Social-Change Nexus: Some Old Theories and New Hypotheses
Author(s): Ted Robert Gurr Source: Comparative Politics, Vol. 5, No. 3, Special Issue on Revolution and Social Change (Apr., 1973), pp. 359-392 Published by: Ph.D. Program in Political Science of the City University of New York Stable URL: http://www.jstor.org/stable/421270 Accessed: 03/02/2010 13:47 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=phd. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Ph.D. Program in Political Science of the City University of New York is collaborating with JSTOR to digitize, preserve and extend access to Comparative Politics. http://www.jstor.org T h e Re volution- Social-Ch ange N e x us Some Old T h e orie s and N e w Hypoth e s e s T e d Robe rt Gurr T h e analys is of th e ne x us be twe e nre volutionand s ocial ch ange may be approach e d in th re e ways . One is th e que s tion-be gging de finitionof re volutionas a type of s ocial ch ange -abrupt ch ange in fundame ntal patte rns of be lie f and action. T h is "de finitional" approach is que s tion be gging ins ofar as it fails to de al with s uch as pe cts as th e conditions unde r wh ich s ocial ch ange be come s "re volutionary" rath e r th an "non- re volutionary," and h ow it doe s s o. T h e s e cond and th ird approach e s as s ume s ome caus al conne ctionbe twe e nre volutionand ch ange . One of th e s e s pe cifie s kinds or s e que nce s of ch ange th at le ad to viole nce and re volution; we may call it th e "e tiology" approach . T h e oth e r, "ins tru- me ntal" approach cons ide rs th e cons e que nce s of re volution, and le s s e r forms of viole nce , for s ocial ch ange . T h e s e approach e s , as de s cribe d h e re , are polar type s and not ne ce s - s arily accurate ch aracte rizations of particular th e orie s . Se ve ral ge ne ral ch aracte ris tics of th e type s be ar me ntioning, h owe ve r. T h e de finitional approach is a manife s tationof traditional s ch olars h ip, in wh ich pro- ce dure s of de finition, s pe cification of type s , and th e continge nt cate gori- zationof h is torical cas e s of "re volution" are e nds in th e ms e lve s . Among th e ir more abs urd cons e que nce s are s ch olas ticde bate s about wh e th e r a give n s e t of e ve nts was "re ally" a re volutionor not. T h is approach is none mpirical, for its ge ne ralizations are not inte nde d for ve rification; and it will not be furth e r cons ide re d in th is article . T h e "e tiological" approach h as its roots in s uch s tudie s as Crane Brinton's clas s ic T h e Anatomy of Re volution 1 th at atte mpt to s ort out th e caus e s or proce s s e s , or both , of h is toricre volutions . Conte mporary e tio- logical th e orie s are principally conce rne d with ide ntifyingtype s of ch ange or s e ts of pre conditions th at are various ly s aid to be ne ce s s ary, s ufficie nt, or probabilis tically pre dis pos ing conditions of political viole nce 'N e w York, 1938. 359 Comparative Politics April 1973 ge ne rally, or re volution s pe cifically. T h e s e th e orie s are ordinarily s ubje ct to e mpirical validation; s ome of th e m h ave be e n s pe cifically formulate d with th is obje ct in mind. A late nt purpos e of many e tiological th e orie s s e e ms to be th e control of viole nce and re volution. T h e ir auth ors ofte n -th ough not ne ce s s arily-proce e d from th e as s umption th at re volution is unde s irable , and may be fore s talle d if s cie ntifically unde rs ood. T h e s e as s umptions do not make s uch th e orie s inh e re ntly cons e rvative or re - pre s s ive ; s ome h ave s trong "re formis t" as s umptions and implications . We migh t als o note th at Karl Marx 's th e ory of re volution is e tiological, and s e lf-e vide ntly prore volutionary.2 T h e "ins trume ntal" approach to re volution is all but an e mpty cate - gory. T h e ins trume ntality of various forms of viole nce is ofte n alle ge d by re volutionarie s and re volutionary th e oris ts . Marx fore cas ts a progre s - s ion of re volutionary s truggle s culminating in clas s le s s utopias . T h e re are als o s ome ide ograph ic s tudie s of th e e ffe cts of viole nce and re volu- tion on particular s ocie tie s . But it is s till large ly true , as Harry Ecks te in wrote ne arly a de cade ago, th at "almos t noth ing care ful and s ys te matic h as be e n writte n about th e long-run s ocial e ffe cts of inte rnal wars ," not e ve n on s uch bas ic que s tions as "h ow political le gitimacy and s ocial h ar- mony may be re s tore d afte r viole nt dis ruption, wh at make s inte rnal wars acute or ch ronic, and wh at th e comparative cos ts (and probabilitie s ) are of re volutionary and e volutionary trans formations ." 3 T h e firs t obje ct of th is article is to e x amine wh at s ome conte mporary e tiological th e orie s s ay about s ocial ch ange as a caus e of viole nce ge n- e rally, and re volution s pe cifically. Its s e cond obje ct is to cons ide r th e limite d implications of th e s e e tiological th e orie s for ans we ringth e ins tru- me ntal que s tion. Finally, I propos e s ome ge ne ral de te rminants of th e "s ucce s s " of various kinds of viole nt conflict, as a furth e r s te p toward de alingwith th e kinds of que s tions Ecks te in h as rais e d. I. Some Conce pts A fe w pre liminary, de finitional points mus t be made . By viole nce I me an de libe rate us e s of force to injure or de s troy ph ys ically, not s ome more ge ne ral cate gory of coe rcive actions or policie s , and not ins titutional arrange me nts th at de me an or frus trate th e ir me mbe rs . T h is de finition is inde pe nde nt of age nts , obje cts , or conte x ts of viole nce . T h e caus e s of th e e x te nt of viole nce pe r s e may be of intrins icinte re s t, and inde e d th e y are 2 On th e une x amine d ide ological pre mis e s and implications of conflict th e orie s , s e e T e rry N ardin, Viole nce and th e State : A Critique of Empirical Political T h e ory (Be ve rly Hills , 1971). ' Harry Ecks te in, "On th e Etiology of Inte rnal Wars ," His tory and T h e ory, IV (2/1965), 136. 360 T e d Robe rt Gurr for s ome of th e e tiological th e oris ts cons ide re d be low. A more common conce rn of e tiological th e orie s is wh y particular viole nt e ve nts occur: "viole nce " is us e d th e ore tically as a s h ort-h and te rm for, or de fining prope rty of, e ve nts various ly labe le d riots , re be llions , inte rnal wars , tur- moil, re volutions , and s o on. A re late d e tiological approach is to focus ons ome s e t of fre que ntly-but-not-ne ce s s arily-viole nt inte ractions be twe e n groups , e .g., "conflict be h avior" or "clas s conflict"; de finitions and e x - planations of "viole nce " are s ubs idiary to th e e x planation of conflict ge ne rally. In e valuating th e s e th e orie s , and advancingmy own h ypoth - e s e s , I will us e "viole nce " in th e abs tract s e ns e s pe cifie d above . Wh e n dis cus s ing e ve nts or inte ractions involving th e us e of viole nce , I will us e th e ph ras e "viole nt conflict." As to "re volution," it is de fine d in diffe re nt ways from diffe re nt th e o- re tical pe rs pe ctive s . I h ave me ntione d its de finitionas a s pe cie s of abrupt ch ange . Etiological th e orie s various ly us e th e te rm to re fe r to (a) a motive or obje ctive of a group of pe ople ; (b) a s tyle or form of action; (c) an outcome of action; and (d) ch ange s continge nt upon action. Le t us cons ide r th e s e brie fly. a. Individuals , groups , and organizations are s aid to be "re vo- lutionary" if th e y are (th ough t to be ) committe d to accomplis h - ings we e ping, fundame ntal ch ange s . Socie tie s in wh ich s uch groups are wide s pre ad may be s aid to be in a "re volutionary s ituation," wh e th e r or not ove rt conflict or viole nce h as occurre d. b. Conce rte d action aime d at trans forminga s ocial s ys te m or ove rth rowinga re gime is s ome time s calle d "re volution," with out re fe re nce to its impact or outcome . T h us , re volutionary activis ts may proclaim at th e ons e t of viole nce th at "th e re volution h as be gun." Similarly, s ch olars may cate gorize a continuing viole nt conflict as a "re volution"; in th is way, we s pe ak of th e "Vie t- name s e re volution" wh ile h os tilitie s are s till in progre s s . Once outcome s are known, s uch conflicts may be de note d "s ucce s s ful" or "uns ucce s s ful" (or "atte mpte d") re volutions . All th e s e us age s pre s ume th e e x is te nce of re volutionary motive s (above ), adding to th e m th e occurre nce of ove rt action aime d at th e ir attainme nt. c. T h e imme diate outcome of viole nt conflict is s ome time s th e crite rion for "re volution." If th e "outs " s ucce e d in dis placing th e "ins ," a re volution h as occurre d; oth e rwis e , th e actions of th e would-be re volutionarie s are de s cribe d as a"re be llion," "upris ing," "puts ch ," or s ome s uch te rm. One anomaly of th is us age is th at "re volution" may h appe n with out re volutionary inte nt or action. T h e re are a numbe r of ins tance s -for e x ample , El Salvador in 361 Comparative Politics April 1973 1944 and th e Sudanin 1964-wh e re ge ne ral s trike s and riots ove r imme diate grie vance s induce d rule rs to re s ign, th us giving dis s i- de nts une x pe cte d victorie s th at we re late r h aile d as "re volutions ." d. Se izure of powe r may be dis tinguis h e d from th e s ubs e que nt atte mpt to ach ie ve re volutionary goals ; "re volution" is re garde d as th e s truggle toward or th e attainme nt of th os e goals . Brinton, amongoth e rs , de s cribe s th e e ve nts s ubs e que nt to th e s e izure of powe r as th e "proce s s of re volution." Conte mporary re volution- arie s ge ne rally re cognize th e s e izure of powe r as only th e firs t s te p in "re volution," and re fe r to th e ir s ocioe conomicand political ch ange s in te rms analogous to th e Cuban's la re volucione n march a. Finally-and h e re we come full circle in th e de finitional cycle -s ome advocate s of "re volution" me an by th at te rm th e attainme nt of s ubs tantial s ocial ch ange with out th e us e of s ubs tan- tial viole nce or aconve ntional s e izure of powe r. All of th e s e us age s s e e m to me to h ave s ome validity, or at le as t s ufficie nt curre ncy th at it is foolis h to s ay th at one of th e s e th ings is "re volution" and anoth e r is not. In s ubs e que nt dis cus s ionI avoid th e un- qualifie d te rm "re volution," and us e ins te ad one of th e followingph ras e s , wh ich corre s pond rough ly to th e above us age s : (a) re volutionary move - me nts ; (b) re volutionary conflicts ; (c) political re volution; and (d) re v- olutionary ch ange . In th e final s e ctionof th is article I s h all dis tinguis h re volutionary move me nts from move me nts ch aracte rize d by oth e r kinds of motive s . A fe w words ne e d to be s aid als oabout "s ocial ch ange " as aconce pt. Abs tractly, I re gard it as any colle ctive ch ange in th e me ans or e nds of h umanaction. More s pe cifically, followingT alcott Pars ons and Edward Sh ils , and N e il Sme ls e r,4 we canth ink of h umanactionas be ingde te r- mine d by (a) pe ople 's value s , i.e ., th e ir value d goods and conditions of life ; (b) pe ople 's norms about h ow th os e value s are appropriate ly pur- s ue d; (c) th e patte rne d forms of action-ins titutions -by wh ich pe ople organize or are organize d for action; and (d) pe ople 's s ituations , th e circums tance s -e nvironme nt, re s ource s , te ch nology-th at facilitate or h inde r th e ir purs uit of particular value s . Any ch ange in any of th e s e de - te rminants of actionis "s ocial ch ange ." So de fine d, it is aportmante au conce pt, and I s o inte nd it. For any me aningful analys is one mus t dis - tinguis h a numbe r of as pe cts and dime ns ions of ch ange . Amongth e more obvious one s are : 4T alcott Pars ons and Edward A. Sh ils , T oward a Ge ne ral T h e ory of Action (Cambridge [Mas s .], 1951); N e il J. Sme ls e r, T h e ory of Colle ctive Be h avior (N e w York, 1963). 362 T e d Robe rt Gurr a. T ype of ch ange . Wh ich variable s among wh ich of th e above de te rminants of action ch ange ? b. Ex te nt of ch ange . How many ch ange s in th e above variable s occur and are th e y ch ange s in de gre e or s te p (th re s h old) ch ange s ? c. Scope of ch ange . Wh ich groups in a s ocie ty are affe cte d by wh ich ch ange s , and to wh at e x te nt? d. Patte rn of ch ange . Is ch ange nonrandom, and, if s o, wh at kind of tre nd or cycle doe s it re pre s e nt? e . Rate of ch ange . How s low or rapid is e ach s pe cifiable ch ange ? Atte mpts to formulate compre h e ns ive th e ory about th e caus e -and- e ffe ct re lations h ips be twe e nviole nce and s ocial ch ange would s e e m to re quire an e normous amount of prior conce ptual and th e ore tical mate - rials . T h e s e would include e tiological th e orie s of viole nce , conce ptualiza- tion and th e orie s of conflict proce s s e s , analytic s ch e me s for de s cribing ch ange , and th e orie s s pe cifying inte rre lations among dime ns ions of ch ange . Partial and compe ting th e orie s of th e s e type s e x is t s e parate ly, but th us far th e y afford ano more inte grate d pe rs pe ctive onth e viole nce - ch ange ne x us th andid th e re ports of th e apocryph al blind me nabout th e true nature of th e e le ph ant-and for an ide ntical re as on: all h ad dif- fe re nt points of de parture . II. Social Ch ange as a Caus e of Viole nce and Re volution All th e e tiological th e orie s with wh ich I am familiar attribute viole nt con- flicts and re volutionary move me nts to s ome s pe cifie d kinds of s ocial ch ange . T h e ge ne ral re lations h ip is s o clos e to tautological th at it ne ve r s e e ms que s tione d: viole nt conflicts and re volutionary move me nts occur in time s of ch ange , not s tas is . Some th ing ch ange s , e ve nif it is only an old e lite loos ingits grip on th e ins trume ntalitie s of force . An e x ample is provide d by th e paradigm of re volutionary caus ationh e ld by mos t Ame ricans ch olars in th e firs t h alf of th is ce ntury. Value ch ange s of s ubs tantial e x te nt and s cope occur, at apace toorapid for comme ns urate ch ange inins titutions . Pe ople cannolonge r ach ie ve many of th e ir goals ; th e re s ult is 'wide s pre ad s tre s s , various ly calle d "balke d dis pos itions ," "re pre s s ion," or "cramp." A pe riod of "milling" and "agitation" s e ts in, le adingtoward cons e ns us on grie vance s and th e de ve lopme nt of anti- e lite norms ; and "s olidifie d publicopinion" de ve lops . Gove rnme nts prove too incompe te nt or re calcitrant to re me dy th e s ituation. Wh e ne ve r fur- th e r ins titutional or s ituational ch ange we ake ns th e ability of th e ruling 363 Comparative Politics April 1973 e lite to re s is t, s ome pre cipitant will s park re volutionary conflict.5 In brie f, a s ucce s s ion of s ocial ch ange s is s aid to "caus e " viole nce . Pre vale nt th e orie s te nd to be more s pe cific about th e kinds of ch ange th at dis pos e to viole nt conflict. In fact, it is pos s ible to clas s ify th e m ac- cording to th e ir me tath e ore tical approach to th e ch ange -viole nce ne x us . Some are conce rne d with th e s ocial s ource s of ps ych ological ch ange s th at dis pos e to viole nce ; th e s e are th e s ocial-ps ych ological th e orie s . A s e cond cate gory cons is ts of e x planations th at e mph as ize s ocial-s tructural ch ange , s ome wh at in th e tradition of th e e arlie r th e orie s me ntione d above . T h e th ird type cons ide re d h e re is group conflict th e ory: viole nce is s aid to flow from th e e fforts of s ocial or functional clas s e s to maintain or improve th e ir pos itions re lative to oth e rs . One comme nt be fore e x - amining s ome of th e s e th e orie s : e ach de als with th e viole nce -inducing e ffe cts of s ome particular kinds of s ocial ch ange , wh ich are argue d on various inductive and a priori grounds to caus e viole nt conflict. With th e e x ce ption of Sme ls e r,6 none of th e m be gins with a ge ne ral analys is of s ocial ch ange , from wh ich migh t be de rive d a compre h e ns ive s e t of s tate - me nts about th e caus al s e que nce from ch ange to viole nce . Social-ps ych ological th e orie s Some th e oris ts be gin with th e s e e mingly s e lf-e vide nt pre mis e th at dis conte nt is th e root caus e of viole nt conflict. Principal e x pone nts of th is vie w, in addition to mys e lf, are Jame s C. Davie s and Ivo and Ros alind Fe ie rabe nd.7 My ve rs ion of th e pre mis e is th at th e pote ntial for colle ctive viole nce in a nation or s malle r com- munity varie s with th e inte ns ity and s cope of s ocially induce d dis con- te nt among its me mbe rs . T h e pre mis e is e s s e ntially a ge ne ralization of XT h e ph ras e s in quotations are th os e us e d re s pe ctive ly by Lyford P. Edwards , T h e N atural His tory of Re volutions (Ch icago, 1927); Pitirim A. Sorokin, T h e Sociology of Re volution (Ph ilade lph ia, 1925); Ge orge Pe tte e , T h e Proce s s of Re volution (N e w York, 1938); Re x D. Hoppe r, "T h e Re volutionary Proce s s : A Frame of Re fe re nce for th e Study of Re volutionary Move me nts ," Social Force s , XXVIII (March 1950), 270-79; and Louis Gotts ch alk, "Caus e s of Re volution," Ame rican Journal of Sociology, I (July 1944), 1-8. 0 Sme ls e r, T h e ory of Colle ctive Be h avior. 7 T h e principal th e ore tical s tate me nts by th e s e auth ors are T e d Robe rt Gurr, Wh y Me n Re be l (Prince ton, 1970); Jame s C. Davie s , "T oward a T h e ory of Re vo- lution," Ame ricanSociological Re vie w, XXVII (Fe bruary 1962), 5-19; Jame s C. Davie s , "T h e J-Curve of Ris ingand De clining Satis factions as a Caus e of Some Gre at Re volutions and a Containe d Re be llion," in Hugh Davis Grah am and T e d Robe rt Gurr, e ds . Viole nce in Ame rica: His torical and Comparative Pe rs pe ctive s (N e w York, 1969), ch ap. 19; Ivo K. Fe ie rabe nd and Ros alind L. Fe ie rabe nd, "Aggre s s ive Be h aviors with in Politie s , 1948-1962: A Cros s N ational Study," Journal of Conflict Re s olution, X (Se pte mbe r 1966), 249-71; and Ivo K. Fe ie ra- be nd, Ros alind L. Fe ie rabe nd, and Be tty A. N e s vold, "Social Ch ange and Political Viole nce : Cros s -N ationalComparis ons ," in Grah am and Gurr, e ds ., ch ap. 18. 364 T e d Robe rt Gurr th e frus tration-ange r-aggre s s ionprinciple from th e individual to th e s o- cial le ve l. All th e s e e mpirical th e orie s e laborate on e s s e ntially th e s ame bas ic pre mis e by s pe cifying wh at kinds of s ocial conditions and proce s s e s of ch ange incre as e s ocial dis conte nt to th e th re s h old of viole nt conflict. Davie s attribute s re volutionary conflict to one s pe cific patte rn of ch ange th at h e calls th e "J-curve ": "re volution is mos t like ly to take place wh e n a prolonge d pe riod of ris ing e x pe ctations and ris inggratifi- cations is followe d by a s h ort pe riod of s h arp re ve rs al, during wh ich th e gap be twe e n e x pe ctations and gratifications quickly wide ns and be come s intole rable . T h e frus tration th at de ve lops . . . s e e ks outle ts in viole nt action. Wh e n th e frus tration be come s focus e d on th e gove rnme nt, th e viole nce be come s coh e re nt and dire ctional. If th e frus tration is s uf- ficie ntly wide s pre ad, inte ns e , and focus e d on gove rnme nt, th e viole nce will be come are volution. ..." 8 I pos it two additional patte rns of ch ange th at cre ate a pote ntial for viole nt conflict, th ough not ne ce s s arily re volu- tionary conflict. One is th e s o-calle d "re volutionof ris inge x pe ctations ," wh e re by me n be come ange re d be caus e th e y acquire ne w or inte ns ifie d e x pe ctations wh ich cannot be s atis fie d by me ans at th e ir dis pos al; th e pe rce ive d gap be twe e n e x pe ctations and capability is "re lative de priva- tion," wh ich ge ne rate s dis conte nt. T h e s e cond migh t be calle d a "capa- bility de cay" patte rn; th e s ource of pe ople 's dis conte nt is th e ir de clining capacity to s atis fy s table e x pe ctations .9 T h e Fe ie rabe nds and Be tty N e s - voId add to th e s e bas ic mode ls , dis tinguis h ing, for e x ample , s e ve ral J- curve type patte rns of ch ange , rapid and minimal ch ange patte rns , and a fluctuation ch ange patte rn.10 T h e s e kinds of th e orie s quickly e ngage us in two additional que s tions : Wh at ch ange s according to th e s e patte rns , and wh y? T h e re is approx i- mate cons e ns us among th e s e th e oris ts about wh at ch ange s . On th e one h and, wh at pe ople e x pe ct out of life ch ange s (Davie s : "e x pe cte d ne e d s atis faction"; Gurr: "value e x pe ctations "; Fe ie rabe nds and N e s vold: "s ocial e x pe ctations ," "pre s e nt e x pe ctations of future gratifications "). On th e oth e r h and, wh at pe ople do ge t or th ink th e y can ge t out of life ch ange s (Davie s : "actual ne e d s atis faction"; Gurr: "value capabilitie s "; Fe ie rabe nds and N e s vold: "s ocial ach ie ve me nt"). All th e oris ts re cognize th at me n s e e k many diffe re nt type s of value s . I us e a clas s ification of value s wh os e th re e ge ne ric type s are we lfare , powe r, and inte rpe rs onal (be longingne s s , re s pe ct, s tatus ) value s . Davie s propos e d a fourfold clas s ification compris ing ph ys ical ne e ds , s ocial-affe ctional ne e ds , e qual e s te e m or dignity ne e ds , and, finally, s e lf-actualization. More ove r, h e 8 Davie s , "T h e J-Curve ," 547. 9 Gurr, ch ap. 2. 10Fe ie rabe nd, Fe ie rabe nd, and N e s vold. 365 Comparative Politics April 1973 as s e rts th at th e s e h ave aunive rs al h ie rarch y of importance : once ph ys i- cal ne e ds are s atis fie d, s ocial-affe ctional ne e ds be come dominant; wh e n th e y are s atis fie d th e ne e d for dignity pre dominate s ; and s o forth . T h is h ypoth e s is h as a ps ych ological bas is . If it is manife s t in colle ctive affairs , it is of gre at importance for any analys is of th e conne ctions be twe e n ch ange and viole nce . Cons ide r th e fre que nt obs e rvationth at incre as ing mate rial we ll-be ing s e e ms s o ofte n as s ociate d with th e ge ne ration of re volutionary de mands . T h os e de mands may we ll re fle ct e me rging ne e ds . At pre s e nt, th e h ypoth e s is is s upporte d only by afe w plaus ible , but far from de finitive , cas e s tudie s . Wh y e x pe ctations and ach ie ve me nt ch ange ove r time is a th orny que s tion, one wh os e ans we rs involve a re gre s s ionup an e ve rwide ning "funne l of caus ality." One of Davie s ' ans we rs was jus t outline d: ne w ne e ds e me rge . T h e Fe ie rabe nds ' and N e s vold's ans we r is s pe cific to th e conte mporary world: th e y take th e s ocioe conomicmode rnization proce s s as a give n, and s ugge s t th at it is s imultane ous ly th e s ource of incre as ing e x pe ctations among"mode rnizing" groups and a th re at to th e ach ie ve - me nts of "traditional" groups . More ove r, th e attainme nt of mode rn goals is like ly to be th warte d by th e traditionals , and vice ve rs a. "T h e farth e r th e proce s s of trans ition progre s s e s , th e more like ly and th e more inte ns e th e conflicts be twe e nmode rnand e s tablis h e d patte rns . T h e s ituation [is ] a mas s ive conflict, re fle cte d in myriad individual ps ych e s of diffe re nt s trata.. . and infe ctingdiffe re nt domains of th e s ocial proce s s ." 11 In an ope rational te s t of th is th e ory, time -lagge d comparis onof many conte m- porary nations s h ows ge ne rally th at th e h igh e r th e le ve ls of s ocial and mate rial mode rnity, th e le s s th e political ins tability and viole nt conflict, but th at th e gre ate r th e rate of ch ange toward th os e conditions , th e gre ate r th e dis ruption. In oth e r words , th e trans itional nations -re la- tive ly unmode rnize d but rapidly ch anging-are th e mos t s ubje ct to viole nt conflict. T h e e vide nce is s ugge s tive , not de finitive . An e qually im- portant kind of s tudy re mains to be done : th e corre lationof indice s of s pe cifickinds of ch ange with th e incide nce of viole nt conflict ove r time in particular countrie s . T h is type of longitudinal s tudy is woe fully lack- ing in s tudie s of conflict and viole nce . T h e one e x ce ptionth at be ars me ntioningis Pitirim Sorokin's e x traordinary analys is of ch ange , s ocial dis turbance s , and war acros s twe nty-five ce nturie s of Europe anh is tory. He finds e vide nce th e re infor along-te rm, cyclical s h ift from mate rialis tic to ide ological culture s and back again, e ach pe riod of trans itionbe ing accompanie d by inte ns ifie d viole nce . From th is point of vie w, th e con- te mporary "drive to mode rnity" is only a ph as e of a large r wave of ch ange . T h e th e ory h as powe rful implications for our topic. It doe s not "Ibid., 507. 366 T e d Robe rt Gurr s e e m to h ave be e n th e s ubje ct of any furth e r conce ptual or e mpirical work.l2 I h ave take n a more inductive and microanalyticapproach to th e que s tion of s pe cifying s ource s of ch ange in colle ctive e x pe ctations and capabilitie s . A numbe r of re lative ly s pe cific conditions th at are ide nti- fiable e ith e r incre as e e x pe ctations or de cre as e capabilitie s in ways th at ge ne rate incre as e d pote ntial for viole nt conflict. For e x ample , e ach group's pas t rate of ch ange in abs olute pos ition, up or down, and its de - cline re lative to oth e r groups are s trongpos itive de te rminants of its po- te ntial for viole nce . So is lack of a value , like powe r, th at prove s ne ce s - s ary to prote ct value s alre ady attaine d, s uch as we ll-be ing and s tatus . Conve rs e ly, th e gre ate r th e range of alte rnative s ope n to a group's me m- be rs and th e gre ate r th e availability of re s ource s in th e s ocie ty, th e le s s is th e pote ntial for viole nt conflict.l3 All of th e s e variable s are s ubje ct to ch ange ove r time ; th e y could be analyze d as s pe cific ins tance s of s ome of th e dime ns ions of s ocial ch ange cite d e arlie r. T h e fact th at th e y are not h igh ligh ts th e ir ad h occh aracte r, and points up again th e de s irability of working toward be tte r inte grate d th e orie s of s ocial ch ange and viole nt conflict. T h e re are oth e r ps ych ological approach e s to th e analys is of viole nt conflict. A re ce nt s ympos ium by ps ych iatris ts analyze s individual and colle ctive viole nce as forms of adaptive or copingbe h avior, and argue s th at th is inte rpre tationis cons is te nt with avarie ty of narrowe r ps ych o- logical and ps ych iatricth e orie s about th e caus e s of viole nce .'4 Victor Wolfe ns te inh as atte mpte d aps ych odynamicinte rpre tationof th e origins of re volutionary le ade rs h ip.'5 Political alie nationis th e s ubje ct of e x te n- s ive th e ore tical and e mpirical work by David Sch wartz, wh o is con- ce rne d with both its ps ych os ocial de te rminants and with th e ps ych ologi- cal variable s th at de te rmine wh e th e r it is dire cte d intopas s ive or re be l- lious be h avior.'6 In none of th e s e ins tance s , h owe ve r, h as much atte n- tion be e n give nto th e patte rns of s ocial ch ange th at h ave wide s pre ad ps ych ological e ffe cts of th e kinds cons ide re d. T h e s e are s pe cifically l Pitirim A. Sorokin, Social and Cultural Dynamics , Vol. l11: Fluctuations of Social Re lations h ips , War and Re volutions (N e w York, 1937). A re analys is of Sorokin's data, th us far unpublis h e d, h as be e n made by Profe s s or Paul Smoke r of th e Unive rs ity of Britis h Columbiaand Unive rs ity of Lancas te r (England). 13Gurr, ch aps . 4 and 5. +David N . Danie ls e t al., Viole nce and th e Struggle for Ex is te nce (Bos ton, 1970). 15 E. Victor Wolfe ns te in, T h e Re volutionary Pe rs onality: Le nin, T rots ky, Gandh i (Prince ton, 1967). 1l David C. Sch wartz, Political Alie nation and Political Be h avior (Ch icago, forth coming). 367 Comparative Politics April 1973 ps ych ological th e orie s , in contras t with th e th re e s ocial-ps ych ological th e orie s dis cus s e d above . Social-s tructural th e orie s T h e s ocial-s tructural th e orie s of viole nt con- flict diffe r in e mph as is rath e r th an kind from th e s ocial-ps ych ological th e orie s . T h e ir common pre mis e is th at s ome fundame ntal s ocial dis - location, various ly calle d "s train" or "dys function," is th e ne ce s s ary pre condition for re volutionary conflict. Wh e re as th e s ocial-ps ych ological th e oris ts of viole nce be gin with aggre gate ps ych ological s tate s , th e n work both "backward" to th e ir s ocial de te rminants and "forward" to th e ir cons e que nce s , th e s ocial-s tructural th e oris ts link s pe cifie d kinds of s ocial ch ange dire ctly to th e ir colle ctive ly viole nt outcome s with out s ubs tantial re fe re nce to any inte rve ning ps ych ological variable s . A s imilarity be - twe e n th e two type s of th e ory is th e ir analogous conce ptualization of conditions th at inte rve ne be twe e n th e ps ych ological or s ocial pre condi- tions and th e actual occurre nce of viole nt conflict. All of th e m s pe cify s ome as pe cts of gove rnme nt or ins titutional arrange me nts ge ne rally th at facilitate or de fle ct th e unde rlyingimpe tus to viole nce . I will s ummarize two of th e s e s ocial-s tructural th e orie s and me ntion s e ve ral oth e rs . Sme ls e r's T h e ory of Colle ctive Be h avior incorporate s a ge ne ral con- ce ptual analys is of s ocial ch ange . It is principally conce rne d with s h ow- ingh ow various kinds of s tructural s train produce "colle ctive be h avior," wh ich is de fine d as "mobilizationon th e bas is of abe lie f wh ich re de fine s s ocial action." 17 More concre te ly, "colle ctive be h avior" include s panics , craze s , h os tile outburs ts (includingriots ), and norm- and value -orie nte d move me nts . Hos tile outburs ts and value -orie nte d move me nts compris e mos t viole nt conflicts -th ough note th at Sme ls e r is conce rne d with ac- counting for th e ir non- or antis ys te m compone nt, not with e x plaining conflict more ge ne rally. Sme ls e r ide ntifie s s ix s e ts of s ocial de te rminants wh os e various de gre e s , type s , and concurre nce produce diffe re nt kinds of colle ctive be h aviors : 1. s tructural conducive ne s s - s tructural ch aracte ris tics th at pe rmit or e ncourage colle ctive be h avior, wh ich are e ffe ctive only in com- bination with 2. s tructural s train - "ambiguitie s , de privations , conflicts , and dis - cre pancie s ," s uch as "re al or anticipate d e conomic de privation"; 3. growth and s pre ad of ge ne ralize d be lie f - wh ich provide s pote n- tial actors with an inte rpre tationof th e s ituation and s pe cifie s ap- propriate re s pons e s ; 17 Sme ls e r, 8. 368 T e d Robe rt Gurr 4. pre cipitatingfactor - a s pe cific e ve nt th at trigge rs group action; 5. mobilization of participants for action - us ually by ale ade r; and 6. ope ration of s ocial control - "th os e counte r-de te rminants wh ich pre ve nt, inte rrupt, de fle ct, or inh ibit th e accumulation of th e [above ] de te rminants ," including th os e wh ich minimize condu- cive ne s s and s train, and th os e mobilize d afte r a colle ctive e pis ode be gins .18 Structural s train is th e mos t important of th e s e , as is e vide nt in th is h ypoth e s is : "Some form of s train mus t be pre s e nt if an e pis ode of col- le ctive be h avior is to occur. T h e more s e ve re th e s train, more ove r, th e more like ly is s uch an e pis ode to appe ar." 19 All th e s ix de te rminingcon- ditions are s ubje ct to ch ange ; (3) and (5) de note particular kinds of ch ange . T h e mos t s ignificant s ocial ch ange s , h owe ve r, are pre s umably th os e th at caus e s e ve re s tructural s train. Sme ls e r offe rs s e ts of cate gorie s for typologizing s train. One s e t of th e s e cate gorie s is h is compone nts of s ocial action, wh ich I me ntione d above : th e y re fe r to pe ople 's value s , norms , organization, and "s ituational facilitie s ." A cros s -cutting s e t of cate gorie s re fe rs to th e le ve ls of s pe cificity of e ach compone nt of action. T o us e norms as an e x ample , a particular bus ine s s 's code of ope rations is more s pe cific th an th e bus ine s s community's norms about bus ine s s h one s ty and de ce ncy, and it is more s pe cific th an le gal code s about contracts and prope rty. Se ve n le ve ls of s pe cificity are ide ntifie d for e ach compone nt of s ocial action, ge ne rating twe nty-e igh t cate gorie s accord- ing to wh ich various kinds of s ocial ch ange s may be pige onh ole d. Sme ls e r doe s not provide any dynamic analys is of ch ange pe r s e . His dis cus s ion of de te rminants of value -orie nte d move me nts , for e x ample , me re ly lis ts kinds of s trainlike "inade quacy of knowle dge of te ch nique s to grapple with ne w s ituations ," "s e ve re ph ys ical de privation," and "th e normative dis organization th at war occas ions ." 20 T h e dynamic com- pone nt to th e th e ore tical frame work conce rns th e linkage s be twe e n type s of s train and th e nature of s ubs e que nt colle ctive be h avior. Brie fly, s train of a give n le ve l and type s timulate s th e cre ation of ge ne ralize d be lie fs at a h igh e r le ve l, wh ich re de fine s ocial action in s uch a way th at th e s train is modifie d. Sme ls e r s pe cifie s proce s s e s by wh ich th is occurs and s ug- ge s ts wh ich type s of s train le ad to, s ay, "h ys te rical be lie fs ," "h os tile be lie fs ," and s o forth . Sme ls e r's th e ory th us offe rs a ge ne ral approach to analyzing conne ctions be twe e n s ocial ch ange and viole nt conflict; its cate gorie s of ch ange , at le as t, are s ys te maticrath e r th an ad h oc. But it doe s not advance cle arly formulate d, dynamic h ypoth e s e s , much le s s "Ibid., 15-18. "9 Ibid., 48. I Ibid., 338-40. 369 Comparative Politics April 1973 s ugge s t h ow th e y migh t be ope rationally te s te d. Such h ypoth e s e s could be de rive d from th e th e ory; as writte n, th ough , it is principally a typology wh ich is more s uitable for de s cribing and inte rpre tings pe cific e ve nts th an for s timulating furth e r th e ore tical de ve lopme nt and as s e s s me nt. Wh e re as Sme ls e r is conce rne d with colle ctive be h avior ge ne rally, Ch alme rs Joh ns on h as propos e d a th e ory to account s pe cifically for re v- olutionary conflict.21 T h e th e ory can be s ummarize d rath e r brie fly. It take s account of four de te rmining variable s . T h e firs t ne ce s s ary caus e of re volutionary conflict is a dis e quilibrate d s ocial s ys te m, one in wh ich e ith e r s ys te ms of value s and "s ymbolicinte rpre tations of s ocial action," or th e s ocie ty's patte rn of adaptation to th e e nvironme nt, ch ange s uffi- cie ntly th at s ocie ty's functional re quire me nts can no longe r be fulfille d. Elite s face d with th is s ituation may or may not atte mpt to re dre s s th e dis e quilibrium. If th e y prove intrans ige nt or unable to do s o, th e y los e le gitimacy-th e s e cond ne ce s s ary caus e of re volution. T h e y may s till continue in powe r for s ome time by re lying on coe rcion. T h e final, s uf- ficie nt caus e of re volution-Joh ns on calls it an "acce le rator"-is th e e lite 's los s of control ove r th e ins trume nts of coe rcion. T h e military may be de fe ate d in war, or be come incre as ingly ambitious or dis affe cte d from th e rule rs , or be ch alle nge d to re volutionary combat, but, h owe ve r it h appe ns , th e acce le rator pre cipitate s re volutionary conflict. (T h e true de gre e of military loyalty and e ffe ctive ne s s is th e major de te rminant of th e outcome of th at conflict.) All four variable s are e vide ntly s ubje ct to ch ange . Value s and adapta- tion patte rns may ch ange ; s ubs tantial s h ifts in e ith e r or both caus e dys - functions . Joh ns on ide ntifie s four kinds of dys functions , de rive d from Pars ons ' s pe cification of s ocial s ys te ms ' functional ne e ds . T h e dys func- tions are incoh e re nt s ocialization, inappropriate e ns e mble of role s , dis - s e ns us on goals , and-tautologically-failure to re s olve conflicts pe ace - fully. T h is typology is analogous to Sme ls e r's much more de taile d typol- ogy of kinds of ch ange . Joh ns on goe s one s te p furth e r back, to s ugge s t as imple typology of s ource s of ch ange , with e x ample s . T h e s e are : (1) e x - oge nous conditions caus ing value ch ange s , e .g., inte rnational "de mon- s tration e ffe cts "; (2) e ndoge nous value -ch anging conditions , e .g., th e ris e of ne w re ligious move me nts ; (3) e x oge nous s ource s of e nvironme ntal ch ange , e .g., fore ign trade , te ch nology, and conque s t; and (4) e n- doge nous s ource s of e nvironme ntal ch ange , e .g., te ch nological innova- tions .22 Joh ns on, like Sme ls e r, doe s not formulate h ypoth e s e s about pos s ible s ys te maticconne ctions amongtype s or ch aracte ris tics of ch ange 2l Ch alme rs Joh ns on, Re volution and th e Social Sys te m (Stanford, 1964); Ch alme rs Joh ns on, Re volutionary Ch ange (Bos ton, 1966). 22Joh ns on, Re volutionary Ch ange , pp. 64-70. 370 T e d Robe rt Gurr and type s or e x te nt of dys function; th is part of Joh ns on's th e ory is e s - s e ntially an e x e rcis e in typology. T h e th e ory's dynamic e le me nts conce rn th e cons e que nce s of dys function for re volutionary conflict. T h e e lite 's re s pons e can vary from "cons e rvative ch ange " to "intrans ige nce "; its le gitimacy de pe nds on wh ich re s pons e it take s . Joh ns on fails to s pe cify or e ve n to s ugge s t th e de te rminants of th e s e re s pons e s or h ow th e y migh t ch ange ove r time . T h e intrans ige nt e lite 's s ubs e que nt re liance on force s e ts th e s tage for th e ope ration of th e "acce le rators ," a clas s of particu- lar type s of ch ange s . Joh ns on dis tinguis h e s th re e kinds : conditions wh ich re duce th e e ffe ctive ne s s of th e arme d force s dire ctly; conce rte d be lie fs of oppos itional groups th at th e y can win; and s trate gically calculate d viole nt cons piracie s . He goe s on to s ugge s t h ow th e forms and proce s s e s of re volution and its outcome s de pe nd on various combinations of th e s pe cifie d conditions , plus additional variable s introduce d on an ad h oc bas is .23 T h e re is no s ys te matic atte mpt to re late th is dis cus s ion to fore - goingcate gorie s of s ocial ch ange or dys function, th ough , as with Sme ls e r, anumbe r of h ypoth e s e s of th is s ort could be de rive d from th e dis cus s ion. I me ntione d above th at both th e s ocial s tructural th e orie s and th os e re vie we d in th e pre vious s e ction ide ntify as pe cts of s ocie ty wh ich facili- tate or inh ibit th e impe tus to viole nt conflict. T h e kinds of conditions cite d are re markably s imilar. Ide ological factors are ge ne rally re cog- nize d, not only by Sme ls e r: Davie s s pe cifie s th at h os tility mus t be focus e d on gove rnme nt be fore re volutionary move me nts occur; 24 and I ide ntify cultural and political s ource s of be lie fs th at pe rform th is focus s ing function.25 Sme ls e r's conce rn with "s tructural conducive ne s s " and "mo- bilization of participants for action" is paralle le d by my h ypoth e s is in Wh y Me n Re be l th at "th e magnitude of political viole nce varie s . . . with th e ratio of dis s ide nt ins titutional s upport to re gime ins titutional s upport to th e point of e quality, and inve rs e ly be yond it." 26 I go on to s pe cify s ome ge ne ral de te rminants of ins titutional s upport and orie ntation. Joh n- s on's inte re s t in e lite intrans ige nce and re volutionary organization are as pe cts of wh at Sme ls e r and I dis cus s in more ge ne ral te rms . Finally, th e s e th e orie s , with out e x ce ption, ide ntify coe rcion as an inte rve ning variable . Sme ls e r and th e Fe ie rabe nds de al re s pe ctive ly with a s ys te m's e x te nt of "s ocial control" and "coe rcive ne s s " as inh ibitors of viole nce . Joh ns on and I are both conce rne d with th e re lative balance be twe e n re volutionary and e lite capacitie s for e mployingforce . T h e principal diffe re nce s be twe e n th e two groups of th e orie s re vie we d 23Ibid., ch ap. 7 and 8. 24 Davie s , "T h e J-Curve ." 25 Gurr, ch ap. 6 and 7. 20 Ibid., p. 276. 371 Comparative Politics April 1973 th us far are : (1) dis agre e me nt about wh e th e r and h ow th e ps ych ological le ve l of analys is ough t to be take n into account; and (2) wh e th e r a typological or propos itional approach to e x plaining viole nt conflict ough t to be take n. T h e ne x t s e ction cons ide rs s ome conflict th e orie s of caus a- tion th at h ave a s ubs tantially diffe re nt approach . Some oth e r s ocial-s tructural th e orie s migh t be me ntione d. Some e arly twe ntie th -ce ntury th e orie s are e s s e ntially of th is s ort, for e x ample , th os e of Lyford Edwards and Ge orge Pe tte e .27 More re ce ntly David Wille r and Ge orge Zolls ch anattribute re volutionary move me nts to th e e x is te nce of e x ige ncy, a cons e que nce of dis cre pancy be twe e n th e s tructural pos ition of individuals in s ocie ty and th e ir inte re s t pos ition. Wide s pre ad e x ige ncie s are trans forme d in re volutionary move me nts to th e e x te nt th at e x ige ncie s are articulate d, affe cte d groups are organize d to facilitate action, and th e ir me mbe rs pe rce ive th e re gime as th e s ource of pe rs is tinge x ige ncy.28 Group conflict th e orie s All th e th e orie s cons ide re d s o far are con- ce rne d principally with th e motive s of one party to viole nt conflicts : th e "re be ls ," th e "dis ruptors ," th os e wh o ch alle nge an e x is ting s tatus quo or e quilibrium. T h e re is a compe ting paradigm for th e ory, wh os e origins can be trace d at le as t back to Aris totle , th at as ks h ow and wh y groups in s ocie tie s come into conflict. From th is point of vie w th e inte re s ts and conflict be h avior of "e lite s " and "re gime s " are as important a s ubje ct for inquiry as th os e of any oth e r s ocial group. T h e bas ic pre mis e of group conflict th e orie s is th at viole nt conflict and re volution aris e out of group compe tition ove r value d conditions and pos itions . As Ch arle s T illy puts it in a wide ly quote d s tate me nt, "Me n s e e king to s e ize , h old, or re align th e le ve rs of powe r h ave continually e ngage d in colle ctive viole nce as part of th e ir s truggle s . T h e oppre s s e d h ave s truck in th e name of jus tice , th e privile ge d in th e name of orde r, th os e in be twe e n in th e name of fe ar." 29 Powe r is not th e only is s ue of conflict, of cours e ; s o are we ll-be ing, s tatus , communal and ide ological purity, and s o forth . A corollary of th is pre mis e is th at viole nt conflict is are curre ntfe ature of s ocie tie s . As th e compos ition, inte re s ts , and re lative pos itions of groups ch ange , conflict occurs and s o doe s viole nce . T h e re are many s uch th e orie s . T wo s ome wh at diffe re nt type s are re vie we d be low: th os e wh ich e mph as ize conflictful e le me nts in group diffe re ntiation ge ne rally, and 27 Se e footnote 5. 28 David Wilie r and Ge orge K. Zolls ch an, "Prole gome nonto a T h e ory of Re vo- lutions ," in Ge orge K. Zolls ch anand Walte r Hirs ch , e ds . Ex plorations in Social Ch ange (Bos ton, 1964), pp. 125-51. 29 Ch arle s T illy, "Colle ctive Viole nce in Europe anPe rs pe ctive ," in Grah am and Gurr, ch ap. 1. 372 T e d Robe rt Gurr th os e wh ich e mph as ize h orizontal or clas s cle avage s . T h e forme r are more ge ne ral th an, but do not ne ce s s arily s ubs ume , th e latte r. Group diffe re ntiationand conflict N ich olas T imas h e ff and Ralf Dah r- e ndorf are among th e th e oris ts wh o attribute conflict to th e e x is te nce of ge ne ral group diffe re ntiation. T imas h e ff take s th e more "traditional" ap- proach , writing of "ne ce s s ary and s ufficie nt" conditions for viole nt con- flict, but h is obje ct is more ge ne ral: to account for th e occurre nce of both re volution and war. He pos tulate s four s ufficie nt conditions : two partie s or groups h ave h igh ly value d, incompatible goals ; normative in- h ibitions agains t viole nce are we ake ne d among one or both partie s to th e conflict; one or both los e s h ope of winning by nonviole nt me ans of conflict re s olution; and e ach th inks it h as s ome ch ance of winning by viole nt me ans . T imas h e ff catalogs rath e r th an ge ne ralize s about th e nature and s ource s of goal conflicts in pre re volutionary s ituations . He lis ts dis pute s ove r th e nature or functioning of gove rnme nt, ch anging clas s dis tributions of s ocial or e conomic value s , and ris inge x pe ctations . Such conflicts mus t be "s e rious , or many, or both " to caus e re volution. He s ugge s ts th at th e failure of th e coordinating function of th e s tate caus e s th e groups in conflict to de s pair of pe ace ful s olutions ; h e doe s not s pe culate s ys te matically wh y th at coordinating function migh t fail. As te ns ions grow, inh ibitions agains t viole nce te nd to de cre as e . T h e final s tage s re s e mble Joh ns on's th e ory of re volutionary caus ation. T h e gove rnme nt, wh ich is by de finition one of th e partie s to re volutionary conflict, s e e s its ch ance s of winning as re s ting with its control of th e arme d force s . T h e re volutionarie s may th ink th at th e y can win be caus e th e y s ubs cribe to an ide ology th at convince s th e m th e time is ripe , or be caus e of tactical calculations , or out of de s pair th e y may take action agains t all odds .30 Dah re ndorf's th e ory of group conflict is more pre cis e and s oph is ticate d th an T imas h e ff's . It s uffe rs cons ide rably from any atte mpt at brie f s um- marization. One major th e s is of Dah re ndorf's work is th at group con- flict is pe rvas ive in all s ocie tie s at all time s ; it cannot be re s olve d, only- s ome time s -re gulate d. He is more conce rne d with wh at h e calls clas s conflict th an T imas h e ff, and h e cons ide rs clas s conflict a s pe cie s of group conflict. Clas s conflict is de fine d as any conflict be twe e n th e s upe rordi- nate and s ubordinate s trataof auth ority s tructure s . T h is is cons ide rably diffe re ntfrom and broade r th an Marx 's de finitionof clas s e s by re fe re nce to groups ' re lations h ip to th e me ans of production (s e e be low). For s oFrom N ich olas S. T imas h e ff, War and Re volution (N e w York, 1965), ch aps . 4 and 7. 373 Comparative Politics April 1973 Dah re ndorf th e de te rminant of "clas s " is diffe re ntial allocation of au- th ority, wh ich ch aracte rize s a much wide r varie ty of ins titutions th an jus t th e s tate . T wo principal as pe cts of conflict about wh ich Dah re ndorf ge ne ralize s are its inte ns ity (th e e ne rgy e x pe nditure and de gre e of involve me nt of conflicting partie s ) and its viole nce . T h e inte ns ity of clas s conflict is s aid to de cre as e with th e e x te nt of clas s organization; with th e e x te nt to wh ich various clas s and group conflicts are dis s ociate d rath e r th an coinci- de nt; and with "th e e x te nt th at th e dis tributionof auth ority and th e dis - tributionof re wards and facilitie s in an as s ociation are dis as s ociate d...." T o clarify th e las t s tate me nt, Dah re ndorf is s aying th at conflict is le s s inte ns e if th e h olde rs of powe r in an as s ociation do not als o h old h igh e s t s tatus , mate rial we ll-be ing, and th e like . T h e viole nce of clas s conflict is s aid to de cre as e with th e e x te nt of clas s organization, with th e de cline of abs olute de privation, and (tautologically?) with th e e x te nt of e ffe ctive conflict re gulation. All th e s e de te rmining conditions are s ubje ct to ch ange , but th e s ource s of th os e ch ange s are not much de alt with in th e th e ory. Cons ide rable nonformalize d cons ide ration is give n to th e e me rge nce and h is torical de ve lopme nt of clas s e s . Dah re ndorf's ce ntral purpos e , h owe ve r, is to s h ow h ow s tructural ch ange s in s ocie ty are de te rmine d by group conflict. He is th us th e only one of th e viole nce and conflict th e oris ts we h ave e x amine d wh o is s ignificantly conce rne d with ge ne ral s ocie tal cons e que nce s of conflict: conflict is th e inde pe nde nt variable , not th e de pe nde nt one , in th e th e ory.3' We will cons ide r th e s e caus al conne c- tions brie fly in s e ction III of th is article . T h e re are nume rous oth e r th e orie s conce rne d with th e origins , pro- ce s s e s , and me lioration of group conflict, but fe w are s ubs tantially con- ce rne d with viole nt, re volutionary conflict. Ins ofar as th e s e th e orie s ac- count for conflict ge ne rally, th e y may be s aid to account for its viole nt manife s tations as we ll; th e crucial que s tion for us is wh y conflict doe s or doe s not take viole nt forms , and in re s pons e to wh at if any kinds of ch angings ocial conditions . Some of th e conflict th e orie s I h ave in mind are Ke nne th E. Boulding's Conflidt and De fe ns e : A Ge ne ral T h e ory,32 Anatol Rapoport's Figh ts , Game s , and De bate s ,33 and T h omas C. Sch e ll- ing's T h e Strate gy of Conflict.34 Oth e rs are re vie we d and s ynth e s ize d by Raymond Mack and Rich ard Snyde r.35 T h e s e th e orie s various ly attribute 31 From Ralf Dah re ndorf, Clas s and Clas s Conflict in Indus trial Socie ty (Stan- ford, 1959), ch aps . 5 and 6; quotationfrom p. 239. 32N e w York, 1962. 33Ann Arbor, 1960. 3Cambridge [Mas s .], 1963. 88 Raymond W. Mack and Rich ard C. Snyde r, "T h e Analys is of Social Conflict: 374 T e d Robe rt Gurr th e trans itionfrom nonviole nt to viole nt conflict to th e we akne s s of inte - grating or re gulatory proce dure s , and to calculations of participants th at viole nce is an e ffe ctive me ans to th e ir e nds . T h e y are not, h owe ve r, particularly fruitful as s ource s of s ys te matic ge ne ralizations about con- ne ctions be twe e n s ocial ch ange and viole nt conflict, e x ce pt pe rh aps at a ve ry abs tract le ve l. Se e Mack and Snyde r's re mark th at "s ocial ch ange -its rate and dire ction-is an ultimate s ource of conflict be caus e , as th e factual s ocial orde r unde rgoe s trans ition, ne w incompatibilitie s and antagonis tic inte re s ts aris e ." 36 T h e orie s of clas s conflict T h e s e ve ral th e orie s of viole nt conflict con- s ide re d h e re attribute viole nt conflict, at le as t in its re volutionary form, to dis paritie s in th e s h are s of value d goods h e ld by diffe re nt, h orizontally s tratifie d clas s e s . Aris totle h e lpe d e s tablis h th is tradition of th e orizing in h is e fforts to e x plain th e circums tance s of political re volutions in th e Gre e k city-s tate s . T h e principal caus e of re volution, h e propos e d, is th e common pe ople 's as piration for e conomic or political e quality wh e n th e y lack it, and th e as piration of oligarch s for gre ate r ine quality th an th e y h ave . Me n's de s ire s for one value d condition are corre late d with th e ir attainme ntin re s pe ct to oth e rs . T h us , if th e common pe ople are e qual in powe r and s tatus to th e ir rule rs , th e y will like ly s e e k e quality in mate rial we ll-be ing. Me mbe rs of an oligarch y, if th e y h ave s upe rior mate rial we ll- be ing, will, conve rs e ly, te nd to s e e k s upe rior powe r and s tatus as we ll. T h e imme diate caus e s of re volution are th e s e as pirations afte r e quality (or ine quality); th e principal or more re mote caus e s of re volution are th e s ocial conditions wh ich produce s uch dis pos itions . Like many late r th e oris ts , Aris totle catalogs nume rous conditions th at h ave th e s e e ffe cts but doe s not ge ne ralize about th e m to th e de gre e th at h e ge ne ralize s about re volutionary caus ation. Some are re lative ly s taticconditions , s uch as th e conte mpt of a s ubordinate clas s for its mas te rs , and e th nic dive r- s ity. Oth e r conditions are type s of ch ange , s uch as a dis proportionate in- cre as e in th e re lative s ize of one clas s , or th e improvingcondition of one clas s or group; th e latte r is s aid to ins pire re volution in two ways : by fos te ring e nvy in oth e r clas s e s , and by cre ating as pirations for furth e r ine quality in th e advancingclas s .37 T h e re is a re s e mblance h e re to th e conte mporary "J-curve " and re lative de privation approach e s to e x plain- ing viole nt conflict. T oward an Ove rvie w and Synth e s is ," Journal of Conflict Re s olutions , I (June 1957), 212-48. 3 Ibid., 227. 37 From Aris totle , T h e Politics , Book V. 375 Comparative Politics April 1973 We may ne x t s ke tch Marx 's th e ory of th e origins of re volutionary con- flict, alth ough it is s o we ll known th at it is pe rh aps gratuitous to do s o. Marx finds re volution e s s e ntially a function of e conomic ch ange , s pe - cifically th e de ve lopme nt of contradictions be twe e n productive force s of s ocie ty and th e re lations of clas s e s to production. T h e re are a s ucce s s ion of h is torically ine vitable s tage s of e conomic organization, th e pe nultimate of wh ich , bourge ois capitalis m, give s way to th e clas s le s s s ocie ty of th e worke rs . Re volution marks th e trans itions among s tage s . T h is is Marx 's own s ummary: In th e s ocial production of th e ir me ans of e x is te nce me n e nte r into . . . productive re lations h ips wh ich corre s pond to a de finite s tage of de ve lopme nt of th e ir mate rial productive force s . T h e aggre gate of th e s e productive re lations h ips cons titute s th e e conomics tructure of s ocie ty, th e re al bas is on wh ich a juridical and political s upe rs tructure aris e s . . . . T h e mode of production of th e mate rial me ans of e x is te nce con- ditions th e wh ole proce s s of s ocial, political and inte lle ctual life . ... At a ce rtains tage of th e ir de ve lopme nt th e mate rial productive force s of s ocie ty come into contradictionwith th e e x is tingproductive re la- tions h ips , or, wh at is but a le gal e x pre s s ion of th e s e , with th e prope rty re lations h ips with in wh ich th e y h ad move d be fore . From forms of de ve lopme nt of th e productive force s th e s e re lations h ips are trans - forme d into th e ir fe tte rs . T h e n an e poch of s ocial re volution ope ns . With th e ch ange in th e e conomicfoundationth e wh ole vas t s upe r- s tructure is more or le s s rapidly trans forme d. ... A s ocial s ys te m ne ve r pe ris h e s be fore all th e productive force s h ave de ve lope d for wh ich it is wide e nough ; and ne w, h igh e r productive re lations h ips ne ve r come into be ingbe fore th e mate rial conditions for th e ir e x is te nce h ave be e n brough t to maturity with in th e womb of th e old s ocie ty its e lf. ... In broad outline , th e As iatic, th e ancie nt, th e fe udal and th e mode rn bourge ois mode s of productioncan be indicate d as progre s s ive e poch s in th e e conomics ys te m of s ocie ty. Bourge ois productive re lations h ips are th e las t antagonis ticform of th e s ocial proce s s of production.38 T h e s pe cificcontradictionth at brings about th e re volutionary re place - me nt of th e capitalis t s ys te m of productionis th e progre s s ive immis e riza- tion of th e working clas s . Capitalis m mus t ne ce s s arily and incre as ingly e x ploit worke rs , Marx argue d, be caus e profit rate s te nd to de cline (for as s e rte d e conomic re as ons not re vie we d h e re ). As th e y de cline , e m- ploye rs pre s s worke rs toward and th e n be low th e s ubs is te nce le ve l. T h is e conomic e x ploitation is accompanie d by incre as ingpolitical oppre s s ion, a cons e que nce of capitalis m's ne e d for s table control ove r th e me ans of 38 Karl Marx , A Contribution to th e Critique of Political Economy, as quote d in William Ebe ns te in, Mode rn Political T h ough t: T h e Gre at Is s ue s (N e w York, 1954), 367-68. 376 T e d Robe rt Gurr production and ove r th e worke rs th e ms e lve s . Re volutionary move me nts be gin wh e nworke rs be come cons cious th at th e ir colle ctive mis e ry can be re lie ve d only by a total trans formationof th e e conomicand its de - pe nde nt political s ys te m. Marx at le as t cannot be accus e d of "catalog- ing" s ocial ch ange s th at dis pos e to re volutionary conflict: h e pre cis e ly ide ntifie s th e type s and s e que nce s of ch ange th at caus e re volution. But wh e th e r th is e x planation fits all manife s tations of re volutionary conflict e qually we ll is anoth e r matte r. Marx , at le as t, would argue th at all con- flict is e conomic-bas e d clas s conflict. In th is conne ction, we canme ntion th at one of Mao T s e -tung's major contributions to Marx is t th e ory is h is e mph as is on th e conflictful and re volutionary pote ntial of powe r diffe r- e nce s amongclas s e s ; th e s ource s and dynamics of political oppre s s ion are th us anadditional are afor re volutionary th e orizing.39 Both Aris totle and th e Marx is t th e oris ts we re conce rne d primarily with ide ntifying th e clas s bas e s of re volutionary conflict. Joh an Galtung h as propos e d a"s tructural th e ory of aggre s s ion" th at ge ne ralize s be yond th e narrow conce ption of "clas s " as a party to conflict, and th at als o pro- pos e s to iaccount for a range of aggre s s ive be h aviors from viole nt crime th rough riots and re volutionto war. Accordingto Galtung, it is th e h ie rarch ical pos itionof th e individual or group re lative to oth e rs th at de te rmine s aggre s s ive be h avior. His ce ntral h ypoth e s is is th at "Aggre s - s ionis mos t like ly to aris e in s ocial pos itions in rank-dis e quilibrium." 40 "Rank dis e quilibrium" is aconditionin wh ich an individual/group/na- tionh as a re lative ly h igh e r le ve l of attainme nt of s ome value d conditions th anoth e rs . A group wh ich h as h igh we alth , me dium powe r, and low s tatus , for e x ample , is "dis e quilibrate d" and will s e e k to attainh igh powe r and s tatus as we ll. Wh e th e r it doe s s o viole ntly or not will de pe nd on wh e th e r oth e r me ans of e quilibrationh ave be e n trie d and found wanting, and wh e th e r "th e culture h as s ome practice in viole nt aggre s - s ion." 41 A bas ic s imilarity be twe e nth is argume nt and Aris totle 's s h ould be e vide nt. But Galtungs ays th at e quilibrium is ne ce s s arily s ough t, wh e re as Aris totle s ays th at th e de s ire for cons is te nt e quality, or in- e quality, is afunctionof me n's conce ption of jus tice . And Galtungap- plie s th e principle to all dis e quilibriaamongall h igh ly value d conditions , wh e re as Aris totle pos its re volutionary cons e que nce s only of dis e quilibria be twe e nth e e conomicand political pos itions of a clas s . We may als o note afundame ntal contradictionbe twe e n Galtung's argume nt and Dah r- 39 Se e Franz Sch urmann, "OnRe volutionary Conflict," Journal of Inte rnational Affairs , XXIII (1/1969), 36-53. "Joh an Galtung, "A StructuralT h e ory of Aggre s s ion," Journal of Pe ace Re - s e arch , (2/1964), 95-119; quotationfrom 98. " Ibid., 99. 377 Comparative Politics April 1973 e ndorf's th e ory of clas s conflict. Galtung s ays dis e quilibrium inte ns ifie s aggre s s ion, wh ile Dah re ndorf h ypoth e s ize s th at "th e lowe r th e corre la- tion is be twe e n auth ority pos itions and oth e r as pe cts of s ocioe conomic s tatus , th e le s s inte ns e are clas s conflicts like ly to be ," 42 or, in oth e r words , dis e quilibrium be twe e n a group's powe r pos ition and its pos ition on oth e r dime ns ions s h ould minimize conflict inte ns ity. T h e compe ting h ypoth e s e s s h ould be ame nable to e mpirical te s t. Give n our conce rn with th e s ocial ch ange -viole nce ne x us , Galtung's th e ory s ugge s ts th at we look for s ys te matic s ource s of dis e quilibria. He make s a fe w ge ne ral re marks on th e s ubje ct, s ugge s ting, for e x ample , th at s ocial ch ange is "s tructural" in th e s e ns e th at it te nds to introduce ne w, us ually dis e quilibrate d, rankings . "Or it may dis tribute ne w re - s ource s more e ve nly, wh ich is anoth e r way of s aying th at s ome comple te or ne arly-comple te unde rdogs will ris e on one or more dime ns ion due to mas s e ducation, pros pe rity, unive rs al s uffrage . . . . T h e re s ult is dis - e quilibrium with cons e que nt aggre s s ion until more e quilibrate d combi- nations of rank-s e ts are ach ie ve d...." 43 He als o s pe culate s on th e con- ditions for re volution. One formula, not an e x clus ive one , is to e x pand h igh e r e ducation, make fe w ne w e lite pos itions available , ins titute mas s e ducation, but make no oth e r s ocial ch ange s -all of wh ich in combina- tion cre ate s h arp and wide s pre ad dis e quilibria. T o th e s e conditions h e adds , ad h oc, s e ve ral oth e rs including a J-curve , boom-and-bus t e co- nomic patte rn, plus ide ology and ch aris maticle ade rs h ip.44 For our pur- pos e s Galtung's th e ory h as one dis tinct advantage ove r many oth e rs cons ide re d h e re : it points dire ctly to rath e r s ys te matic, ope rational pro- ce dure s for e x amining th e ch ange -viole nce re lations h ip. In any give n s ocie ty, one migh t e x amine all s ocial ch ange s th at alte re d th e re lative rankingof e ach ide ntifiable group on s ome s pe cifie d s e t of dime ns ions , th e n atte mpt to link th e s e to th e e x te nt and form of s ubs e que nt mani- fe s tations of viole nce . Some obs e rvations I h ope it is cle ar th at th e s e s ummarie s are only a s ample of th e e tiological th e orie s of viole nce and conflict, alth ough mos t of th e th e orie s I re gard as "important" are me ntione d. T h e dominant impre s s ion th e y give may be one of confus ion, for a gre at many diffe re nt kinds and as pe cts of s ocial ch ange are cite d as "caus e s ." Some propos e cate gorie s in wh ich to clas s ify caus e s , s ome afford ros te rs of e x ample s , s ome focus on patte rns or dime ns ions of s ocial ch ange s uch as its rate 4Dah re ndorf, p. 218. 4' Galtung, 112. 4 Ibid., 108-9. 378 T e d Robe rt Gurr and s cope . T h e obje cts of e x planation, th e de pe nde nt variable s , are als o dive rs e . Some th e orie s de al with e x te nt of aggre s s ion or viole nce , s ome with conflict, s ome with "re volution." Se ve ral approach e s to de aling with th is profus ion of th e orie s canbe s ugge s te d. I as s ume th is obje ctive : to make ge ne ral, te s table s tate me nts about th e e ffe cts of s ocial ch ange on viole nce and conflict. One bas ic pre mis e is th at th is can be s t be atte mpte d by be ginning with a ge ne ral conce ptual s ch e me for de s cribing s ocial ch ange , th e n working out its various patte rns and cons e que nce s , rath e r th an taking a particular type or as pe ct of viole nt conflict and working back out a caus al funne l to its as s orte d caus e s . T h e latte r kind of th e ory is a h igh ly us e ful input for th e large r e nte rpris e s ugge s te d h e re , but, as th e re vie w jus t comple te d s ug- ge s ts , it is not afruitful way to de al ge ne rally with s ocial ch ange . One approach migh t be calle d "conce ptual re duction." T h e th e orie s re vie we d h e re s tipulate many caus al as pe cts of ch ange th at re s e mble one anoth e r. T h e s e variable s and re lations h ips could be clas s ifie d according to an analytics ch e me , compris ing, for e x ample , th e dime ns ions of s ocial ch ange lis te d in s e ction I, and th e n combine d and ge ne ralize d into a "s ynth e tic" th e ory of ch ange -conflict-viole nce . Anoth e r, more de ductive approach would re quire us to s tate as s umptions about th e nature and ge ne ral e ffe cts of e ach dime ns ionof s ocial ch ange , th e n to de rive a s e t of more s pe cifich ypoth e s e s , wh ich in turns h ould be s uite d (a) for comparis onwith th e h ypoth e s e s of th e more ad h octh e orie s , and (b) for e mpirical validation. It would be e s pe cially fruitful to e s tablis h de finitive te s ts be twe e nh ypoth e s e s s o de rive d and th os e s tipulate d inth e e tiological th e orie s . T h is s ugge s ts ath ird, more inductive , approach to s ys te matizing th e ch ange -viole nce re lation. Various pairs of contradictory h ypoth e s e s can be found in, or de rive d from, e x tant th e orie s ; th e oppos itionof Galtung's bas ich ypoth e s is to one of Dah re ndorf's propos itions was note d above . De finitive e mpirical te s ts migh t be s e t up for s uch oppos ingh y- poth e s e s , and as e mpirically s upporte d re lations h ips we re winnowe d out, th e y could be s ubs ume d to a s ys te maticinve ntory of ve rifie d ch ange - conflict re lations h ips -one wh ich would e ve ntually le ad to, or take on th e proportions of, a compos ite th e ory. Each of th e s e approach e s to th e ory cons tructionh as its advocate s , and much more could be s aid about th e proble m of h ow to tidy up aninte l- le ctual lands cape litte re d with partial th e orie s . Rath e r th an doings o h e re , I propos e to cons ide r now th e oth e r s ide of th e ch ange -viole nce ne x us : th e orie s about th e e ffe cts of viole nt conflict on s ocial ch ange . T h e re th e lands cape is much diffe re nt, th e s ubje ct of cons ide rable con- je cture and wis h ful as s umption, but not much s ys te maticth e ory. 379 Comparative Politics April 1973 III. Viole nce and Re volution as Source s of Social Ch ange I s ugge s te d th at s ocial ch ange can be analyze d according to th e e x te nt, s cope , patte rn, and rate of ch ange in th e four type s of de te rminants of s ocial action: value s , norms , ins titutions , and facilitie s (s e ction I). A ge ne ral approach to th e "ins trume ntal" linkage be twe e n viole nce and ch ange is to h ypoth e s ize and de s cribe h ow diffe re nt kinds and inte ns itie s of viole nt conflict affe ct th e de te rminants of action, and h ow wide ly, s wiftly, and s o forth th e y do s o. Marx and Dah re ndorf are th e only th e oris ts dis cus s e d th us far wh o do anyth ings ys te matic of th is s ort. Marx quite concre te ly s tate s th at clas s re volution is a function of value ch ange s -clas s cons cious ne s s -wh ich le ads to s pe cifie d ins titutional re arrange me nts . T h os e re arrange me nts in- clude a ne w s e t of productive re lations h ips , a continge nt s e t of ne w political and s ocial re lations h ips , and, at th e final, pos tcapitalis t s tage , a with e ring away of th e s tate . Marx th us attribute s s pe cific e ffe cts to a ve ry s pe cific kind of conflict. Dah re ndorf, by contras t, argue s mos t ab- s tractly th at clas s conflict produce s s tructural ch ange s in th e as s ociations in wh ich it occurs . He attribute s ch aracte ris tics of th e s e s tructural ch ange s to s e ve ral conflict variable s . T h e radicalne s s of s tructural ch ange (i.e ., its cons e que nce s ) is s aid to be a function of conflict inte ns ity; th e s udde n- ne s s of s tructural ch ange is a function of th e e x te nt of viole nce in clas s conflict.45 Some of th e oth e r th e oris ts s urve ye d h e re cons ide r s ome more limite d cons e que nce s of viole nt conflict, in particular its fe e dback e ffe cts on th e variable s s aid to caus e viole nce . One bas ic fe e dback re lations h ip on wh ich a numbe r of th e oris ts agre e is th at viole nt conflict te nds to be come e nde mic, to fe e d on its e lf.46 I h ave s ugge s te d in s ome de tail wh y th is s h ould be th e cas e , and unde r wh at circums tance s . An imme diate e ffe ct is th at th e us e of viole nce by one party to a conflict s trongly dis pos e s th e oth e r party to re taliate in kind; viole nce and counte rviole nce te nd to e s calate until one or both partie s ' capacity for viole nce is e x h aus te d. I bas e th is propos ition on th e pre mis e th at pe ople h ave an inh e re nt dis - pos ition, irre s pe ctive of cultural diffe re nce s , to re s pond viole ntly to vio- le nt attacks .47 A more indire ct way in wh ich viole nt conflict affe cts future viole nce is by cre atingor re inforcinggroup jus tifications for future viole nce . I dis tinguis h be twe e n normative jus tifications , i.e ., th e be lie f th at viole nce is an approve d mode of action; and utilitarianjus tifications , th e be lie f th at viole nce is a us e ful me ans for th e attainme nt of group value s . N ormative s upport for viole nt conflict is propos e d to vary with 45Dah re ndorf, pp. 231-35. 4' Se e Ecks te in, 150-51. 4Gurr, ch ap. 8. 380 T e d Robe rt Gurr th e magnitude of pas t viole nce ; utilitarian s upport varie s with th e pas t s ucce s s of viole nce . Some conflict th e oris ts concur th at "viole nce bre e ds viole nce ," but account for th e re lation s ome wh at diffe re ntly, arguing th at th e occurre nce of viole nce in conflict s ituations te nds to unde rmine th e e ffe ctive ne s s of conflict-re gulatingproce dure s s uch as ne gotiation and me diation. A re late d propos ition is th at inte rgroup viole nce incre as e s intragroup coh e s ive ne s s and h e nce s h arpe ns line s of divis ion be twe e n conflicting groups .48 Various th e oris ts attribute s ome more pos itive kinds of s ocial ch ange to conflict. Conflict among groups is propos e d to s tre ngth e ngroup co- h e s ive ne s s and s e parate ne s s , as note d be fore ; to re duce te ns ion and de viation with in th e group; to clarify group obje ctive s ; and to h e lp e s tablis h group norms .49 T h e s e more or le s s be ne ficial ch ange s are attribute d to conflict ge ne rally, th ough not s pe cifically, nor ne ce s s arily to viole nt conflict. Els e wh e re , I h ave propos e d th at political viole nce (viole nt conflict involving political groups ) te nds to re s olve its e lf if th e dis s ide nt group th e re by obtains re s ource s and opportunitie s by wh ich it can re s olve its dis conte nts .50 T h e propos ition is not th at "winning e nds viole nce ," wh e n "winning" me ans th at one group dis place s anoth e r downward; th is me re ly make s it like ly th at th e dis place d group rath e r th an th e winningone will initiate th e ne x t round of conflict. T h e propo- s ition re s ts rath e r on th e pre mis e th at mos t s ocie tie s h ave unus e d or unde rutilize d s tocks of re s ource s and te ch nique s , wh ich in th e h ands of dis conte nte d groups can be us e d to improve th e ir abs olute if not re lative pos ition in th e dis tributionof value d goods and conditions . Finally, I mus t me ntion th e utopian and mille nnial e x pe ctations th at s ome re volutionary th e oris ts and ph ilos oph e rs h ave h ad about th e e ffe cts of re volutionary conflict. I h ave in mind s uch me n as th e more radical Fre nch ph ilos oph e s , T h omas Paine , Karl Marx , Pie rre -Jos e ph Proudh on, Ge orge s Sore l, Louis Augus te Blanqui, Le on T rots ky, Mao T s e -tung, Frantz Fanon, Erne s to Gue vara, Re gis De bray-th e lis t could e as ily be e x te nde d. T h e s e me n various ly cons ide re d and advocate d re volu- tionary conflict as a ne ce s s ary condition for s ocial ch ange . Almos t with out e x ce ption th e y h ope d for a more e galitarian, jus t, and unop- pre s s ive s ocial orde r, and mos t s aw in re volutionary conflict a ne ce s s ary condition to th at e nd. T h e re is cons ide rable plaus ibility in th e conte n- tion th at viole nt, re volutionary conflict is a ne ce s s ary condition to th e 48 T h is is mos t fully de ve lope d by Le wis Cos e r, T h e Functions of Social Conflict (N e w York, 1956). 49 Se e Mack and Snyde r, 228; Cos e r, pas s im; and Be rnard J. Sie ge l, "De fe ns ive Cultural Adaptation," in Grah am and Gurr, ch ap. 22. s oGurr, pp. 348-57. 381 Comparative Politics April 1973 utopian s ocial obje ctive s th e s e me n h ave s ough t; fe w "top dogs " will- ingly re linquis h powe r or acce pt le ve lling. But th e s e th e oris ts are h ope - le s s ly unre alis ticto th e e x te nt th at th e y re gard re volutionary conflict as a s ufficie nt condition for s ocial trans formation. Marx 's doctrine of h is - torical ine vitability was in h is applications amos t s oph is ticate d mille nnial argume nt. For Blanqui and s ome of th e anarch is ts , th e utopian outcome of re volutionary conflict was a little -e x amine d pre mis e , almos t an act of faith . T h e le as t utopian, and mos t re alis ticof re volutionary advocate s are th os e like T rots ky, Le nin, and Mao T s e -tung, wh o pe rs onally face d th e h ars h re alitie s of atte mpting to carry out re volutionary ch ange . By virtue of th e ir as s umptions about re volutionary be ne fice nce , th e ph ilos oph e rs of th e pre ce dingparagraph are no more fruitful s ource s of ge ne ralizations about h ow re volutionary conflict ch ange s s ocie ty th an are th e more "obje ctive " conte mporary th e oris ts wh o confine th e ms e lve s principally to re volutionary caus ation. With th e e x ce ption of s ome prom- is ing be ginnings in th e work of Dah re ndorf and Le wis Cos e r, we are pre tty much with out th e ore tical guide line s . One partial th e ore tical con- tribution of th is s ort is offe re d in th e ne x t s e ction. IV. Some De te rminants of th e Succe s s of Viole nt Conflict I am conce rne d h e re with one particular que s tion about th e viole nce - ch ange ne x us : Unde r wh at conditions is viole nt conflict like ly to be s uc- ce s s ful? T h e que s tion make s s e ns e , of cours e , only if as ke d from th e vie wpoint of th os e e ngage d in viole nce . It is a "political" kind of que s - tion to as k and atte mpt to ans we r, at le as t more obvious ly political th an a compre h e ns ive atte mpt to s h ow th e e ffe cts of all kinds of viole nce on all as pe cts of s ocial ch ange . My rationale for atte mpting it is th at it doe s offe r apurch as e on th e large r que s tion; th at it is important to s ome s e rious conte mporary is s ue s ; and, frankly, th at it is e as ie r th an th e large r e ffort. Motive s for viole nce in s ocial move me nts T o e valuate th e "s ucce s s " or "failure " of particular viole nt acts re quire s s ome cate gorization of type s of motive s for th os e acts , and, e mpirically, e vide nce on th e motive s manife s t in occurre nce s of viole nt conflicts . T h e cate gorie s s h ould s ub- s ume all us e s of viole nce , at le as t all colle ctive one s , and not be re - s tricte d to th e motive s of one or anoth e r group; s pe cifically, th e y s h ould be applicable to viole nce us e d by private groups in conflict with oth e rs , and by public groups in conflict with one anoth e r and with private groups . Four ge ne ral cate gorie s of motive s are dis cus s e d be low: s e lf- 382 T e d Robe rt Gurr as s e rtion, de fe ns ive , re formis t, and re volutionary. It is my e x te ns ion of a typology us e d by T illy.51 Subordinate cate gorie s could be de vis e d for e ach , but th is will not be done h e re . Se lf-as s e rtion T h e re is an e le me nt of s e lf-as s e rtion in almos t all acts of viole nce : a de s ire to s atis fy ange r, obtain re ve nge , as s e rt pride , cre ate fe ar in oth e rs . T h e motive is appare nt in actions both of rule rs and of rule d. Frus tration-aggre s s ion th e oris ts argue th at aggre s s ion (of wh ich viole nce is one form) is an inh e re ntly s atis fying re s pons e to ange r.52 Fanon s ays th at oppre s s e d pe ople s re dre s s th e ir infe riority th rough vio- le nce . "At th e le ve l of th e individuals , viole nce is a cle ans ing force . It fre e s th e native from h is infe riority comple x and from h is de s pair and in- action; it make s h im fe arle s s and re s tore s h is s e lf-re s pe ct." T h e e ffe cts of viole nce are s aid to h ave e qually pos itive colle ctive e ffe cts : "for th e colonis e d pe ople th is viole nce , be caus e it cons titute s th e ir only work, inve s ts th e ir ch aracte rs with pos itive and cre ative qualitie s . T h e practice of viole nce binds th e m toge th e r as awh ole ...." 53 Fanon's own ps ych ia- tric cas e s tudie s cas t doubt on th e accuracy of th e s e ge ne ralizations , by portraying th e grie vous ps ych ic cos ts of viole nce on its practitione rs ,54 but th e pre cis e accuracy of h is claims is not at is s ue h e re . T h e point is th at many angry and oppre s s e d me n, in Fanon's Alge ria and e ls e wh e re , h ave acte d viole ntly to s atis fy s trongps ych ic ne e ds (not me re ly be caus e Fanon s aid th e y s h ould). Similar e le me nts of s e lf-as s e rtion may be s e e n on th e part of rule rs . Slave owne rs s ome time s us e d e x traordinary viole nce on s lave s for triflingoffe ns e s , a primary motive s e e mingly be ing to de mons trate th e ir abs olute mas te ry. Abs olutis t rule rs in many s ocie tie s h ave us e d s um- mary e x e cutions for s ligh t offe ns e s and e ve n with out offe ns e as an as s e r- tion of mas te ry. T h e h is torical ch ronicle s of Europe an abs olutis m offe r individual ins tance s . Some incre dible e x ample s of ch ronic butch e ry of th is s ort in me die val African kingdoms are pre s e nte d and analyze d by E. V. Walte r.55 De fe ns ive viole nce Pe ople wh o us e viole nce almos t always h ave more th an imme diate motive s of s e lf-as s e rtion. Wh e n viole nce is us e d as part 51 T illy. 52 Re vie we d in Gurr, ch ap. 2. 5 Frantz Fanon, T h e Wre tch e d of th e Earth (N e w York, 1966), p. 73. 54 Ibid., pp. 201-51. 6 E. V. Walte r, T e rror and Re s is tance : A Study of Political Viole nce (London and N e w York, 1969). 383 Comparative Politics April 1973 of an e ffort to maintain a group's pos ition, to pre s e rve th e s tatus quo, I call it de fe ns ive . Mos t viole nce us e d by gove rnme nts is de fe ns ive in th is s e ns e . Almos t all policie s of all gove rnme nts are unpopular with s ome citize ns , and ph ys ical force -viole nce -is us ually re garde d as th e las t re s ort of rule rs face d with noncompliance . Counte rins urge ncy is a fa- miliar type of de fe ns ive viole nce by gove rnme nts , us e d to s uppre s s re v- olutionary and s e ce s s ionis t move me nts . T h rough out h uman h is tory th e mos t common kind of viole nt move - me nt by private groups h as probably be e n de fe ns ive re s is tance to e x - te rnally impos e d ch ange . Ex ample s include re s is tance to fore ign con- que s t, to th e e x pans ion of ce ntral gove rnme nt powe rs at th e e x pe ns e of local autonomy and privile ge s , to th e compe tition and pre s s ure s of ris ing clas s e s , and to e conomic immis e rization impos e d by e mploye rs . In th is country many private groups s uch as vigilante s h ad de fe ns ive motive s for viole nce . T h e vigilante groups , wh ich we re e s tablis h e d to cre ate and maintain public orde r, flouris h e d th rough out th e nine te e nth ce ntury in almos t all parts of th e Unite d State s e x ce pt N e w England. Many frontie r and farme rs ' re be llions -th e Wh is ke y Re be llion, Sh ay's Re be llion, th e Gre e n Corn Re be llion, among oth e rs -we re de fe ns ive in nature . T h e mos t notable e x ample in Ame rican h is tory is th e s e ce s s ion of th e South in 1861 and th e e ns uing Civil War. T h e South in fact fough t to de fe nd its traditions and privile ge s agains t th e e ncroach me nts of fe de ral auth ority.56 Re formis t and re volutionary viole nce T h e re formis t motive for viole nce is ade s ire for limite d ch ange , th e re volutionary motive a de s ire for wide - s pre ad, th orough goingch ange . More pre cis e ly, th e re volutionary motive is to ch ange fundame ntally th e patte rns of auth ority, th at is , to ch ange th e bas ic ins titutions and proce dure s of s ocie ty. Its s atis faction us ually re quire s a s ubs tantial ch ange in th e value s of s ocie ty, a ch ange in th e ope ratingnorms of ins titutional life , and re place me nt of th e e lite s wh o manage ins titutions . T h e re formis t motive is to ch ange wh at e x is ting in- s titutions do, th at is , to ch ange th e ir ope ratingnorms and more s pe cif- ically s ome of th e ir practice s and policie s . Both re formis t and re volu- tionary motive s for viole nce are progre s s ive or "forward looking," as dis - tinct from "backward-looking" de fe ns ive move me nts . T h os e wh o h old s uch motive s and us e viole nce as a me ans for ach ie ving th e m want to cre ate s ome th ingne w, not to re s tore s ome th ingold. An Ame rican com- MFor a ge ne ral inte rpre tation of Ame rican political viole nce as de fe ns ive , s e e Rich ard E. Rube ns te in, Re be ls in Ede n: Mas s Political Viole nce in th e Unite d State s (Bos ton, 1970). 384 T e d Robe rt Gurr paris on s h ould make th is cle ar. T h e Wh is ke y Re be llion in Pe nns ylvania in th e 1790s was de fe ns ive , a farme rs ' move me nt of re s is tance to an in- flux of fe de ral officials and th e impos ition of ne w tax e s . Dorr's Re be llion in Rh ode Is land in 1842 was re formis t; th e de mand was for e x te ns ion of th e s uffrage for worke rs , in th e face of e conomic de pre s s ion. Re formis t and re volutionary motive s are not abs olute ly dis tinct: th e y diffe r in th e de gre e of ch ange s ough t, and th us cons titute two pole s of a continuum. T o ach ie ve re forms it is ofte n ne ce s s ary to make s ome funda- me ntal ins titutional ch ange s . And th os e wh o want re volutionary ch ange s us ually do s o be caus e th e y s e e k mas s ive ch ange s in th e practice s of e x is t- ing ins titutions , not be caus e th e y value ins titutional ch ange as an e nd in its e lf. Social move me nts and motive s for viole nce I am s pe cifically conce rne d with th e us e of viole nce on th e part of s ocial move me nts . I am re fe rring to th e coale s ce nce of a large numbe r of pe ople to take conce rte d e fforts to s olve a s e t of common proble ms .57 A s ocial clas s is not as ocial move - me nt in th is s e ns e , th ough it can give ris e to one . N e ith e r is a gove rn- me nt, th ough it may originally h ave aris e nfrom a s ocial move me nt (e .g., th e N e w England Puritans ), and may fos te r s ocial move me nts (e .g., "private " re s is tance to inte gration in th e Ame rican South during th e 1950s and e arly 1960s , abe tte d by local gove rnme nts ). Mos t s ocial move me nts draw th e ir me mbe rs h ip from particular functional or h ie r- arch ic groups . T h e trade union, fe minis t, and black powe r move me nts are familiar e x ample s . Oth e rs , like th e e cology and "law and orde r" move me nts in th e conte mporary Unite d State s -if th e las t is inde e d a move me nt-bring toge th e r me mbe rs of dis parate groups and clas s e s . Social move me nts can be organize d around an infinity of obje ctive s , and can purs ue th os e obje ctive s us ing a gre at varie ty of tactics . Fe w move me nts will be h omoge ne ous in obje ctive s ; h e nce th rough th e ir me mbe rs th e y may re pre s e nt motive s of s e ve ral or e ve n all four of th e type s me ntione d above . At th e mos t ge ne ral le ve l of analys is , h owe ve r, it s h ould be both pos s ible and us e ful to indicate wh e th e r aparticular move - me nt is pre dominantly de fe ns ive , re formis t, or re volutionary, and, failing th is , to indicate wh at th e approx imate "mix " is . T h e curre nt wome n's move me nt in th e Unite d State s , for e x ample , h as both re formis t and re volutionary obje ctive s , and, as is fre que ntly th e cas e , th os e obje ctive s are re pre s e nte d in s e parate organizations : th e N ational Organizationof 57 For a s imilar de finition, s e e Hans T och , T h e Social Ps ych ology of Social Move me nts (N e w York, 1965), p. 5. 385 Comparative Politics April 1973 Wome n is pre dominantly re formis t, wh ile th e conge rie s of groups calle d Wome n's Libe ration are pre dominantly re volutionary. A move me nt may s pe cify viole nce as a primary or s e condary tactic, or may e x plicitly rule it out. A move me nt may als o be th e obje ct of viole nt oppos ition from oth e r groups e ve n th ough its me mbe rs ne ve r advocate nor take viole nt action th e ms e lve s . I propos e , h owe ve r, th at th e h ypoth e s e s advance d be low conce rning th e de te rminants of th e s uc- ce s s of move me nts are applicable to move me nts involving viole nce irre - s pe ctive of th e circums tance s in wh ich viole nce occurs . Wh e th e r a move - me nt initiate s viole nce , is victimize d by it, or be come s involve d in vio- le nt conflict with anoth e r group, th e ge ne ral h ypoth e s e s about its s ucce s s s h ould te nd to h old. One las t qualification is ne e de d be fore proce e ding to th e h ypoth e s e s . I h ave no ge ne ralizations to offe r about th e "s ucce s s " of th e viole nce of s e lf-as s e rtion. It is an imme diate motive , wh os e s atis faction is found in, or imme diate ly following, th e act. If an individual or group us ing s e lf- as s e rtive viole nce "ge ts away with it" with out dire imme diate cons e - que nce s , th e y s ucce e d. T h e more inte re s ting kinds of que s tions are th e de te rminants of th e s ucce s s or failure of viole nce wh e n us e d in th e pur- s uit of long-run obje ctive s -in oth e r words , viole nce us e d in th e atte mpt to ach ie ve s ubs tantial s ocial ch ange . A ge ne ral propos ition T h e gre ate r th e viole nce us e d agains t pe ople wh o be lie ve th e y are in th e righ t, th e gre ate r th e like lih ood of th e ir re - s is tance , to th e e x te nt of th e ir capacitie s . T h is propos ition applie s to all s ituations in wh ich viole nce is us e d for s ocial purpos e s , and it unde rlie s th e h ypoth e s e s to follow. T h e rationale is s traigh tforward; th e us e or th re at of viole nce agains t pe ople is an attack, a frus tration and, unle s s it is of an abs olute ly ove rwh e lming, life -e ndange ring nature , it s timulate s ange r and a de s ire to re taliate . Eve n wh e n fe ar pre dominate s , wh ich it may do at firs t attack, th e longe r run e ffe ct is th at ange r incre as e s as fe ar s ubs ide s . T h e propos ition migh t be calle d th e "backlas h " principle . It applie s to any kind of targe te d viole nce , wh ate ve r group or ins titution initiate s it, and h as one important ramification th at be ars me ntioning h e re . Pe ople are s ymbol us e rs , and re s pond as much to s ymbols as to dire ct action. T h us , th e ne ws th at viole nce h as be e n us e d agains t "your kind," agains t any group or ins titution with wh ich pe ople ide ntify, te nds to ange r all th os e wh o s o ide ntify, wh e th e r or not th e y are dire ctly affe cte d. T h e s ucce s s of de fe ns ive viole nce Viole nce us e d in e fforts to maintain orde r or oth e rwis e to de fe nd a s tatus quo in th e allocation of value s is 386 T e d Robe rt Gurr like ly to h ave th e inte nde d e ffe cts to th e e x te nt th at th re e conditions h old: (1) th e purpos e s for wh ich viole nce is us e d are ge ne rally acce pte d; (2) viole nce is ge ne rally acce pte d as a le gitimate me ans ; and (3) th os e agains t wh om viole nce is us e d are too fe w and powe rle s s to incre as e th e ir re s is tance . T h is tripartite h ypoth e s is s h ould apply to any colle ctive us e of viole nce by or agains t a s ocial move me nt. It attribute s th e s ucce s s of a viole nt de fe ns ive move me nt le s s to its tactics th an to th e s ocial circums tance s in wh ich it ope rate s . An e x ample of a s ucce s s ful, nongove rnme ntal de fe n- s ive move me nt in Ame rica is provide d by th e Re cons truction South . Be twe e n 1865 and 1878 South e rnwh ite s s ucce e de d in re ve rs ing almos t all th e e ffe cts of th e 14th and 15th Ame ndme nts ; black South e rne rs we re not re e ns lave d, but th e te rror and intimidationinflicte d on th e m and th e ir h andful of wh ite Re publican s upporte rs in th e South was in many ways wors e th an s lave ry its e lf. T h e ch ronicle of mob action, vigilante -s tyle be atings , s h ootings , burnings , and ope n re be llion agains t N orth e rn au- th oritie s and black militia is too long, dis mal, and we ll docume nte d to re pe at h e re .58 T h e s ucce s s of counte rre cons tructionwas as s ure d by th e fact th at all th re e of th e h ypoth e tical conditions we re me t. 1. Almos t all wh ite South e rne rs be lie ve d th at blacks h ad to be re - s ubjugate d, oth e rwis e South e rncivilization was doome d. 2. Viole nce was ge ne rally acce pte d by South e rne rs as a le gitimate me ans to th is e nd. T h e y h ad jus t fough t a much more viole nt war for s imilar purpos e , wh ile ne ws pape r e ditorials and public s tate - me nts by officials and politicians of th e e ra re pe ate dly condone d s uch viole nce . 3. South e rn blacks and th e ir N orth e rn s ympath ize rs we re too we ak to re s is t or re ve rs e th e e ffe cts of viole nce for any le ngth of time . Re publican political s upporte rs , black and wh ite , we re a minority almos t e ve rywh e re in th e South , and, e ve n wh e re th e y we re not, th e y ofte n lacke d me ans to re s is t. T h e mainte nance of Fe de ral law and orde r de pe nde d s ubs tantially on black militiaunits commande d by wh ite office rs , at le as t until cons e rvative wh ite s tate gove rn- me nts we re re e s tablis h e d. T h e s e units we re ne ith e r we ll traine d nor we ll e quippe d; th e ir office rs we re s ome time s ope nly murde re d, th e ir me n attacke d by arme d bands of wh ite s . Fe de ral policy was ge ne rally not to s e nd troop re inforce me nts . T h e mos t te lling fac- tor, pe rh aps , was th e lack of wh ite coope ration in e fforts to appre - h e nd th e wh ite s wh o carrie d out campaigns of conce rte d te rror; 68 Of many s tudie s s e e , for e x ample , Joh n Hope Franklin, Re cons truction(Ch i- cago, 1961). 387 Comparative Politics April 1973 wh ite South e rne rs could us e viole nce with almos t comple te im- punity agains t we akly de fe nde d blacks . T h e cas e is illus trative , not de finitive . A numbe r of additional cas e s of both s ucce s s ful and uns ucce s s ful de fe ns ive move me nts would be re - quire d to de te rmine its ge ne ral applicability. T h e s ucce s s of re formis t viole nce I as s ume h e re th at mos t re formis t move me nts h ave no more th an mode rate powe r to attainth e ir own e nds ; th e y are ordinarily in th e pos ition of trying to pe rs uade , or force , powe r- ful "oth e rs " to make ch ange s on th e ir be h alf. In th e s e circums tance s , th e followingh ypoth e s is s ugge s ts th e conditions of th e ir s ucce s s . Viole nce us e d in e fforts to pe rs uade or coe rce "powe rful oth e rs " to ch ange is like ly to h ave th e inte nde d e ffe cts to th e e x te nt th at: (1) th e "oth e rs " acce pt or at le as t do not re je ct outrigh t th e de s ira- bility of th e ch ange s ough t; (2) th e "oth e rs " h ave th e capacity to make th e ch ange ; and (3) th e re action of th e "oth e rs " agains t viole nce doe s not ove rride th e ir willingne s s to ch ange . T h e fe minis t move me nt in England from th e 1890s to 1918 provide s an e x ample of s ucce s s ful re form, in wh ich all th re e conditions appe ar to h ave be e n me t. It h ad as its s ole obje ctive th e righ t of wome n to vote and oth e rwis e to participate in politics . T h e principle divis ion in th e move - me nt was be twe e n th e tactically mode rate Wome n's Fre e dom Le ague and th e militant s uffrage tte s of th e Wome n's Social and Political Union. T h e tactics of th e latte r include d an e igh t-ye ar s e que nce of incre as ingly dis ruptive mas s de mons trations , confrontations with officials , and face - s lapping brawls with police , de ve loping into a conce rte d te rroris t cam- paign. At its h e igh t, be twe e n 1911 and 1914, th e s uffrage tte s carrie d out gre at numbe rs of s tonings , fire bombings of public buildings , railway s tations , and officials ' h ome s , window bre akingof th e type now calle d "tras h ing," and s ome wh ippingand h atch e t attacks on officials . By 1918 all wome n ove r th irty we re e nfranch is e d and allowe d to s tand for Parlia- me nt. T h e conditions of s ucce s s de ve lope d s e riatum: 1. T h e purs uit of woman's s uffrage be gan in th e 1860s , firs t agains t th e dis inte re s t, th e n th e de ris ion, of politicians . Some fifty ye ars of agitation on th e is s ue gradually incre as e d male s upport, as e vi- de nce d in incre as ingly favorable Hous e of Commons vote s on th e is s ue . T h ough no Gove rnme nt would de clare its e lf in favor of woman's s uffrage until 1917, by th at ye ar public and political opinion cle arly acce pte d th e principle . 2. T h e Englis h gove rnme nt unque s tionably h ad th e auth ority to grant 388 T e d Robe rt Gurr woman's s uffrage ; Commons h ad in fact take n away th e righ t to vote h e ld by a minis cule numbe r of wome n in 1832. 3. T h e public re action agains t s uffrage tte viole nce s e e ms s ignificantly to h ave de laye d th e granting of th e re form. Alth ough Commons h ad vote d favorably on s uffrage in 1909, th e Gove rnme nt allowe d th e Hous e of Lords to kill it. T h e s ubs e que nt four ye ars of viole nt prote s t we re accompanie d by incre as e d re luctance of s ucce s s ive gove rnme nts to de al with th e is s ue . T h e outbre ak of World War I provide d a re s pite ; with in a month of its ons e t all militant actions we re s us pe nde d, and many of th e s uffrage tte s took up aux iliary military s e rvice . Give n th e lull in viole nt prote s t, and th e war its e lf, in 1917 th e prime minis te r de clare d h ims e lf for s uffrage and an e le ctoral re form was promptly pas s e d by th e ove rwh e lming margin of 364 to 23.59 T h e s e th re e conditions of re formis t s ucce s s can be us e d to e valuate both conte mporary and h is torical move me nts , as a comparis on of two conte mporary Ame rican move me nts de mons trate s : T h e civil righ ts for black Ame ricans , and th e antiwar, antimilitaris m move me nts . N e ith e r of th e s e move me nts h as be e n pre dominantly viole nt, but e ach h as provide d th e occas ion for cons ide rable viole nce of prote s t and re s is tance . T h e que s tion is h ow th is viole nce affe cts th e ir s ucce s s . 1. De s irability of ch ange . Cons ide ringonly th e fe de ral le ve l of gov- e rnme nt, a s ubs tantial diffe re nce s e e ms appare ntin th e de s irability of th e two type s of ch ange . Fe de ral officials and th e majority of congre s s me n h ave s ubs tantially acce pte d th e de s irability of e ffe c- tive civil righ ts for blacks ; fe w fe de ral officials and only a minority of congre s s me n h ave adopte d a s ubs tantially antimilitaris ticvie w. As one cons e que nce , th e re h as be e n a gre at de al of rule making and adminis trative activity on be h alf of civil righ ts , but only a gradual and limite d s h ift in military policy. 2. Capacity to ch ange . On th is variable , th e balance is re ve rs e d. T h e fe de ral gove rnme nt h as only limite d capacity to de al with th e fundame ntal proble m of civil righ ts , wh ich is compris e d of th e in- graine d racis t attitude s and practice s of awide s pe ctrum of Ame r- ican ins titutions . T h e s lowne s s of e ffe ctive civil righ ts progre s s is a function of wide s pre ad public re s is tance more th an of gove rn- me ntal dis inte re s t. On th e antimilitaris m is s ue , h owe ve r, th e fe de ral gove rnme nt cle arly h as th e powe r to ch ange th e Unite d State s ' military policie s ove rs e as as we ll as to re tre nch ove rall military 69 T h is s ummary is drawnlarge ly from Roge r Fulford, Vote s for Wome n (Lon- don, 1957). 389 Comparative Politics April 1973 e x pe nditure s and activitie s ; it is th e de s irability of ch ange th at h as not be e n wide ly acce pte d, e x ce pt to s ome de gre e on Vie tnam. 3. Re action agains t viole nce . T h e re is e ve ry e vide nce th at gh e tto riots in th e mid-1960s and th e dramatize d te rroris m and re volutionary pos turing of s ome black militants h ad s ubs tantially unde rmine d th e partial public s upport gaine d by th e civil righ ts move me nt in th e late 1950s and e arly 1960s . T h is backlas h e ffe ct was re fle cte d in opinion poll data, in th e N ix on Adminis tration's policie s of "be nign ne gle ct," in wide s pre ad political oppos ition to programs th ough t to be ne fit blacks primarily, and in a varie ty of more s pe - cific and localize d ways . Similarly, th e us e of viole nce by th e s uicidal le ft of th e antiwar move me nt re s ulte d in a wide ly-fe lt backlas h at colle ge s tude nts and youth ge ne rally. In s ummary, th e civil righ ts and antiwar move me nts h ave both be e n no more th an partly s ucce s s ful, be caus e in e ach ins tance th e y h ave con- fronte d a "mix " of favorable and unfavorable conditions for th e ir s ucce s s . T h e s ucce s s of re volutionary viole nce Re volutionary obje ctive s can be purs ue d with out th e us e of re volutionary viole nce . In th e Unite d State s today, many groups h ave be e n cons cious ly working for re volutionary ch ange s by pe ace ful me ans , to a de gre e th at I th ink is with out h is torical paralle l. India duringGandh i's s atyagrah acampaigns s e e ms to offe r th e only approx imate analogy. T h e que s tion h e re , h owe ve r, is wh at th e re - quire me nts are for s ucce s s ful re volutionary ch ange wh e n pros e cute d by traditional, viole nt me th ods of re volutionary conflict.60 I s ugge s t th e following conditions : Viole nce us e d to ach ie ve re volutionary ch ange is like ly to le ad to th e de s ire d e ffe cts to th e e x te nt th at th re e conditions h old: (1) th e re volutionarie s ove rcome th e re s is tance of th e ir oppone nts ; (2) th e re volutionarie s re main committe d to th e ir plans for ch ange afte r gainingpowe r; and (3) th e re volutionarie s h ave th e re s ource s and s ocial e ngine e ringcapacitie s to bringabout s ocial ch ange . T h e firs t of th e s e conditions s tipulate s , in e ffe ct, th at th e re volution- arie s h ave to "win"-i.e ., a political re volution mus t occur be fore th e y can e ve n atte mpt to bringabout ch ange . Conditions for winningcan als o be s tipulate d, de rive d from th e nume rous e tiological th e orie s re vie we d above . T h e re volutionarie s re quire s ubs tantial popular, th ough not ne c- e s s arily majoritarian, s upport. In addition, th e y mus t h ave a de gre e of 00 T h e e s s e ntial conditions of nonviole nt re volutionary s ucce s s are probably s imi- lar to th os e s pe cifie d e arlie r for re formis t move me nts as s ociate d with viole nce . 390 T e d Robe rt Gurr organizational s upport from th e ir followe rs at le as t e quivale nt to th at of th e ir oppone nts . Finally, and mos t imme diate ly, th e y re quire military me ans s ufficie nt to s tand off th e incumbe nts . Re volutionary commitme nt, s trate gics kill, favorable te rrain, e x te rnal s upport, and gue rrilla or te rroris tictactics can compe ns ate to a cons id- e rable de gre e for wh at th e re volutionarie s may lack in numbe rs and e quipme nt. Almos t always , h owe ve r, re volutionary victory re s ults not from a re gime 's de fe at on th e battle fie ld pe r s e , but from th e e ros ion of th e military's willingne s s to continue to s upport th e re gime . T h e Com- munis t Ch ine s e de fe at of th e N ationalis t re gime in 1949 is one of th e rare e x ce ptions to th is principle . Fide l Cas tro's victory in Cubain 1959 was a good de al more typical. T h e Batis ta re gime los t much of its popular s upport be caus e of its incre as ingly te rroris tic and incons is te nt practice of viole nce agains t urban and rural pe ople s . T h e ins titutional balance s h ifte d toward th e Cas tro force s , not be caus e of re lative ly limite d organizational de ve lopme nt in th e rural are as th e y controlle d, but be - caus e of th e dis affe ction of labor and middle -clas s organizations with th e re gime . T h e military balance was turne d not by any gre at re be l vic- tory-th e ir numbe r unde r arms in De ce mbe r 1958 was no more th an 1,000-but by th e pre cipitous de cline in army morale . Had Batis ta not gone into e x ile on De ce mbe r 31, le avingHavanaope n to Cas tro, h e would h ave be e n ove rth rownve ry s h ortly by one of th e coup groups al- re ady active in th e military.61 One oth e r point s h ould be made about "winning." T h e political triumph of th e re volutionarie s is not ne ce s s arily th e e nd of re volutionary conflict. Counte rre volutions are always apos s ibility. In Hungary in 1919, a five -month -old Socialis t re volutionary re gime was ove rth rown in a counte rre volutionth at h ad fore ign military s upport. Counte rre volutionin Rus s ia, 1919-21, was uns ucce s s ful de s pite fore ign inte rve ntion, but it de vas tate d th e country. T h e abs e nce of organize d, viole nt atte mpts to re - ve rs e a re volution doe s not ne ce s s arily me an acquie s ce nce to th e re volu- tionary caus e , e ith e r. Mas s ive e migration, noncoope ration, and clande s - tine s abotage by th os e wh o s upporte d th e old e lite can be a de vas tating burde n on "s ucce s s ful" re volutionarie s . One of th e Cuban re volution's gre ate s t liabilitie s afte r 1959 was th e wh ole s ale de fe ction of pe ople with th e profe s s ional, te ch nical, and adminis trative s kills ne e de d for e conomic and political trans formation. Re volutionary le ade rs may ch ange th e ir obje ctive s , too. T h e y may prove more inte re s te d in e x ploiting th e "re volution" for th e ir pe rs onal be ne fit th an in pros e cutingch ange . Or th e y may e ngage th e ir re s ource s 01 Se e , for e x ample , Boris Golde nbe rg, T h e CubanRe volutionand LatinAme rica (N e w York, 1965). 391 Comparative Politics April 1973 and e ne rgie s in fore ign conflict, e ith e r out of ne ce s s ity or an e x ce s s of re volutionary ze al. Some , like th e Cubanle ade rs in 1959, and pe rh aps to a le s s e r de gre e th e Ch ine s e le ade rs in 1949, may ch oos e re volutionary policie s th at alie nate many of th os e wh o initially s upporte d th e m. Re vo- lutionary le ade rs cannot e ntire ly avoid bre aking faith with th e ir followe rs , s imply be caus e of th e dive rs ity and mutual incons is te ncy of th e ir re volu- tionary as pirations . In Cuba, h owe ve r, th e re volutionary "be trayal" was both s ubs tantial and unus ual, s ubs tantial be caus e it h urt mos t of th e re volution's middle and s kille d working-clas s s upporte rs , unus ual be caus e th e le ade rs s ough t much gre ate r ch ange th an originally propos e d, rath e r th an th e re volutionary norm of circums cribe d ch ange . Probably th e mos t common s ource of failure among th os e wh o win political re volutions is th e ir limite d capacity to manage ch ange . At th e point of victory almos t e ve ry re volutionary re gime confronts e ith e r an ine rt s ocial mas s of be lie fs , practice s , and ins titutions th at obdurate ly re s is t ch ange , or at be s t an inch oate s ocial s ys te m rive n by dis s e ns us and paralyze d by ins titutional collaps e . (T h e latte r is be tte r be caus e , to e x - te nd a familiar re volutionary analogy, it is e as ie r to make an ome le tte if th e e ggs are alre acid broke n.) In ne ith e r circums tance are re volutionary le ade rs like ly to h ave te s te d me th ods for bringing about th e kinds of s ocial ch ange th e y s e e k. T h e h is tory of e ve ry atte mpt at manage d re vo- lutionary trans formationwith wh ich I am familiar is compris e d of amix - ture of imme diate gains , bootle s s e x pe rime ntation, and policy dis as te rs . Only occas ionally is th e re a gradual, long-run re dire ction of s ocial, pro- ductive , and political tre nds s ome wh at in th e original re volutionary dire ction.62 T h e accuracy of th e s e fe w h ypoth e s e s about th e utility of viole nce as ale ve r for s ocial ch ange can be e mpirically as s e s s e d. I s ugge s t th e m only as a firs t cut at th e proble m. T h e y s h ould be s ubs tantially e laborate d and made more pre cis e . Eve n at be s t, th e y can make only a s mall con- tribution to th e large r proble m of tracingth e wh ole ne twork of conne c- tions be twe e n viole nt conflict and s ubs e que nt s ocial ch ange . 62 Much th e s ame can be s aid of mos t large -s cale nonre volutionary e fforts at s ocial e ngine e ring; vide th e re s ults of mos t of th e Joh ns onAdminis tration's "Gre at Socie ty" programs . T h e proble m is not s o much th at th e re forme rs or re volu- tionarie s lack ze al, or e ve n re s ource s ; lack of ope rationalknowle dge and re s is tance by th os e wh o are to be e ngine e re d are us ually more vitiating. 392