a theory - semantic scholar...a theory of revolutionraymond tanter department of political science,...

A theory of revolution

RAYMOND TANTER

Department of Political Science, University of Michigan

and

MANUS MIDLARSKY

Department of Political Science, University of Colorado

Overview

The focus of this study is an empiricalexamination of some causes of revolution.lWhile prior studies have delineated dimen-sions of conflict behavior within nations

(Rummel, 1963, 1966, 1967; Tanter, 1965,1966) in order to discover the variables

most representative of internal conflict, a

purpose of the present inquiry is to discovertheir theoretical significance. A typology ofrevolution is presented and two possiblecauses of revolution-changes in economicdevelopment and level of education-are

examined regarding the extent of their asso-ciation with certain characteristics of revo-lution. Regional differences are found whichimply a fundamental distinction between atleast two of the categories in the typology.

Finally, hypotheses are offered to accountfor some of the findings of this study.

A Categorization of RevolutionsPrior to a discussion of theoretical signifi-

cance, it may be useful to inquire into the

meaning of revolution itself. This exercisein definition appears necessary primarily be-cause prior usage of the term has beensomewhat ambiguous. To the Hegelian, therevolutionary idea is equated with irresist-ible change-a manifestation of the worldspirit in an unceasing quest for its own ful-fillment. Similarly, the Marxist, althoughopposed to Hegelian idealism, sees revolu-tion as a product of irresistible historical

forces, which culminate in a struggle be-tween the bourgeoisie and the proletariat.Hannah Arendt (1965, pp. 34-40), on theother hand, interprets the revolutionary ex-perience as a kind of restoration, wherebythe insurgents attempt to restore libertiesand privileges which were lost as the resultof the government’s temporary lapse into

despotism. Indeed, aspects of the AmericanRevolution as well as some recent anti-

colonial revolutions may be amenable to

Arendt’s interpretation. The insurgentsmight view the colonial elite as strangerswho have usurped the freedoms which,

1 This study is a part of a series on revolu-tions and their implications for international re-lations and especially some consequences forinternational business communities. Acknowl-

edgments are due to Harold Guetzkow of theInternational Relations Program at NorthwesternUniversity for his suggestions during formativestages of this inquiry. Thanks are due to theFord Foundation for a fellowship given to thejunior author and to the National Science Foun-dation for its assistance to the senior author.

Helpful comments were made by Richard A.Brody, James C. Davies, and Robert C. North.

CONFLICT RESOLUTION VOLUME XI NUMBER 3

265

TABLE 1

CHARACTERISTICS OF FOUR TYPES OF REVOLUTION

mythically or otherwise, once belonged tothe people now in revolt.

Tocqueville (1955, p. 8), on the otherhand, has employed a more empirical ap-proach to the problem of revolution, and hasdefined it as an overthrow of the legally con-stituted elite, which initiated a period of in-tense social, political, and economic change.Crane Brinton (1952, pp. 3-4) has continuedthis empirical thrust by differentiating be-tween the coup d6tat, as a simple replace-ment of one elite by another, and majorrevolutions such as the French or Russian,which were accompanied by social, political,and economic changes. Similarly, with re-spect to Latin America, George Blanksten(1962, p. 72) suggests that we should dis-tinguish between the coup d6tat and revo-lutions such as the Mexican experience,which eventually had profound conse-quences for the structure of that society.The distinction between two forms of

revolution may provide a basis for the de-velopment of further classifications. For

example, Harold Lasswell and Abraham

Kaplan (1950, p. 252) present a further re-finement in the classification of revolution

by the introduction of a three-category ty-pology in which they differentiate betweenpalace revolutions, political revolutions, andsocial revolutions. Edwin Lieuwen (1960,

pp. 22-24) constructs a similar classifica-

tion, but instead of the palace revolution hediscusses &dquo;caudillismo&dquo; (predatory milita-

rism), which is a common form of the coupd6tat in Latin America. These three formsof revolution appear to reflect an increasingdegree of change initiated by the successfulinsurgents, and may be placed on a rank-order of increasing political or social change.James Rosenau, in fact, constructs such aclassification of revolution.

Personnel wars are defined by Rosenau asthose which are fought over the occupancyof existing roles in the structure of politicalauthority (Rosenau, 1964, pp. 63-64). Anexample of this type is the palace revolutionor Latin American caudillismo. A second

category is what Rosenau calls the authoritywars, or those in which the insurgents com-

pete not only for the occupancy of roles inthe political structure, but for their arrange-ment as well. Struggles to replace dictator-ships with democracies would be classifiedas authority wars. The final classification isthat of the structural wars, in which the goalof the insurgents is the introduction of socialand economic changes in the society. Warsinvolving Communist factions would fall

under this heading. In addition, Rosenaunotes that structural wars contain elements

of both personnel and authority wars.

266

Changes in the occupancy of governmentroles, as well as the arrangement of these

roles, would automatically take place if theinsurgents were successful in a structuralwar. Similarly, the authority war is person-nel-oriented, because the arrangement of

political roles would seldom, if ever, bealtered without a change in the occupancyof these roles. In Rosenau’s ranking of thethree types of internal war, personnel warswould occupy the lowest rank with regardto the degree of societal change; the au-

thority wars would occupy an intermediaterank, and the structural wars would receivethe highest rank.

Samuel Huntington (1962, pp. 23-24)has suggested a classification of revolutionin which four categories are enumerated:the internal war, the revolutionary coup, thereform coup, and the palace revolution.

Huntington’s use of the concept &dquo;internalwar&dquo; differs from the meaning attributed tothat concept in earlier systematic studies.

For that reason the term mass revolution2will be substituted for internal war as used

by Huntington. The terms mass revolutionand palace revolution correspond respec-

tively to Rosenau’s structural and personnelwars, while the revolutionary and reformcoups both may be placed under the head-ing of the authority wars. Kemal Ataturk’srevolution in Turkey, for example, illustrateswhat Huntington might call a revolutionarycoup, whereas the 1955 coup in Argentinamight be classified as a reform coup. The

major difference between the two forms isin the degree of change initiated in thestructure of the political authority. The

&dquo;Young Turks&dquo; implemented a complete re-vision of the political authority which led toa truncation of the Ottoman Empire and theestablishment of a republic. The revolt

against Peron, on the other hand, was an at-tempt at reform, in that Peron’s mismanage-ment of the economy and the dissatisfactionof major political forces, such as the RomanCatholic Church, led to a revolt against whathad become an oppressive political execu-tive.

The existence of several types of revolu-tion suggests that we might be able to isolatedifferent characteristics of revolution. Karl

Deutsch (1964, pp. 102-104) proposes thatthe degree of mass participation in a revo-lution, as well as its duration, may be essen-tial to an adequate description of the revolu-tionary experience. A third characteristic

may be the number killed as a result of therevolution. Given a high degree of commit-ment by the insurgents and the incumbents,the number of persons killed both duringand after the revolution may be a measure

of intensity. This measure will be discussedmore fully at a later point. Finally, the in-tentions of the insurgents may be critical tothe form of the revolution as well as to its

eventual outcome. If the successful insur-

gents are ideologically committed to certaingoals, then they may initiate changes in thesocietal structure to effect the realization of

these goals. If, on the other hand, the in-

surgents have no particular ideological orien-tation, then they might intend to replace theincumbents in the structure of political au-thority without recourse to changes in thesocietal structure.

Table 1 contains a rank-order of the cate-

gories of revolution based on their positionwith respect to mass participation, duration,domestic violence, and the intentions of the

2 One may also distinguish between varioustypes of mass revolution. Lawrence Stone

(1966, pp. 162-63), for example, presents a

typology which differentiates among the Jac-querie, the Anarchistic Rebellion, the JacobinCommunist Revolution and the Militaristic MassInsurrection. However, a purpose of this studyis to compare the larger categories of revolution(e.g., mass revolution vs. palace revolution)rather than to engage in a detailed examinationof subcategories.


267

TABLE 2FOUR TYPES OF REVOLUTIONa

a Unsuccessful revolutions are indicated below the the dotted line.b The date in parentheses refers to the year in which the insurgents were either defeated or seized the roles in

the structure of political authority.c The Algerian revolution of 1962 refers to the anti-colonial revolt of the Algerian independence movement,

whereas the French revolution of 1958 applies to the fall of the French Fourth Republic.

insurgents. We may now suggest a defini-tion of revolution which would be designedto include the categories in Table 1, as wellas to allow for the possibility of a continuum.A revolution may be said to exist when a

group of insurgents illegally and/or force-fully challenges the governmental elite forthe occupancy of roles in the structure of

political authority. A successful revolutionoccurs when, as a result of a challenge tothe governmental elite, insurgents are even-tually able to occupy principal roles withinthe structure of the political authority. This

is not to say that once the insurgents haveoccupied these roles, the structure of politi-cal authority will remain unchanged. As

suggested above, changes in the personnelof the governmental elite often are the pre-condition for meaningful changes in the

political and social structure. If the insur-

gents intend major political and social

changes, they must first occupy these roles

within the political structure. This defini-

tion, then, sets a lower bound or minimumcriterion for the existence of revolution.

Table 2 provides examples of the fourtypes of revolution. This table is not meant

to be exhaustive, but serves to provide a list-ing of those revolutions judged by the

authors to be illustrative of the particularcategories.3 The dotted line in Column 1

separates successful mass revolutions fromthose which were unsuccessful. In addition,instances of both successful and unsuccess-ful anti-colonial revolutions are included to

indicate the possible similarity of this typeto the general mass revolution. Having pro-vided such illustrative material, a next stepis to provide the operational referents forthe examination of causes of revolution.

3 Other writers such as Gross (1958), Hunt-ington (1962), and Silvert (1961) are in sub-stantial agreement with this classification.

268

Empirical ReferentsTwo of the four measures-duration and

domestic violence-may be interpreted as

defining characteristics of revolution, andtaken together may be seen as a measure ofrevolutionary intensity.4 Domestic violence,for example, is present in the majority ofrevolutions. If the population is indifferentto changes in the personnel of the govern-mental elite, then a palace revolution maybe initiated by insurgents. They may en-counter opposition only from the incumbentelite. Because of the lack of large-scale,organized conflict behavior, the number ofdeaths would be at a minimum, and theduration of such a revolution would be short.

If, however, the insurgents proclaim theirintention to initiate changes in the social

structure, then large segments of the popu-lation may be alienated from the insurgentcause. The number of deaths, and possiblythe duration of the revolution, may increaseproportionally to the degree of societal

change envisaged by the insurgents. Simi-

larly, the duration of a revolution has beencited as a possible measure of revolutionary

TABLE 3DEATHS AND DURATION FOR SUCCESSFUL

REVOLUTIONS, 1955-60

a The data sources for Column 1 are the raw data forRummel (1963), Tanter (1965), and the New YorkTimes. Two countries which experienced successfulrevolutions, Sudan ( 1958 ) and Laos (1960), were

excluded because of missing data on deaths.b Per million population, 1950-62 (Russett et al.,

1964, pp. 97-100).c Data relevant to the duration of revolutions were

drawn from the New York Times.d Both the 1957 and 1958 Thai revolts were initiated

by the same individual, the commander-in-chief of thearmy, Sarit Thanarat. In addition, the 1958 coup tookplace with the express agreement of the &dquo;ousted&dquo; Kitti-kachom government. For these reasons, this latter in-stance is omitted from the subsequent analysis although,for the sake of completeness, it is included in this listing.

intensity (Rosenau, 1964, pp. 76-77). ManyLatin American palace revolutions are ofshort duration, and the number killed is

relatively small. Indeed, bloodless coupsoften occur in which no one is killed, but anew set of elites assumes the major roles inthe political authority structure. Revolu-tions such as the French and the Russian,however, are of longer duration and a

greater number of persons are killed.The two measures-duration and deaths

from domestic violence-are presented in

Table 3. The first column lists all success-ful revolutions which occurred between

4 The degree of mass participation as well asthe intentions of the insurgents are less satisfac-tory as empirical referents. Mass participationmight range from popular dissatisfaction with-out direct involvement, on the one hand, to

overt complicity and active support of the in-surgents, on the other. Thus, measurement ofmass participation presents a great deal of dif-ficulty unless the researcher is present at thesite of the revolution and is able to measure

public opinion at the time the revolution takesplace. Similarly, the intentions of the insurgents,although important as a descriptive variable,nevertheless present important problems of mea-surement. Intentions and ideological directionsare subject to changes during the process ofrevolution. Fidel Castro, for example, pro-claimed his support for the goals of the Cubanmiddle class during the recent Cuban revolu-tion, yet once in power he allied himself withthe Communists.


269

1955 and 1960. The criterion for the in-clusion of a revolution was that the revoltended sometime within that period.5 Col-umn 2 lists the number killed as a result ofdomestic group violence, per million popu-lation. These figures are for the period1950-62. An alternative would be to con-struct an index on the basis of the numberkilled during the period of the revolutionitself. However, the number killed duringa revolutionary coup lasting a few days maynot be indicative of the characteristics ofthat revolution. For example, thirty personswere killed during the Iraqi revolution of1958. Two of these were the king, Faisal,and his prime minister, Nuri. The intensityof feeling against the government was mani-fested when the populace dragged the king’sbody through the streets of Baghdad. The

death of thirty people would seem to be aninadequate reflection of this degree of in-tensity. However, if we examine the num-ber of deaths prior to the revolution and

a f ter its occurrence, we find that Iraq ex-perienced the second highest number of

deaths per million population of any revolu-tion during the period 1955-60. Similarly,Venezuela, which experienced the overthrowof the Jiminez dictatorship, also had one ofthe highest numbers of deaths as a result ofdomestic violence. The inclusion of the

number of deaths prior to and after a revo-lution might be indicative of the frustration

built up after a long period of repression.A revolution generally occurs after a periodof instability, and it is suggested in this

paper that the form of a revolution is de-

pendent on the degree of political instabilitywhich exists prior to its occurrence.

Two Correlates of Revolution

Having provided operational referentsfor the concept of revolution, we may nowturn to an examination of possible correlates.The problem of identifying the precondi-tions of revolution has been of long concernto political theorists. For example, Plato

(1951, p. 54) proposed that differences ineconomic interests led to factionalism in

politics and contributed to the instability ofthe city-state. Poverty, according to Plato,produces revolution, meanness, and villainy,while riches produce luxury, idleness, andvillainy. In substantial agreement with this

position, Aristotle (1962, p. 59) also pro-

posed that poverty may be a cause of politi-cal revolution. Tocqueville (1955, pp. 22-23), however, dissented from this emphasison poverty. He suggested that the Frenchpeasant prior to 1789 enjoyed a considerablyhigher degree of economic independencethan did the remainder of the Europeanpeasantry. Because of this independenceand security, those aspects of feudalism stillremaining in French society, such as thecorvée (a form of periodic forced labor),appeared all the more odious and con-

temptible.Indeed, long-term economic depressions

were not present prior to the outbreaks ofeither the American or Russian revolutions.After an exhaustive study of four majorrevolutions, Crane Brinton (1952, p. 264)concluded that one uniformity in the oc-

currence of these revolutions was that thesocieties under investigation were all on the

upgrade economically before the outbreakof revolution. Some statistics relevant to the

5 The duration of a revolution is bounded bytwo points in time: (1) when active hostilitiesfirst broke out against the regime in power, and(2) when the insurgents occupied the roles inthe structure of political authority. A period ofunrest prior to the victory of the insurgents wasincluded in the determination of the values inColumn 3. Riots and clashes between the policeand populace are examples of these disturbances.In addition, if two or more changes in the per-sonnel of the governmental elite took placewithin a three-month period of time, the periodbetween changes was included in the duration.

270

French economy prior to 1789 may be in-structive. In the 6lectioit of Melun from

1783-85, the amount of uncultivated landwas reduced from 14,500 to 10,000 arpents.Rouen in 1787 produced cotton cloth worth50,000 livres, which was double the produc-tion of a generation before. In the dozen

years since the death of Louis XIV, Frenchtrade had increased nearly 100,000,000livres (Brinton, 1952, p. 31).The English Revolution of the 17th cen-

tury and the American Revolution were pre-ceded by similar economic experiences. Ac-cording to Brinton, early Stuart Englandwas notably prosperous, as were the Ameri-can colonies prior to 1775 (1952, p. 32).Russia also was making significant economicprogress prior to the outbreak of World WarI. Table 4 illustrates this growth in theRussian economy.The theories of Plato and Aristotle appear

to contradict those of Tocqueville and

Brinton. The first set asserts that povertyleads to revolution while the second set

claims that revolutions are preceded by asignificant increase in economic develop-ment. James C. Davies (1962, pp. 6-7) sug-gests that a partial synthesis of these ap-proaches may provide a more comprehensiveexplanation than either of the two taken

alone; major revolutions may be precededby steady long-term increases in economicdevelopment, followed by a sharp reversaljust before the outbreak of the revolution.In more general terms, economic develop-ment might be viewed as a single aspect ofa society’s achievement. That is, political,economic, and cultural development may to-gether comprise a single concept which, forour purposes, may be called achievement.

A second concept closely associated with

achievement is termed aspirations. The re-sults of political, economic, or cultural at-

tainment ordinarily are visible to the gen-eral populace. Newly granted freedoms, as

TABLE 4RUSSIAN EXPORTS AND IMPORTS PER CAPITA:

1895-1913a

a Mazour, 1962, p. 323.

well as roads, factories, museums, and ahigher standard of living, which generallyaccompany a high rate of achievement, alsomay tend to increase aspirations. It is positedthat the rates of change of the achievementsand aspirations are correlated. If achieve-ment is increasing at a given rate, then thepopulace would most likely aspire to the

acquisition of social commodities at thesame rate as they had been previously ac-quired.

If we were to plot the increase in aspira-tions over time, the slope (rate of change)of this variable would approximate the rateof change of achievement. The similaritiesin the slope of the two plots is illustrated inFigure 1. Now if, instead of increasing, therate of achievement were to decrease, wemay posit a third concept, expectations. Therate of change of expectations would ap-proximate that of the decrease in the rate ofachievement. Rates of expectations may beaffected more by immediate reality than arerates of aspirations. Expectations representa change in outlook caused primarily by animmediate decrease in the production of

social commodities. Aspirations are more inthe nature of a hope and an optimism gen-erated by long-term past performance.6 Thedistance between the two concepts (revolu-

6 The distinction between aspirations and ex-pectations may be analogous to the Feierabends’(1966, pp. 256-57) use of the terms "wantformation" and "want satisfaction."


271

FIG. 1. Relationship between rate of achievement, aspirations, and expectations. (Adapted andmodified from Davies, 1962, p. 6.)

tionary gap) may be seen as a measure ofthe potentiality for the occurrence of a vio-lent revolution. The larger the revolutionarygap, the longer and the more violent therevolution may be. This hypothesis is repre-sented in Figure 1.As evidence for the low rate of expecta-

tions prior to major revolutions, Davies citesthe poor agricultural harvests of 1788-89 inFrance which followed the economic ad-vances of the prior decade. The Englisheconomy had suffered a reversal prior to

1688, while there was a tightening up ofbusiness conditions in America in 1774-75.

Conditions in Tsarist Russia in 1917 providea striking example of a breakdown in theeconomic machinery immediately prior to

the outbreak of a revolution. Industrial pro-duction had declined rapidly. Coal mininghad fallen to a level that threatened the

nation’s transportation. The peasants had

grown sick of the war, and instead of trans-

porting much-needed grain to Russia’s urban

centers, they were hoarding food in the ex-pectation of future disasters. The economywas in a state of virtual collapse when the

revolution broke out (Mazour, 1962, pp.

553-55). Apparently, these four major revo-lutions provide evidence for Davies’ hy-pothesis. But whether this hypothesis holdstrue in all such instances is a question to beinvestigated by means of a systematic analy-sis across a larger universe of revolutions.A second correlate of revolution might be

the level of educational attainment prior tothe outbreak of the disturbance. SeymourM. Lipset (1959, pp. 73-80) has found thatthe occurrence of revolutions is associated

negatively with the level of educational at-tainment. Lipset categorized nations ac-

cording to their stability and degree ofdemocratic or totalitarian control and foundthat Latin American dictatorships, whichare notoriously amenable to the occurrenceof palace revolutions, are also societies whichrank lowest on his scale of educational

attainment. Evidence supporting Lipset’sfinding is provided in the Dirnensions of fNations (Rummel, Sawyer, Tanter, and

Guetzkow, forthcoming, 1967). These

writers find that three educational variablesare correlated negatively with the occur-

272

rence of revolution. One of the three-theratio of number enrolled in primary schoolsto the total population aged 5-14-gave thehighest negative correlation (r = -.84, N =

66). It appears from the data of the mid-

1950s that revolutions will occur more fre-

quently in societies with the lowest levelsof educational attainment.

Having stated the principal concepts andtheir measures, it may be useful to summa-rize the definitions prior to a presentation ofthe results:

( 1 ) A revolution is operationally definedby domestic violence and duration.

(2) Achievement and aspirations are de-fined by the rate of change of GNP/CAPover time (gross national product divided bypopulation).

(3) Expectations are defined by the dropor reversal in the rate of change of GNP/CAP.

(4) Educational level is defined by theprimary school ratio-number enrolled in

primary school divided by total populationaged 5-14.

Given these definitions, the following hy-potheses are suggested:

(1) The higher the rate of increase ofGNP/CAP preceding the revolution and thesharper the reversal immediately prior to

the revolution, the greater the duration andviolence of the revolution.

(2) The lower the level of educational at-tainment prior to the revolution, the greaterthe duration and violence of the revolution.

These hypotheses were tested acrossseventeen cases of successful revolutions for

1955-60.

Data and Results

Data for the years 1955-60 are presentedin Table 5. Unsuccessful and colonial revo-

lutions are excluded. The variation over

TABLE 5

DOMESTIC VIOLENCE, TIME RATE OF CHANGE OFGNP/CAP, AND PRIMARY ENROLLMENT RATIO

a Same variable as in Table 3, Column 2.b The sources for this data column are the UN Sta-

tistical Yearbooks, 1955-1962. The data are presentedin time-series form with base year 1953 = 100.

C Primary enrollment divided by total populationaged 5-14. These ratios were drawn from the WorldSurvey of Education II, UNESCO, 1958, Table 15, pp.58-60. The majority of these values are for the years1950-1954, with Brazil, Cuba, and Iraq for 1950-1953, Haiti and Jordan for 1952-1954, and Turkey for1950-1952.

d The rate of change of GNP/CAP for Cuba can beseen in Figure 2.

. The values for Iraq and El Salvador were notincluded in the standardized tabular form in the datasources, and were therefore calculated separately bymeans of net national product and population data.

f The values for Syria and Thailand are based onfour and six data points respectively, due to limiteddata availability.

g The dashes in the columns indicate that the datawere not available.

time in GNP/CAP is presented in Column3. The figures in this column are regressioncoefficients which were calculated for the

seven-year period preceding each revolu-tion.7 Here, the regression coefficient is a

7 The period of seven years prior to the out-break of the revolution is an arbitrary choice,determined chiefly by the availability of com-parable data. It is possible that the investiga-tion of longer periods might yield different re-sults.


273

quantitative representation of a given vari-able’s rate of change over time ( slope ) .8Gross national product per capita wasplotted against time for each nation, andthe slopes were approximately linear. The

values of these coefficients are comparable,since the same base year (1953 = 100) waschosen for all cases.9 Column 4 lists the pri-mary school ratios.The product-moment correlation coeffi-

cients are presented in Table 6. These co-efficients were calculated first for the total

number of cases in Column 1, and thenseparately for the Asian and Latin Americancases. The number of cases on which each

coefficient is based is written in parenthesesbelow the value of that coefficient. Before

calculating the coefficients, curves were

TABLE 6

CORRELATION COEFFICIENTS BETWEEN CHANGESIN WEALTH, LEVEL OF EDUCATION, AND

DOMESTIC VIOLENCEa

a Domestic violence was logarithmically transformedfor the first two values in the second row. No trans-formations were necessary for the calculation of the

remaining values.The dependent variable in the calculation of these

coefficients was domestic violence.

plotted and, where necessary, domestic vio-lence was logarithmically transformed to

compensate for the effect of an outlier

( Cuba ) .Since these calculations are based on an

entire population of revolutions in the givenperiod, it must be noted that the sample ofcases is not random. Although the conven-tional significance levels are thus not ap-

plicable, they are utilized to provide somecriterion for the tenability of hypotheses.All correlation coefficients are considered

significant at the .05 level of confidence;i.e., five times in 100 the coefficient couldbe due to chance. A second interpretationof the correlation coefficient is the percent-age of variance explained-the square of

the coefficient. This mode of presentation isalso employed in the following discussion.The values in the first two data columns

of Table 6 are statistically nonsignificant.The figures in Column 3, however, are some-what more indicative of the presence of re-

lationships. As Hayward R. Alker (Russettet al., 1964, pp. 322-23) has suggested, cul-tural differences between regions may ac-count for an apparent lack of significant re-lationships. When the total number of casesis considered, domestic violence in Asian

countries in general, and Middle Eastern

8 The data in Table 5 are presented with cer-tain reservations. The regression coefficients inthe second data column were calculated fromUN statistical yearbooks, and in that source it isnoted that gross national product data are in-

dicative only of general trends. Indeed, differentyearbooks reported varying values of GNP/CAPfor the same country and the same year. In

order to partially resolve these inconsistencies,the regression coefficients were calculated usinga single time series (from one yearbook) for eachcoefficient. In addition, the criterion for choos-ing a time series was that it was reported in theyearbook published one year subsequent to theoccurrence of the revolution (e.g., the 1959Yearbook for Pakistan [1958]). For those casesin which the time series was incomplete, suchas Syria or Thailand, a subsequent yearbookwas chosen which contained the most completetime series.

9 The term "base year" refers to the basis ofthe UN Statistical Yearbook’s calculation of thetime-series figures. In all of the series, 1953 is

set equal to 100 and the values for the otheryears are expressed relative to that year. Thisterm does not refer to the starting point for thecalculation of the regression coefficients. Thechoice of a starting year for the calculation ofthese coefficients is based solely on the fact thatthe chosen year is seven years prior to the out-break of the revolution.

274

countries in particular, may have an entirelydifferent political meaning from the samedegree of violence in Latin American coun-tries. This division of the countries in Table5 into Latin American, Asian, and MiddleEastern is a first approximation to a controlfor cultural differences. Only France, of thecountries included in Table 5, did not fallinto any one of these categories.When Asian revolutions are considered,

we find that the degree of association be-tween rate of change of GNP/CAP anddomestic violence is significant at p < .01 1

and accounts for approximately 88 percentof the variance. The proportion of studentsenrolled in primary schools is correlated

negatively with domestic violence, and thecoefficient is nonsignificant at p < .10. Ap-proximately 58 percent of the variance is

accounted for by this relationship. For theMiddle Eastern countries, the correlationbetween the rate of change of GNP/CAPand domestic violence is significant at p <.05 and accounts for approximately 92 per-cent of the variance. The rate of change ofGNP/CAP prior to the revolution is almostperfectly associated with the level of domes-tic violence. The correlation between the

primary enrollment ratio and domestic vio-lence is nonsignificant at p < .10 and ac-counts for 85 percent of the variance.The coefficients for the Asian and Middle

Eastern revolutions appear to provide someof the explanation of variance for allthe cases. However, because of the smallnumber of cases, these results should be ac-

cepted with caution. Before we can assumethat 92 percent of the variation of domesticviolence in the Middle East is accounted for

by the rate of change of economic develop-ment, we clearly must have further verifica-tion efforts based on a longer time-period.

Deaths from domestic violence is one

measure of revolution; another is the num-ber of days of the revolution. A set of cor-

relation coefficients was calculated for dura-

tion, the rate of change of GNP/CAP, andthe level of the primary enrollment ratio;the results were not significant at p < .05.

Perhaps one reason for this lack of a rela-tionship is the extreme difficulty in mea-

suring duration. As Brinton (1952, pp. 238-50) notes in his study of the Russian Revolu-tion, the notion of duration may be bettersuited to the making of qualitative judg-ments rather than quantitative distinctions.For example, there is a large degree of un-certainty as to the duration of the RussianRevolution. On the one hand, the period inthe early 1920s may be viewed as a &dquo;Ther-midor,&dquo; hence a denouement of the revolu-tion. Later, on the other hand, the SovietUnion under Stalin experienced a reign ofterror which rivalled the intensity of the

French Revolution at its height. The Sovietsmay have undergone a &dquo;permanent revolu-tion&dquo; throughout the Stalin period, and onlyafter Khrushchev’s condemnation of Stalin

did the revolution subside into a period ofnormalcy. Thus there are at least two possi-ble measures of the duration of the Russian

Revolution. In his attempt to define the

duration of the English Revolution, Brintonencountered similar difficulties. Because of

the close-knit series of events prior to theoutbreak of the Civil War in 1642, the dura-tion of this revolution is essentially a matterof subjective determination (Brinton, 1952,p. 75).

Brinton’s analysis and the results of thisstudy suggest that the duration variable is

difficult to measure. Moreover, when com-

paring durations we often commit the fallacyof employing different units of analysis inthe same investigation. By recording dura-tions we are, in effect, ignoring the eventswhich occurred during these periods. A

three-day revolutionary coup which is foughtwith bitter intensity should not be equatedwith a three-day palace revolution which is


275

FIG. 2. Plot of gross national product per capita (GNP/CAP) versus time for Cuba, 1947-53. (Inthe construction of this figure, the inception of the Cuban revolution was taken to be Castro’s first

armed landing in 1953.)

bloodless and leads only to a change in thepersonnel of the governmental elite. A moreadequate measure might be the number ofevents of a similar type compressed into agiven time-interval. Rather than treatingduration as a continuous variable, however,we may more profitably view it as discrete

at a given threshold. Below a certain value

of this variable, the duration of a revolutionmay be irrelevant to the explanation of

either the antecedents or the consequencesof its occurrence. Above that value, the

duration of a revolution may be a meaning-ful characteristic.

The longer the revolution persists, the

greater the polarization of the society aroundthe warring factions. Traditional associa-

tions are broken. Factions which at thestart of the conflict might have been able tosettle their differences by negotiation nowfind that, with the passage of time, enmitieshave increased to the extent that only un-conditional surrender by one of the factionscan terminate the revolution. Reference to

Table 3 reveals that the Cuban revolution,with a duration of six years, is at the same

time the only instance of mass revolution inthe chosen period. The duration of the

276

Cuban revolution is more than ten times

greater than any of the other values in Col-umn 3, and the &dquo;threshold&dquo; may be some-where between one and five years. Future

attempts at the determination of this value

may be fruitful for a better explanation ofthe revolutionary process.

A Mass Revolution: The Case of CubaWe may recall that Davies’ hypothesis

suggests that the higher the long-term rateof achievement and aspirations precedingthe revolution, and the sharper the reversal(expectations) immediately prior to the revo-lution, the greater the intensity as measuredby duration and domestic violence. Rate ofachievement is defined as rate of change ofGNP/CAP, whereas the reversal is equatedwith expectations. The rate of change of

GNP/CAP for Cuba was not included in

Table 5. Instead, Figure 2 plots GNP/CAPagainst time for the seven-year period pre-ceding Castro’s first armed incursion againstBatista in 1953.

If we compare Figure 1 (p. 271) with

Figure 2 we see a marked similarity. Ahigh rate of achievement is followed by areversal which, according to the theory, cor-responds to the expectation of the society.In addition, the Cuban revolution, with

2,900 deaths per million and a duration ofsix years, falls into the mass revolution cate-

gory-the only instance of such a revolutionin the period 1955-60. Although Cuba is asingle case, it compares favorably withBrinton’s analysis, in that the four countriesin his study experienced economic condi-tions similar to that of Cuba prior to theonset of revolution.The Cuban revolution is the only instance

in the period under consideration where asevere reversal actually occurred immedi-ately prior to the revolution. We thus finda confirmation of the existence of a &dquo;revolu-

tionary gap&dquo; in the Cuban example-our

one example of a mass revolution-but areleft without an adequate test for the otherinstances of revolution. How do we mea-sure expectations in cases of revolutionarycoups, reform coups, and palace revolutions?

Differences in Expectations: TheGini Index

An operational definition of expectationsmight be the degree of inequality in incomeor land distribution.10 If the rate of achieve-ment is high and unequally distributed, largesectors of the population might be arousedby this appearance of new wealth in a

limited segment of the population. Aspira-tions, which are based on long-term achieve-ments, may tend to be higher, whereas ex-pectations-more a function of immediate

reality-might be lower because of the pre-vailing economic inequalities. Where the

&dquo;revolutionary gap&dquo; between aspirations andexpectations is largely due to a high level ofinequality in land distribution, we would ex-pect an increase in the probability of revolu-tion.

The Gini index is a measure of inequality.It measures the area between the Lorenz

curve of any distribution and the line of per-fect equality. For high inequality, the areabetween the Lorenz curve and the line of

perfect equality would be correspondinglylarge; hence, the Gini index would also havea large value. For lower degrees of in-

equality, the reverse is true.Table 7 presents the Gini indices for all

nations included in the data source (Russettet al., 1964 p., 239ft). Table 7 (a) lists the

values for nations which experienced suc-cessful revolutions in the period 1955-60,whereas Table 7 ( b) presents the values for

10 It should be noted that this treatment of"expectations" is cross-sectional in design, as

opposed to the longitudinal nature of Davies’theory.


277

those nations which did not have successfulrevolutions. The indices are for land, whosedistribution is probably correlated with in-come distribution. A one-tailed t-test of the

significance of difference was performedand a t equal to 2.641 was calculated. With48 degrees of freedom, this value is signifi-cant at p < .05. Thus degree of land in-equality might be interpreted as an opera-tional form of expectations; at any rate,revolutions occurred in those societies witha higher degree of land inequality. Russett

(1964, p. 272) also reports a relationshipbetween the Gini index of land inequalityand domestic violence of r = .44 across 44

cases.

Conclusions

The analysis of inequalities of land distri-bution and the Cuban case study provideempirical support for the hypothesized rela-tionship between the revolutionary gap andthe domestic violence characteristic of revo-lutions. However, while the correlational

analysis provides support for the hypothesisin the case of Asian revolutions, it was not

substantiated for Latin American revolutionsin this period.The results suggest two alternatives: (1)

If we assume that the level of domestic vio-lence is indeed an indicator of the type ofrevolution in Latin America, then we areforced to conclude that in Latin Americathere is virtually no relationship betweenthe rate of change of GNP/CAP and theform of revolution. This alternative leads toa reliance on an explanation based on re-gional differences. (2) If we assume thatdomestic violence bears little relationship tothe palace revolution in Latin America (mostof the Latin American revolutions in Table5 are of the palace variety), then the rela-tionship between the rate of change of

GNP/CAP and the other forms of revolu-tion is not invalidated. Those revolutions

TABLE 7GINI INDICES OF LAND INEQUALITY’

a From Russett et al., 1964, pp. 239-40.

occurring in Asia are either reform coupsor revolutionary coups; thus we may suggestthat the palace revolution is little affected

by domestic group violence, whereas theother forms of revolution may indeed be soaffected. The rate of change of GNP/CAPand level of education may then be theoreti-

cally significant for the study of revolutionsother than the palace variety. Now we mayask why the palace revolution should be

relatively unaffected by domestic violence.George Blanksten (1958, p. 143) has sug-

gested that one reason for the repeated&dquo;palace revolutions&dquo; in Latin America is thatthe class-oriented social structure providesthe upper class with a virtual monopolyof participation in the political authority.Indians are almost entirely excluded from

278

participation in politics, and the mestizos(descendants of Spanish-Indian marriages)are also inactive in politics but emerge attimes as followers of a particular politicalfaction. Class distinctions in Latin America

then act as a relatively impermeable barrierbetween the elite (the socioeconomic classfrom which holders of high political officeare drawn) and the mass (all those who areexcluded from this highest socioeconomiccategory). Because of this barrier, distur-

bances in the society may seldom affect the&dquo;normal&dquo; operations of government in mostLatin American nations. Thus palace revo-lutions may frequently take place uncon-nected with violent changes within the so-ciety. We find that the rate of change ofGNP/CAP bears little or no relationship todomestic violence in Latin America, but-at the same time-domestic violence bearslittle relationship to the palace revolution,which until now has been the most frequentform of revolution in this region.The Asian countries included in this study

may not experience the same degree of classrigidity as do the Latin American nations.Barriers between the mass and the elite

might not be so rigid or impermeable.Therefore, economic changes which affectdomestic violence also have an effect on the

form of the revolution. A similar argumentcan be developed for the effect of educationon the form of revolution. As a generalproposition we suggest that the more im-permeable the barriers between the massand the elite, the less the effect of domesticviolence on the form of revolution. The con-verse may be equally valid. The more per-meable the barriers between the mass and

the elite, the greater the effect of mass activ-ity on the form of revolution.The social structure of France prior to

1789 provides an example of these relation-ships. According to Tocqueville (1955, pp.97-108) the seventeenth and eighteenth

centuries witnessed a hardening of classlines between the nobility, the small but

growing bourgeoisie, and the peasantry. Be-

fore 1780, class lines had become so rigidthat Tocqueville referred to the nobility asa caste within the general class structure ofFrench society. We may also note thatFrance during this period was free of revo-lutionary disturbances. At the same time,the English were thought of as politicallyunstable, whereas the French were consid-ered to be fortunate in having a governmentwhich exemplified political stability.

In the decade prior to 1789, however,there was a significant decrease in the rigid-ity of barriers between classes. Local parle-ments were once again assigned a politicalfunction which they had virtually relin-

quished centuries before. The meetings ofthe parlements provided a social setting forthis increasing fluidity across class lines.

The case of the French Revolution may

provide a comparison within a single coun-try : during one period, barriers between theelite and the mass prevented social distur-bances from disrupting the &dquo;normal&dquo; opera-tions of government; later, according to

Tocqueville, this barrier became increasinglypermeable and mass revolution occurred.

Recent theories of mass society have also

emphasized the importance of the relation-ship between the elite and the mass. LeBon(1947, pp. 214 f f ) and Ortega y Gassett(1940, pp. 97 f f ) stress that for the govern-mental elite to perform its proper functionof judicious decision-making, it should be

protected from the effects of mass persua-sion, which frequently manifests itself in

violent forms. On the other hand, HannahArendt’s view (1958, pp. 315-24) is that

the mass, if unprotected from the elite, mayundergo a process of &dquo;atomization&dquo; in whichindividuals are dissociated from each other

and become easy prey for the totalitarian

designs of the elite.


279

William Kornhauser (1959, pp. 39 f f )suggests that if organized group activity in apluralistic society can act as an intermediarybetween the elite and the mass, then neitherwill be accessible to undue penetration bythe other. The intermediary group associa-tions can then be viewed as a second formof the social barrier between the elite andthe mass.

Both notions, that of class distinctionsand that of the pluralistic group process,may be amenable to future research as to

their effect on the incidence and form ofrevolution. With the possible exception ofFrance, none of the countries in Table 5

have strongly developed pluralistic societies.However, the state of their pluralistic de-velopment may be related to the form of

revolution.

Finally, what do these results and our in-terpretation say for the scheme of classifica-tion of revolutions presented initially? Thatis, we suggested that revolutions may fallon a rank-order: mass revolution, revolution-ary coup, reform coup, and palace revolu-tion. In Asian societies, such an orderingwas suggested by the relationship betweenchanges in wealth, in educational level, andin domestic violence. Latin American so-

cieties, with a preponderance of palace revo-lutions, did not exhibit these relationships.The palace revolution appears, then, to bea breed apart from the others.Our definition was predicated on the

palace revolution as a lower bound abovewhich lies a rank-order or continuum ofmore intense changes in both governmentand society. Since the data appear to ex-clude the palace revolution from the con-cept of a continuum, the definition mighthave to be revised so that the &dquo;lower bound&dquo;is raised and the palace revolution is ex-

cluded from the definition.

This would appear reasonable if we exam-

ine the insurgents’ orientations. A continuum

of revolution is essentially one of commit-ment to change. Revolutions are not piecesof historical machinery, as the Hegelians orMarxists may contend, but are deeply asso-ciated with the ideological orientations ofthe participants. Indeed, this study suggeststhat domestic violence may vary with the

degree of change intended by the insurgents.Therefore, the continuum we speak of is

essentially a continuum of human desires forchange, which-translated into a specificform of political action-make a revolution.The insurgents in a palace revolution, how-ever, intend only to occupy roles in the

existing authority structure. Changes in thesocial system or in the structure of politicalauthority are not included in their program.Thus, if we are measuring a continuum ofpolitical and social change, the palace revo-lution may not contribute to these measure-

ments.

REFERENCES

ARENDT, HANNAH. The Origins of Totalitarian-ism. Cleveland: Meridian, 1958.

—. On Revolution. New York: Viking,1963.

ARISTOTLE. Politics. (Ernest Barker, ed. and

translator.) New York: Oxford UniversityPress, 1962.

BLANKSTEN, GEORGE. "Revolutions." In H. E.

DAVIS (ed.), Government and Politics in

Latin America. New York: Ronald Press,

1958, 119-46.—. "Latin American Revolutions." In

The 1962 Carolina Symposium: Today’s Revo-lutions. Chapel Hill: University of North

Carolina, 1962, 71-79.BRINTON, CRANE. The Anatomy of Revolu-

tion. New York: Vintage, 1952.DAVIES, JAMES C. "Toward a Theory of Revo-

lution," The American Sociological Review,27 (Feb. 1962), 5-13.

DEUTSCH, KARL W. "External Involvement inInternal Wars." In H. ECKSTEIN (ed.), Inter-nal War. Glencoe: Free Press, 1964, 100-10.

FEIERABAND, IVO K., and ROSALIND L. FEIERA-BAND. "Aggressive Behavior Within Polities,1948-62: A Cross-National Study," Journal

280

of Conflict Resolution, 10, 3 (Sept. 1966),249-69.

GROSS, FELIX. The Seizure of Political Power.New York: Philosophical Library, 1958.

HUNTINGTON, SAMUEL. "Patterns of Violencein World Politics." In S. HUNTINGTON (ed.),Changing Patterns of Military Politics. NewYork: Free Press, 1962, 17-50.

KORNHAUSER, WILLIAM. The Politics of MassSociety. Glencoe: Free Press, 1959.

LASSWELL, HAROLD, and ABRAHAM KAPLAN.Power and Society. New Haven: Yale Uni-versity Press, 1950.

LEBON, GUSTAVE. The Crowd: A Study of thePopular Mind. London: T. Fisher Unwin,1896.

LIEUWEN, EDWIN. Arms and Politics in LatinAmerica. New York: Praeger, 1960.

LIPSET, SEYMOUR M. "Some Social Requisitesof Democracy," American Political Science

Review, 53 (March 1959 ), 69-105.MAZOUR, ANATOLE G. Russia: Tsarist and

Communist. New York: Van Nostrand, 1962.ORTEGA Y GASSET, JOSE. The Revolt of the

Masses. New York: W. W. Norton, 1940.PLATO. The Republic. ( A. D. Lindsay, transla-

tor.) New York: Dutton, 1951.ROSENAU, JAMES N. "Internal War as an Inter-

national Event." In J. N. ROSENAU (ed.),International Aspects of Civil Strife. Prince-

ton : Princeton University Press, 1964, 45-91.

RUMMEL, RUDOLPH J. "Dimensions of ConflictBehavior Within and Between Nations," Gen-eral Systems Yearbook, 8 (1963), 1-50.

—. "Dimensions of Conflict BehaviorWithin Nations, 1946-59," Journal of Con-flict Resolution, 10, 1 (March 1966), 65-73.

—, JACK SAWYER, RAYMOND TANTER, andHAROLD GUETZKOW. Dimensions of Nations,forthcoming, 1967.

RUSSETT, BRUCE M., HAYWARD R. ALKER, JR.,KARL DEUTSCH, and HAROLD LASSWELL.World Handbook of Political and Social Indi-cators. New Haven: Yale University Press,1964.

SILVERT, KALMAN H. Reaction and Revolutionin Latin America. New Orleans: Hauser

Press, 1961.STONE, LAWRENCE. "Theories of Revolution,"World Politics, 18 (Jan. 1966), 159-76.

TANTER, RAYMOND. "Dimensions of ConflictBehavior Within and Between Nations, 1958-

60," Journal of Conflict Resolution, 10, 1

(March 1966 ), 41-64.—. "Dimensions of Conflict Behavior

Within Nations, 1955-60: Turmoil and In-ternal War," Papers, Peace Research Society,3 (1965), 159-83.

TOCQUEVILLE, ALEXIS DE. The Old Regimeand the French Revolution. New York:

Doubleday, 1955.


a theory - semantic scholar...a theory of revolutionraymond tanter department of political science,...

Documents