THE POLITICAL USES

OF FORCE

The Four Functions of Force

ROBERT J. ART

In view of what is likely to be helcm' liS, it is \'ita! to think carefillh Hnd

about the uses and limits of military power. That is the puqx)se this ess<t\:. It is i
intenclecl as a backdrop for policy clebates, not a prescription of sp('('i g ., "nhel(,S. It
conscious I:' ('schews elaborate detail on rC(luisite lllilitary forc('s
a ., nand focl1ses instead Oll what military pO\\'t'r has and has not dOnl\ ean
cannot do, E"en' model of how the world has nolicv implications. But not
What thell, are the

WHAT ARE THE USES OF FORCE?

The goals that states pursue range widel:' and \'a]\ consicierably from case to case,
Military power is more useful f(ll' realizing some than others, tjJn1H1h it is
allv considercd of some use b\· most states for
.I •
of the C!<10als that
accurate, to elllllllerate
end
accuracy is not a \11tue per Sf' for In
cost of anahtica) utility. (A
. rather th~11 compile an ~xhat~sthe list of
I hm'e selected four categories that thelllst'lves anahticall~' <,xhaust th:,
force CaIl sen'e: defense, cOlllpellellc~, and "swagg<'Iing.

From "To \\hlt Ends Military PmH'r" I,,· Rohl'lt J. Art. inllltcmaliOlwl SCi'll 1'11'1- Yo/. ..j i Spling H)i)(ll.
pp...··:35. Portions of the te\t and tlw /()"tnot..., lim,,' beell omitted.
131

l
 '"'' all rour tUllctiow; are necessarih- well or

postlln'. In usually olll\' the ?;reat 'powers havC'

directed at the <lch-ersary:, population andior industrial infrastructure. The effectivC'­
ness of the threat UrOlI a state's abili~' to comiu('{' a potential adwrsary that
that can SC'IYe HJOrt' than two functions at onee. E\'en
\1s-il-\'is slIlaller powers, not \is-,I-vis the other
om'S. it has both the \\ill and power to punish him sen'rely if he lmd(AJiakes the umksir­
Sure of the C'Hpabilities 0/';1 statc's military forees must he
Deterrence therefore employs /()l'C'C DC'llcefullv, It is tl!(··
another state, not with re/(']'('tlt'e to some absolute scale, A state lImc can that is the CSS('Il{,(,
another state Can also dell'uel against it and uSlIally dder it. A state that can definition has failed. A (leterrent threat is
against another state canllot thereb~ automatically deter or compel it. A stale (',m made \\ith the intl'nl that it will not han' to he carried out.
deter another state witholtt !rm'illg the abili~ to eithcr c1dptlcl against Or compel it. made 10 prevent aetions frolll being undeltaken. If thc Ihreat has to
A state that can swagger \ is-i'His another lllay or may Dot be able to peri()J'Jll any of menled, the action has alreadv been llllCleltaken. Hence deterrence call be
the other three lllllctions rclathe to it. "'here feasible, ddense is the goal that all successful onl\' if the retaliatol'\' threats ha\'e not beell
states aim [()r first If de/c'use is not possihle, cIeterrt'IlCt' is generally the next prior- D('lerrel~('e and defellse ,~re alike in that hoth are in\<'lldecl to the state
or its closest allies from physical attacks, The pllrpose of both is dissn<lsioll­
is th(" function most difficult to pin down nnnJ:tically; deterrence,
one whose achic'\'("lllellt is the most difficult to dClllonstrate; compellenee, the persuading others /lot to undertake actions harmful to on('self. The defensin' ns('
easiest to delllnmh''''f> ]",...,"­ .1 hardest 10 achieve. The of force dissuades by COIt\ illcing nn ackersm;' that he cannot ('ong uer one's militar~'
forces. The deterrent use of force dissuades by comillcing the that his
popnlation and territory will suffer tenihle damage if he initiates the II ndesirable
action. Defense dissuades b: prt'senting an ull\'allqllisltable lllilitaIY forcC', Deter­
a state will the cettainl\'
,
of retalialol'\'
,
de\'astntion.
or actual attach,I', but not deterrence, and ddern'lll'l' is possible without
a stalc' can ckpJo.\· its f(w('es in A state can have the mil itm;' wherewithal to repel an iJlvasion without also
attack has o('C'urn'd to repel it, or shike to devastation to the im'ader's population or territmy Similarly,
imlllinent or il1C'\itable. The deff'llsi\-e lise of iorcC' call thus a stale can ha\e the wherewithal crediblY to threaten an adversan with such denls­
and phvsical emplmlllent and both repellent (second) strikes and
strikes. If a state sl rik0s first whell it beliC'\'es all attack upon it is imminent, it is
Iautlching a preempti\'C' blow. If it strikes first W!J011 it belie\'cs all attack is
inevitable but not mOlllelltary, it is launching a prC'\errlive blow. Preemptiv(' and phYSical attacks can be
pren'nth'e blows an' undertaken whell a state calculates, that otllcTs plan to the illcenth'e to de\elop a rdaliator:'
attack it and, second, that to delay ill striking ont-llsi\el~ is against its interests, A itself. however, will try to (Ie\elop an
state nrepmntc in ordtT to wrest the ad\'<lntage or the strike fi'om an Opponent. state will leave its population and territor: open to attack
redress the situation. \\1wtll('r it giwJl state can defelld or deter or do both \'is-a-\'is
attack hee';]"''' it that others will attack it
another depends upon 1\\'0 factors: (1) the qUHntitatin' balance between it
thpl'"f~"" " u .. - 1 bal­
and its achersar\': and (2) the qualitath'(' halance of forces, that is, whether the
extant military teelmology b\'ors the offense or the defense. I\\'() Llctors arE'
attack \"ill OCC1Ir. For nrppmnli~n therefore re(lllire car<:>ful anal; sis of the case at haml.
at the most; lise of force is the deplo:'ment of militm;' power so as to be
case of prl:'crnption, the state all ac!n·rsaI;- from doing something that he has aln'acly under­
attack; ill the case of pren'ntioll, hilll to do sornethin~ that be h,l~ Hot vet undertaken, COl1lpt,llenc(',
the timing of its altack. For both cases, it is the belief in . that can cease, or h(,(,OJl te
the OHC'IlSiw, uefenshe attack For hath cases, the lllax.im, "thl:' b("st force either
is a good off<'llse," makes good sense.
The deterrcllt lise oftrJrCe is the cleplo\'rnent of militm; power so as to be "bll:' to
prevent an ach-ersHI; ii"olll doing something tll<"lt olle cloes llot want hilll to do and
that he migbt otllC'l'\\ise be tempteci to do hy threatening him witll unacceptable
iflw dot's it. Deterrence is Ihm the tlrreat of retaliation, Its plIJp)se is
to prevent sOlllf'tllinCT n,,,L.; .... LL ii"oll] hapPc'l1ing. The threat of pUllishment is in the two d<'cades
In the first case, the U nitecl
 Stutes startet! bOlllbill~ North \'ietnalll in order to compel
hm; a bard time 1Inderstanding what it is that OIle wished him tu do. his C'ompliance'
Vietcong {()ref's in Sonth Vietnam, [n tlw latter case, German\" built a battlefleet
that ill ,~n engagement threatelled lo (Tipple England's in ()nl~'r to compel her to
with one's wishes is made more difficult, There is. ho\\,ever, no inherent reason why a
compellent a('tion mnst be vaguer than a deterrent thrpat with n:gnnl to how ch',;rl'
make a geneml political settk1Jlf'Ht adv<lHtagpous to Genmuw. In hoth Cases. one
the adversarv nmlerstamb what is wanted frolll him, "Do not <tttack nH,'" is not all\'
state initiated SOIlH' t:lw of action against another preCisely so as to he abk' to stop
it, to bargain it away for tlte appropIiate respomE' from til(' "put upon" stale. clearer in its'ultimate llH'alling tban "stop attacking IllV ftiPlltl." A state can Iw as (,Ol~­
fused or as clear about what it wishes to prevent as it can be ahout wltat it wishes to
The distillCtiOIl betweell compellellce and deterrence is onp between the
acti\'e aBd passive use of force, The Sllceess of a deterrent threat is llleasured bv its stop. The dmity. or lack of it. of the objl'cti\'es of
not hm'ing to be llsed, The SUccess of a cOll1pellellt actioll is measured by does not vary according to whpther the given action is compellellt or
rent in naturt\ bllt rather ac('onling to a welter of pmiicuhllities associated \\1th the
dosd\' and qUickly the <lch-ersarv confc)J']m to one's stipulated \\isl1(='s, In thl' ease of
given action. SOlHP ohjecth'es. tor example, are inhPrE'l1tly (·karer ilnd bellce easier
dptern:'llc(:\ Olle is trYing to demonstrate a lIegatiH'. to show why some­
to perceive than others. Some statesllIen cOlllnmnieate more dearly than others,
did not hanDen. It can never be clear whether one's actiolls were crucial to, or
Some states have more power to tuing to bear for a gh'en ohjecti\'e than others. It is
another state chose /lot to do sOlllE'thillg. I n the ('ase
clear sC'4uel1ce of actions and reactions lends a the sppcifics a givcll situation, not allY intIillSic difference b(twC'en
that determines the darity with whieh ,Ill objective is
olle's actiolls. Figure I iIIustratcs the
cesstill compellpllce, stute B can claim that its pressllre elsewhere tor the reason as to whv compellence is
to achien' than detcrrence. It lips, 1I0t in what olle asks
eOllrse of action. In successful deterrence, state B has 110 clt<lnge in statC' As
but in 'IOte one asks, \Vitl! (letprrenc(', statp Basks somethinQ of
ior to pOint to, but instead must rpsort to claiming that its threats werl'
for thE' COlltilllLity ill As beha\ior, Statp A 111m hm'e c1wngpd its beha\'Jor tor rt'Cl­ stat(=' A ill this fashion: "Do not take action X; for if \on
SOilS other than stale B's compel/ent action. State A may h,lW' continupd with its
head with this (,luI)." With
fashion: "1 am no\\' going to basI; \'0\1 0\'(:'1' the head with this club anel will cOlltinue
t()r reasons other thall stalp B's delplTent threat. "PrO\ing" the
to do so nntil you do what I want." In the ronner cas(', state A (',m easilv clem' with
of B's influellce on A for either case is not easy. btlt it is lllore
to claim that B illfluencl'd A wllel} there is a change ill As b(·havior than whcll great plamibilitv allY intelltiOlt of hming planlled to take action X. I~l the' latter
there is not. Explaining w]l\' something did not happell is mOfP ditIicnlt thall case, state A cannot clem' either that it is engaged ill a givcn COllrse of action or that
, ~)J}1C'thing diLL it is being subjeded to pressure by state B, If they arpto he
COillpellence may he p<lsier to ckmollstrate than dcterrence, but it is
actions require a state to alter its behmior in a manner
achieve. Schellillg m'h'11eS that c01llpellcllt actions tcndlo he vagW'f in
respollsP to an e(}llalh- visible j(Jl'('t{ul initiative taken hy
threats and /()r that reason more difIicl!lt to attain, If an
to eompellent adiom. deterrent threats are hoth easier to appt'ar to haw'
or easier to acquiesce to without great loss of face. III contrast to deterrent
compellent actions more din'cth' engage tht, nrestirrc nml the nassions of'
A state, Lpss prestige is lost in not C' " ,

/ t
A A
llue to J)reSSllrt' from another. III the case ofCOlnpelll'llce, a state has
its prestige allll resources to a giwn Iin(' of cOllduct that it is HOW
asked to gi\'e np. This is Jlot so for deterrence. Thus. eompellcllce is intrinskall~'
B
hanler to attaiu than deterrence, not hec<tuse its objecti\'cs are vaguer. but because
it
A B

COMPElLENCE DETERRENCE at
(1) A is doing something that B cannot tolerate (1) A is presently not doing anything that B finds it to do something specifie. The ohjecthes for s\\'aggt'ling are more dif­
(2) B initiates action against A in order to get him intolerable
to stop his intolerable actions amI prohl('matic titan that. SwaggeriHg almost <11\\,<l\:S im'oh-es
(2) B tells A that if A changes his behavior and
(3) A stops his intolerable actions and B stops his use of hHce and is pxpresscd usuallv ill olle of two \\'<1\'S: displaying OllP'S
does something intolerable, B will punish him
(or both cease simultaneously)
(3) A continues not to do anything B finds intoler­ at militarv C'xercises anc1natiollnl demonstmtions and lnl\ing or huild­
able i.ng the: em's most prestigious weapolls. The swagger use of force is the ;no~t
FIGURE 1 • It aims to enhance the natiollal pride of a people or to satisf\' tlw n~,'''nn''
of its ruler, ,f.. state or statesman swaggers in order to look and
 to he takpil sprioush" lw othl'rs in the coulldis of intenwtional deci­ One holds that its molin>s \\'ere aggressin..-territorial aggrandizement, eC'oll()lllic

to enhance the nath;n \' illlng(> ill the evps of others. If its image is gain, and eJenttion to tilt' statns of a world empire. Allotlwr bolds that her lllotin's

the nation's defense. dHern>1lt. and (-ollliwlIent capabilities may' also were preveJltin:' and lWllCt' ddellsh'e, She struck lirst Iwe,llIse she fean'c! ('lIeir­

hut S\\'a,TgerilH' is Ilot llIul,'rh1.-"n "n1"I,· .... " .. _.. ····'Illarih· f(ll' dement, slo\\' strangulation, and tlwll inedtabh' attack 1)\ her h\'o

prestige "on tlw hors, foes whorn sill' It'lt were daily increasing their militan'

\'t'aruillg of states and was. She struck wbile she had the chullC'(' to \\ill.

to be enjOYed It is not simple to clpcid(' whieh school is thl' more Ilearh correct beclllise hoth
end. .. evidence to build a powerfid case. ASS1I11IC for the 1l101llt'IiL
he totallv is closer to the truth. There are tl1('11 two possihilities to consider:
(1) Germany launched an llitack bt'c<lllS(, it lcaS the l'ase that her foes werc

to attack lwr ultimately, and Germany had the e\'leJpnce to pnAe it; or (2) CerHUlll\'

she had n>aslmable evidellct' nflwr foes' intclil to attack lwr e\'entually, but
fact her evidencp was wrong lwcclUse she lllisperceiH'd tlIeir intent from thpir
on Il''''!lUlllll\ lIlies. If swaggenng causes otller states to actions. If the first was the case, thell \W Il1 list ask this qucstion: How
one's interests more sE'riollsh- illtn <[CCOUII!. tbell the general interests of the state was Germany's diplomacy in the fifteen years before H)lA, aggressin' and
will henefit. Ewn ill its instrumental role. hO\\'('\'('r, swaggering is undertaken ing as it was, ill breeding hostility in her neighbors') Cerm<lll\ attacked ill tIlt'
fix <m~' given end than fell' alll'llcis. The s\\'aggerillg fimetion of militarY knowledge that the: would E'YelltmIllv han' struck her. hilt if her
at OIlE' ;md tilt' salile tillle the most compn,hellshe and the most
matic record was a significant h1<:tor ill causing tltem to 1m these plans, must we
most H'rsatile in its effects awl the least focused in its
conclude> that Germ<lllV in ID14 was merel: acting del(>nsiwI::? MllSt we confine
our judgment about the dcfensiw' or aggressi\'e IUltnre of the ad to the llIolith or
even the year ill \vhieh it ncclIITcd? JI' not. how IllallY \ ears back ill histor: do wp go
in order to make a judgment? If tlw second was tlte cast\ then \\e Jllllst ask this
in international question: If Germany attacked ill the helief, mist<lkelily as it turns out. that she
would be attacked. must \\'(' conclnde that Germam \\'as ,Ictillg ddc'llsh'ely') l\,IlIst
and s\yag,\!;pring-thest' are the four gcn­ ahout the defc>nsiw or aggressin' na/nre of the act ,.
among them ana­ about others' illtent, without relc~renl't' to their actual intene
'. is E'<lsier than apphing them in prac!ice. This is due to two It is not (:,<ls:' to anS\\'er these questiolls. Fortunatek we do not have to.
factors. First. we neEd to know the motine'S hc'hinel an act in order to judge its pnr­ them is enough hecause it illustrates that all assessnlent of tIl(:' /el'iti/lli/Cfl of n
the problt>m is that Illoth'es cannot be readik inferred from actions state's motivE'S in nsillg fc)l'('E' is illtpgralto the task of
stseral moti\t's can be senpd by arc. One GUillot, that is, specify Ill()ti\'(~s withont at the S;lll](, timp
infer the lIloti\·ps of a state ii'om what it PUDI1CIV or ments about tbeir legitimacy. Tilt' root cause oj' this need lie's in the natlIn'
Such statE'lllents shoulclllot Ilecessarih- be taken ilt action. Tn anarchy P\t'r~ shlte is a \',Ilid judgp of the
role that bluff" and dissimulation there is no supranational authority to enforce
cOl1('octpd with domestic lack of' uniyersal standards, we are
context and to makp indhi(
When individual judgment is ~wr('ised, imlhicl1wls Illil\ well differ. Ddinitin"
answers are more like!: to be tht' ('\ceptioll rather than the rule.
\Vhere does all of this Ipa\p us';:> Our rom categories tell us what are the 1'0111'
possible purposes for which states can power. The attributps of
as pp<lcellll cOlllpdlent olles. Del(~nsiH> attacks are nearh- indistinguishable frolll liS to the 1\J)f'S of evidence for But because the context
c:olllpellent ones. Is he who attacks first the ddE'ueler or thp cOlllpellpr') Detf'rrence these four
swaggeling both inmkc the acquisition and displav of all era\
weapons. Are sllch weapons aCljllirt'd to f'nlWllce prestige or to
SeconrL to make lllatkrs \H)rSe, consider are, \\·ltich one it is, is an exercise in judgment. all exercise
an attack upon France and Russia at tIlt' end of as much upon the purticulars of the giH'n case as it does UpOll Ihe
\\'orld \\'ar 1. There are two schools of features (,f the
 TABLE I • THE PURPOSES OF FORCE
Type Purpose Mode Targets Characteristics

Defensive Fend off attacks Peaceful and Primarily Defensive

and/or reduce physical military can have dissuasion
of an Secondarily value;
attack industrial Defensive preparations
can look
First strikes can be taken
for defense.
Deterrent Prevent adversary Peaceful Primarily Threats of retaliation
,'I
from initiating an civilian made as not to have
action Tertiarily to be carried out;
military Second strike preparations
can be viewed as first
strike preparations
Get adversary Peaceful and All three with to but
to stop dOing physical no clear hard to achieve;
ranking actions can
be Justified on
defensive grounds.
Peaceful None Difficult to describe
because of instrumental
and irrational nature;
can be

: II