For exampl e, i n a compl ex
t ool whose par t s ar e made t oget her may be l ear ned as a mor e or
l ess i nsepar abl e bl ock, so t hat i t s i ndi vi dual component s
sel dom " r ecombi ne. " An adapt i ve i nnovat i on i n one par t of a
t ool may cause t he hi t ch- hi ki ng of non- adapt i ve var i at i on and
st yl i st i c f eat ur es wi t h r egar d t o ot her par t s.

Hi t ch- hi ki ng, however , does not necessar i l y r equi r e any
l i nkage i n t he physi cal sense, onl y an i ni t i al st at i st i cal
associ at i on. Because of t hi s, genes can easi l y hi t ch- hi ke wi t h
cul t ur al i nnovat i ons or vi se ver sa. Thus Ammer man and Caval l i -
Sf or za ( 1985) expl ai n t he gr adi ent of cer t ai n human gene
f r equenci es i n Eur ope as a r esul t of t he genes hi t ch- hi ki ng on
t he wave- l i ke spr ead of agr i cul t ur e f r omt he Mi ddl e East west -
nor t hwest war ds begi nni ng about 9, 000 BP. I t i s a di f f i cul t t o
di st i ngui sh t hei r hypot hesi s f r omt he sel ect i ve expl anat i on
t hat t hese gr adi ent s ar e l ar gel y adapt at i ons t o cl i mat i c
gr adi ent s, si nce measur es of cl i mat e ( e. g. , i sot her ms) l ar gel y
par al l el t he i sol i nes f or t he dat es of t he agr i cul t ur al wave.
The hi t ch- hi ki ng hypot hesi s has been f r equent l y i nvoked i n
var i ous f or ms t o expl ai n t he spr ead of l anguages, especi al l y
I ndo- Eur opean ( Renf r ew, 1987; Mal l or y, 1989) . Renf r ew' s
hypot hesi s i s t hat I ndo- Eur opean hi t ch- hi ked f r oman or i gi nal
f ocus i n Anat ol i a, l i ke Ammer man and Caval l i - Sf or za' s genes,
wi t h spr ead of agr i cul t ur e. Mal l or y di scusses t he mor e
t r adi t i onal hypot heses t hat l i nk I ndo- Eur opean t o l at er
i mpr ovement s i n t he use of hor ses i n war f ar e. Gi ven t hat
l anguage var i at i on i s pr ot ot ypi cal l y st yl i st i c, wi t h no
f unct i onal di f f er ence bet ween al t er nat i ve wor ds, et c. , t he
pat t er ni ng of l anguage i n t i me and space i s a power f ul
conf i r mat i on of t he i mpor t ance of t he hi t ch- hi ki ng ef f ect .
Such cul t ur al hi t ch- hi ki ng, of cour se, i s t he sour ce of what
i s known as " Gal t on' s pr obl em, " i n whi ch cor r el at i on pr oduced
by adapt i ve f or ces cannot be di st i ngui shed f r omcor r el at i ons
pr oduced by shar ed hi st or y. Deet z and Det hl ef sen ( 1967) have
ar chaeol ogi cal l y document ed a f or mof hi t ch- hi ki ng i n New
Engl and gr avest one st yl es t hat i s evi dent l y r el at ed t o
shi f t i ng t r ade net wor ks.


Patterning as the Result of Signaling

As we not ed i n r ef er ence t o Wi essner ' s ( 1983) wor k,
ant hr opol ogi st s and ar chaeol ogi st s commonl y at t r i but e
commmui cat i ve f unct i ons t o st yl i st i c var i at i on, and at l east
t he mor e symbol i c cases of st yl e i n ar t i f act s do commonl y
appear t o f unct i on as expr essi ve or embl emat i c communi cat i on.
What i s l ess wel l appr eci at ed i s t hat t he pr ocesses t hat
af f ect t he evol ut i on of communi cat i ve el ement s of st yl e go
wel l beyond si mpl e r andomi nnovat i on and st at i st i cal dr i f t .
Rat her , sever al di f f er ent di r ect i onal evol ut i onar y f or ces wi l l
af f ect st yl i st i c var i abl es.

The i ssue i s not a si mpl e one. Consi der t he pr ot oypi cal
symbol i c communi cat i on syst em, human l anguage. Al l human
l anguages ar e f unct i onal l y equi val ent ( var i at i ons i n t echni cal
vocabul ar y asi de) . I t doesn' t mat t er whi ch one we speak, but
i t i s i mpor t ant f or pur poses of ef f i ci ent communi cat i on t hat
we f ol l ow l ocal convent i ons of semant i cs and synt ax. Thus, t o
pr eser ve f unct i on, we mi ght expect f or ces t hat act t o l i mi t
i ndi vi dual - l evel i nnovat i on ( e. g. , t he conf or mi st t r ansmi ssi on
bi as) t o domi nat e t he evol ut i on of l anguage. I f so, we woul d
expect l anguage evol ut i on t o be qui t e conser vat i ve when, i n
f act , i t i s f ai r l y r api d. Mut ual l y uni nt el l i gi bl e di al ect s
ar i se i n separ at e popul at i ons i n a f ew hundr ed year s ( Ruhl en,
1994) . At f i r st gl ance, t hi s r api d t ower - of - babel evol ut i on
seems t o be i n def i ance of t he communi cat i on f unct i on of
l anguage. Why don' t human communi cat i on syst ems behave i n a
much mor e conser vat i ve f ashi on? For t hat mat t er , why i sn' t
l anguage a har d- wi r ed human uni ver sal ? The hi ghl y conser ved
basi c st r uct ur e of our genet i c code behaves as expect ed, but
our l anguage does not .

The most pl ausi bl e cur r ent hypot hesi s t o expl ai n t he
r api d evol ut i on of human symbol i c syst ems i s t hat t hei r mai n
f unct i on i s t o communi cat e embl emat i c i nf or mat i on about gr oup
member shi p and about appr opr i at e gr oup behavi or i n cases wher e
i ndi vi dual s ar e f r equent l y exposed t o soci al i nt er act i on wi t h
member s of anot her gr oup.

Det ai l ed mi cr oevol ut i onar y st udi es of di al ect change by
soci ol i ngui st s ( Labov, 1980) suppor t t hi s i dea. I n many ar eas
of t he cont empor ar y wor l d, mi cr odi al ect change i s r api d enough
t o be det ect ed bet ween gener at i ons. Di al ect change seems t o be
set i n mot i on by soci ol ogi cal pr ocesses, f or exampl e,
compet i t i on bet ween et hni c gr oups. I n one of Labov' s cases,
t he Whi t e di al ect of Phi l adel phi a appear s t o have ar i sen i n
r esponse t o t he i nf l ux of Sout her n Bl acks dur i ng and af t er
WWI I . I n anot her case, on Mar t ha' s Vi neyar d, t he evol ut i on of
I sl ander di al ect appear s t o be dr i ven by I sl ander desi r e t o
emphasi ze an i dent i t y separ at e f r ommai nl and t our i st s t owar d
whomeconomi c necessi t y compel s an uncomf or t abl e l evel of
def er ence. Cohen ( 1974) devel oped a ver y si mi l ar hypot hesi s t o
expl ai n t he evol ut i on of i deol ogi cal and cer emoni al syst ems,
such as t he adopt i on of Fr eemasonr y by Si er r a Leone Cr eol es i n
t he f ace of pol i t i cal compet i t i on f r omt r adi t i onal l y
di senf r anchi sed gr oups.

By pr eser vi ng et hni c i dent i t y t hi s sor t of pr ocess does
f ost er soci al sol i dar i t y but need not be vi ewed i n pur el y
st r uct ur al - f unct i onal t er ms because t he symbol i c behavi or s
t hat i dent i f y gr oup member shi p may of t en be associ at ed
behavi or s t hat ar e f unct i onal l y adapt i ve. Boyd and Ri cher son
( 1985: Ch. 8, 1987; Ri cher son and Boyd, 1989) have exami ned
t hi s possi bi l i t y wi t h model s of t he evol ut i on symbol i c
cul t ur al t r ai t s i nspi r ed by dat a such as Labov' s and Cohen' s.
I n t he si mpl est syst ems t hey have st udi ed, popul at i ons ar e
char act er i zed by a symbol i c t r ai t , such as di al ect , whi ch i s
sel ect i vel y neut r al , and an or di nar y adapt i ve char act er , such
as a subsi st ence t echni que. Bot h t r ai t s wer e model ed as
quant i t at i ve char act er s. They suppose t hat chi l dr en acqui r e
t he symbol i c var i ant when young by unbi ased i mi t at i on of a
l ocal adul t . I n a second epi sode of i mi t at i on as " t een- ager s, "
i ndi vi dual s acqui r e t hei r subsi st ence t r ai t by obser vi ng and
i mi t at i ng a wi der r ange of i ndi vi dual s. They bi as t hi s second
deci si on about whomt o i mi t at e i n f avor of i ndi vi dual s bear i ng
a symbol i c var i ant si mi l ar t o t hei r s. Af t er a per i od of
exper i ment at i on, t hese " t een- ager s" compar e t he success of
di f f er ent behavi or al combi nat i ons ( symbol i c pl us adapt i ve) and
r ej ect l ess successf ul combi nat i ons i n f avor of mor e
successf ul ones. The cr i t er i on of " success" i s ar bi t r ar y i n
t he model , but can cer t ai nl y be i nt er pr et ed as adapt i ve
success.

Accor di ng t o t hi s model , i n a spat i al l y var i abl e wor l d i n
whi ch opt i mal subsi st ence behavi or i s ver y di f f er ent i n
di f f er ent envi r onment s, a cor r el at i on can bui l d up bet ween
t he subsi st ence t r ai t and a symbol i c mar ker , so l ong as t he
r at e of mi gr at i on of peopl e i n t he f i r st symbol i c epi sode of
cul t ur al t r ansmi ssi on i s l ess t han i n t he second epi sode wher e
subsi st ence t r ai t s ar e al so t r ansmi t t ed. Once a cor r el at i on
accumul at es bet ween a symbol i c t r ai t and a f avor ed subsi st ence
t r ai t , t her e i s a subst ant i al advant age t o usi ng t he symbol i c
mar ker as a gui de f or whomt o i mi t at e. Fi gur e 2 i l l ust r at es
t hat t wo popul at i ons usi ng t he symbol i c r ul e can wi l l di ver ge
wi t h r egar d t o t he i ndi cat or char act er unt i l unt i l t he mean
val ues of t he adapt i ve char act er ar e opt i mal , wher eas t he
adapt at i on t o a var i abl e envi r onment by non- symbol i c
popul at i ons i s adver sel y af f ect ed by mi gr at i on and l eads t o
l ess successf ul adapt at i ons. Boyd and Ri cher son ar gue t hat
such model s ar e consi st ent wi t h t he hypot hesi s of a wi despr ead
advant age f or t he use of af f ect - l aden embl emat i c symbol
syst ems t o r egul at e cul t ur al t r ansmi ssi on.

What Wi essner cal l s assertive style may r esul t f r omt he
bui l dup of cor r el at i ons bet ween syl i st i c and f unct i onal
var i abl es. Sever al car ef ul st udi es of cont empor ar y popul at i ons
( I r ons, 1979; Bor ger hof f Mul der , i n pr ess) have shown a st r ong
cor r el at i on bet ween pr est i ge, as def i ned by t he l ocal
i deol ogi cal syst em, and weal t h and r epr oduct i ve success. Thi s
cor r el at i on al so suggest s t hat t he use of symbol i cal l y def i ned
st at us as a gui de t o whomt o i mi t at e woul d be f unct i onal , as
Fl i nn and Al exander ( 1982) ar gued. Boyd and Ri cher son ( 1985:
Ch. 8) r evi ew sever al ot her l i nes of evi dence f or t he
i mpor t ant r ol e of usi ng mar ker t r ai t s i n choosi ng f r omwhomt o
acqui r e cul t ur al var i ant s. Empi r i cal mi cr oevol ut i onar y st udi es
of expr essi ve ar t st yl es i n t he moder n West have been
conduct ed by Mar t i ndal e ( 1975, 1990) . He gi ves an i nt er est i ng
account of how psychol ogi cal pr ocesses mi ght dr i ve t he t r ends
and cycl es he di scover s i n hi s dat a.

The pr ocesses t hat bui l d cor r el at i ons bet ween ar bi t r ar y
st yl i st i c f eat ur es and adapt i ve char act er s mi ght be t er med
" act i ve hi t ch- hi ki ng. " Agai n, i t woul d not be easy t o
di st i ngui sh t he pat t er ns gener at ed by t hi s pr ocess f r om
si mpl er ones, especi al l y f r omt he ef f ect s of " passi ve" hi t ch-
hi ki ng. The pot ent i al f or conf oundi ng i s even mor e ser i ous
her e si nce t hose aspect s of t he symbol syst emt hat ar e most
subj ect t o dr i f t t end t o make t he best adapt i ve mar ker s. Thi s
i s because t he act i ve hi t ch- hi ki ng ef f ect i s ver y weak when
t he symbol i c di f f er ence bet ween popul at i ons i s smal l . The
bi ased i mi t at i on ef f ect wor ks on t he cor r el at i on bet ween
mar ker and adapt i ve char act er s ( Boyd and Ri cher son, 1987) .
Accor di ngl y, i f t he i ni t i al var i at i on bet ween popul at i ons i s
smal l , t he cor r el at i on bet ween symbol i c and adapt i ve
char act er s bui l t up by mi gr at i on bet ween t hemwi l l al so
necessar i l y be smal l . When t her e i s a subst ant i al cor r el at i on
bet ween symbol i c mar ker and adapt i ve char act er i n a gi ven
envi r onment , i ndi vi dual s who have t he common symbol i c val ue
associ at ed wi t h an advant ageous var i ant of t he adapt i ve
char act er ar e doubl y advant aged i n cul t ur al t r ansmi ssi on
r el at i ve t o ot her t ypes. Cont ar i wi se, bef or e some var i at i on i n
t he symbol i c char act er ar i ses, t he t endency t o bui l d f ur t her
cor r el at i on wi l l be ver y weak. For exampl e, i f an ancest r al
popul at i on di vi des and becomes segr egat ed i n di f f er ent
habi t at s, t he bi as pr ocess can bui l d up a symbol i c di f f er ence
bet ween t hemonl y ver y sl owl y. I f r andom, dr i f t - l i ke pr ocesses
ar e st r ong i n some char act er s, t hey wi l l pr ovi de t he f i r st
per cept i bl e di f f er ences bet ween t he descendant popul at i ons.
The act i ve hi t ch- hi ki ng ef f ect i s t hen l i abl e t o sei ze j ust
t hese t r ai t s and bui l d a cor r el at i on bet ween t hemand adapt i ve
char act er s. I t i s under such condi t i ons, f or exampl e, t hat
chance l ocal var i at i on i n t he f r equency of f unct i onal l y
equi val ent t echni cal al t er nat i ves mi ght become t he basi s f or
symbol i c var i at i on bet ween gr oups.

The asser t i ve use of st yl e by i ndi vi dual s, par t i cul ar l y
by pr est i gous i ndi vi dual s, seems t o be t he cause of r api d
evol ut i on of pot ent i al mar ker t r ai t s. Labov s account of
di al ect evol ut i on and Mar t i ndal e s account of ar t i st i c
evol ut i on bot h depend upon a cer t ai n, l i mi t ed, t ast e f or t he
novel whi ch dr i ves l i ngust i c and ar t i st i c evol ut i on i n spi t e
of t he f or ces of conf or mi t y t hat ar e r equi r ed t o keep such
syst ems f unct i oni ng as medi a of communi cat i on. Language cannot
change much i n any one gener at i on wi t hout di sr upt i ng
communi cat i on bet ween i ndi vi dual s, but t he smal l i nnovat i ons
made by t he l eader s of l i ngui st i c change l ead t o r at her
dr amat i c changes i n a r el at i vel y f ew gener at i ons. We
hypot hesi ze t hat t he asser t i ve use of st yl e i s t he mot or t hat
bui l ds up var i at i on bet ween semi - i sol at ed gr oups, whi ch can i n
t ur n t hen ser ve as badges of gr oup member shi p. The
evol ut i onar y or i gi ns of t he moder n human sense of st yl e mi ght
wel l l i e i n t he t he advant age of usi ng st yl i st i c var i at i on as
an i ndi cat or of di f f er ences i n adapt i ve char act er s.


Patterning as a result of runaway effects

The pr ocess of usi ng a mar ker t r ai t t o chose whomt o
i mi t at e has pot ent i al l y expl osi ve unst abl e dynami c pr oper t i es
( Boyd and Ri cher son, 1985: ch. 8; Ri cher son and Boyd, 1989) .
I n t he abst r act i t i s easy t o see t hat a dr amat i cal l y
exagger at ed pr est i ge syst emcan be pr ot ect ed agai nst nat ur al
sel ect i on or sel ect i on- der i ved r at i onal choi ce i f success i n
t he pr est i ge syst emof f set s l osses due t o t he mal adapt i ve
consequences of t he exagger at i on. Once enough peopl e use a
speci f i c mar ker t o choose whomt o i mi t at e, anyone f ai l i ng t o
di spl ay t hat i ndi cat or wi l l be ef f ect i vel y be i gnor ed i n t he
t he pr ocess of cul t ur al t r ansmi ssi on. Wher e soci al syst ems ar e
based on an el ement of coer ci on, t he possi bi l i t y of
mai nt ai ni ng ar bi t r ar y, non- f unct i onal behavi or s i s i ncr eased
st i l l f ur t her ( Boyd and Ri cher son, 1992) . The key quest i on i s
how such a set of pr ef er ence r ul es f avor i ng i mi t at i on of
peopl e di spl ayi ng cost l y pr est i ge symbol s, or a wi l l i ngness t o
puni sh devi ant s, can ar i se i n t he f i r st pl ace. As wi t h t he
anal ogous case of mat e- choi ce sexual sel ect i on, t heor et i cal
model s show t hat a syst emof coupl ed pr ef er ence char act er s and
di spl ay char act er s can r un away t o exagger at ed ext r emes
( Lande, 1981) . The mal e- bi ased di spl ay char act er s of many
ani mal s, such as t he f eat her s of peacocks and t he el abor at e
const r uct i ons of bower bi r ds, ar e of t en at t r i but ed t o t he
r unaway ef f ect . We r egar d t he si mi l ar i t y bet ween t he pl umes of
bi r ds and t he f i ner y di spl ayed by pr est i gi ous humans as mor e
t han coi nci dent al . I n t hi s r espect , t he r unaway hypot hesi s
pr ovi des a way of t ur ni ng Sahl i ns' s ( 1976) not i on of " cul t ur al
r eason" i nt o a cogent f or mal ar gument .

The i mpor t ance and mechani cs of t he r unaway ef f ect ar e
hot l y debat ed by evol ut i onar y bi ol ogi st s ( see e. g. Bar t on and
Tur el l i , 1990; Ki r kpat r i ck, et al . , 1990) . Bar t on and
Tur el l i ' s t heor et i cal i nvest i gat i ons suggest t hat t he pur e
r unaway ef f ect i s weak because t he f or ces mai nt ai ni ng t he
cor r el at i on bet ween t he symbol i c di spl ay and pr ef er ence t r ai t s
ar e i nher ent l y weak. What wi l l encour age t he exagger at i on
pr ocess i s some i ndependent adapt i ve advant age accr ui ng t o t he
sel ect i on of i ndi vi dual s wi t h el abor at e mar ker s. For exampl e,
i f economi c success gener at es t he wher ewi t hal t o di spl ay
st at us mor e ef f ect i vel y ( buy t he f anci est car , pay t he br i de-
pr i ce f or t he youngest and most beaut i f ul woman) , t her e wi l l
r emai n a cor r el at i on bet ween or di nar y adapt i ve success and t he
degr ee of exagger at i on of di spl ay. Then i t pays i n bot h t he
or di nar y adapt i ve and pr est i ge games t o chose mat es or ment or s
wi t h t he most exagger at ed syst emof pr est i ge. I n t he l i mi t ,
al l t he gai ns accr ui ng f r omor di nar y adapt i ve adavant ages ar e
di ssi pat ed i n suppor t of t he most el abor at e possi bl e st at us
di spl ays, a sor t of per ver se i nver si on of t he or di nar y
hi t chi ki ng ef f ect . I n such cases, i t may be sai d j ust i f i abl y
t hat t he cul t ur al l y dr i ven symbol i c syst emhas capt ur ed t he
mundane economi c syst em, much as Sahl i ns cl ai ms.

Cl ear l y, t he gr oss pat t er ns pr edi ct ed by t he r unaway and
si gnal i ng hypot heses ar e r at her si mi l ar . Boyd and Ri cher son
( 1985: Ch. 8) ar gue t hat t he or di nar y adapt i ve advant ages of
choosi ng ment or s by means of i ndi cat or char act er s wi l l
mai nt ai n t hi s sor t of choi ce mechani smby nat ur al sel ect i on,
even t hough t he syst emmi sf i r es occasi onal l y and get s caught -
up i n t he r unaway pr ocess. They f ur t her suppose t hat t he
t r ai t s most subj ect t o exagger at i on ar e wi l l gener al l y be one
hi st or i cal l y connect ed wi t h adapt at i on. The gr owi ng of gi ant
yams on Ponapae as a par t of pr est i ge cont est s i s a possi bl e
exampl e. I t seems pl ausi bl e t hat when t he cust omor i gi nat ed,
good f ar mer s di d gr ow l ar ger yams, and t hat l ar ge yams wer e a
good i ndex of yam- gr owi ng t al ent . I f Tur el l i ' s hypot hesi s
appl i es i n t he cul t ur al case, t he gr ower s of gi ant yams may
st i l l be t he best hor t i cul t ur al i st s; t he adapt i ve and r unaway
hypot heses ar e r eal l y wonder f ul l y ent angl ed i n t hi s case.


Conclusion

The mai n at t r act i on of t he st yl e- f unct i on di chot omy i s
t hat i t appar ent l y r educes t he t ask of ar chaeol ogi cal
expl anat i on t o a manageabl e subset of phenomena t hat can be
addr essed by si mpl e causal model s. Some wi l l undoubt edl y r ead
our r ej ect i on of t he di chot omy as compl i cat i ng t he pr obl emof
ar chaeol ogi cal i nt er pr et at i on t o t he poi nt of i mpossi bi l i t y.
Because we r ej ect t he di chot omy and even ci t e t he wor ks of
such aut hor s as Sahl i ns wi t h ( qual i f i ed) appr oval , st i l l
ot her s wi l l r ead us as advocat i ng a f or mof post - st r uct ur al i st
ar chaeol ogy. Nei t her r eadi ng i s war r ant ed. We ar e ent husi ast i c
advocat es of causal and mat er i al i st expl anat i on of soci al
phenomena. We ar e al so advocat es of si mpl e model s of compl ex
phenomena ( Levi ns, 1966) . Tr eat i ng pat t er ned var i at i on i n
ar t i f act s and behavi or s as t hough i t wer e pur el y f unct i onal
and adapt i ve and assumi ng t hat st yl i st i c var i at i on i s noi sy
and i r r el evant may of t en be an accept abl e si mpl i f i cat i on.
Sur el y, human gr oups cannot exi st as goi ng economi c concer ns
unl ess a l ar ge f r act i on of pat t er ned var i at i on i s adapt i ve.
Li kewi se, assumi ng t hat st yl e behaves as i f i t wer e subj ect
onl y t o r andomi nnovat i on and dr i f t pr ovi des an i mpor t ant
t heor et i cal war r ant f or ser i at i on t hat can be hi ghl y usef ul
even when pat t er ns depar t somewhat f r omt he i deal . Wi t hi n t hi s
cont ext , our ar gument boi l s down t o t hese t hr ee si mpl e poi nt s.

Fi r st , t her e i s good r eason t o t hi nk t hat t he st yl e-
f unct i on di chot omy i s f r equent l y an unaccept abl e
si mpl i f i cat i on. I t i s wel l wor t h t hi nki ng about t hi s
possi bi l i t y and what i t i mpl i es about how we shoul d do
ar chaeol ogy. As Wi msat t ( 1980) not es i n hi s def ense of t he use
of si mpl e model s, f ai l ur e t o r ecogni ze t he speci f i c
l i mi t at i ons of wi del y- adopt ed si mpl i f yi ng assumpt i ons can l ead
t o danger ous over conf i dence i n t he r obust ness of our model s
and t hei r r esul t s. As par t of t hi s, we must squar el y f ace t he
di f f i cul t i es i nvol ved i n sol vi ng what physi cal sci ent i st s cal l
t he i nver se pr obl em( mor e f ami l i ar t o ar chaeol ogi st s as t he
pr obl emof " equi f i nal i t y" ) . I t may be har d, somet i mes per haps
i mpossi bl e, t o i nf er t he mi cr o- scal e pr ocesses t hat gave r i se
t o a par t i cul ar macr oscal e pat t er n. Many di f f er ent
evol ut i onar y pr ocesses, f or i nst ance, can cause t he f ami l i ar
bat t l eshi p ( l ent i cul ar ) pat t er n of i ncr ease and decr ease. The
pr obl ems t hi s r ai ses cannot be i gnor ed. I t i s di f f i cul t t o
di st i ngui sh i sochr est i c f r omf unct i onal var i at i on, as near l y
ever yone agr ees. However , i f t he ar gument we have pr esent ed i s
cor r ect , assumi ng t hat f unct i on and st yl e can be separ t ed i nt o
di scr et e cat egor i es wi t h ver y di f f er ent evol ut i onar y
pr oper t i es i s not possi bl e. The pr ocesses of evol ut i on ar e
j ust mor e compl ex t han t hat . What i s cal l ed f or i s a
met hodol ogi cal l y r i gor ous pr ogr amof st udy pat t er ned i n t he
mol d of cont empor ar y mi ddl e r ange st udi es t hat use t i ght l y
cont r ol l ed et hnoar chaeol ogi cal and t aphonomi c i nvest i gat i on t o
di st i ngui sh t he si gnat ur es of di f f er ent pr ocesses t hat t end t o
pr oduce out war dl y si mi l ar ar chaeol ogi cal consequences. I n some
cases t he dat a wi l l be i nsuf f i ci ent t o deci de bet ween
compet i ng al t er nat i ve hypot heses but t her e i s no r eason t o
t hi nk t hat t hi s wi l l be t r ue gener al l y. The i nver se pr obl emi s
cause f or despai r onl y i f i t can be sol ved so i nf r equent l y
t hat t her e i s no hope of bui l di ng a sat i sf act or y pi ct ur e of
t he r el at i ve power of di f f er ent gener al hypot heses. Ot her wi se,
i t i s mer el y a chal l enge t o our i magi nat i on and i ni t i at i ve.

Second, st yl e i s t oo i mpor t ant and t oo i nt er est i ng t o
l eave t o st r uct ur al i st s and post - moder ni st s. A number of
i mpor t ant ar chaeol ogi cal phenomena make much mor e sense i f we
assume t hat st yl i st i c var i at i on i s f unct i onal l y i mpor t ant .
Why was t he Upper Pal eol i t hi c t r ansi t i on a st yl i st i c as wel l
as economi c r evol ut i on? Why does st at e f or mat i on so f r equent l y
i nvol ve t he el abor at i on of r el i gi ous i nst i t ut i ons, i deol ogy,
and t he ar t s? How cost l y ar e symbol i c i nst i t ut i ons, and how
much do t hey di st or t or f ost er adapt at i on ( however t hat mi ght
be oper at i onal i zed) ?

Thi r d, Dar wi ni an t heor y wi l l event ual l y of f er a
pr ocessual account of cul t ur al evol ut i on t hat i s as power f ul
as t he one i t now of f er s f or genet i c evol ut i on. The pr obl em
pr esent l y i s t hat we have l i mi t ed knowl edge of t he oper at i on
of t hese pr ocesses i n t he cul t ur al r eal mand ar e handi capped
i n our abi l i t y t o use t hemi n i nt er pr et i ng past behavi or . On
t he ot her hand, t hi s si t uat i on shoul d be at t r act i ve t o t hose
of us who cont i nue t o shar e t he pr ocessual goal s t hat
i nspi r ed t he New Ar chaeol ogy. Ther e ar e a l ar ge number of
essent i al l y unst udi ed pr ocesses beggi ng t he ki nd of cr i t i cal
exper i ment al and obser vat i on pr ogr amadvocat ed t hen and si nce
by Bi nf or d and ot her s.

Ar chaeol ogi st s, who ar e somet i mes dr i ven t o do t he wor k
t hat shoul d mor e pr oper l y f al l t o et hnogr apher s
( et hnoar chaeol ogy) , shoul d appr eci at e t hat ar chaeol ogy must
pl ay a di st i nct i vel y cr i t i cal r ol e i n under st andi ng t he
pr ocesses of cul t ur al evol ut i on. The synchr oni c st udy of
symbol syst ems and t hei r evol ut i on on t he mi cr o t i me scal e i s
sur el y cr i t i cal , but ar chaeol ogi st s, hi st or i ans, and
pal eoant hr opol ogi st s have a monopol y on dat a f r omt he l onger
t i me scal es over whi ch t he evol ut i onar y pr ocesses gener al l y
wor k t hemsel ves out . The model s of t he adapt i ve r ol e of
symbol i c mar ki ng of et hni c gr oups r evi ewed above, f or exampl e,
ar e necessar i l y si l ent about j ust what sor t s of adapt i ve
di f f er ences bet ween gr oups mi ght be pr ot ect ed by t hi s
mechani sm. That depends on how easi l y cor r el at i ons bet ween
var i ous ki nds of t r ai t s can be bui l t up i n t he f ace of
mi gr at i on. I t i s unl i kel y t hat shor t - t er mst udi es wi l l be as
convi nci ng i n t hi s r egar d as t he act ual l ong- r un dat a.

A bi t of our own wor k i l l ust r at es t he ki nd of pr ocess
r el at ed i nf or mat i on avai l abl e f r oml ong r ecor ds. Bet t i nger and
Baumhof f ( 1982) have exami ned t he case of t he spr ead of Numi c
speaker s acr oss t he Gr eat Basi n of t he West er n US f r omabout
700 t o 200 year s ago. I n t hi s case t he evi dence suppor t s t he
i dea t hat t he et hni c boundar y bet ween Numi c peopl es and t hei r
pr e- Numi c pr edecessor s must have l i mi t ed t he spr ead of a
soci al - or gani zat i onal var i abl e, not a di r ect t echnol ogi cal
var i abl e. The r ecor d i ndi cat es t hat t he same t echnol ogy and
t ool t ypes wer e ever ywher e avai l abl e, yet t he spr ead of
st yl i st i c el ement s associ at ed wi t h Numi c- speaker s, i ncl udi ng
at l anguage and i deol ogy, i s associ at ed wi t h hi gher densi t i es,
di f f er ences i n l ocat i on of set t l ement s, and quant i t at i ve
var i at i on i n f r equenci es of t he var i ous t ool t ypes. Si nce t he
mor e pl ant - i nt ensi ve Numi c st r at egy r equi r ed a l ar ger r ol e f or
pl ant st or age and women' s l abor , i t i s pl ausi bl e t hat t he key
t o Numi c success was a nor mat i ve compl ex t hat condoned t he
hoar di ng of pl ant r esour ces, gave women a gr eat er r ol e i n
deci si on- maki ng, and r educed t he aut onomy of hunt er s.

The ut i l i t ar i an consequences of mundane t echnol ogy ( e. g. ,
seed- beat er s) ar e gener al l y r at her obvi ous, and hence move
easi l y acr oss boundar i es. Soci al nor ms have mor e compl ex and
f ar - r eachi ng ef f ect s t hat ar e of t en di f f i cul t f or act or s t o
under st and and mor e cl osel y t i ed t o af f ect - l aden i deol ogi cal
syst ems. Per haps onl y an et hni c i sol at e i s l i kel y t o t ake
unusual st eps away f r omobl i gat e shar i ng and t owar d gender
egal i t ar i ani sm. Once t he et hni c advance begi ns, an et hni c
boundar y can expl ai n why t he l osi ng gr oup per si st s i n
r et ai ni ng i t s behavi or despi t e t he obvi ous di sadavant ages.

The Numi c case may or may not be cor r ect l y i nt er pr et ed,
and even i f cor r ect l y i nt er pr et ed, i t may not be
r epr esent at i ve. I t does have t he vi r t ue of suggest i ng t est abl e
hypot heses: ( 1) That et hni c or ot her st yl e- mar ked boundar i es
ar e i mpor t ant i n t he or i gi n and spr ead of cer t ai n t ypes of
i nnovat i ons, and ( 2) t hat t echni cal i nnovat i ons per se ar e
l i kel y t o spr ead i r r espect i ve of st yl e- mar ked boundar i es,
wher eas mor e subt l e aspect s of adapt at i on ( as j udged by
abi l i t y t o suppor t hi gher popul at i on densi t i es f or exampl e)
ar e l i kel y t o r equi r e boundar i es t o or i gi nat e, and ar e t hen
l i kel y t o spr ead associ at ed st yl es by act i ve hi t ch- hi ki ng. I t
i s ar chaeol ogi st s who ar e i n a posi t i on pr obe t he l ong- t er m
pat t er ns of cor r el at i on bet ween di f f er ent ki nds of t r ai t s, and
hence t o make an essent i al cont r i but i on t o t he ver y basi c
soci al sci ence pr obl emof under st andi ng j ust what i s t he
si gni f i cance of moder n humans' massi ve pr eoccupat i on wi t h
st yl e, and how i t i s t hat we came t o r epl ace popul at i ons wi t h
an ( appar ent l y) mor e ut i l i t ar i an out l ook.
Ref er ences

Ammer man, A. J . and L. L. Caval l i - Sf or za. 1985. The Neol i t hi c
Tr ansi t i on and t he Genet i cs of Popul at i ons i n Eur ope.
Pr i ncet on: Pr i ncet on Uni ver si t y Pr ess.

Ar t hur , W. B. 1990. Posi t i ve f eedbacks i n t he economy.
Sci ent i f i c Amer i can ( Febr uar y) : 92- 99.

Bar t h, F. 1987. Cosmol ogi es i n t he Maki ng: A Gener at i ve
Appr oach t o Cul t ur al Var i at i on i n I nner New Gui nea. Cambr i dge:
Cambr i dge Uni ver si t y Pr ess.

Bar t on, N. H. and M. Tur el l i . 1991. Nat ur al and sexual
sel ect i on on many l oci . Genet i cs 127: 229- 255.

Bet t i nger , R. L. 1980. Expl anat or y- Pr edi ct i ve Model s of
Hunt er - Gat her er Behavi or . Advances i n Ar chaeol ogi cal Theor y
and Met hod 3: 189- 255.

Bet t i nger , R. L. 1991. Hunt er - Gat her er s: Ar chaeol ogi cal and
Evol ut i onar y Theor y. New Yor k: Pl enum

Bet t i nger , R. L. , and M. A. Baumhof f . 1982. The Numi c Spr ead:
Gr eat Basi n Cul t ur es i n Compet i t i on. Amer i can Ant i qui t y
47: 485- 503.

Bi nf or d, L. R. 1962. Ar chaeol ogy as Ant hr opol ogy. Amer i can
Ant i qui t y 48: 217- 225.

Bi nf or d, L. R. 1965. Ar chaeol ogi cal Syst emat i cs and t he St udy
of Cul t ur e Pr ocess. Amer i can Ant i qui t y 31: 203- 210.

Bi nf or d, L. R. 1989. St yl es of St yl e. J our nal of
Ant hr opol ogi cal Ar chaeol ogy 8: 51- 67.

Boyd, R. and P. J . Ri cher son. 1985. Cul t ur e and t he
Evol ut i onar y Pr ocess. Chi cago: Chi cago Uni ver si t y Pr ess.

Boyd, R. and P. J . Ri cher son. 1987. The evol ut i on of et hni c
mar ker s. Cul t ur al Ant r opol ogy 2: 65- 79.

Boyd, R. and P. J . Ri cher son. 1992a. How mi cr oevol ut i onar y
pr ocesses gi ve r i se t o hi st or y. I n: Hi st or y and Evol ut i on.
M. H. and D. V. Ni t ecki , eds. Al bany: St at e Uni ver si t y of New
Yor k.

Boyd, R. and P. J . Ri cher son. 1992b. Puni shment al l ows t he
evol ut i on of cooper at i on ( or anyt hi ng el se) i n si zabl e gr oups.
Et hol . Soci obi ol . 13: 171- 95.

Campbel l , D. T. 1965. Var i at i on and sel ect i ve r et ent i on i n
soci ocul t ur al evol ut i on. I n: Soci al Change i n Devel opi ng
Ar eas: A Rei nt er pr et at i on of Evol ut i onar y Theor y, H. Bar r i nger ,
G. I . Bl ankst en, and R. W. Mack, eds. Cambr i dge Mass. :
Schenkman. Pp. 19- 49.

Caval l i - Sf or za, L. L. and M. W. Fel dman. 1981. Cul t ur al
Tr ansmi ssi on and Evol ut i on: A Quant i t at i ve Appr oach.
Pr i ncet on: Pr i ncet on Uni ver si t y Pr ess.

Cl egg, M. T. , and R. W. Al l ar d. 1972. Pat t er ns of genet i c
di f f er ent i at i on i n t he sl ender wi l d oat , Avena barbata. Pr oc.
Nat l . Acad. Sci . USA 69: 1820- 4.

Cohen, A. 1974. Two- Di mensi onal Man: An Essay on t he
Ant hr opol ogy of Power and Symbol i smi n Compl ex Soci et y.
Ber kel ey: Uni ver si t y of Cal i f or ni a Pr ess.

Cosmi des, L. and J . Tooby. 1992. Cogni t i ve adapt at i ons f or
soci al exchange. I n: The Adapt ed Mi nd: Evol ut i onar y Psychol ogy
and t he Gener at on of Cul t ur e. J . H. Bar kow, L. Cosmi des, and J .
Tooby, eds. New Yor k: Oxf or d Uni ver si t y Pr ess. Pp. 163- 228.

Dawki ns, R. 1986. The Bl i nd Wat chmaker : Why t he Evi dence of
Evol ut i on Reveal s a Uni ver se Wi t hout Desi gn. New Yor k: Nor t on.

Deet z, J . F. , and E. S. Det hl ef sen. 1967. Deat h s Head,
Cher ub, Ur n and Wi l l ow. Nat ur al Hi st or y 76: 29- 37.


Dunnel l , R. 1978. St yl e and Funct i on: A Fundament al Di chot omy.
Amer i can Ant i qui t y 43: 192- 202.

Dunnel l , R. 1980. Evol ut i onar y Theor y and Ar chaeol ogy.
Advances i n Ar chaeol ogi cal Theor y and Met hod 3: 35- 99. New
Yor k: Academi c.

Dur ham, W. H. 1991. Coevol ut i on: Genes, Cul t ur e, and Human
Di ver si t y. St anf or d: St anf or d Uni ver si t y Pr ess.

Eber har d, W. G. 1986. Sexual Sel ect i on and Ani mal Geni t al i a.
Cambr i dge, Mass. : Har var d Uni ver si t y Pr ess.

Fl i nn, M. V. and R. D. Al exander . 1982. Cul t ur e t heor y: The
devel opi ng synt hesi s f r ombi ol ogy. Human Ecol . 10: 383- 400.

Gi l l espi e, J . H. 1987. Mol ecul ar evol ut i on and t he neut r al
al l el e t heor y. Oxf or d Sur veys Evol . Bi ol . 4: 10- 37.

Gi l l espi e, J . H. 1991. The Causes of Mol ecul ar Evol ut i on. New
Yor k: Oxf or d Uni ver si t y Pr ess.

Hami l t on, W. D. and M. Zuk. 1982. Her i t abl e t r ue f i t ness and
br i ght bi r ds: a r ol e f or par asi t es. Sci ence 218: 384- 387.

Har r i s, M. 1968. Ri se of Ant hr opol ogi cal Theor y. New Yor k:
Cr owel l .

Hei ner , R. A. 1983. The or i gi n of pr edi ct abl e behavi or . Amer .
Econ. Rev. 73: 560- 595.

I r ons. W. 1979. Cul t ur al and bi ol ogi cal success. I n;
Evol ut i onar y Bi ol ogy and Human Soci al Behavi or : An
Ant hr opol ogi cal Per spect i ve. N. Chagnon and W. I r ons,
eds. Nor t h Sci t uat e, Mass. : Duxbur y. Pp. 257- 72.

Ki mur a, M. 1968. Evol ut i onar y r at e at t he mol ecul ar l evel .
Nat ur e 217: 624- 6.

Ki mur a, M. 1983. The Neut r al Al l el e Theor y of Mol ecul ar
Evol ut i on. Cambr i dge: Cambr i dge Uni ver si t y Pr ess.

Ki r ch, P. V. 1980. The Ar chaeol ogi cal St udy of Adapt at i on:
Theor et i cal and Met hodol ogi cal I ssues. Ar chaeol ogi cal Theor y
and Met hod 3: 101- 156.

Ki r kpat r i ck, M. 1989. I s bi gger al ways bet t er ? Nat ur e 337:
116- 117. [ Bet t er , Ann Rev Ecol Syst ]

Ki r kpat r i ck, M. , T. Pr i ce, and S. J . Ar nol d. 1990. The Dar wi n-
Fi sher t heor y of sexual sel ect i on i n monogamous bi r ds.
Evol ut i on 44: 180- 193.

Ki r kpat r i ck, S. , C. D. Gel at t , and M. V. Vecchi . 1983.
Opt i mi zat i on by si mul at ed anneal i ng. Sci ence 220: 671- 80.

Kr oeber , A. L. 1948. Ant hr opol ogy. New Yor k: Har cour t , Br ace,
and Wor l d.

Kr oeber , A. L. , and J Ri char dson. 1940. Thr ee Cent ur i es of
Women s Dr ess Fashi ons: A Quant i t at i ve Anal ysi s. Uni ver si t y of
Cal i f or ni a Ant hr opol ogi cal Recor ds 5( 2) .

Labov, W. ( ed. ) . 1980. Locat i ng Language i n Ti me and Space.
New Yor k, Academi c Pr ess.

Lande, R. 1981. Model s of speci at i on by sexual sel ect i on on
pol ygenei c t r ai t s. Pr oc. Nat l . Acad. Sci . USA 78: 3721- 3725.

Leonar d, R. D. , and G. T. J ones. 1987. El ement s of an
I ncl usi ve Evol ut i onar y Model f or Ar chaeol ogy. J our nal of
Ant hr opol ogi cal Ar chaeol ogy 6: 199- 219.

Levi ns, R. 1966. The st r at egy of model bui l di ng i n popul at i on
bi ol ogy. Amer . Sci ent . 54: 421- 431.

Lumsden, C. J . and E. O. Wi l son. 1980. Tr ansl at i on of epi genet i c
r ul es of i ndi vi dual behavi or i nt o et hnogr aphi c pat t er ns. Pr oc.
Nat l . Acad. Sci . USA 77: 4382- 6.

Mal l or y, J . P. 1989. I n Sear ch of t he I ndo- Eur opeans: Language,
Ar chaeol ogy and Myt h. London: Thames and Hudson.

Mangl esdor f , Paul C. 1974. Cor n. Cambr i dge: Har var d Uni ver si t y
Pr ess.

Mar t i ndal e, C. 1975. Romant i c Pr ogr essi on: The Psyi chol ogy of
Li t er ar y Hi st or y. New Yor k: Hal st ead.

Mar t i ndal e, C. 1990. The Cl ockwor k Muse: The Pr edi ct abi l i t y of
Ar t i st i c Change. New Yor k: Basi c Books.

Maynar d Smi t h, J . 1991. Theor i es of sexual sel ect i on. Tr ends
i n Ecol ogy and Evol ut i on 6: 146- 151.

Nei man, F. 1993. What Makes t he Ser i at i on Cl ock Ti ck? Paper
Pr esent ed at t he Annual Meet i ng of t he Soci et y f or Amer i can
Ar chaeol ogy, St . Loui s.

Nel son, R. R. and S. G. Wi nt er . 1982. An Evol ut i onar y Theor y of
Economi c Change. Cambr i dge, Mass. : Har var d Uni ver si t y Pr ess.

O Br i en, M. J . , and T. D. Hol l and. 1992. The Rol e of
Adapt at i on i n Ar chaeol ogi cal Expl anat i on. Amer i can Ant i qui t y
57: 36- 59.

O Connel l , J . F. , K. T. J ones, and S. Si mms. Some Thought s on
Pr ehi st or i c Ar chaeol ogy i n t he Gr eat Basi n. I n D. B. Madsen
and J . F. O Connel l , eds. , Man and Envi r onment i n t he Gr eat
Basi n, pp. 227- 240. Washi ngt on, D. C. : Soci et y f or Amer i can
Ar chaeol ogy Paper s 2.

Oht a, T. 1976. Rol e of ver y sl i ght l y del et er i ous mut at i ons i n
mol ecul ar evol ut i on and pol ymor phi sm. Pheor . Pop. Bi ol . 10:
254- 275.

Opl er , M. E. 1964. The Human Bei ng I n Cul t ur e Theor y. Amer i can
Ant hr opol ogi st 66: 507- 528.

Phi l l i ps, P. , J . A. For d, and J . B. Gr i f f i n. 1951.
Ar chaeol ogi cal Sur vey i n t he Lower Mi ssi ssi ppi Al l uvi al
Val l ey, 1940- 1947. Paper s of t he Peabody Museum25: 219- 233.
Cambr i dge: Har var d Uni ver si t y.

Pomi ankowski , A. 1988. The evol ut i on of f emal e mat e
pr ef er ences f or mal e genet i c qual i t y. Oxf or d Sur veys i n
Evol ut i onar y Bi ol ogy 5: 136- 184.

Renf r ew, C. 1987. Ar chaeol ogy and Language: The Puzzl e of
I ndo- Eur opean Or i gi ns. New Yor k: Cambr i dge Uni ver si t y Pr ess.

Ri cher son, P. J . and R. Boyd. 1989. The r ol e of evol ved
pr edi sposi t i ons i n cul t ur al evol ut i on: Or , human soci obi ol ogy
meet s pascal s wager . Et hol . Soci obi ol . 10: 195- 219.

Ri dl ey, M. 1993. Evol ut i on. Bost on: Bl ackwel l .

Ruhl en, M. 1994. The Or i gi n of Language: Tr aci ng t he Evol ut i on
of t he Mot her Tongue. New Yor k: Wi l ey.

Ryan, M. J . , J . H. Fox, W. Wi l czynski , and A. S. Rand. 1990.
Sexual sel ect i on i n t he f r og Physalaemus pustulosus. Nat ur e
343: 66- 7.

Sacket t , J . R. 1982. Appr oaches t o St yl e i n Li t hi c
Ar chaeol ogy. J our nal of Ant hr opol ogi cal Ar chaeol ogy 1: 59- 112.

Sacket t , J . R. 1985. St yl e and Et hni ci t y i n t he Kal ahar i : A
Repl y t o Wei ssner . Amer i can Ant i qui t y 50: 154- 159.

Sacket t , J . R. 1986. St yl e, Funct i on, and Assembl age
Var i abi l i t y: A Repl y t o Bi nf or d. Amer i can Ant i qui t y 51: 628-
634.

Sahl i ns, M. 1976. Cul t ur e and Pr act i cal Reason. Chi cago:
Chi cago Uni ver si t y Pr ess.

Si mon, H. A. 1959. Theor i es of Deci si on- maki ng i n economi cs.
Amer . Econ. Rev. 49: 253- 283.

Sel ander , R. K. 1976. Genet i c var i at i on i n nat ur al popul at i ons.
I n: Mol ecul ar Evol ut i on, F. J . Ayal a, ed. Sunder l and: Si nauer .
Pp. 21- 45.
St r i nger , C. and C. Gambl e. 1993. I n Sear ch of t he
Neander t hal s. New Yor k: Thames and Hudson.

Sugden, R. 1986. The Economi cs of Ri ght s, Co- oper at i on and
Wel f ar e. Oxf or d: Basi l Bl ackwel l .

St ewar d, J . H. 1938. Basi n- Pl at eau Abor i gi nal Soci opol i t i cal
Gr oups. Bur eau of Amer i can Et hnol ogy Bul l et i n 120. Washi ngt on,
D. C.

Vayda, A. P. , and R. A. Rappapor t . 1968. Ecol ogy: Cul t ur al and
Non- Cul t ur al . I n J . A. Cl i f t on, ed. , I nt r oduct i on t o Cul t ur al
Ant hr opol ogy, pp. 477- 497. Bost on: Hought on.

Wat t , W. B. 1977. Adapt at i on at speci f i c l oci . I . Nat ur al
sel ect i on on phosphogl ucose i somer ase of Colias but t er f l i es:
Bi ochemi cal and popul at i on Aspect s. Genet i cs 87: 177- 94.

Wei ssner , P. St yl e and Soci al I nf or mat i on i n Kal ahar i San
Pr oj ect i l e Poi nt s. Amer i can Ant i qui t y 48: 235- 276.

West - Eber har d, M. J . 1983. Sexual sel ect i on, soci al
compet i t i on, and speci at i on. Quar t . Rev. Bi ol . 58: 155- 183.

Whi t e, L. A. 1959. The Evol ut i on of Cul t ur e. New Yor k: Far r ar ,
St r aus.

Wi l de, D. J . 1978. Gl obal l y Opt i mal Desi gn. New Your k: McGr aw
Hi l l .

Wi msat t , W. 1980. Reduct i oni st i c r esear ch st r at egi es and t hei r
bi ases i n t he uni t s of sel ect i on cont r over sy. I n Sci ent i f i c
Di scover y, I I . Case St udi es. T, Ni ckl es ed. Dor dr echt : Rei del .
Pp. 213- 259.

You might also like