CONTENTS iii

HUM AN AGGRESSION
SECOND EDITION

Russell G. Geen

OPEN UNIVERSITY PRESS

Buckingham • Philadelphia
iv HUM AN AGGRESSION

Op en Un iversity Press
Celtic Cou rt
22 Ballm oor
Bu ckin gh am
MK18 1XW

em ail: en q u iries@op en u p .co.u k

world wid e web: www.op en u p .co.u k

an d
325 Ch estn u t Street
Ph ilad elp h ia, PA 19106, USA

First ed ition p u blish ed 1990

Rep rin ted 1999

First Pu blish ed in th is secon d ed ition 2001

Cop yrigh t © Ru ssell G. Geen 2001

All righ ts reserved . Excep t for th e q u otation of sh ort p assages for th e

p u rp ose of criticism an d review, n o p art of th is p u blication m ay be
rep rod u ced , stored in a retrieval system , or tran sm itted , in an y form or by
an y m ean s, electron ic, m ech an ical, p h otocop yin g, record in g or oth erwise,
with ou t th e p rior written p erm ission of th e p u blish er or a licen ce from
th e Cop yrigh t Licen sin g Agen cy Lim ited . Details of su ch licen ces (for
rep rograp h ic rep rod u ction ) m ay be obtain ed from th e Cop yrigh t Licen sin g
Agen cy Ltd of 90 Totten h am Cou rt Road , Lon d on , W 1P 0LP.

A catalogu e record of th is book is available from th e British Library

ISBN 0 335 20472 4 (h b) 0 335 20471 6 (p b)

Library o f Co n gress Catalo gin g-in -Publicatio n Data

Geen , Ru ssell G., 1932–
Hu m an aggression / Ru ssell G. Geen . — 2n d ed .
p . cm . — (Map p in g social p sych ology series)
In clu d es bibliograp h ical referen ces an d in d exes.
ISBN 0 -335-20472-4 — ISBN 0-335 -20471-6 (p bk.)
1. Aggressiven ess. I. Title. II. Series.

BF575.A3 G43 2001

302.5′4 —d c21
2001021146

Typ eset by Grap h icraft Lim ited , Hon g Kon g

Prin ted in Great Britain by Bid d les Lim ited , Gu ild ford an d Kin gs Lyn n
 CONTENTS vii

CONTENTS

Preface xi

ONE In tro ductio n to th e study o f aggressio n 1

Defin ition of aggression 2
Problem s of defi nition 2
Intent to harm 2
Motivation of the victim 2
A working defi nition of aggression 3
The form of aggression 3
Affective an d in stru m en tal aggression 3
Affective aggression 4
Instrum ental aggression 5
Proactive and reactive aggression 5
The two factors in affective aggression 6
Origin s an d an teced en ts of aggression 8
Biological bases of aggression 8
Social learning 16
A n ote on m eth od s of stu d y 18
Su m m ary 19
Su ggestion s for fu rth er read in g 20

TWO Th e pro vo catio n o f aggressio n 21

Fru stration an d aggression 21
The frustration–aggression hypothesis 21
Frustration and social confl ict 24
Frustration and arousal 26
Stress, arousal and aggression 26
Frustration and coping 27
In terp erson al attack an d aggression 27
Com parison of frustration and attack 27
Intent of the aggressor 28
viii HUM AN AGGRESSION

V iolation of norm s 28
The cognitive–neoassociationist m odel of affective aggression 30
En viron m en tal con d ition s an d aggression 32
Am bient tem perature 32
Noise 36
Disp laced aggression 37
Su m m ary 39
Su ggestion s for fu rth er read in g 40

THREE In terven in g pro cesses in aggressio n 41

An ger 41
Excitation transfer 42
Anger as a socially constructed state 44
Anger as a risk factor for violence 45
The suppression of anger 46
Cogn itive p rocessin g 46
From im pulse to cognition 47
Attributional m ediators of aggression 47
Understanding the harm ful act 48
Judging what ought to have happened 50
Social inform ation processing 50
The hostile attribution bias 52
Generation of solutions to confl ict 52
Reciprocality of aggression and social com petence 53
Social scripts 54
The General Affective Aggression Model 55
Alcohol, cognitive processing and aggression 58
Su m m ary 59
Su ggestion s for fu rth er read in g 60

FOUR Mo derato r variables in aggressio n 61

Sex d ifferen ces in aggression 61
Biology or culture? 61
Norm s, expectations and beliefs 62
Provoking conditions 62
Beliefs about consequences 63
Mode of aggressing 64
Person ality an d aggression 67
Stability of aggressive behaviour 67
Assessm ent of aggressiveness 69
Personality variables as m oderators of aggression 70
Cu ltu ral in flu en ces on aggression 75
National culture 75
Subcultures of violence 76
Su m m ary 77
Su ggestion s for fu rth er read in g 79

FIVE Aggressio n in life an d so ciety 80

Violen ce in in tim ate relation sh ip s 80
Proxim al variables 81
Dispositional and contextual variables 82
 CONTENTS ix

Sexu al violen ce 86
Arousal and inhibition 86
Personality variables as m oderators 87
Cognitive processing 90
Violen ce in th e fam ily 91
The cycle of violence 91
Observation of violence between parents 92
Dom estic roots of delinquency 93
Bu llyin g 95
Characteristics of bullies and victim s 95
Reactions to being bullied 96
The social context of bullying 96
Long-term effects of being bullied 97
Su m m ary 97
Su ggestion s for fu rth er read in g 99

SIX Aggressio n in en tertain m en t 100

Motion p ictu res an d television 100
Evidence from laboratory research 101
Evidence from non-laboratory research 104
Processes in m edia-elicited aggression 108
Vid eo gam es 111
V iolence of gam e content 111
Effects of video gam es on players 112
Sp ectator aggression at sp ortin g even ts 115
Popularity of violent sports 115
Effects of violent sports on spectators 116
Su m m ary 118
Su ggestion s for fu rth er read in g 119

SEVEN Ho stility, h ealth an d adjustm en t 121

Ph ysiological effects of an ger an d h ostility 122
Physiology of anger 122
Anger, hostility, hypertension and coronary disease 122
Aggression an d p h ysiological recovery 126
Direct aggression 126
Status of the victim 127
Aggression guilt 127
The nature of the response 128
Th e beh aviou ral con seq u en ces of aggression 130
Catharsis in psychoanalytic theory 131
Evidence bearing on the catharsis hypothesis 132
Aggression and reinforcem ent 134
Is there an alternative to aggression? 135
Su m m ary 136
Su ggestion s for fu rth er read in g 137

EIGHT Po stscript 138

Overview 138
x HUM AN AGGRESSION

Prosp ects for th e fu tu re 141

Methodology 142
Theory 142
Applications 143

Glossary 144
References 148
Index 168
CONTENTS xi

PREFACE

Books abou t h u m an aggression ten d to be of two typ es. Som e are com -
p osed accord in g to a p articu lar p oin t of view. Th ese books organ ize an d
exp licate research fin d in gs from a p ersp ective d eterm in ed by th e th eory
th at u n d erlies th e ap p roach . Oth er books are broad er, m ore in clu sive an d
less th eory-d riven . Th ey stru ctu re m aterial alon g top ical lin es su ch as th e
d evelop m en t of aggression , th e social an d en viron m en tal d eterm in an ts
of aggressive beh aviou r an d strategies for th e con trol of aggression . Th is
book falls som ewh ere between th e two. As a volu m e in a series of sh ort
texts, it does n ot p reten d to be exh au stive in its review. As a book in ten ded
for classroom u se, it d oes n ot asp ire to m ake an y in n ovative con tribu tion s
to th eory. In stead , it reviews a bod y of literatu re th at h as accu m u lated in
th e stu d y of on e typ e of h u m an aggression an d it organ izes th e fin d in gs
of th e literatu re alon g th e lin es of a sim p le p rocess m od el.
Th is book is th e secon d ed ition of Hu m an Aggression , an d its h as been
organ ized to reflect th e state of kn owled ge on th is su bject at th e begin n in g
of th e twen ty-first cen tu ry. Aggression is d efin ed as an act of h arm d oin g
in flicted con sciou sly an d in ten tion ally u p on an u n willin g victim . Trad i-
tion ally, aggression h as been con cep tu alized as p rim arily eith er affective
or in stru m en tal in n atu re, i.e. as eith er an ou tbu rst of beh aviou r an im -
ated by an ger or a relatively affect-free action carried ou t becau se it serves
som e oth er p u rp ose. Th is d istin ction form s th e basis for a m ore recen t
on e between p roactive aggression , wh ich is in stru m en tal in n atu re, an d
reactive aggression , wh ich is a resp on se to som e p rovokin g circu m stan ce.
Affective aggression is fu rth er d efin ed as a resp on se to an aversive ch an ge
in th e p erson ’s en viron m en t – a p rovocation – an d n ot to an yth in g in ter-
n ally gen erated . Th e relative em p h asis of th is book is on th e an alysis of
aggression th at is reactive an d affective, becau se th is is th e ap p roach
taken by m ost stu d en ts of h u m an aggression an d violen ce. Neverth eless,
it will also be clear from a p eru sal of th is book th at in terest in th e in stru -
m en tal n atu re of aggression h as been in creasin g in recen t years, an d th at
in stru m en tal m otives (e.g. p ower an d d om in an ce) are h eavily in volved in
xii HUM AN AGGRESSION

th e exp lan ation of certain everyd ay p roblem s of violen ce, su ch as sp ou se

abu se an d bu llyin g.
It will also be obviou s th at m u ch of th e research an d th eorizin g on
affective aggression is ad d ressed to th e id en tification of p rocesses th at
in terven e between provokin g con dition s an d aggressive reaction s. Here, too,
certain th eoretical tren d s can be seen . Th e p rocesses th at m ed iate aggres-
sion are d escribed in term s of affect, cogn ition an d activation or arou sal.
Cogn ition h as becom e th e m ajor p rocess on wh ich m ost m ed iation al
m odels are bu ilt, with affect an d arou sal as im p ortan t p arallel m ech an ism s.
Th e n atu re of th e cogn itive em p h asis h as ch an ged som ewh at from wh at
it on ce was. In th e 1980s, th e cogn itive an alysis of aggression was based
m ain ly on th e attribu tion of m ean in g to th e p rovokin g even t. More re-
cen tly th e role of cogn ition h as been cou ch ed in term s of su ch m atters as
th e accu racy of social in form ation p rocessin g, with p rocessin g d eficits
servin g as th e u n d erlyin g an teced en t of aggression in p rovokin g situ ation s,
an d th e m ech an ism s by wh ich en viron m en tal con d ition s p rim e aggressive
th ou gh ts, feelin gs an d d isp osition s to act. Th e level of cogn itive an alysis
h as th erefore becom e m ore d etailed an d h as, as a con seq u en ce, in creased
in p red ictive p ower.
Th e im p ortan t fu n ction of backgrou n d , or m od erator, variables in
aggression occu p ies a cen tral p osition in cu rren t research an d th eory.
Th e p rim ary m od erator variables id en tified in th e literatu re are sex, bio-
logical in h eritan ce, socio-cu ltu ral backgrou n d an d p erson ality. In th e first
ch ap ter of th is ed ition , a carefu l d istin ction is m ad e between th ree sets
of variables th at m ake u p th e an teced en ts of aggression : th e p rovokin g
situ ation s, th e backgrou n d m od erators wh ich tem p er th e reaction to th ese
situ ation s an d th e m ean s of aggressin g th at are available to th e aggrieved
in d ivid u al. Un d erstan d in g th e cau ses of h u m an aggression req u ires th at
we keep th ese th ree classes of variables con cep tu ally sep arated , th at we
recogn ize th e ways in wh ich th ey in teract an d th at we recogn ize th e ways
in wh ich th ey are em bed d ed in a th eory of aggression . Pop u lar accou n ts
of sp ectacu lar violen t even ts, like m ass m u rd ers in sch ools an d restau r-
an ts, freq u en tly con fou n d th ese sep arate con tribu tors, with som e observers
blam in g social con dition s, oth ers citin g in dividu al differen ces or th e m en tal
in stability of aggressors an d still oth ers con cen tratin g on th e easy avail-
ability of gu n s. Th ere is n o q u estion th at all of th ese factors are in volved ,
bu t an accu rate an alysis of th e even t req u ires th at each m u st be seen in
term s of its p rop er fu n ction an d n ot as an altern ative ‘cau se’ th at exp lain s
everyth in g.
Th e stu d y of th e relation of th eories of aggression to th e solu tion of
p ractical p roblem s rep resen ts an oth er m ajor em p h asis in cu rren t work.
In creasin g atten tion is n ow bein g d evoted to th e carefu l stu d y of su ch
everyd ay con cern s as sp ou se abu se, sexu al violen ce again st wom en an d
bu llyin g in sch ool. In terest also rem ain s h igh in th e exten t to wh ich th e
m ass en tertain m en t m ed ia p u rvey th e sym bols of violen ce to th e p u blic
an d in wh eth er th ese sym bols evoke aggression in observers. In th e p ast
th is p roblem h as been con stru ed as in volvin g on ly th e effects of violen ce
in m otion p ictu res an d television , bu t in th e p ast d ecad e p u blic con cern
h as focu sed u p on two related p h en om en a: th e rap id ly exp an d in g m arket
for vid eo gam es, m an y of wh ich p resen t violen t con ten t to th e p layer,
PREFACE
CONTENTS xiii

an d th e risin g tid e of violen ce am on g sp ectators at rou gh an d aggressive

sp ortin g even ts. In all th ese areas of ap p lied research , we n ow see exp lan a-
tion s of th e fin d in gs p resen ted in term s of th e th eories an d con cep tu al
m od els th at h ave been d evelop ed th rou gh m ore basic research . In fact,
on e of th e m ore p rom isin g d evelop m en ts in th e stu d y of aggression is th e
d egree to wh ich sop h isticated th eoretical m od els are n ow in voked in d is-
cu ssion s of su ch p ractical m atters as th e role of tem p eratu re in aggression ,
th e cau ses of d om estic violen ce an d wh at h ap p en s to p eop le wh en th ey
watch m ayh em on th eir TV screen s.
Th is secon d ed ition of Hum an Aggression con sists of eigh t ch ap ters. In
th e first (‘In trod u ction to th e stu d y of aggression ’), certain m atters of
gen eral im p ortan ce are d iscu ssed , e.g. th e m ean in g of aggression , th e
m eth od s by wh ich it is stu d ied an d th e backgrou n d p rovid ed by biological
in h eritan ce an d social learn in g. Th e secon d ch ap ter (‘Th e p rovocation of
aggression ’) goes in to th e situ ation al circu m stan ces th at p rovoke affective
aggressive beh aviou r. Th ese circu m stan ces in clu d e fru stration , p ain fu l
stim u lation , in terp erson al attack an d su ch en viron m en tal con d ition s as
in ten se h eat or cold an d u n con trollable n oise. A n ew con cep tu al m od el
th at lin ks all th ese an teced en ts to ch an ges in n egative affect is d escribed
in d etail.
Ch ap ter 3 (‘In terven in g p rocesses in aggression ’) d escribes th e in terven -
in g processes th at m ediate th e con n ection between th e provocation an d th e
aggressive resp on se. It is d ivid ed in to two section s, th e first d ealin g with
an ger an d th e secon d with cogn itive p rocessin g. In clu d ed are su ch top ics
as excitation tran sfer, an ger as a social con stru ct, th e attribu tion al processes
in aggression , th e social in form ation processin g m odel an d th e role of social
scrip ts in aggression . Backgrou n d variables are d iscu ssed in Ch ap ter 4
(‘Mod erator variables in aggression ’). Becau se th e gen etic-biological in h eri-
tan ce of th e p erson is d iscu ssed in Ch ap ter 1, th e treatm en t of m od erators
in Ch ap ter 4 covers sex, p erson ality an d socio-cu ltu ral backgrou n d .
Ch ap ters 5, 6 an d 7 are d evoted to a d iscu ssion of som e of th e ways in
wh ich th e th eories an d con cep ts reviewed in th e first fou r ch ap ters h ave
been ap p lied to aggressive beh aviou r in n atu ral settin gs. In Ch ap ter 5
(‘Aggression in life an d society’), fou r top ics are reviewed . Th e first is
violen ce in in tim ate relation sh ip s, with an em p h asis on th e violen ce of
m en towards fem ale partn ers. Th e secon d is sexu al violen ce again st wom en .
Th e th ird is th e effect of aggression with in th e fam ily, in clu d in g th e
in tergen eration al ‘cycle of violen ce’, th e effects of p aren tal figh tin g on
ch ild ren an d som e of th e roots of d elin q u en cy in early fam ily exp eri-
en ces. Th e fou rth is bu llyin g am on g sch oolch ild ren , in clu d in g som e of
th e social asp ects of bu llyin g an d certain of th e lon g-ran ge ou tcom es.
Ch ap ter 6 (‘Aggression in En tertain m en t’) organ izes th e evid en ce from
laboratory an d n on -laboratory research regard in g th e effects of observ-
in g violen ce in th e cin em a an d on television accord in g to a n u m ber of
variables an d p rocesses th at m ed iate th ese effects. Th e ch ap ter also in -
clu d es su m m aries of evid en ce on th e effects of violen t vid eo gam es on
gam e p layers an d a d iscu ssion of som e of th e effects of observin g rou gh
an d aggressive sp ortin g even ts. Ch ap ter 7 (‘Hostility, h ealth an d ad ju st-
m en t’) exam in es two issu es. On e is th e effect of an ger an d h ostility u p on
p h ysical h ealth , in clu d in g su ch q u estion s as wh eth er th e exp ression or
xiv HUM AN AGGRESSION

rep ression of aggressive em otion s can h ave ad verse effects on th e card io-
vascu lar system . Th e secon d issu e is th at of wh eth er aggressin g wh en
an gry is followed by a p eriod of less aggression , accord in g to th e fam iliar
cath arsis h yp oth esis.
Th e fin al ch ap ter (‘Postscrip t’) is a sh ort su m m ary of som e of th e m ain
th em es of th e book an d a su ggestion of wh ere th e stu d y of aggression
ap p ears to be h ead ed .
Th is book was written for classroom u se. It is ad d ressed p rim arily to
u n d ergrad u ate stu d en ts wh o h ave h ad at least an in trod u ctory cou rse in
p sych ology an d are th erefore acq u ain ted with th e fu n d am en tals of th e
field . At th e u n d ergrad u ate level it cou ld serve as th e basic textbook for a
cou rse. Becau se of th e gen eral orien tation of th e book to basic research on
aggression , it m ay also be u sefu l for grad u ate stu d en ts wh o d esire an
overview of th at research . At th e grad u ate level it will p robably best serve
th e fu n ction of a p rim er u sed in con n ection with p rim ary sou rce m ater-
ials. Stu den ts u sin g th e book as eith er a textbook or a secon dary sou rcebook
m ay fin d u sefu l th e an n otated referen ces listed as su ggestion s for fu rth er
read in g at th e en d of each ch ap ter. It sh ou ld also be n oted th at, alth ou gh
th e m ajor focu s of th e book is on a social p sych ological ap p roach to
h u m an aggression , it also con tain s m aterial of relevan ce for th e areas of
p h ysiological, d evelop m en tal an d p erson ality p sych ology.
No con certed attem p t is m ad e in th is book to recom m en d strategies
for th e con trol or elim in ation of aggression . Th e book is in ten d ed on ly
for p u rp oses of in trod u cin g th e read er to wh at is kn own abou t h u m an
aggression at th e p resen t tim e from research th at h as been con d u cted on
th e p roblem . Th u s, th ere is exten sive d iscu ssion of th e resu lts of stu d ies
bu t n o ch ap ter on p u blic p olicy. My in ten tion is to in form th e read er. It
is h op ed th at an y d ecision s th at are even tu ally m ad e with resp ect to th e
p roblem of h u m an aggression , an d wh at can be d on e abou t it, will be
based on th e in form ation th at research h as given u s.
Russell G. Geen
Colum bia, Missouri, USA
 INTRODUCTION TO THE STUDY OF AGGRESSION 1

ON E

INTRODUCTION TO THE STUDY

OF AGGRESSION

Few p eop le wou ld d en y th at ‘aggression ’ is com m on p lace in con tem p orary

society. For som e, su ch as th ose livin g in th e Mid d le East, th e Balkan s,
p arts of cen tral Africa an d an y m axim u m secu rity p rison in th e Un ited
States, aggression an d vio len ce are exp erien ced d aily an d in in ten sely
p erson al ways. For oth ers th e p h en om en on is kn own m ain ly in in d irect
ways, su ch as th rou gh m otion p ictu res an d television . However, even th ose
fortu n ate en ou gh to h ave been sp ared th e d irect exp erien ce of leth al viol-
en ce m ay occasion ally en cou n ter som eth in g p erceived to be aggression in
a less h arm fu l form , like verbal in su lt, rou gh p h ysical con tact or h ostile
rejection . Aggression , wh eth er h arm fu l to life an d lim b or m erely p ain fu l
to th e ego, seem s to be a real an d im p ortan t p art of th e h u m an con d ition .
Th e broad an d in clu sive way in wh ich th e word h as been u sed , h ow-
ever, m akes th e system atic stu d y of aggression d ifficu lt. To p u t th e m atter
sim p ly: d oes it m ake sen se to u se th e sam e word to refer to su ch d issim ilar
even ts as a gan glan d m u rd er, th e bom bin g of a restau ran t, a figh t at a
football m atch an d a cu ttin g rem ark at a cocktail p arty? Even if th e term
is u sed on ly in casu al everyday speech , its u sefu ln ess is obviou sly lim ited by
th e n u m erou s m otives an d form s of exp ression th at ch aracterize ‘aggres-
sive’ beh aviou r. W h en we attem p t to u se th at word in scien tific d iscou rse,
th e p roblem becom es even m ore seriou s. Scien ce d ep en d s on p recision
an d clarity of defin ition s. From th at stan dp oin t, we m igh t do well to forget
abou t a u n itary con cep t of ‘aggression ’ an d to search in stead for fu n ction al
relation sh ip s between sp ecific acts an d th eir eq u ally sp ecific cau ses. Th e
variou s beh aviou rs n ow su bsu m ed by th e word ‘aggression ’ cou ld u n dou bt-
ed ly be stu d ied as in d ivid u al p h en om en a d efin ed in term s of th eir own
an teced en t con d ition s, in terven in g p rocesses, an d ou tcom es. Neverth eless,
su ch stu d ies wou ld obscu re th e p ossibility th at th e aggressive beh aviou rs
n oted above, h owever d ifferen t th ey m ay seem , p ossess som e com m on
featu res. If we are to m ake an y sen se of th e wh ole id ea of aggression , it is
th ese featu res th at we m u st seek. W e m u st th erefore begin by d efin in g ou r
term .
2 HUM AN AGGRESSION

DEFINITION OF AGGRESSION

Problems of definition

On e m igh t th in k th at p eop le wou ld be in su bstan tial agreem en t on d efin -

in g som eth in g as im p ortan t an d p ervasive as aggression , bu t su ch is n ot
th e case. Probably m ost p eop le, in clu d in g p sych ologists, wou ld agree
in gen eral with th e d efin ition give by Bu ss (1961: 1): aggression is ‘a
resp on se th at d elivers n oxiou s stim u li to an oth er organ ism ’. Certain ly
wh at we ord in arily call aggression d oes in volve aversive stim u lation of
som e sort an d in ten sity, wh eth er it be in th e form of a bu llet, a bom b
fragm en t, a p h ysical blow, or som e m ore su btle act like an in su lt or an
u n d eserved criticism . W e d o n ot en cou n ter th e real p roblem of statin g
wh at aggression is u n til we attem p t to form u late a d efin ition th at goes
beyon d sim p le h arm d oin g beh aviou r. Aggression is n ot as sim p le as a
p u rely beh aviou ral d efin ition wou ld in d icate. Oth er elem en ts m u st be
ad d ed , an d th ese elem en ts create certain com p lexities.

Intent to harm

On e con stru ct th at m ost p eop le wou ld p robably con sid er n ecessary to an

ad eq u ate d efin ition of aggression is intent to h arm th e victim . Th e n otion
of in ten tion is exp licit in th e d efin ition of aggression given by on e in flu -
en tial grou p of p sych ologists: ‘Aggression is . . . an act wh ose goal-resp on se
is in ju ry to an organ ism ’ (Dollard et al. 1939: 11). Use of th e term ‘goal-
resp on se’ in d icates m otivation an d strivin g, so th at aggression becom es
th e en d resu lt in a d eliberate series of action s. Dollard an d h is colleagu es
sp ecifically ru le ou t th e p ossibility th at aggression can be th e accid en tal
con seq u en ce of beh aviou r lackin g in ten t to h arm : ‘On e p erson m ay in ju re
an oth er by sh eer accid en t. Su ch acts are n ot aggression , becau se th ey are
n ot goal-resp on ses’ ( Dollard et al. 1939: 11). Th e id ea th at in ten t is re-
q u ired in aggression h as been d ism issed by som e on th e grou n d s th at it is
a m en tal con cep t th at lacks objectivity an d d efies rigorou s an alysis (e.g.
Bu ss 1961). However, d esp ite th e p roblem s th at it raises, th e con cep t of
in ten t seem s u n avoid able. For exam p le, a team of firefigh ters attem p tin g
to rescu e som eon e trap p ed in a bu rn in g bu ild in g m ay cau se th e p erson
p ain , bu t th e h arm d on e is h ard ly d eliberate an d m ay even be n ecessary
to save th e p erson ’s life. Th e firefigh ters’ beh aviou r wou ld th erefore n ot
be con sid ered aggressive by m ost observers.

M otivation of the victim

A secon d ch aracteristic of aggression th at sh ou ld go in to a d efin ition

h as been stated by Baron an d Rich ard son : ‘Aggression is an y form of
beh avior d irected toward th e goal of h arm in g or in ju rin g an oth er livin g
bein g wh o is m otivated to avoid su ch treatm en t’ (Baron an d Rich ard son
1994: 7). Th e u n willin g p articip ation of th e victim is often overlooked in
 INTRODUCTION TO THE STUDY OF AGGRESSION 3

d iscu ssion s of aggression , p erh ap s becau se it is d ifficu lt to im agin e situ -

ation s in wh ich a p erson d oes n ot m ake at least som e effort to escap e or
avoid bein g attacked . Still, situ ation s d o arise in wh ich a p erson m ay
tolerate, an d even seek ou t, p u n ish m en t to aton e for gu ilt. In su ch a case
th e willin g recip ien t wou ld p robably n ot accu se th e p u n ish in g agen t of
bein g an aggressor. Baron an d Rich ard son (1994) also con sid er su icid e to
be h arm fu l beh aviou r th at is obviou sly n ot resisted by th e victim , an d
th ey th erefore con clu d e th at su icid e can n ot be classified as aggression .

A working defi nition of aggression

W ith all of th e foregoin g in m in d , we can n ow su ggest a workin g d efin i-

tion of h u m an aggression : aggression is th e delivery of an aversive stim u lu s
from on e p erson to an oth er, with in ten t to h arm an d with an exp ectation
of cau sin g su ch h arm , wh en th e oth er p erson is m otivated to escap e or
avoid th e stim u lu s. It is ad m itted th at th is d efin ition m ay n ot cover all
exam p les an d th at it can be attacked on several p oin ts. Nor d oes it even
begin to d eal with m an y of th e variables in volved in aggression . For
in stan ce, it d oes n ot m en tion th e role p layed by em otion s in m an y aggres-
sive action s. It d oes n ot accou n t for th e com p lex cogn itive ju d gem en ts
th at often p reced e aggression . It d oes n ot take in to accou n t th e fact th at
aggression is often recip rocal, or th at it m ay som etim es serve as a valu able
‘safety valve’ for a p erson exp erien cin g extrem e levels of an ger. Th ese an d
m an y oth er su ch p roblem s all im p in ge u p on a basic d efin ition of aggres-
sion , yet th ey go beyon d a sim p le on e an d in fact com p rise th e m an y
variables th at con stitu te m u ch of th e m aterial of th is book.

The form of aggression

W h en th ey h ear th e word ‘aggression ’ m ost p eop le p robably ten d first

to th in k of p h ysical force – a fist-figh t, an assau lt with a weap on , a lou d
verbal retort or som e oth er form of in ten se an d p u n itive action en acted
in th e cou rse of con flict between two p eop le. Actu ally, accord in g to th e
d efin ition we h ave ad op ted , aggression m ay be carried ou t in any beh a-
viou r actu ated by an in ten t to h arm an oth er p erson again st th at p erson ’s
wish es. Sp read in g viciou s gossip abou t som eon e in h op es of ru in in g th at
p erson ’s rep u tation wou ld be con sid ered aggression . Likewise, d am agin g
or destroyin g an oth er’s p rop erty can be a h igh ly effective way of aggressin g
again st th at p erson . Even som eth in g as su btle an d con trolled as a social
sn u b can be a p owerfu l sou rce of h arm to th e victim , a h arm th at is
clearly in ten d ed by th e p erson d eliverin g it.

AFFECTIVE AND INSTRUM ENTAL AGGRESSION

Th e exam p les of aggression th at h ave been given so far su ggest m an y

d ifferen t kin d s of aggressive beh aviou r in asm u ch as th ey reveal varyin g
4 HUM AN AGGRESSION

cau ses an d m otives. It is com m on in th e stu d y of aggression to id en tify

su ch varieties of assau ltive beh aviou r on th e basis of an teced en ts an d
in terven in g p rocesses. Th u s we d iscrim in ate am on g su ch action s as ‘an gry
aggression ’, ‘in stru m en tal aggression ’ an d ‘learn ed’, ‘im itative’ or ‘biogen ic’
aggression becau se th ese beh aviou rs d iffer from each oth er in im p ortan t
ways regard in g th eir u n d erlyin g m otives. In th e an im al kin gd om , several
kin d s of aggression are rou tin ely observed , su ch as m atern al, territorial
an d p red atory. Am on g h u m an s we m ay also n ote certain d istin ction s,
su ch as an gry retaliation , self-d efen ce, calm an d cold -blood ed retribu tion
an d violen ce carried ou t for p u rp oses of coercion , p u n ish m en t an d p rofit.
Neverth eless, it is cu stom ary am on g th ose wh o stu d y h u m an aggression
to classify aggressive beh aviou r in to two categories: an gry or affective
aggressio n , in wh ich h arm in g th e victim is th e m ain m otive for th e act;
an d in strum en tal aggressio n , wh ich m ay or m ay n ot in volve stron g
em otion s bu t is m otivated by con cern s m ore im p ortan t to th e aggressor
th an th e h arm d oin g itself.

Affective aggression

Aggression is often accom p an ied by stron g n egative em otion al states. Th e

em otion th at we call an ger is u su ally arou sed by som e pro vo catio n . An ger
is m ost often th ou gh t of as an in terven in g con d ition th at in stigates, an d
th en gu id es, affective aggressive beh aviou r aim ed p rim arily at in ju rin g
th e p rovokin g p erson . It is accom p an ied by d istin ctive p attern s of activity
in th e cen tral an d au ton om ic n ervou s system s, in clu d in g activation of th e
h yp oth alam u s, in creased blood flow to th e m u scu latu re, h eigh ten ed blood
p ressu re an d p u lse rate, p u p illary d ilation an d d ecreased flow of blood to
th e viscera ( Joh an sson 1981).
Th e id ea th at a flash of an ger can in sp ire retaliatory aggression is easy
en ou gh to grasp . Som etim es, h owever, retaliation com es so lon g after
p rovocation th at we fin d it d ifficu lt to attribu te th e action to an ger, an
em otion th at is relatively sh ort-lived for m ost p eop le, even th ou gh th e
retaliatory asp ect of th e an ger is still ap p aren t. Frijd a (1994) h as com -
m en ted on th e p ossible em otion al state in volved in th is sort of ‘d elayed -
resp on se’ affective aggression . It h as m an y of th e p rop erties of an ger: it is
a state of im p u lse, it d isp oses th e p erson to action , it is often accom p an -
ied by bod ily aro usal, it can becom e a p reoccu p ation th at takes atten tion
away from oth er m atters. Bu t it d iffers from an ger in oth er ways, on e of
wh ich is th e often exten d ed d u ration between p rovocation an d resp on se.
Frijd a su ggests th at th is con d ition is n ot an em otion , per se, bu t a com p lex
cogn itive state h avin g close lin ks to em otion , th at h as grown ou t som e
p u rer em otion al state. Su ch a state m ay be labelled a ‘sen tim en t’: th e
em otion of an ger toward s th e oth er p erson becom es in tim e tran sform ed
in to th e sen tim en t of hatred, wh ich ou tlives th e origin al an ger. Lon g-term
feu d s an d gru d ges rep resen t cases in wh ich p eop le are aggressed again st
becau se th ey are h ated , n ot becau se th ey h ave d on e an yth in g in p articu -
lar to elicit an ger in th e aggressor. Frijd a’s su ggestion s ad d an im p ortan t
an d u su ally overlooked d im en sion to th e an alysis of affective aggression
an d in d icate th at fu tu re stu d y of m otives for retaliation is in ord er.
INTRODUCTION TO THE STUDY OF AGGRESSION 5

Instrumental aggression

Beh aviou r n eed n ot h ave a stron g em otion al basis to be aggressive, h ow-

ever. Peop le often attack oth ers with in ten t to h arm bu t with ou t n ecessar-
ily feelin g an y m alice toward s th em ; th e aggression is sim p ly a m ean s to
som e oth er en d . On e su ch en d is self-d efen ce, wh ich m ost cou rts of law
recogn ize as a valid ju stification for acts of violen ce. Aggressive m ilitary
action is u su ally carried ou t for con trol of territory or to d issu ad e th e
en em y from attackin g on e’s own troop s. An oth er typ e of in stru m en tal
aggression is th e attem p t to establish co ercive po w er over oth ers th rou gh
violen ce or th e th reat of violen ce (Ted esch i an d Felson 1994), Fin ally, in
on e of th e m ost wid ely cited research stu d ies in recen t tim es, Milgram
(1963) sh owed th at p eop le were cap able of com m ittin g acts of violen ce
again st oth er h u m an bein gs sim p ly in obed ien ce to com m an d s from a
p erson with au th ority. It m u st be n oted th at th e d istin ction between
affective an d in stru m en tal aggression is n ot a rigorou s on e. Th e two kin d s
of aggression are n ot m u tu ally exclu sive, an d som e acts of aggression h ave
both affective an d in stru m en tal p rop erties. For exam p le, a m oth er wh o
becom es exasperated at h er ch ild’s beh aviou r an d u ses corporal pu n ish m en t
m ay be m otivated to m od ify th e ch ild ’s beh aviou r (an in stru m en tal u se of
aggression ) wh ile still reactin g to th at beh aviou r with an ger. For p u rp oses
of u n d erstan d in g h er beh aviou r, on e m u st d eterm in e wh ich of th e two
categories describes h er principal m otivation al state an d an alyse th e relevan t
variables accord in gly.

Proactive and reactive aggression

A n u m ber of recen t stu d ies of aggression d raw a d istin ction between

reactive an d pro active aggressio n (e.g. Crick an d Dod ge 1996). Th e first
of th ese term s refers to aggressive beh aviou r th at is en acted in resp on se
to p rovocation , su ch as an attack or an in su lt, an d it is m an ifested in
both self-d efen sive an d an gry action s. Th e latter term refers to aggres-
sion th at is in itiated with ou t ap p aren t p rovocation , su ch as we see in
bu llyin g beh aviou r. Su ch beh aviou r is n ot evoked by an ger, h o stility or
th e n eed to d efen d on eself, bu t by oth er m otives th at relate to obtain in g
good s, assertin g p ower, assu rin g th e ap p roval of referen ce grou p s an d
oth er su ch goals. Reactive an d p roactive aggression are th e eq u ivalen t of
wh at earlier th eorists called affective an d in stru m en tal aggression . Th e
affective –in stru m en tal (or reactive–p roactive) d ifferen tiation will com e u p
in certain sp ecific con texts in th is book, an d th e read er sh ou ld bear in
m in d th e d ifferen ces between th e two kin d s of aggression .
Th e rem ain d er of th is book is d evoted p rim arily to a d iscu ssion of th e
p rocesses in volved in affective aggression . Th is em p h asis is in n o way
in ten ded to im ply th at in stru m en tal aggression is u n im portan t. However, it
h as n ot been stu d ied in n early th e sam e d ep th as h as affective aggression .
Th ere is, for exam p le, n o large bod y of literatu re an alysin g th e variables
in volved wh en on e p erson h u rts an oth er for m on ey, n or d o we h ave
exten sive d ata from con trolled stu d ies on th e m ed iators of self-d efen ce.
Som e stu d ies h ave begu n to in q u ire in to th e an teced en ts of each typ e of
6 HUM AN AGGRESSION

aggression , an d th ey will be cited . On th e oth er h an d , we h ave a great

d eal of in form ation on affective aggression an d th e p rocesses th at con trib-
u te to it.

The two factors in affective aggression

Even tu ally, all d iscu ssion s of aggression get arou n d to cau ses. W h y d o
som e p eop le h u rt, crip p le or kill oth ers in con texts th at in on e way or
an oth er in volve an gry reaction s to real or im agin ed p rovocation s? Th e
alleged cau ses of h u m an aggression ten d to be varied an d com p lex, an d
d iscu ssion s of th em are m ost often in con clu sive. Con sid er an exam p le.
On 20 Ap ril 1999, at 11:25 in th e m orn in g, two m ale stu d en ts en tered
th eir h igh sch ool in Littleton , Colorad o, arm ed with sawn -off sh otgu n s,
sem i-au tom atic weap on s an d h om em ad e p ip e bom bs. After firin g an esti-
m ated 900 rou n d s at fellow stu d en ts, staff m em bers an d an arm ed secu rity
gu ard , an d tossin g n u m erou s bom bs, th e two en d ed th e ord eal by killin g
th em selves. In all, twelve stu d en ts an d on e teach er d ied in th e attack, in
ad d ition to th e killers.
Alm ost im m ed iately, th e n ews m ed ia began rep ortin g op in ion s on th e
cau ses of th e m ass m u rd er exp ressed by citizen s’ grou p s, n ewsp ap er an d
television com m en tators, p rofession al p sych ologists an d p olitical figu res.
Th e im p u ted cau ses were p red ictably variable. Grou p s th at p rom ote legis-
lation for con trol of gu n s cited th e easy availability of weap on s in th e
Un ited States. Som e p sych ologists n oted th at th e killers were p art of a
stu d en t su bcu ltu re th at was rejected an d rid icu led by th eir m ore p op u lar
sch oolm ates, an d th at th ey h ad on occasion exp ressed h atred for su ch
class ‘stars’. It was also n oted th at th e two h arbou red racist an d n eo-Nazi
sen tim en ts, an d th at th ey h ad tau n ted on e of th eir victim s, an African -
Am erican , with a racial ep ith et before sh ootin g h im . Con servative religiou s
grou p s traced th e killin g to a breakd own in trad ition al valu es an d fam ily
stru ctu re. Th is social breakd own was attribu ted n ot on ly to a d eclin e in
p aren tal su p ervision bu t also to su ch m ass cu ltu ral effects as violen t son g
lyrics, m otion p ictu res, television an d vid eo gam es. It was also n oted th at
on e of th e stu d en ts h ad obtain ed d etailed in form ation on th e m an u fac-
tu re of bom bs over th e In tern et. Man y p eop le ten d ed to em p h asize on e or
th e oth er of th ese ‘cau ses’ to th e relative exclu sion of th e oth ers, an d to
gen eralize th eir con clu sion s far beyon d th e d etails of th e case.
As we review th ese an teced en ts of th e Littleton m assacre, n ote th at
th ey can be grou p ed in to th ree categories. On ly on e of th e th em is recog-
n izable as a tru e provocation cap able in itself of evokin g an aggressive
resp on se – th e in su ltin g an d rid icu lin g beh aviou r of som e of th e killers’
classm ates. Two oth er exp lan ation s are actu ally related m ore to m ethods of
aggression th an to cau ses – th e easy p rocu rem en t of gu n s an d access
to th e in form ation n ecessary to m an u factu re bom bs. Th e th ird category
p ertain s to certain variables th at m ay h ave in creased th e likelih ood th at
th e two killers wou ld beh ave aggressively u n d er p rovokin g con d ition s, i.e.
wh at we m igh t call background variables: th eir h istory of exp osu re to
sym bols of violen ce in th e m ass m ed ia an d electron ic gam es, th eir fam ily
situ ation s, th eir attitu d es toward s violen ce, th eir gen eral valu es, th eir
INTRODUCTION TO THE STUDY OF AGGRESSION 7

p erson alities. Non e of th ese latter variables in itself gu aran tees th e com -
m ission of overt aggression , bu t, by th e sam e token , n o p rovocation is
likely to m ake a p erson a m u rd erer u n less th at p erson is d isp osed by th ese
backgrou n d con d ition s to becom e on e. An d , fin ally, it is obviou s th at th e
likelih ood of aggressive beh aviou r is red u ced if th e m ean s of aggressin g
are n ot read ily available.
Th eorizin g on th e n atu re an d cau ses of h u m an aggression m u st begin
with th e recogn ition of th e d istin ction am on g th e th ree classes of vari-
able: p rovocation , m eth od an d backgrou n d or d isp osition . Society will
n ot solve an y of its p roblem s with violen ce by treatin g th e th ree as if th ey
were eq u ivalen t. For on e th in g, th e th ree relate to d ifferen t tim e-fram es
for th e con trol of violen t beh aviou r. It is p robable th at th e best lon g-ran ge
strategy for con trollin g an d red u cin g violen ce wou ld be th e creation of a
n ation al cu ltu re in wh ich aggression d oes n ot en ter to su ch a large d egree
in to social an d p erson al valu e system s an d in wh ich im p ression able ch il-
d ren are n ot bom bard ed with violen t am u sem en ts an d en tertain m en ts.
It is also p robable th at greater efforts to p rom ote toleran ce an d m u tu al
resp ect across racial, eth n ic an d social class bou n d aries m igh t d ecrease th e
p revalen ce of th e sort of rage th at m otivated th e you n g m en in Colorad o.
Few p eop le will argu e with su ch p rop osition s for lon g-ran ge reform s. How-
ever, it seem s in argu able th at th e sim plest im m ediate strategy for preven tion
of su ch violen ce is to red u ce d rastically th e n u m ber of gu n s in th e h an d s
of th e gen eral p op u lation . It is legitim ate to argu e for ch an ges in th e valu e
system of th e cou n try an d to ad vocate greater m u tu al love an d resp ect,
bu t it is also d isin gen u ou s to set u p th ese goals as altern atives to gu n
con trol wh en we are talkin g abou t th e p ractical red u ction of violen ce.
Psych ologists are, h owever, m ore con cern ed with isolatin g th e cau ses of
aggression th an with en gin eerin g th e m ean s of con trol. W h en we sp eak
of th e cau ses of aggression , in d ep en d en t of th e relative ease with wh ich it
is carried ou t, we are d ealin g with two of th e classes of variable m en -
tion ed above – th ose p ertain in g to th e social an d p erson al backgrou n d of
th e in d ivid u al, an d th ose p ertain in g to p rovocative an d an ger-in d u cin g
situ ation s. Th is book is organ ized alon g lin es of su ch a two-p rocess ap -
p roach . Su ch an ap p roach is ad m itted ly on ly a p relim in ary statem en t
abou t affective aggression . It is, fu rth erm ore, sim p le to th e p oin t of trivi-
ality. However, we gain som eth in g by u sin g su ch a fram ework for aggres-
sion . On e som etim es h ears th at becau se aggression is p art of ‘h u m an
n atu re’ th ere is little th at can be d on e abou t it. Th erefore, th e argu m en t
goes, society sh ou ld forget social reform an d oth er am eliorative m easu res
th at are ad d ressed , am on g oth er th in gs, to th e con trol an d red u ction of
violen ce. W e d o n ot d en y th at th e p oten tial for aggression is p resen t in
th e con stitu tion of h u m an bein gs. However, by d efin in g aggression as a
reaction to situ ation s, we can h ave som e reason for h op e th at p rop er
social action m ay at least lim it, if n ot rem ove altogeth er, th e likelih ood of
violen ce. In ad d ition , we can seek to create social con d ition s th at p rom ote
beh aviou r th at is in com p atible with aggression .
On th e oth er h an d , on e som etim es h ears th at all aggression is th e resu lt
of bad social con d ition s, cu ltu ral d ep rivation s, u n h ap p y ch ild h ood exp eri-
en ces an d so on . Su ch argu m en ts ten d to rem ove all resp on sibility for
violen ce from th e p erp etrators an d to lay th e blam e en tirely on oth ers.
8 HUM AN AGGRESSION

Su ch argu m en ts d o n ot exp lain wh y m ost p eop le wh o exp erien ce th e

variou s con d ition s th ou gh t to be th e sole cau ses of violen ce are n ot h igh ly
aggressive. W e wou ld argu e th at su ch p eop le, th ou gh in stigated to aggress
as m u ch as oth ers, d o n ot p ossess th e backgrou n d ch aracteristics of ag-
gressors. If su ch is th e case, society’s task is to d iscover th e con d ition s
u n d er wh ich su ch relatively u n aggressive ten d en cies are fostered an d to
seek to im p lem en t th em m ore wid ely.

ORIGINS AND ANTECEDENTS OF AGGRESSION

Con sid erable con troversy exists over th e origin s of aggression in h u m an s.

In th is con troversy we see rep eated th e old issu e of n atu re versu s n u rtu re
in p sych ology, with m an y of th e old q u estion s raised again . Is h u m an
beh aviou r th e resu lt of gen etically in h erited biological d rives an d im -
p u lses or is it acq u ired th rou gh exp erien ce an d learn in g? Is aggression
a n orm al p art of h u m an beh aviou r carried over from ou r p rim ate an ces-
tors or is it an aberration of h u m an n atu re origin atin g in social an d
en viron m en tal con d ition s? Does belief in a biological basis for aggres-
sion p reclu d e attem p ts at social con trol of violen ce? Are con flict an d war
in evitable? Evid en ce can be ad d u ced for both th e biological an d th e beh a-
viou ristic exp lan ation s of aggression . Th ere is little d ou bt th at m u ch of
h u m an aggression is learn ed . Beh aviou ral biologists d o n ot d en y th is.
However, n eith er can it be d en ied th at h u m an s sh are with lower an im als
certain d isp osition s to aggress th at are tran sm itted gen etically. In th e
followin g two section s we review som e of th e evid en ce bearin g on th ese
viewp oin ts.

Biological bases of aggression

Beh avio ur gen etics
Th e p ossibility th at som e h u m an aggression m ay be attribu table to h ered i-
tary factors is su p p orted by a growin g n u m ber of stu d ies. At on e tim e,
esp ecially in th e Un ited States wh en beh aviou rism d om in ated p sych ology,
few p sych ologists believed th at h u m an beh aviou r h ad h ered itary origin s.
It was p ractically a tru ism th at all beh aviou r is learn ed . Nowad ays th is
p rem ise is n ot as wid ely accep ted . Hered itary aggression in lower an im als
h as n ever been seriou sly q u estion ed , an d in recen t years th e id ea th at at
least som e p art of h u m an aggressiven ess is in h erited h as been gain in g
in creasin gly in accep tan ce.
An obviou s p roblem th at arises in th e stu d y of th e gen etic bases of
aggression in h u m an s is m eth od ological. Hu m an rep rod u ction can n ot
be con trolled th rou gh selective breed in g in th e sam e way as th at of
lower organ ism s, so altern ative m ean s of an alysin g gen etic in flu en ce
m u st be fou n d . On e su ch m eth od in volves com p arison between p airs of
twin s. Twin stu d ies an alyse sim ilarities between m em bers of id en tical
(m o n o zygo tic) an d n on -id en tical, or fratern al (dizygo tic), p airs. Th e
m eth od rests on th e fact th at wh ereas d izygotic twin s sh are a com m on
INTRODUCTION TO THE STUDY OF AGGRESSION 9

Table 1.1 Correlation s of twin typ e (m on o- or d izygotic) with

p erson ality variables

Scale Monozygotic Dizygotic

Altru ism 0.53 0.25

Em p ath y 0.54 0.20
Nu rtu ran ce 0.49 0.14
Aggressiven ess 0.40 0.04
Assertiven ess 0.52 0.20

Source: Ru sh ton et al. (1986).

en viron m en t bu t are n ot id en tical in gen etic m ake-u p , m on ozygotic twin s

sh are a com m on en viron m en t an d are also id en tical in h ered ity. If a trait
h as som e h ered itary basis, th erefore, it sh ou ld be sh ared by m on ozygotic
twin s to a greater exten t th an by m em bers of d izygotic p airs. Correlation s
of th e trait between p airs of twin s of each typ e are com p ared, an d eviden ce
of h igh er correlation s am on g m on ozygotics is taken as evid en ce of som e
h eritability associated with th e trait.
Th e evid en ce from twin stu d ies on th e role of in h erited biological
factors in h u m an aggression is m ixed an d in con clu sive. Th e m ost con -
vin cin g stu d ies h ave been th ose in wh ich com p arison s h ave been m ad e
on th e basis of self-rep orts of aggressiven ess on p erson ality in ven tories
(e.g. McGu e et al. 1993). In on e su ch in vestigation , Ru sh ton et al. (1986)
ad m in istered q u estion n aires assessin g five p erson ality traits to m ore th an
500 p airs of ad u lt twin s. Th ree of th e traits (altru ism , em p ath y an d
n u rtu ran ce) were assu m ed to be n egatively related , an d two (aggressive-
n ess an d assertiven ess) p ositively related , to aggressive beh aviou r. Th e
results (Table 1.1) sh owed th at correlation s were h igh er am on g m on ozygotic
twin s th an am on g dizygotic p airs for each of th e five traits. Sim ilar fin din gs
of h eritability h ave been rep orted in stu d ies in volvin g m eth od s oth er
th an p erson ality in ven tories. In on e su ch stu d y, m oth ers’ d escrip tion s
of th e beh aviou r of th eir twin ch ild ren revealed a sign ifican t h eritability
factor for aggressiven ess (Gh od sian -Carp ey an d Baker 1987).
On th e n egative sid e, several stu d ies in volvin g self-rep ort in ven tories
h ave yield ed eith er sm all h eritability effects or n o su ch effects at all (e.g.
Carm elli et al. 1990). In m ost su ch cases, in ven tories d ifferen t from
th e on es sh owin g p ositive effects were u sed , raisin g th e p ossibility th at
h eritability effects m ay d ep en d to som e exten t on th e way on wh ich
aggressiven ess is d efin ed an d m easu red . Nor is th e twin stu d y m eth od
th e on ly on e available for assessm en t of h eritability. An oth er m eth od
th at h as been u sed in volves assessm en t of sim ilarities in aggressive beh a-
viou r between ad op ted ch ild ren an d th eir biological p aren ts com p ared
to sim ilarities between su ch ch ild ren an d th eir ad op tive p aren ts. Med n ick
et al. (1984) com p ared th e crim in al beh aviou rs of ad op ted p erson s
with sim ilar beh aviou rs in th eir biological p aren ts an d fou n d a h igh level
of con cord an ce between th e two for n on -violen t crim es bu t n ot for
violen t on es.
10 HUM AN AGGRESSION

Th e p ictu re th at we get from research on h eritability of aggression is

th erefore n ot clear. In a review of 24 stu d ies coverin g a wid e ran ge of
m eth od s, Miles an d Carey (1997) fou n d th at evid en ce for h eritability of
aggression d ep en d s on several variables, su ch as th e age of th e sam p le
an d wh eth er aggression is q u an tified in term s of p aren t- an d self-rep ort or
observation of beh aviou r. Ou tcom es also seem to d ep en d on h ow aggres-
sion is d efin ed . W e still d o n ot h ave su fficien t evid en ce from an y typ e of
stu d y to d raw stron g con clu sion s. Rep lication s across sam p les an d re-
search p rogram m es, u sin g sim ilar op eration s an d m eth od s, are bad ly
n eed ed . For th e tim e bein g, h owever, th e beh aviou r gen etic m eth od h as
sh own th at h eritability u n d ou bted ly p lays som e p art in h u m an aggression
(Plom in et al. 1990).

Eth o lo gy an d evo lutio n ary bio lo gy

Exp lan ation s of h u m an aggression based on th e scien ce of beh aviou ral
biology, or eth o lo gy, can be traced back at least to Kon rad Loren z’s 1966
book On Aggression. Em p h asizin g th e p lace of h u m an bein gs with in th e
an im al kin gd om , Loren z exp lain ed aggression as beh aviou r triggered by
specific extern al stim u li followin g a progressive accu m u lation of aggression -
sp ecific en ergy with in th e p erson . Aggression is followed by a cath artic
d ecrease in su ch en ergy an d th e begin n in g of a n ew bu ild -u p . Alth ou gh
d irect evid en ce for su ch p rocesses h ad been obtain ed th rou gh th e observa-
tion of aggressive beh aviou r in n on -h u m an sp ecies, som e eth ologists h ave
argu ed th at sim ilar p rocesses form th e basis for aggression in h u m an s.
Evid en ce ad d u ced in su p p ort of th is claim is u su ally based on th e m eth od
of an alogy, i.e. th e d em on stration th at certain p attern s of beh aviou r in
on e sp ecies are sim ilar to p attern s in th e action s of an oth er becau se both
h ave evolved u n d er sim ilar en viron m en tal d em an d s (Rajecki 1983).
More recen tly, stu d en ts of an im al beh aviou r h ave argu ed th at in feren ces
abou t h u m an aggression can be d rawn from th e stu d y of an im als th at are
closely related to h u m an s in term s of gen etic en d owm en t, sp ecifically
p rim ates. For exam p le, stu d ies of p rim ate organ ization h ave sh own th at
social com p eten ce is closely tied to an an im al’s d evelop m en t of th e cap ac-
ity for u sin g aggression ap p rop riately (Higley et al. 1994). To be socially
com p eten t, th e an im al m u st learn som etim es to in h ibit aggression , su ch
as wh en th e you n g m ale is en cou n tered by a p owerfu l an d d om in an t
ad u lt, bu t at oth er tim es to exp ress aggression rath er th an to with d raw
from ch allen ges. Th e con d ition s u n d er wh ich th e you n g p rim ate is reared
p lay an im p ortan t role in d eterm in in g wh eth er su ch d iscrim in ation s are
learn ed . Mon keys th at are reared eith er in isolation or am on g p eers on ly,
with ou t m atern al con tact, gen erally sh ow d eficits in th is sort of learn in g
an d, as a con sequen ce, often m an ifest socially in appropriate aggression . Th e
u p sh ot of th is beh aviou r is th at th ey ten d to be socially rejected . Sim ilar
rejection of offsp rin g becau se of eith er aggressiven ess or victim ization by
oth ers h as also been sh own am on g h u m an s (e.g. Dish ion et al. 1994).
An oth er biological ap p roach to h u m an aggression em p h asizes th e evolu -
tion ary h istory of th e h u m an sp ecies (Geary 1998). Th e case for th is
ap p roach h as been m ad e p rim arily in a series of stu d ies by Daly an d h is
associates. For exam p le, th ese in vestigators id en tify m ale sexu al jealou sy
INTRODUCTION TO THE STUDY OF AGGRESSION 11

as a m ajor factor in h om icid al violen ce am on g you n g m en . Su ch jealou sy

is a con seq u en ce of th e evolu tion of certain p sych ological ten d en cies th at
en h an ce th e m ale’s con fid en ce in h is p atern ity of offsp rin g born e by th e
wom an or wom en with wh om h e h as sexu al relation s. Th e n eed to h ave
con fid en ce th at alien m ales h ave n ot fath ered th e offsp rin g m otivates th e
m ale n ot on ly to d om in ate an d con trol th e fem ale (as n oted below) bu t
also to com p ete aggressively with oth er m ales for available resou rces th at
fu rn ish rep rod u ctive op p ortu n ities. In m od ern society su ch resou rces are
m an ifested n ot in n estin g sites or feed in g territories, as th ey are in lower
an im als, bu t in in tan gibles su ch as statu s an d social p ower. Th e d isp lays
of aggressive m achism o sh own , for exam p le, in you th gan gs rep resen t
m an ifestation s of statu s- an d p ower-d riven beh aviou r. Daly an d W ilson
(1988a) ap p ly th is ration ale to exp lain in g th e com m on ly observed fin d in g
of h igh rates of h om icid e am on g u rban m en in th eir twen ties. Given an
evolved p rop en sity for com bative com p etition , wh ich sh ou ld be at m ax-
im u m stren gth in th e years of greatest rep rod u ctive p ower, even m in or
p rovocation s m ay elicit violen ce from su ch m en . Th e sam e th eoretical
reason in g is also u sed to exp lain wh y m en kill wom en m ore freq u en tly
th an wom en kill m en . ‘Male sexu al p rop rietarin ess’, Daly an d W ilson
(1988b: 521) con clu d e, ‘is th e d om in an t issu e in m arital violen ce. In
stu dies of “m otives” of sp ou sal h om icide, th e leadin g iden tified su bstan tive
issu e is in variably “jealou sy” . . . th e h u sban d ’s p rop rietary con cern with
h is wife’s fid elity or h er in ten tion to q u it th e m arriage led h im to in itiate
th e violen ce in an overwh elm in g m ajority of cases.’
Argu m en ts from evolu tion ary biology h ave gen erally been ign ored or
d ism issed by social p sych ologists. On e reason is th at social p sych ologists
trad ition ally h ave con sid ered socio-cu ltu ral variables to be m ore p owerfu l
d eterm in an ts of aggression in h u m an s th an p sych o-evolu tion ary on es.
In resp on se to th is objection , som e evolu tion ary biologists in sist th at th e
d evelop m en t of h u m an cu ltu re is itself sh ap ed by evolved con strain ts.
Cu ltu re, accord in g to th ese observers, is th e p rod u ct of ‘th e p sych ology
of th e cu ltu re-bearin g organ ism ’ (Daly an d W ilson 1989: 108), an d th is
p sych ology is sh ap ed by evolu tion ary h istory. An oth er sou rce of social
p sych ologists’ resistan ce to evolu tion ary p rop osals is th e im p ression
th at accep tan ce of aggression an d violen ce as ‘n atu ral’ beh aviou rs n ecessi-
tates accep tan ce or en d orsem en t of su ch action s, an allegation sp ecifically
d en ied by evolu tion ists (see Th orn h ill an d Th orn h ill 1992 for a d etailed
d iscu ssion of th is issu e). Som e social p sych ologists are also scep tical of th e
evid en ce th at h as been ad d u ced for th e evolu tion ary argu m en t to d ate
(e.g. Baron an d Rich ard son 1994: 19 –20). It is too soon to tell wh eth er
p sych o-evolu tion ists an d social p sych ologists will even tu ally fin d m u ch
com m on grou n d .

Ho rm o n es an d aggressio n
Th e im p ortan ce of an d rogen ic h o rm o n es – testo stero n e bein g p erh ap s
th e best kn own – in aggression am on g n on -h u m an an im als h as lon g been
recogn ized . Th ese h orm on es h ave two effects on an im als th at m ay in flu -
en ce aggression . First, th ey can in flu en ce th e bod ily d evelop m en t an d th e
stru ctu re an d fu n ction of th e n ervou s system in th e fetu s p rior to birth .
12 HUM AN AGGRESSION

Su ch in flu en ces are called organizing effects. Secon d, th ey can cau se ch an ges
in th e m ood s an d beh aviou rs of an im als after birth , ou tcom es th at are
called activating effects. Desp ite evid en ce th at testosteron e activity h as
been sh own to facilitate aggression between m ales of several vertebrate
sp ecies an d figu res, esp ecially at th e tim e of rep rod u ctive activity (Arch er
1988), wh eth er sim ilar effects occu r in h u m an m ales is still bein g d ebated .
Som e in vestigators d en y th at testosteron e h as an y im p ortan t fu n ction in
h u m an aggression , in sistin g th at th e d om in an ce of social an d cogn itive
factors in h u m an beh aviou r ou tweigh an y in flu en ce th at h orm on es m ay
h ave (e.g. Cam p bell et al. 1997).
Desp ite su ch d isclaim ers, th ere is evid en ce th at th e aggression of
h u m an s sh ows som e h orm on al effects sim ilar to th ose seen in an im als.
On ly a few stu d ies with h u m an s h ave in vestigated th e effects of p ren atal
ad m in istration of an d rogen s on aggression in ch ild h ood an d th e scan t
evid en ce th at is available is m ixed . Given su ch am bigu ou s fin d in gs, Arch er
(1991) con clu d ed th at we h ave little evid en ce p oin tin g to a p ossible
organ izin g role of pren atal an drogen s on aggression . Eviden ce of an activat-
in g fu n ction for testosteron e in aggressive in h u m an s h as been som ewh at
m ore im p ressive ( J. Arch er 1994). For exam p le, research by Dabbs an d h is
associates h as ten d ed to sh ow a reliable p ositive relation sh ip between
testosteron e level an d aggression in varyin g sam p les of m ales, su ch as
p rison in m ates (Dabbs et al. 1987) an d com bat veteran s (Dabbs an d Morris
1990). Horm on al activity in m ales h as also been sh own to be correlated
with ratin gs of aggressiven ess m ad e by observers or by th e aggressors
th em selves. Su sm an et al. (1987) fou n d th at in a sam p le of ad olescen t
boys th e levels of several h orm on es were p ositively correlated with d e-
scrip tion s of th e boys, m ad e by th eir m oth ers, as ‘d elin q u en t’, ‘n asty’ an d
‘rebelliou s’, wh ile Glad u e (1991) fou n d th at m en with relatively h igh
levels of testosteron e activity were m ore likely to d escribe th em selves in
aggressive term s th an m en sh owin g lower levels.
On e q u estion th at is im p licit in all th is research is wh eth er h orm on es
raise levels of gen eral aggressiven ess or in stead create a h igh er level of
read in ess to aggress in resp on se to su fficien t p rovocation . For in stan ce,
d oes a h igh level of testosteron e activity m ake a p erson m ore likely to
p ick figh ts with oth ers or to seek op p ortu n ities for beh avin g violen tly, or
d oes it m erely in crease th e likelih ood th at th e p erson will aggress wh en
attacked ? Several stu d ies h ave sh own th at levels of testosteron e m ay
p red isp ose a p erson to beh ave aggressively in p rovokin g situ ation s. For
exam p le, Olweu s et al. (1980) fou n d a sign ifican t association between
testosteron e an d self-rep orted p h ysical an d verbal aggression in a sam p le
of ad olescen t boys, an d n oted th at th e relation sh ip was m ore ch aracter-
istic for scale item s th at d escribed aggression as a resp on se to p rovocation
th an for th ose item s in wh ich aggression was d escribed in oth er ways.
A series of stu d ies carried ou t in Hollan d by Van Goozen an d h er associ-
ates h as p rovid ed clear evid en ce th at testosteron e activity is related to
reactive aggression bu t n ot to overall aggressive beh aviou r. In th ese stu d -
ies, testosteron e level was exp erim en tally m an ip u lated am on g p erson s
u n d ergoin g sex-ch an ge treatm en ts. Ad m in istration of testosteron e to a
sam ple of fem ale-to-m ale (FM) tran ssexu als was accom pan ied by n o ch an ge
in gen eral self-rep orted aggressive beh aviou r, bu t by sign ifican t in creases
INTRODUCTION TO THE STUDY OF AGGRESSION 13

in self-rep orted p ron en ess to resp on d aggressively to a series of h yp oth et-

ical con flict situ ation s (Van Goozen et al. 1994). Corresp on d in g relation s
between an ger p ron en ess an d testosteron e were fou n d by Van Goozen et
al. (1995a) in a stu d y com p arin g FM tran ssexu als u n d ergoin g an d rogen
th erap y with a sam p le of m ale-to-fem ale (MF) tran ssexu als th at was bein g
given an ti-an d rogen ic agen ts. Th e clearest fin d in g of th e in vestigation
was an in crease in an ger p ron en ess am on g FM tran ssexu als followin g
treatm en ts, an d a corresp on d in g d ecrease in an ger p ron en ess am on g MFs,
with an ger p ron en ess again d efin ed in term s of reaction s to sp ecific h yp o-
th etical situ ation s. Horm on e activity h as been sh own to in flu en ce an ger
even in situ ation s wh ere p rovocation is n ot exp erien ced bu t is m erely
im agin ed . Evid en ce of greater an ger in FM tran ssexu als th an in MF an d
fem ale com p arison grou p s was fou n d in an exp erim en t in wh ich su bjects’
feelin gs were assessed after th ey h ad role-p layed a h yp oth etical fru stratin g
situ ation (Van Goozen et al. 1995b).
Correlation s between h orm on e activity an d aggression are u su ally in ter-
p reted as sh owin g th at h orm on e level is th e an teced en t variable an d
aggression th e ou tcom e. However, som e stu d ies in d icate th at cau se an d
effect m ay work in th e op p osite d irection , an d th at testosteron e level m ay
som etim es be in creased by exp erien ces related to aggression , sp ecifically
th ose in wh ich com p etitive an d assertive beh aviou r occu rs. Research with
p rim ates h as sh own th at th e level of m ale testosteron e ch an ges as an
an im al’s statu s ch an ges, risin g wh en th e m ale ach ieves or defen ds a dom in -
an t p osition in th e grou p an d fallin g wh en h e is d om in ated by an oth er
m ale. In a stu d y d esign ed to exten d th is p h en om en on to h u m an s, Mazu r
an d Lam b (1980) fou n d th at th e win n ers in ten n is m atch es sh owed in -
creased levels of testosteron e d u rin g th e h ou r ju st after victory, wh ereas
th e losers sh owed d ecreased levels, p rovid ed th at th e victory h ad been
clear an d con vin cin g. W in n ers of closer m atch es d id n ot sh ow a sim ilar
in crease in testosteron e. Com p arable fin d in gs rep orted by Salvad or et al.
(1987) sh owed th at you n g m ale ju d o com p etitors wh o h ad recen tly been
m em bers of a win n in g team d isp layed h igh er testosteron e levels th an d id
n on -com p etin g con trols, wh ereas m em bers of a team th at h ad lost sh owed
lower levels th an con trols.
Fu rth erm ore, th e ch ief cau se of in creased h orm on al activity in su ch
ath letic settin gs m ay n ot be th e h igh level of activity th at th ey con tain ,
bu t th e in ter-m ale com p etition th at th ey en gen d er. Glad u e et al. (1989)
fou n d th at even win n ers of a sed en tary an d relatively n on -p h ysical labor-
atory reaction -tim e task sh owed h igh er testosteron e levels followin g com -
p etition th an losers. Glad u e an d h is colleagu es th erefore sh owed th at th e
effect fou n d by Mazu r an d Lam b an d by Salvad or et al. is n ot restricted to
stren u ou s ath letic com p etition an d , m oreover, th at th e m ere p ercep tion
of victory is som etim es su fficien t to in flu en ce testosteron e level.
It sh ou ld be clear from th is brief su rvey th at n o sim p le cau sal relation -
sh ip exists between levels of an d rogen s an d aggression in h u m an m ales.
Testosteron e activity ap p ears to be related to com p etitiven ess, strivin g for
m astery an d ach ievem en t of a d om in an t statu s, an d all th ese activities are
often associated with con flict an d aggression . However, h orm on al activity
also ap p ears to p red ict aggression an d related action s best wh en situ ation al
con d ition s m ake in terp erson al con flict an d p rovocation m ost likely. Th u s,
14 HUM AN AGGRESSION

h orm on al activity wou ld seem to be best regard ed as a d isp osin g or ‘back-

grou n d ’ variable in aggression elicited by aversive situ ation al con d ition s.

Brain m ech an ism s

Th e role of certain brain stru ctu res in aggression am on g lower an im als h as

lon g been kn own to research ers, lead in g to th e su p p osition th at sim ilar
brain m ech an ism s m ay con tribu te to aggression in h u m an s. Two stru c-
tu res h ave been p articu larly im p licated – th e lim bic system an d th e cereb-
ral cortex. Th e lim bic system is a collection of in terrelated n eu ral gan glia
th at h ave d irect lin ks to both h igh er an d lower cen tres of th e n ervou s
system . It is a p h ylogen etically ‘old ’ p art of th e brain , an d it is closely
related to em otion al exp erien ce. In th e case of an ger, for in stan ce, cen tres
in th e lim bic system receive in p u ts from th e sen sory system s regard -
in g p rovocation s, an d th ese in p u ts im m ed iately in itiate p rim itive an ger
reaction s. Su bseq u en t in teraction s between th e lim bic system an d h igh er
cen tres m ay refin e an d d irect th e em otion al state, p ossibly lead in g to
aggression , bu t th e origin of th e em otion is at a lower level. Pan ksep p
(1998: 42) m akes th is p oin t in th e con text of a gen eral m od el of em otion :

Alth ou gh it is self-evid en t th at extern al even ts p rovoke ou r feelin gs,

em otion s actu ally arise from th e activities of an cien t brain p rocesses
th at we h ave in h erited from an cestral sp ecies. Extern al stim u li on ly
trigger p rep ared states of th e n ervou s system . Th e fu n ction of an cien t
em otion al system s is to en ergize an d gu id e organ ism s in th eir in ter-
action s with th e world , bu t th eir p ower arises from th eir in trin sic
n atu re in th e brain .

On e ap p roach to stu d yin g th e in volvem en t of th e lim bic system in

aggression con sid ers th e activity of th e cen tral n eu rotran sm itter seroton in ,
or 5-h yd roxytryp tam in e (5-HT). Dim in ish ed levels of seroton in h ave been
fou n d to be correlates of reactive aggressiven ess in h u m an s (Coccaro 1992).
Th e ap p aren t reason for th is is th at im p overish ed levels of 5-HT activity
in th e n ervou s system lead to both a h eigh ten ed ten d en cy to react actively
to aversive stim u li – su ch as by fleein g or attackin g an oth er organ ism –
an d a decreased ability to con trol activity un der such con dition s. Eich elm an
(1995) h as id en tified th e first of th ese ou tcom es as th e basis for th e fin d in g
th at ph ysical pain can in duce aggression towards a bystan der (e.g. Berkowitz
et al. 1981). Th e latter ou tcom e u n d erlies th e ability to resp on d to a th reat
or p rovocation with p assive avoid an ce, i.e. by avoid in g an y activity th at
cou ld lead to p u n ish m en t. Eich elm an (1995) regard s th ese two ten d en -
cies, both related to seroton in d ep letion , as th e p h ysiological bases for
‘im p u lsive’ aggression . Rats th at h ave u n d ergon e lesion s to th e sep tal
n u clei, a p art of th e lim bic system located in th e forebrain , reveal m arked
in creases in sh ock-in d u ced aggression (Eich elm an 1971), a fin d in g wh ich
su ggests th at th e lim bic system is in volved in seroton in regu lation an d
th at reactive aggression is activated by alteration s in seroton in -m ed iated
im p u lse con trol. An y lesion s in th e sep tal region wou ld th erefore in terfere
with seroton in regu lation an d p red isp ose an an im al (or p erson ) to becom e
INTRODUCTION TO THE STUDY OF AGGRESSION 15

Reduction in 5-HT
Septal lesion Enhanced active avoidance Aggression
Decreased passive avoidance

Provocation

Figure 1.1 Hyp oth etical m od el of effects of sep tal lesion s on affective
aggression .
Source: based on Eich elm an (1995).

m ore aggressively reactive to p rovocation (Figu re 1.1). In ad d ition , treat-

m en ts with d ru gs th at en h an ce or in h ibit seroton in activity h ave been
sh own to p rod u ce effects on aggressiven ess th at are con sisten t with th e
gen eral m od el d escribed h ere. On e su ch treatm en t in volves m an ip u lation
of th e p erson ’s level of p lasm a tryp top h an , a p recu rsor of 5-HT. Restriction
or lim itation of tryp top h an will resu lt in a relatively low level of 5-HT
activity an d sh ou ld th erefore be associated with su bseq u en t h igh levels of
aggressiven ess. In a stu d y sh owin g th is effect, Fin n et al. (1998) fed a
m ixtu re of am in o acid s to fastin g m ale su bjects, with som e of th e su bjects
receivin g a n u trition ally balan ced m ixtu re an d th e oth ers receivin g a m ix-
tu re d evoid of tryp top h an . Am on g su bjects wh o h ad sh own a h igh level
of p re-treatm en t h ostility, th e exclu sion of tryp top h an from th e d iet led
to levels of h ostility h igh er th an th ose sh own by su bjects wh o h ad taken
th e balan ced m ixtu re. No effects of tryp top h an m an ip u lation were fou n d
am on g n orm ally less h ostile su bjects.
Th e activity of th e cerebral cortex also p lays an im p ortan t p art in h u m an
aggression . As is d iscu ssed m ore fu lly in th e n ext ch ap ter, in creasin g
atten tion h as been p aid in recen t years to th e im p ortan ce of cogn itive
p rocesses in th e relation of p rovocation to aggression . Th ese cogn itive
p rocesses are m ed iated by th e activity of th e cortex. Th ey in volve aware-
n ess of th e p rovocation an d th e m ean in g attribu ted to it, th e ju d gem en ts
th at are m ad e con cern in g th e m otives of th e p erson resp on sible, th e recall
of strategies u sed in th e p ast for d ealin g with su ch situ ation s an d exec-
u tive strategies for th e en actm en t of su itable beh aviou rs. Th e en d resu lt
of th is com p licated p rocessin g is th at th e p erson d oes n ot sim p ly react to
p rovocation at th e p rim itive ‘old brain ’ level bu t in stead con trols, m od u -
lates an d elaborates th e resp on se in su ch a way th at it is m ore su itable to
th e social situ ation in wh ich it occu rs.
Now, it sh ou ld follow from th is th at an y d am age or d ysfu n ction occu r-
rin g in th e cerebral cortex cou ld , by d isru p tin g th e n orm al flow of cogn it-
ive activity, in terru p t th ese con trollin g an d m od u latin g effects an d leave
th e ou tcom e of p rovocation to th e lower an d m ore p rim itive cen tres. A
growin g bod y of evid en ce su p p orts th is con clu sion . Two typ es of evid en ce
m ay be cited . Th e first com es from stu d ies of brain im agin g by m ean s of
com p u ter-assisted to m o graph y an d p ositron em ission tom ograp h y (PET),
a tech n iq u e th at p rod u ces im ages of th e brain from wh ich fu n ction s su ch
as cerebral blood flow an d glu cose m etabolism can be in ferred . In on e
16 HUM AN AGGRESSION

stu d y in volvin g th e PET-scan m eth od ology with fou r p sych iatric p atien ts
with h istories of violen t beh aviou r, Volkow an d Tan cred i (1987) fou n d
evid en ce of abn orm al blood flow an d m etabolic p roblem s in th e left
tem p oral lobes of all fou r p atien ts. Sim ilar fin d in gs h ave been rep orted by
Miller et al. (1997) from a larger p atien t sam p le. Twen ty-two p atien ts wh o
sh owed red u ced blood flow to th e fron tal-tem p oral region of th e brain
were com p ared with 22 wh o sh owed a p arallel red u ction to th e tem p oral-
p arietal region on a n u m ber of aggressive an ti-social beh aviou rs. W h ereas
ten of th e 22 sh owin g fron tal-tem p oral an om alies revealed su ch beh avi-
ou rs, on ly on e of th ose in th e tem p oral-p arietal grou p d id so.
Th e oth er bod y of evid en ce con n ectin g aggression to cortical d ysfu n c-
tion com es from in vestigation s in wh ich correlation s h ave been sh own
between variou s in d icators of aggressiven ess an d less d irect m easu res of
d eficits in cogn itive p rocessin g. Th e latter, called ‘soft sign s’ of fron tal-
lobe d isord er, in clu d e su ch sh ortcom in gs as in volu n tary m ovem en ts, d if-
ficu lties in th e p erform an ce of rap id altern atin g m ovem en ts an d difficu lties
in d etectin g sim u ltan eou s stim u lation from two sou rces (Stein et al. 1995).
Th ese sign s can be d etected in a n eu rological assessm en t. An oth er tech -
n iq u e of assessin g soft sign s in volves th e u se of stan d ard tests of ability to
p rocess in form ation , su ch as to d eterm in e an d carry ou t an organ ized
seq u en ce of resp on ses (e.g. Petrid es 1985; Petrid es an d Miln er 1982). On e
su ch test th at h as been u sed exten sively in research is th e Sp atial Con d i-
tion al Associative-Learn in g Task (SCALT), wh ich assesses th e ability to learn
association s am on g n ewly p resen ted stim u li (Petrid es 1985). Perform an ce
on th e SCALT h as been sh own to p red ict ad u lt aggressive beh aviou r in a
laboratory settin g (Gian cola an d Zeich n er 1994) an d to be associated with
figh tin g in you n g boys (Ségu in et al. 1995).
Evid en ce from both im agin g stu d ies an d in vestigation s u sin g soft sign s
of fron tal lobe d ysfu n ction th erefore sh ows th at su ch d ysfu n ction is asso-
ciated with relatively h igh levels of aggressiven ess. Th is con clu sion is
con sisten t with th e h yp oth esis th at a breakd own in in h ibitory cortical
con trol over lower em otion al cen tres is th e in terven in g p rocess. Su p p ort
for th is h yp oth esis h as been rep orted by Lau an d Pih l (1996), wh o fou n d
th at su bjects in th e lower q u artile of a large sam p le given th e SCALT –
th ereby m an ifestin g relatively h igh fron tal lobe im p airm en t – sh owed
h igh er levels of aggression wh en p rovoked th an th ose in th e u p p er q u ar-
tile even wh en th ey were given m on ey for beh avin g n on -aggressively.
Th ese aggressive su bjects ap p aren tly were relatively u n able to p rocess cu es
for th e in h ibition of aggression .

Social learning

Th e an teced en ts of aggression th at we h ave con sid ered so far in th is

section h ave all been grou n d ed in th e h u m an bein g’s biological system .
At on e tim e, d u rin g th e ascen d an cy of beh aviou rism , su ch variables wou ld
n ot h ave been assign ed m u ch of a p lace in aggression ; su ch beh aviou r
wou ld h ave been exp lain ed alm ost en tirely as som eth in g acq u ired th rou gh
con d ition in g an d learn in g. Tod ay, of cou rse, we recogn ize th at th e roles
of learn ed an d in n ate factors in h u m an aggression can n ot be d escribed in
INTRODUCTION TO THE STUDY OF AGGRESSION 17

an eith er/ or way. Virtu ally every psych ologist wh o in vestigates th e problem
recogn izes th at both are in volved an d th at differen ces in viewp oin t in volve
th e relative em p h asis p laced on each . To set ‘n atu re’ again st ‘n u rtu re’ in
d iscu ssin g h u m an aggression is to create a false d ich otom y. Elsewh ere in
th is book it is su ggested th at both learn in g an d h ered ity are best u n d er-
stood as backgrou n d variables th at create a level of p oten tial for aggres-
sion with ou t bein g d irect an teced en ts. Aggressive beh aviou r is a resp on se
to con d ition s in th e situ ation th at p rovoke th e p erson ; even wh en on e is
d isp osed to aggress an d cap able of beh avin g aggressively, a sp ecific situ -
ation m u st elicit th e act. Th e p robability th at su ch beh aviou r will occu r,
an d also th e in ten sity of th e beh aviou r, will vary accord in g to both th e
n atu re of th e p rovocation an d th e level of p oten tial for aggression set by
th e several backgrou n d variables. Certain ly p eop le born with d isp osition s
to be violen t will be m ore aggressive wh en attacked th an th ose lackin g
su ch d isp osition s, an d p eop le wh o h ave acq u ired stron g aggressive ten d -
en cies th rou gh so cial learn in g will react m ore aggressively th an th ose
wh o h ave n ot. Hered ity an d social learn in g are com p lem en tary factors in
h u m an aggression .
Th e social learn in g th eory of aggression em erged in th e 1960s, largely
as a resu lt of th e th eorizin g of Albert Ban d u ra an d h is associates. Th e
ap p roach h as u n d ergon e several elaboration s sin ce its was first p resen ted
an d it con tin u es to exert a stron g in flu en ce. It em p h asizes th e acq u isition
an d m ain ten an ce of aggressive resp on se ten d en cies. Alth ou gh it d oes n ot
ru le ou t p rovocation s as im p ortan t con tribu tors to aggression , th e social
learn in g ap p roach treats su ch even ts as con d ition s u n d er wh ich learn ed
aggressive beh aviou rs m ay be en acted . Likewise, th e th eory in clu d es a
recogn ition of biological factors in aggression with ou t regard in g su ch fac-
tors as d irect cau ses of aggressive beh aviou r. In stead , th e th eory assu m es
th at a p erson ’s gen etic an d biological en d owm en t creates a p oten tial for
aggression , wh ile th e sp ecifics of aggressive beh aviou r – its form s an d
freq u en cy, th e situ ation s th at evoke it an d th e targets toward s wh ich it is
d irected – are acq u ired th rou gh exp erien ce (Ban d u ra 1983).
Social learn in g con sists of th e acq u isition of resp on ses th rou gh observa-
tion an d th e m ain ten an ce of beh aviou rs th rou gh rein forcem en t. Th e n or-
m al ch ild observes n u m erou s in stan ces of aggression both in real life
situ ation s at h om e, in sch ool an d on th e streets, an d in th e fan tasy
world of television an d m otion p ictu res. By observin g th e con seq u en ces
of aggression for th e actors, th e ch ild grad u ally acq u ires a ru d im en tary
kn owled ge of certain ru les of con d u ct (e.g. th at on e m ay som etim es
obtain som eth in g d esirable by u sin g force). In th is way a rep ertoire of
aggressive beh aviou rs is bu ilt. W h eth er th ese beh aviou rs are acted ou t
d ep en d s on th e con tin gen cies th at th e ch ild p erceives for h is or h er beh a-
viou r. If su itable in cen tives for aggression are p resen t, th e p robability of
aggression is likely to be h igh . In ad d ition , th e ch ild exp erien ces reward -
in g an d p u n ish in g con seq u en ces followin g aggression ; aggressive beh a-
viou rs acq u ired th rou gh observation are likely to be carried ou t on ly if th e
ch ild h as been reward ed for su ch action s.
Ban d u ra (1986) h as stated th at th e p rocesses of social learn in g d ep en d
on th e ch ild ’s form in g m en tal rep resen tation s of even ts in th e social
en viron m en t. Reward s an d p u n ish m en ts for aggression are rep resen ted in
18 HUM AN AGGRESSION

th e form of expectancies of fu tu re ou tcom es of aggression an d th e u tility or

value th at aggression h as for th e in d ivid u al. For in stan ce, a ch ild with a
h istory of attain in g valu able en ds th rou gh bu llyin g oth er ch ildren at sch ool
soon com es to believe th at fu rth er bu llyin g in th e sam e settin gs will
con tin u e to d eliver worth wh ile reward s. As a resu lt, aggression acq u ires
a h igh d egree of valu e an d is likely to be rep eated wh en ap p rop riate
situ ation al cu es are p resen t. Fu rth erm ore, in ad d ition to form in g exp ect-
an cies of th e likelih ood th at aggression will be reward ed , called outcom e
exp ectan cies, th e rein forced ch ild also d evelop s a sen se of con fid en ce
in h is or h er ability to execu te th e n ecessary aggressive beh aviou rs, called
self-effi cacy exp ectan cies. Evid en ce for both kin d s of exp ectan cy was re-
p orted by Perry et al. (1986), wh o fou n d th at ch ild ren d escribed as h igh ly
aggressive by th eir p eers exp ressed greater con fid en ce in th eir ability to
carry ou t aggressive solu tion s to in terp erson al con flicts th an less aggres-
sive ch ild ren , as well as greater con fid en ce th at aggression wou ld p rod u ce
tan gible reward an d wou ld be su ccessfu l in forestallin g fu tu re con flicts.
In a stu d y of th e valu e com p on en t of h abitu al aggressiven ess, Bold izar
et al. (1989) fou n d th at p eer-rated aggressiven ess p red icted p ositive ou t-
com es th at ch ild ren associated with th e ou tcom es of aggression . Ch ild ren
rated as h igh in aggressiven ess attach ed greater p ositive valu e th an d id
less aggressive ch ild ren to ‘con trol of th e victim ’ resu ltin g from aggression
again st th e latter, an d less n egative valu e on su ch ou tcom es as th e victim ’s
su fferin g, th reat of retaliation , rejection by p eers an d n egative feelin gs
abou t th em selves. In sh ort, ch ild ren wh o were h igh ly aggressive saw m ore
good ou tcom es arisin g from aggression , an d fewer bad on es, th an less
aggressive ch ild ren .

A NOTE ON M ETHODS OF STUDY

Before we go on to a review of stu d ies of h u m an aggression , a word on

m eth od ology is in ord er. Mu ch of th e research reviewed in th e ch ap ters to
follow com es from exp erim en tal laboratories. Th e laboratory stu d y in
aggression u su ally in volves a sim p le two-p erson in teraction in wh ich on e
p erson is led to believe th at h e or sh e is th e victim of a wilfu l in fliction of
an aversive treatm en t by th e oth er. Even tu ally th e first p erson is allowed
to retaliate in som e way. Aggression is op eration ally d efin ed in term s of
th e overall m agn itu d e of th e retaliation . Most often th e p rovocateu r is an
exp erim en tal accom p lice p layin g a well reh earsed role.
Th e q u estion is often raised as to wh eth er th e exp erim en tal m eth od
is valid for th e stu d y of aggression . Does a h igh ly con trived laboratory
settin g really p rovid e in sigh t in to th e d yn am ics of h u m an violen ce? Or
wou ld we be better off to elim in ate aggression exp erim en ts an d to stu d y
actu al violen ce in real-life settin gs? Su ch q u estion s are n ot easily an -
swered . Th e aggression exp erim en t h as on e ad van tage over oth er m eth od s
in th at it allows a h igh d egree of con trol, p recision an d op eration al d efin i-
tion . It is, th erefore, id eally su ited to testin g cau sal h yp oth eses. For th is
reason , exp erim en ts h ave a h igh d egree of wh at is called ‘in tern al valid -
ity’. On th e oth er h an d , con d ition s an d variables th at m ay assu m e great
 INTRODUCTION TO THE STUDY OF AGGRESSION 19

im p ortan ce in exp erim en tal stu d ies are often less p revalen t in th e world
ou tsid e th e laboratory. Bu sh m an an d An d erson (1998) con d u cted an ex-
ten sive m eta-an alysis of stu d ies of a n u m ber of aggression -related variables
con d u cted both in sid e an d ou tsid e th e laboratory an d con clu d ed from
th eir fin d in gs th at ‘wh en th e con cep tu al variables an d p rocesses are th e
sam e, p arallel resu lts sh ou ld be obtain ed in th e laboratory an d th e real
world ’ (Bu sh m an an d An d erson 1998: 43). Th e q u estion of wh eth er or
n ot con trolled exp erim en ts h ave extern al valid ity, i.e. relevan ce for ‘real
world ’ violen ce, to go alon g with th eir accep ted in tern al valid ity is th ere-
fore a m atter of sou n d th eorizin g an d m eth od ology. Bu sh m an an d
An d erson (1998: 43) su m m arize th e case well:

W h en carefu l con cep tu al an alyses of both typ es of situ ation s are con -
d u cted an d wh en solid em p irical research m eth od s are em p loyed ,
fin d in gs abou t th e relation s between con cep tu al variables will gen er-
alize from th e laboratory to th e real world , an d vice versa.

SUM M ARY

1 Aggression is d efin ed as th e d elivery of a n oxiou s stim u lu s to an oth er

p erson with th e in ten t of h arm in g th at p erson , in th e exp ectation th at
th e aversive stim u lu s will reach its d estin ation , an d with ou t th e con sen t
of th e victim . In h u m an s, aggression in h u m an bein gs takes on e of two
gen eral form s: (a) an gry, or affective; an d (b) in stru m en tal. Nu m erou s
exp lan ation s for affective aggression h ave been form u lated . On e, based
on an evolu tion ary an d biological viewp oin t, is th at h u m an s sh are with
oth er an im als certain gen etically determ in ed ten den cies towards aggress-
ive beh aviou r. An oth er, based on a beh aviou rist p osition , em p h asizes
th e acqu isition of aggression th rou gh experien ce, con dition in g an d learn -
in g. Th e two views are n ot m u tu ally exclu sive. Som e aggressive beh a-
viours in h um an s h ave biological origin s, just as som e are learn ed th rough
observation of oth er p eop le. Fu rth erm ore, aggressive beh aviou r, on ce it
is p art of a p erson ’s rep ertoire, is sh ap ed an d develop ed th rou gh learn in g
p rocesses. Both biological in h eritan ce an d learn ed ten d en cies serve as
p red isp osin g backgrou n d con d ition s for aggression , wh ich is a resp on se
to p rovokin g con d ition s in th e p erson ’s en viron m en t.
2 Con clu sion s regard in g th e role of h ered ity in aggression based on com -
p arison s of id en tical an d n on -id en tical twin s ap p ear to d ep en d on a
n u m ber of sp ecific variables. Th ese in clu d e th e ch aracteristics of th e
sam p les of p eop le tested , th e m eth od s u sed for m easu rin g aggressive-
n ess an d th e way in wh ich aggression is d efin ed . Desp ite th ese q u alifica-
tion s, h owever, a growin g bod y of evid en ce in d icates th at h eritability
m akes som e con tribu tion to in d ivid u al d isp osition s to aggression .
3 Certain sim ilarities between aggression in h u m an s an d th at sh own by
oth er an im al sp ecies, esp ecially p rim ates, h ave been n oted by in vesti-
gators workin g in th e trad ition s of beh aviou ral biology an d th e th eory
of evolu tion . In ter-m ale aggression is attribu ted to sexu al com p etition
d riven by n atu rally selected rep rod u ctive p ressu res an d aggression by
20 HUM AN AGGRESSION

m ales again st fem ales to th e m ale’s p rop rietary con cern with th e latter’s
sexu al fid elity. Th is ap p roach to h u m an aggression h as been criticized
by a n u m ber of social psych ologists wh o rem ain sceptical of th e eviden ce
ad d u ced by p sych oevolu tion ists.
4 Brain m ech an ism s an d th e activity of h orm on es h ave also been im p lic-
ated in h u m an aggression as backgrou n d factors. Elevated levels of
testosteron e h ave been fou n d to covary with aggressiven ess in m ales
by creatin g a h eigh ten ed d isp osition to aggress in resp on se to su itable
p rovocation . Som e stu d ies h ave also sh own th at situ ation s in volvin g
com p etitive an d assertive beh aviou r can lead to elevated testosteron e
levels, su ggestin g th at th e relation between h orm on e activity an d ag-
gression m ay be in p art a recip rocal on e. Both th e lim bic system an d
th e cerebral cortex are lin ked to aggression , th e form er as a p rim itive
cen tre of em otion al reactivity to p rovocation an d th e latter as a h igh er
cen tre exercisin g cogn itive con trols over em otion al resp on d in g. In p ar-
ticu lar, d ysfu n ction in th e fron tal-lobe region of th e cortex is correlated
with aggressive beh aviour an d m ood. Th e activity of th e n eurotran sm itter
seroton in is in volved in aggression , in th at relatively low levels of
seroton in activity – su ch as m ay be in d u ced th rou gh d am age to th e
lim bic system or in h ibitin g d ru gs – are associated with h igh levels of
aggression .
5 Th e social learn in g th eory of aggression exp lain s th e acq u isition , p er-
form an ce an d m ain ten an ce of aggressive beh aviou r th rou gh p rin cip les
of observation al an d in stru m en tal learn in g. Novel aggressive resp on ses
are acq u ired th rou gh observation an d con verted to d isp osition al h abits
by social rein forcem en t. Con sisten t with th e exp ectan cy-valu e th eory
of rein forcem en t, d isp osition al aggressiven ess is associated with h igh
exp ectan cy of bein g reward ed for aggressin g an d with a h igh su bjective
valu e attach ed to su ch reward s.

SUGGESTIONS FOR FURTHER READING

Geen , R. G. an d Don n erstein , E. (eds) (1998) Hum an Aggression: Theories, Research,

and Im plications for Social Policy. San Diego: Academ ic Press. Th is book is a gen eral
in trod u ction to several top ics of cu rren t in terest in aggression research written
by au th orities in each area.
Hollan d er, E. an d Stein , D. J. (ed s) (1995) Im pulsivity and Aggression. New York:
W iley. Several of th e ch ap ters in th is volu m e ad d ress th e role of p h ysiological
m ech an ism s in h u m an aggression .
Stoff, D. M. an d Cairn s, R. B. (ed s) (1996) Aggression and V iolence: Genetic,
Neurobiological, and Biosocial Perspectives. Mah wah , NJ: Erlbau m .