A Factionalism Model For CCP Politics

A Factionalism Model for CCP Politics
Author(s): Andrew J. Nathan

Source: The China Quarterly, No. 53 (Jan. - Mar., 1973), pp. 34-66
Published by: Cambridge University Press on behalf of the School of Oriental and African
Studies
Stable URL: http://www.jstor.org/stable/652506
Accessed: 07-03-2015 12:31 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
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.
Cambridge University Press and School of Oriental and African Studies are collaborating with JSTOR to digitize, preserve
and extend access to The China Quarterly.
http://www.jstor.org
This content downloaded from 128.235.251.160 on Sat, 07 Mar 2015 12:31:42 UTC
All use subject to JSTOR Terms and Conditions
A Factionalism Model for CCP Politics
Andrew J. Nathan
Until the Cultural Revolution, the predominant western view of con-

temporary Chinese elite conflict was that it consisted of "discussion"
(t'ao-lun) within a basically consensual Politburo among shifting
"opinion groups" with no "organized force" behind them.' The purges
and accusations which began in 1965 and apparently still continue,
have shaken this interpretation, and a number of scholars have advanced
new analyses - sometimes explicit, sometimes implicit, sometimes of
general application, sometimes applied only to a particular time span
or segment of the political system. Of these new views, perhaps the
most systematic - and at the same time the one which represents the
least change from the pre-Cultural Revolution "opinion group" model -
is the "policy making under Mao" interpretation, which sees conflict
as essentially a bureaucratic decision-making process dominated by
Mao.2 A similar but less explicit "Mao-in-command" model sees Mao
as "testing" his colleagues' political loyalty and sometimes decisively
beating back efforts to challenge his pre-eminence.3 Deviating further
* I am grateful to Norman Fainstein, Frederic J. Fleron, Jr., Ellen Frost, Mark
Kesselman, Donald W. Klein, Steven I. Levine, Sharon G. Nathan, Michel C.
Oksenberg,Robert H. Silin, Tang Tsou and Ezra F. Vogel for valuable comments,
and to participants in various seminars and colloquia at which the ideas were
tried out. An earlier version was prepared during a research associateship at the
Center for Chinese Studies, University of Michigan, and was presented at the
1971 annual meeting of the American Political Science Association.
1. Franz Schurmann, Ideology and Organization in Communist China (Ber-
keley: University of California Press, 1966), pp. 55-7. Opinion groups twice
solidified into "factions " (" organized opinion groups "), Schurmann notes, in
the Kao-Jao and P'eng Teh-huai affairs. But these were quickly rooted out.
Schurmann's conception seems to include the possibility that certain leaders
will take consistent positions in these debates, ending up sometimes in the majority
and sometimes in the minority. We may thus regard as variants of his model
those interpretationswhich classify leaders as "radicals " or "moderates," " dog-
matists " or "pragmatists,"etc., and view the Chinese political process as a debate
among them.
2. Michel C. Oksenberg," Policy making under Mao, 1949-1968: an overview,"
in John M. H. Lindbeck, ed., China: Management of a Revolutionary Society
(Seattle: University of Washington Press, 1971), pp. 79-115.
3. See Philip Bridgham's three articles on the course of the Cultural Revolu-
tion: "Mao's 'Cultural Revolution': origin and development," The China
Quarterly (CQ) 29 (January-March 1971), pp. 1-35; "Mao's Cultural Revolu-
tion in 1967: the struggle to seize power," CQ 34 (April-June 1968), pp. 6-37;
A Factionalism Model for CCP Politics 35
from the "unorganized opinion group" model are interpretations which

explain political alignments and policy advocacy in terms of leaders'
attachments to various bureaucratic "interest groups" - whether these
be Field Armies, "commissars" vs. "commanders," "the Party" vs.
"the gun," "the legal specialists" vs. "the new cadres," or whatever.4
Other models and variants could be listed.
The purpose of this paper is to delineate and argue for the applic-
ability of a "factionalism" model of Chinese Communist Party (CCP)
elite politics.5 The reader's indulgence is asked as I turn the everyday
term "faction" to a special purpose. Model-building of the sort
attempted here requires as a first step the isolation of a rigorously
defined type of political structure - in this case, a political structure
based upon a particular type of leader-follower relationship. I have
appropriated the term "faction" for the special type of structure
described below partly for lack of a better term, and partly because
I think that the definition I am giving the term falls somewhere near
its core meaning, especially as used in the anthropological literature.
Several other caveats must also be entered. First, the model is not
intended to explain "everything" about CCP elite politics. Like the
other models mentioned above, it seeks to clarify how the Chinese
leaders organize themselves to carry out conflict, how they mobilize
resources (and what resources) for the struggle, under what formal or
informal rules the conflict is carried out, what sanctions are visited
upon the losers, and so forth. Like the other models, it is consistent
with various interpretations about the ultimate motives for conflict,
and "Mao's Cultural Revolution: the struggle to consolidate power," CQ 41

(January-March 1970), pp. 1-25. See also his " Factionalism in the central com-
mittee," in John Wilson Lewis, ed., Party Leadership and Revolutionary Power
in China (Cambridge: Cambridge University Press, 1970), pp. 203-35. The term
" Mao-in-command " is my own.
4. On Field Armies, see William Whitson, "The Field Army in Chinese Com-
munist military politics," CQ 37 (January-March 1969), pp. 1-30. On commissars
versus commanders, see William Whitson, "The military: their role in the policy
process," in Frank N. Trager and William Henderson, eds., Communist China,
1949-1969: A Twenty Year Appraisal (N.Y.: The New York University Press,
1970), pp. 95-122; and Ellis Joffe, Party and Army: Professionalism and Political
Control in the Chinese Officer Corps, 1949-1964 (Cambridge, Mass.: Harvard
University, East Asian Research Center, 1965), and "The Chinese Army under
Lin Piao: prelude to political intervention," in Lindbeck, ed., China, pp. 343-74.
On the Party and the gun, see, among others, Ralph L. Powell, "The Party,
the government and the gun," Asian Survey (AS) X:6 (June 1970), pp. 441-71.
On legal specialists versus new cadres, see Victor H. Li, "The evolution and
development of the Chinese legal system," in Lindbeck, ed., China, pp. 221-255.
For further discussion of the " interest group" and " Mao-in-command " models,
see my " Models of conflict in Chinese politics," Problems of Communism, Vol.
XXI, No. 3 (May-June 1972), pp. 80-3. The "policy making under Mao " and
"interest group" models could nicely be combined, but so far as I am aware
no one has yet done so in print.
5. That is, politics in the upper level of the CCP from 1921 to 1949, and at
the " central level " (see below) of the CPR since 1949.
36 The China Quarterly
which at various times may have encompassed disagreements about the

nature of man," the state of the world,7 the role of the Party 8 or the
means to economic development,9 personal concern about the fate of the
revolution,1?or personality clashes."1Furthermore, the model deals only
with what might broadly be called organizational constraints on political
behaviour, and not with the other sets of constraints - e.g., ideological
and cultural - which provide additional "rules of the game." Third, it
has to be accepted at the outset that the available data on CCP elite
conflict will not be adequate to accept or reject the model decisively -
if they were, we could develop a presumably more suitable model
directly from the data.12 Indeed, the utility of the model is precisely
that, if it passes an initial test of fitting existing data, it suggests ways
of understanding and interpreting the situation that the data are too
slim to suggest by themselves, and thus fills gaps in our knowledge,
on a tentative and speculative basis, through a species of reasoning
through analogy. In the following pages I will, first, present the model
as an abstract proposition; then, try to show that the available data
on politics at the central level are consistent, as far as they go, with
the model; next, suggest how Chinese Communist politics, especially
since the late 1950s, might be interpreted in a way consistent with the
model; and finally, discuss some further implications and problems.
6. See, for example, Richard M. Pfeffer, "The pursuit of purity," Problems of

Communism,Vol. XVIII, No. 6 (November-December 1969), pp. 12-25; Frederick
C. Teiwes, "A review article: The evolution of leadership purges in Communist
China," CQ 41 (January-March 1970), pp. 122-35.
7. See, for example, Franz Schurmann," The attack of the Cultural Revolution
on ideology and organization,"in Ping-ti Ho and Tang Tsou, eds., China in Crisis,
Vol. 1, China's Heritage and the Communist Political System, Bk. Two (Chicago:
The University of Chicago Press, 1968), pp. 558-60.
8. See, for example, Leonard Schapiro and John Wilson Lewis, "The roles of
the monolithic party under the totalitarian leader," in John Wilson Lewis, ed.,
Party Leadership, pp. 114-45; Benjamin I. Schwartz, "The reign of virtue: some
broad perspectives on leader and Party in the Cultural Revolution," ibid. pp. 149-
69; and Stuart R. Schram, "The Party in Chinese Communist ideology," ibid.
pp. 170-202.
9. See, for example, Jack Gray, " Mao's economic thoughts," Far Eastern Eco-
nomic Review, Vol. LXVII, No. 3 (15 January 1970), p. 16.
10. See, for example, Robert Jay Lifton, Revolutionary Immortality: Mao Tse-
tung and the Chinese Cultural Revolution (New York: Vintage Books, 1968).
11. For example, those interpretations which stress Chiang Ch'ing's jealousy
of Wang Kuang-mei (Mrs Liu Shao-ch'i or Mao's resentment of insults suffered
at the hands of Liu Shao-ch'i.
12. The model was first developed to explain political factionalism in early
Republican China in my Ph.D. dissertation, " Factionalism in early republican
China: the politics of the Peking government, 1918-1920," Department of
Government, Harvard University, 1970. It has gone through various revisions in
response to the criticisms of colleagues and to my own reading on both Chinese
and non-Chinese politics. The logical status I would claim for the model, as
distinct from the conditions of its development, is that of an internally consistent
description of what political conflict would be like under certain assumptions.
The Structure of Factions

The starting point is a kind of human behaviour which I shall call
the "clientelist tie." 13 A clientelist tie is a non-ascriptive two-person
relationship founded on exchange, in which well-understood rights and
obligations are established between the two parties. The hallmarks of
a clientelist tie are as follows:
(i) It is a relationship between two people;
(ii) It is a relationship especially selected for cultivation by the
members from their total social networks;
(iii) It is cultivated essentially by the constant exchange of gifts or
services. (This does not imply that the subjective content of the
relationship is cynical or unfriendly: the contrary is normally the
case);
(iv) Since the exchange involves the provision by each partner of
goods or services the other wants, the parties to the tie are
dissimilar; very often they are unequal in status, wealth or power;
(v) The tie sets up well-understood, although seldom explicit, rights
and obligations between the partners;
(vi) It can be abrogated by eithqr member at will; and
(vii) It is not exclusive; either member is free to establish other
simultaneous ties (so long as they do not involve contradictory
obligations).
13. In recent years studies of the political and other functions of clientelist ties
have proliferated. Terminology varies (" dyadic contract," " dyadic alliance,"
" patron-client tie," etc.) but there is little question that various authors are re-
ferring to the same quite clearly defined phenomenon. Among the many discus-
sions of the clientelist ties or their political uses are George M. Foster, "The
dyadic contract: a model for the social structure of a Mexican peasant village,"
in Jack M. Potter, May N. Diaz and George M. Foster, eds., Peasant Society:
A Reader (Boston: Little, Brown, 1967), pp. 213-30; James C. Scott, " Patron-
client politics and political change in Southeast Asia," American Political Science
Review (APSR) LXVI: 1 (March 1972), pp. 91-113; and Carl H. Lands, "Net-
works and groups in Southeast Asia: some observations on the group theory of
politics," paper prepared for delivery at the 1971 annual meeting of the American
Political Science Association, forthcoming in APSR.
Clientelist ties have been highly developed, and have played important roles,
in China, where they have been embedded in and reinforced by a cultural stress
on the mutual rights and obligations involved in social roles. See especially
Benjamin I. Schwartz, " On attitudes toward law in China," reprinted in Jerome
Alan Cohen, The Criminal Process in the People's Republic of China, 1949-1963:
An Introduction (Cambridge, Mass.: Harvard University Press, 1968), pp. 62-70;
and Lien-sheng Yang, "The concept of pao as a basis for social relations in
China," in John K. Fairbank, ed., Chinese Thought and Institutions (Chicago:
University of Chicago Press, 1957), pp. 291-309. In this article I must beg a
number of important questions: what are the specific forms and functions of
clientelist ties in Chinese life and how have these ties changed in the revolutionary
process in China? Some thoughts on these matters are contained in my paper,
"'Connections' in Chinese politics: political recruitment and kuan-hsi in late
Ch'ing and early republican China," prepared for delivery at the 1972 annual
meeting of the American Historical Association.
38 'The China Quarterly
Such ties include patron-client relations, godfather-parent relations,

some types of trader-customer relations, and so forth. Corporate ties,
such as lineage relations, co-membership in an association, or
co-membership in a group of blood brothers exceeding two in number,
are not clientelist ties, although shared corporate membership often pro-
vides an initial contact which leads to the establishment of clientelist tie
as defined here.
The clientelist tie must be clearly distinguished from two other kinds
of relationship with which it is readily confused, the power relationship
and the (generic) exchange relationship. To take the second problem
first, the clientelist tie is founded on exchange. But, according to Peter
M. Blau, all social processes except those that are irrational or non-goal-
oriented or expressive - in other words, all social processes in which
people interact with other people in order to elicit behaviour from
them instrumental to some goal - can profitably be analyzed in terms
of exchange.14And according to Marshal Sahlins, any kind of recipro-
city, including the "negative reciprocity" of an eye for an eye, can
be usefully classified as exchange.15Blau and Sahlins are persuasive on
the utility of their respective analytical systems, but it is essential to be
clear that the clientelist tie as defined here is an exchange relationship
of a limited and specific kind. Embedded in different cultures it takes
somewhat different forms and is more or less explicity recognized,
spelled out, legitimated, and reinforced. But, in any case, it is relatively
stable and persistent, involves well-understood rights and obligations,
and is purposely cultivated by the participants. If nearly all of social
life is to be regarded as exchange, then clientelist ties should be regarded
as a special, quasi-contractual sub-type of exchange relationship.
At the other extreme, the clientelist tie must be distinguished from
the power relationship of "imperative coordination." 16 If for some
reason the subordinate has no real choice but obedience, the conse-
quences for political behaviour of the superior-subordinate relationship
will be quite different from what they would be if the real possibility
of abrogating the tie existed. Since the consequences are so different the
distinction is analytically necessary. It may be objected that, in many
cases, the right of abrogation formally exists but in fact cannot be
exercised, as in the relationship between landlord and tenant on some
Latin American haciendas, or between lord and vassal in feudal
Europe. These relationships, I would argue, should be regarded as
relationships of imperative co-ordination rather than as clientelistic
relations. Of course it is often difficult to tell the difference. There is a
14. Peter M. Blau, Exchange and Power in Social Life (New York: John Wiley
and Sons, 1964), p. 5 and passim.
15. Marshal D. Sahlins, " On the sociology of primitive exchange," in Michael
Banton, ed., The Relevance of Models for Social Anthropology (London: Tavi-
stock, 1965), p. 144 and passim.
16. Max Weber, The Theory of Social and Economic Organization, A. M.
Henderson and Talcott Parsons, trans. (New York: Free Press, 1964), pp. 152-3.
grey area, e.g., as the hold of the landlord begins to weaken but before
it is effectively challenged by that of a local political machine. The
analytic boundary lies somewhere within that grey area.
Clientelist ties in a given society articulate to form complex net-
works which serve many functions, including social insurance,,7 trade,18
and the mobilization and wielding of influence (i.e., political conflict).
It is with this latter function that I am concerned here. What happens
when political conflict is organized primarily through clientelist ties
rather than through formal organizations, corporate lineage units, or
mass or class movements? I would argue that there are three
possibilities.
First, the individual seeking to engage in political conflict may do
so by cashing in on his personal ties to operate as a power broker,
without directly and explicitly involving his partners in any common
or sustained endeavour. Examples include influence-peddling by lawyers
who specialize in arranging access to particular bureaucracies, mediation
of political disputes by middlemen, and the bridging of government/
village gaps by local "linkage figures." 19 The second possibility, which
occurs in a setting of genuine electoral competition, has been called
the " clientelist party," " vertical group," or " machine "-a mass
political organization which buys electoral support with particularistic
rewards distributed through a leader-follower network of clientelist
ties."2
17. Cf. Foster, " The dyadic contract."

18. For suggestive studies of clientelist ties in operation in non-political con-
texts, see, among others, James N. Anderson, " Buy-and-sell and economic per-
sonalism: foundations for Philippine entrepreneurship,"AS IX:9 (September
1969), pp. 641-68; Mary R. Hollnsteiner, " Social structureand power in a Philip-
pine municipality," in Potter, et al., eds., Peasant Society: A Reader, pp. 200-12;
Sidney W. Mintz, " Pratik: Haitian personal economic relationships,"ibid. pp 54-
63; and Robert H. Silin, " Marketing and credit in a Hong Kong wholesale
market," in W. E. Willmott, ed., Economic Organization in Chinese Society "Stan-
ford: Stanford University Press, 1972), pp. 327-52.
19. See, for example, Karl D. Jackson, " Communication and national inte-
gration in Sudanese villages: implications for communications strategy," paper
prepared for presentation at a meeting of the SEADAG Indonesia Seminar in
New York City on 30 March-1 April 1972; Martin and Susan Tolchin, To the
Victor . . . Political Patronage from the Clubhouse to the White House (New
York: Vintage, 1972); William Foote Whyte, Street Corner Society: The Social
Structure of an Italian Slum, Enl. ed. (Chicago: University of Chicago, 1955),
Pt. I1.
20. Distinctions can probably fruitfully be made between " clientelist parties,"
defined as integrating all levels of the political system through clientelist ties,
and " machines," defined as operating strictly on the local level. Among the major
theoretically-orientedstudies of such organizations are James C. Scott, Compara-
tive Political Corruption (Englewood Cliffs, N.J.: Prentice-Hall, 1972), chs. 6-9;
and John Duncan Powell, " Peasant society and clientelist politics," APSR LXIV:2
(June 1970), pp. 411-25.
Scott has gone furthest in specifying the conditions under which clientelist
parties or machines can develop (Corruption, esp. pp. 104-22, and 151-7). The
The third possibility occurs in an oligarchic or relatively small-scale

setting when an individual leader mobilizes some portion of his net-
work of primary, secondary, tertiary ties, and so on,21 for the purpose
of engaging in politics. A machine or clientelist party consists of a
great many layers of personnel, but this third type of clientelist political
structure consists of only one or a few layers.22 I call such a structure-
one mobilized on the basis of clientelist ties to engage in politics and
consisting of a few, rather than a great many, layers of personnel-a
faction.23 Such configurations include what may be called simple or
complex factions, and may control from within or without one or
more "support structures" or power bases, such as clubs, parties,
mobs, newspapers, banks, ministries, armies, and the like. To save
space, I simply indicate some of these possibilities in figure 1, and
problem of the conditions of machine development becomes especially important
when, as here, one is dealing with a mass participationsystem that nonetheless has
oligarchic factions. Obviously a key difference between contemporaryChina and
those political systems in which clientelist parties or machines have developed
is that in China there is no real electoral competition for power. Among the
case studies of political systems organized by clientelist parties are Robert H. Dix,
Colombia: The Political Dimensions of Change (New Haven: Yale University
Press, 1967), esp. chs. 8 and 9; Keith R. Legg, Politics in Modern Greece (Stan-
ford: Stanford University Press, 1969), esp. chs. 6-8; Myron Weiner, Party
Building in a New Nation: The Indian National Congress (Chicago: The Univer-
sity of Chicago Press, 1967); and William Foote Whyte, Street Corner Society, esp.
ch. VI. I am grateful to Michael Bucuvalas and Pedro Cabainfor bringing Legg
and Dix to my attention.
21. For ways of conceptualizinga network from the standpoint of an individual
ego, see J. A. Barnes, " Networks and political process," in Marc J. Swartz, ed.,
Local-Level Politics: Social and Cultural Perspectives (Chicago: Aldine, 1968),
pp. 107-30; and Adrian C. Mayer, "The significance of quasi-groupsin the study
of complex societies," in Michael Banton, ed., The Social Anthropology of Com-
plex Societies, A.S.A. Monograph 4 (London: Tavistock, 1966), pp. 97-122.
22. It seems unnecessary to specify an exact size boundary between factions
and machines since the difference between the two is so large. The difference, of
course, is not just one of size; as a consequence of their differentsizes and different
degrees of selectivity in recruitment, as well as of the different natures of their
respective arenas, machines and factions behave in thoroughly distinguishable
ways.
23. This conception of a faction is similar to that offered by Ralph Nicholas,
"Factions: a comparative analysis," in Michael Banton, ed., Political Systems
and the Distribution of Power, A.S.A. Monograph 2 (London: Tavistock, 1965),
pp. 27-9, and to Land6's concept of the " dyadic following " in his " Networks
and groups." Let me stress again that I have defined "faction " in a technical
sense. By faction I do not mean an " organized opinion group" (cf. Schurmann,
Ideology and Organization, p. 56) contending warlords, or Red Guards. Nor
should the word faction as used here be regardedas a translation of the Chinese
terms p'ai, hsi, tang, or hui. Whether any of these things can be called a faction
by the present definition can only be determined upon close structural analysis.
Although restrictive, the definition advanced here seems to fit a wide range of
configurations found in political systems and sub-systems including governments,
parties, bureaucracies, parliaments, courts and villages in a number of different
geographical areas and historical periods. Some examples are cited in my
dissertation,pp. 372-85.
Figure 1
Simple faction
Complex faction
L Simple factions and

support structures
forgo formal definitions and systematic comparative analysis. What

all these configurations share in common is the one-to-one, rather
than corporate, pattern of relationships between leaders (or subleaders)
and followers. Structurally, the faction is articulated through one or

more nodes; and it is recruited and co-ordinated on the basis of the
personal exchange relationships I have called clientelist ties.
The Structural Characteristics of Factions

To proceed from this conception to the behavioural characteristics
it implies, I must elaborate some of the structural characteristics of
factions.24
Because it is based upon personal exchange ties rather than authority
relations, a faction does not become corporatized after recruitment but
remains structured along the lines of the original ties which formed
the bases of recruitment.25 Upward and downward communications
therefore tend to follow the lines of recruitment. This lends the faction
considerable flexibility: the leader sees the opportunity for political
gain, separately recruits each member into the faction,26 and directs
the activities of each member for the overall good of the faction. The
members need never meet, although they may do so. The members'
activities in disparate locations and institutions can be co-ordinated
through individual communications with the leader.27Indeed, in routine
political situations, regularized co-ordination can be dispensed with
entirely, since the faction as a whole can rely on the members' loyalty
to the leader to insure that each member works to the faction's benefit.
Thus, the faction is capable of the greatest flexibility in seizing political
opportunities and in engaging in a general political strategy on the
basis of scattered positions throughout a political system or an
organization.
On the other hand, such a communications pattern involves certain
liabilities. Upward and downward communications are not delivered
directly to the recipient (in the case of complex factions) but flow
through a series of nodes (sub-leaders). The more steps through which
the information flows, the more time it takes and the more distorted
24. Each structuralcharacteristicdiscussed is not necessarily unique to factions

(for example, guerrilla bands may be equally flexible, and for some of the same
reasons), but none of them is universal and the combination of characteristicsis
distinctive.
25. If a faction becomes corporatized the clientelist relations are submerged
in authority relations and the structureceases to be a faction in this sense. What
might encourage or discourage corporatization of a faction is a question that
cannot be investigatedhere. For the concept of " corporate" used here, see Weber,
Theory of Social and Economic Organization, pp. 145-8.
26. The follower may in turn recruit others as his own followers.
27. For a suggestive exploration of the advantages of this sort of communi-
cations pattern, see Alex Bavelas, "Communications patterns in task-oriented
groups," in Daniel Lerner and Harold D. Lasswell, eds., The Policy Sciences
(Stanford: Stanford University Press, 1951), pp. 193-202.
it is likely to become.28 This is one of several structural characteristics

which tend to set limits to the number of levels to which the faction
can extend without becoming corporatized 29 and to the degree to
which large factions can engage in finely co-ordinated activities. The
faction, in short, is highly flexible but self-limiting in its extensiveness.
Since it is founded upon exchange relationships, a faction depends
for its growth and continuity upon the ability of the leader to secure
and distribute rewards to his followers. It tends to expand and contract
with success or failure, and may even be dissolved when removed
from power. But it can always be reconstituted when it regains the
capability to reward its members. The faction is thus capable of
intermittent but persistent functioning. It takes form out of the general
network in response to an opportunity for political gain, and when it
becomes necessary to retire from politics, as during the dominance
of the political arena by an enemy, it can become relatively dormant.
Its political activities temporarily cease, especially if the members have
scattered; or the activities may continue at a low level of occasional
contact with other opposition groups to scheme for a return to power.
When the enemy is overthrown, the faction may return to full activity
unchanged in form and flexibility.
It also follows that a faction cannot survive its leader. The members
may continue in political activity after the leader's death or retirement;
several members may associate for a time and continue to be known
by the original name of the faction. Members may also join other
factions as temporary allies or they may seek to found their own
faction. But since the set of clientelist ties on which it is founded forms a
unique configuration centered on the leader, the faction can never
be taken over as a whole by a successor. The unique combination of
personnel and strategic political positions held by the faction cannot
be completely reconstructed once the leader is lost.
The more extended the complex faction becomes, the greater the
28. See Barry E. Collins and Bertram H. Raven, " Group structure: attraction,
coalitions, communications, and power," in Gardner Lindzey and Elliot Aron-
son, eds., The Handbook of Social Psychology, 2nd ed. (Reading, Mass.: Addison-
Wesley, 1969), Vol. IV, pp. 137-55.
29. It may be asked why communications patterns and other structuralfeatures
to be discussed below limit the size of factions but not of "clientelist parties."
For one thing, although a clientelist party is founded on patronage dispensed
through clientelist ties, it also takes on elements of formal organization (party
label, headquarters, officers, rules) to enable it to administer its mass electoral
base. It is in this sense not a "pure type" of clientelist structure. Even more
fundamentally, each type of political structure ideal-typically stands in an adap-
tive relationship to its political environment. The clientelist party is adapted to,
and tends to maintain, a mass electoral political system. The faction is adapted
to and tends to maintain an oligarchic or small-scale system. Although I argue
that certain elements of factional structure limit factional size, I could just as
well build the argument in reverse: a political setting which involves relatively
few people makes it possible for men to organize themselves in ways that would
not be suited to large organizations.
number of subordinate leaders it contains and the farther removed

they are from the primary leader. The leader of each simple faction
within the complex faction is primarily responsible to his own followers
for political spoils. The possibility exists that this responsibility may
come into contradiction with the loyalty he owes his own leader-who
is pursuing the interests of a different set of persons. This creates the
tendency for the leaders of lower segments to betray the interests of the
faction as a whole in order to secure greater rewards for the segments
below them, which are capable of operating as distinct factions if they
free themselves from the larger faction.30
Because of this tendency towards breakdown, complex factions are
most likely to develop, and are likely to develop to the largest size,
within bureaucratic formal organization. In formal organizations, first,
the personal loyalty of faction leaders at lower levels to leaders at
higher levels is reinforced by hierarchical authority patterns. Second,
the faction has the benefit of the intra-organizational communications
network to aid in co-ordinating its activities. Third, the effort to gain
control of the organization or to influence its policies requires the
co-operation of the sub-leaders at the various levels and tends to bring
their interests into harmony. In short, the hierarchy and established
communications and authority flow of the existing organization provides
a kind of trellis upon which the complex faction is able to extend its
own informal, personal loyalties and relations.
There is a tendency for vertical cleavages to develop within the
complex faction, running up to the level directly under the highest
leader. Vertical cleavages also develop at each lower level, but remain
latent while they are submerged within the greater rivalry between the
two major segments. (Because the struggle is for access to and influence
over the leader, there is a tendency for these cleavages to be limited
to two.) The conflict between the two major entities within a faction is
kept under tight rein by the faction leader. After his retirement or death,
however, the two entities become two new complex factions.
Internal cleavage tends to be increased by the fruits of victory. First,
the path to victory inevitably involves reaching opportunistic alliances
with factional leaders who are incorporated as allies within the faction
but are not reliable. Second, the increased scale and numerical force
of the faction as it grows enhances the tendency mentioned earlier for
divergent interests to emerge among component sub-leaders and sub-
factions. Even if loyalty prevents an open revolt against the leader,
it permits political clashes and struggles among his subordinates. Third,
if the faction comes near to or achieves victory in a conflict arena as a
whole, the unifying factor of a common enemy ceases to exist, while
divisive factors such as struggle over spoils and efforts by smaller,
enemy factions to buy over component units increase in salience.
30. Cf. the " size principle" as enunciated by William H. Riker, The Theory
of Political Coalitions (New Haven: Yale University Press, 1962), pp. 32-3.
Fourth, the growth of the faction tends to deprive the leader himself of
direct control over component units, weakens his position vis-ac-vis
subordinates, and thus hastens his political retirement and the conse-
quent open split of the faction. In short, division and decline is the
almost inevitable result of success. The only way to avoid such dis-
integration is to refuse to expand beyond the borders of an internally
unified and easily defended factional base.
Finally, it follows from all the above that the faction is limited in
the amount and kinds of power it can wield and generate. A faction
is limited in size, follower commitment, and stability by the principles
of its own organization. Certain other types of conflict structure, for
example highly organized political parties or armies, can, by virtue of
their complex, functionally specialized organization, their clear boun-
daries, and their high degree of control over participants, engage in
feats of mobilization, indoctrination, and co-ordination which are
beyond the capacities of factions.31 A faction of course is limited in
power only so long as it remains a faction; there is nothing inherent
in the existence of a faction to prevent the members, if they need
more power, from organizing themselves in some other way. However,
the fact is that people often do not organize themselves in another
way, but in factions. Why they choose to do so, and what conditions
may cause them to shift to another form of organization, are important
questions beyond the scope of the model.
The Characteristics of Factional Politics

For simplicity, I limit the following discussion to an ideal-typical
political system which is primarily organized by factions. Presumably
factions will behave somewhat differently when they are competing
against other structures which are not factions (for example, clans or
political parties), but the easiest case to deal with is the "pure case"
of an all-factional system. This section completes the model of factional
politics by attempting to show what modes of conflict will be typical of
factions operating in an environment consisting primarily of rival
factions.32
A first set of propositions is based upon the power limitations typical
of factions. Factions, I have argued, enjoy less power capability than
formal organizations because of the limitations on their extent, co-
ordination, and control of followers implied by their basis in the clien-
31. Cf. Amitai Etzioni, Modern Organizations(Englewood Cliffs, N.J.: Prentice-

Hall, Inc., 1964), passim.
32. As was the case with the structural characteristicsof factions, the claim is
not made that the modes of conflict characteristicof factions are each unique to
factional systems, merely that none of them is universal to all political systems
and that the combination of all of them is found only in factional systems.
telist tie, their one-to-one communications structure, and their tendency

towards breakdown. It follows from this that the several factions in a
given factional arena will tend, over time, to enjoy relative power
equality; for no faction will be able to achieve and maintain over-
whelmingly superior power. One faction may for the moment enjoy
somewhat greater power than rival factions, but this power will not
be so much greater that the victorious faction is capable of expunging
its rivals and assuring permanent dominance.33 This is the more so
because the flexibility of the weaker factions and their capability for
intermittent functioning enhance their ability to evade and survive
repression. A faction engaging in conflict with other factions must
therefore operate on the assumption that it will not be able decisively
and finally to eliminate its rivals. The faction which holds power today
can expect to be out of power and vulnerable tomorrow. Politicians
in a factional system are "condemned to live together."34 This enables
us to posit that the following modes of conflict will be typical of
factional systems.
1. Since the impulse to crush one's rivals decisively is stymied by the
limited nature of power, a code of civility arises which circumscribes
the nature of political conflict.35Factions relatively seldom kill, jail or
even confiscate the property of their opponents within the system (the
killing and jailing of persons felt to present a threat to the system
is another matter; see point 12). Indeed, factional systems require
punctiliously polite face-to-face conduct between politicians. As Leites
has written with respect to the French Assembly of the Third and
Fourth Republics, "the vicissitudes of political life exacerbate one's
feelings, but it is imperative that rage be channelled into entirely
appropriate expressions so as not to endanger one's career."3"
2. Since factions are incapable of building sufficient power to rid
the political system of rival factions, they have little incentive to try
to do so. For any given faction, the most important and usually most
immediate concern is to protect its own base of power while opposing
accretions of power to rival factions, while initiatives to increase its
own power and position are of secondary importance. Defensive political
strategies therefore predominate over political initiatives in frequency
and importance.
33. A major reason for differences in the power of factions is the differing power
of their support structures (regional and/or institutional power bases). But op-
posing power bases cannot be entirely eliminated, nor, given the tendency of large,
victorious factions to split, can they be taken over.
34. Nathan Leites, On the Game of Politics in France (Stanford: Stanford
University Press, 1959), pp. 23 and 45.
35. Cf. F. G. Bailey, "Parapolitical systems," in Swartz, ed., Local-Level
Politics, p. 282; Bernard Gallin, " Political factionalism and its impact on Chinese
village school organization in Taiwan," ibid. p. 390; and Melford E. Spiro,
"Factionalism and politics in village Burma," ibid. pp. 410-12.
36. Leites, On the Game, p. 117.
3. When a faction does take a political initiative (which it does only

on those rare occasions when it feels that its power base is secure and its
rivals are relatively off balance), it relies upon secret preparation and
surprise offensive. This minimizes the ability of rivals to prepare their
defensive moves in advance and provides the aggressive faction a
momentum which carries it further than would otherwise be the case
before the defensive moves of rivals stop its progress.
4. In the face of such an initiative, the defensive orientation of the
other factions in the system tends to encourage them to unite against
the initiative. Thus factional political systems tend to block the emerg-
ence of strong leaders. The strong leader constitutes a threat to the
other factions' opportunities for power, and they band together long
enough to topple him from power. (In many political systems this
leads to governmental instability. In France under the Third and Fourth
Republics, for example, where the Government was dependent for its
office upon an alignment of parliamentary factions, the very fact that a
politician was able to organize a Cabinet set in motion the jealousy
and opposition that soon led to its fall. Where the titular government
leader is supra-factional and enjoys little power, however, the formal
governmental leadership of a factional system may remain stable for
long periods.)
5. Since the political life of a factional system consists of occasional
initiatives by constituent factions, followed by defensive alliances
against the initiator, any given faction is obliged to enter into a series of
constantly shifting defensive alliances. Factional alliances cannot remain
stable. Today's enemy may have to be tomorrow's ally.
6. It is therefore impossible for factions to make ideological agreement
a primary condition for alliance with other factions. As I argue below,
factions operate within a broad ideological consensus (point 13) while
exaggerating the small differences that remain among them (point 10).
The struggle for office and influence is unremitting, immediate, and
never decisively resolved. In order to stay in the game, factions must
often co-operate with those with whom they have recently disagreed.
Although factional alignments do not cross major ideological boun-
daries, within those boundaries they are not determined by doctrinal
differences.87
7. When decisions (resolutions of conflict, policy decisions) are made
by the factional system as a whole, they are made by consensus among
37. In the Chinese context, for example, factional alignments did not cross the
ideological boundaries between the late Ch'ing conservatives on the one hand and
the constitutionalistsand revolutionarieson the other, or between the Kuomintang
and the CCP. But within each major ideological current factional alignments were
not (how could they be?) determined by pure, a priori ideological compatibilities.
Ideological stands were developed and revised in the course of politics. For a
case study of the process by which ideological standpoint becomes defined in the
course of political conflict, as rivals force one another to delineate and clarify
their positions, see Benjamin I. Schwartz, Chinese Communism and the Rise of
Mao (Cambridge, Mass.: Harvard University Press, 1952).
the factions. To attempt to take action without first achieving such a

consensus would take the ruling coalition beyond the limits of its
power: the decision could never be enforced. Furthermore, the effort to
enforce a decision would hasten the formation of an opposition coalition
to topple the ruling group from power. Decision by consensus also
has the advantage that action is taken in the company of one's rivals, so
that responsibility cannot be pinned on any single faction.as
8. There is a typical cycle of consensus formation and decline which
characterizes factional systems. The cycle begins with a political crisis.
As the factions contemplate the crisis, "the limits of what every party
(or every clique or individual) may be capable of attaining" becomes
clear to all, and after a lapse of time the crisis becomes "ripe."'
"Imperious necessity . . . make[s] the . . . groups disregard their
positions of principle" which had blocked consensus, and action
becomes possible.39
As a result of the consensus among the factions on the need for
action, a faction or factional alliance achieves office and receives a
mandate to act. The victorious faction takes culturally appropriate
actions to test and solidify the support the other factions have been
obliged to give it. The leader may refuse to take office until the other
factions have publicly committed themselves to him; he may try to
associate the leaders of other factions in the action he proposes to take;
he may allow, or encourage, the crisis to worsen. Ultimately, however, he
acts.
The third phase is the decline of the factional consensus. The actions
taken by the faction in power inevitably have implications for the
relative strength of all the factions in the system. While the actions
carry the system through the crisis which had produced the consensus,
they benefit some factions-usually the one in power and its allies-
more than others. The other factions act to block the effort of the
leading faction to strengthen itself, and the factional consensus deteri-
orates. The factions return to mutual squabbling.
The period of factional conflict often lasts a long while as the factions
manoeuvre for political resources, alliances, and a favourable moment
and pretext for precipitating a new crisis. Eventually a faction feels it is
in a good position to take a political initiative, to precipitate a test
of strength with its major opponent. In many factional systems, this
takes the form of asking the most obstructive opposition factions to
form a government in the expectation that they will fail.40 Whatever
its form, the test of strength initiates a new crisis which begins another
cycle.
A second set of propositions is based upon the fact that factions

consist of a series of clientelist ties. The resources with which the
38. Cf. Leites, On the Game, pp. 48-9.

39. Ibid. ch. 4, esp. pp. 97-8. 40. Cf. ibid. pp. 82-3.
faction carries out political conflict are not corporate, shared resources,
but the personal resources of the individual members-their personal
prestige, official positions, and their own further clientelist ties.
9. To weaken their rivals, factions try to discredit opposition faction
members, dislodge them from their posts, and buy away their allies.
This leads to a politics of personality in which rumour, character
assassination, bribery and deception are used. Passions of jealousy and
revenge are aroused, opportunism and corruption are fostered, and
urgent short-term political goals require the compromise of principles.
These, in short, are the "comic opera" politics or "pure politics" so
characteristic of factional systems.41
10. A further characteristic of factional political conflict may be
called doctrinalism, i.e., the couching of factional struggle for power
in terms of abstract issues of ideology, honour, and face.42 Factions
adopt rigid and minutely defined ideological positions, exaggerate small
differences on abstract questions, and stress the purity of their own
motives. Yet the issues which arouse such fierce and elaborate debate
appear upon close examination to be those with strategic implications
for factional power.43 Although the real distance between cliques in
ideology and programme is small (points 12 and 13), and although
no faction is likely to be able to carry out an innovative political
programme, grand policies and sweeping programmes are articulated
and debated, with small points attracting the most passionate and
lengthy discussion.
Such debate serves several purposes. First, it distinguishes one faction
from another," providing a rationale for the continued struggle among
such otherwise similar entities. Second, it provides an opportunity to
discredit other politicians and to justify oneself on abstract or ideological
grounds. Third, the broad programmes often include inconspicuous
provisions of true strategic political importance. The struggle which is
couched in abstract terms is really over the advantages of a policy to
one side or the other.
41. Cf. James L. Payne, Patterns of Conflict in Colombia (New Haven, Yale
University Press, 1968), pp. 3-24. Payne attempts to explain factionalism in
Colombian politics on the basis of the prevalence of " status," rather than " pro-
gramme," incentives among Colombian politicians. However, if our model is
correct, factionalism can occur in the presence of either type of incentive. Politi-
cians in factional systems will tend to act as if they were motivated by status
incentives because of the importance of personal prestige and personal connexions
as political resources in factional systems. It is immaterial to the model how high-
minded the ultimate motives for conflict are.
42. Cf. Leites, On the Game, pp. 7-34.
43. Cf. Payne, Patterns of Conflict, pp. 249-50.
44. Myron Weiner, Party Politics in India: The Development of a Multi-Party
System (Princeton: Princeton University Press, 1957), pp. 237-40. See also
Lewis A. Coser, The Functions of Social Conflict (New York: The Free Press,
1964), pp. 33-8.
A third set of propositions concerns the size and shape of the

factional system as a whole and the way it relates to its political
environment.
11. Any factional arena is composed of a limited and not very large
number of factions. This is so because in an arena with a very great
number of factions, it will be in the interests of the factions to amalga-
mate, in order to defend against other factions doing the same thing.
The incentives to amalgamate cease to be stronger than those to
engage in conflict only when the total number of factions has been
reduced to the point where most of the constituent factions enjoy
enough strength to launch political initiatives and defend themselves,
while further amalgamation would simply bring in more followers to
share the rewards of the faction without decisively affecting its ability
to survive. It is doubtful whether more than a score or two of factions
can exist in a given factional system or arena. (The limitations on the
number of members in a faction, and on the number of factions in a
system, form a logical circle with the initial assumption of an oligarchic
or small-scale arena.)
12. I have already established that the members of the small factional
elite act within a code of civility which limits the severity of the
sanctions they employ, and under tactical constraints that require
alliances with former enemies and opposition to former allies.45 This
closely knit elite is further united by one overriding shared interest:
that the resources over which they are struggling should be allocated
among themselves and in accordance with the rules of conflict they are
following, rather than to some force from outside the system which
pays no attention to those rules and whose victory would end the
political existence of the factions. The result is a sharp difference
between the modes of intra-elite conflict described in points 1 to 10
and the drastic steps which may be taken by the united factional elite
to resist external enemies or to destroy counter-elites who challenge the
legitimacy of the factional system.46 When, for example, foreign
conquest, rebellion, or a military coup threatens to overthrow a factional
regime, the factions unite behind a suitable leader long enough to
preserve the system, before returning to politics as usual."' If the threat
45. On the unifying effects of conflict, see further Coser, Functions, passim;
and George Simmel, Conflict and the Web of Group-Affiliations, Kurt H. Wolff
and Reinhard Bendix, trans. (New York: The Free Press, 1955), pp. 13-123.
46. The same distinction holds if, as I shall argue was the case with the CCP
until 1949, the group of factions is itself the counter-elite and its enemies are
the ruling elite.
47. Of course, they do not always succeed in preserving the regime. If the
social context of the regime has been changing so that, e.g., new problems demand
solutions or new groups demand access to the system, immobilism (see below)
may prevent the regime from responding successfully. The consequent loss of
legitimacy may make it an easy target for a strong rival. An example is the crum-
bling of the Peking Government before the Kuomintang advance in 1928.
The Clemenceau and Poincare ministries in France are well-known instances
of resistance by a factional system to external threat. See Philip M. Williams,
to the system comes from within, from a factional leader attempting to

break the rules, the efforts of the other factions are directed towards
defeating that attempt and re-establishing the stability of the system.
13. Within the factional elite, it is taboo to question the principle of
legitimacy upon which the factions base their claim to a role in the
larger society. Thus, for example, a factional parliament, regime, or
party may play the role, in the larger society, of a central or local
government, on the basis of a constitutional or charismatic claim to
legitimacy. No matter how much the vicissitudes of struggle oblige the
factions to trample in fact upon constitutional principles, or to disobey
in fact their symbolic leader, these must never be openly questioned or
flouted since that would encourage other forces in the larger society
to "throw the rascals out." 48 Thus politicians in factional systems
compete in expressions of fealty to the constitution or leader and
rationalize every action and every position in terms of their fidelty to it
or him. Care is taken to assure the constitutional or charismatic
continuity of the regime.
14. Issues which arise within the elite are resolved only slowly and
with difficulty. The consensus which is necessary for action is difficult
to achieve because every decision is more advantageous to some
factions than to others. Only the cycle of crisis and consensus brings
action, but it is short-term action to meet the immediate emergency,
and may in any case be followed by contradictory decisions after
the next cycle of conflict. The resulting failure of policy to move
clearly in any one direction is what was called, in the French Third
and Fourth Republics, immobilisme.
15. The immobilism of factional systems, the lack of extreme sanc-
tions employed in their struggles, and their tendency to defend their
existence against rival elites or external threats mean that they are in a
certain sense extremely stable. It does not seem to be true, as some
observers have suggested, that factional systems have an inherent
tendency to break down.49In the absence of outside pressures (in which I
include those from social forces within the society), no force within
the factional system is capable of amassing enough power to overthrow
it. Thus, only continued factionalism can be predicted on the basis of the
fact that a system is already factional.
There will, of course, be other causes acting upon a factional system
which also affect the outcome. Personal, cultural and technical resources
for more complex organizations may be present to a greater or lesser
degree; so may leaders' political vision and the will to move beyond
the factional form of organization; and so may the challenge of
changing social conditions. Members of factional systems may abandon
Crisis and Compromise: Politics in the Fourth Republic, paperback ed. (Garden
City, N.Y.: Anchor Books, 1966), p. 11. 48. Cf. Simmel, Conflict, p. 41.
49. Bernard J. Siegel and Alan R. Beals, " Pervasive factionalism," American
Anthropologist, 62 (1960), pp. 394-417. For a critique, see Nicholas, "Factions,"
pp. 56-7.
factions for other forms of organization (often clientelist political

parties). But if they do so, it will not be because of any dynamic of the
factional system itself. More often than not, factional systems under
challenge from within or without respond by trying to reassert their
equilibrium rather than by changing their internal structure.
Does China Have Factional Politics?

To the extent that this model is applicable to China, one would
expect to find either direct evidence of the existence of factions, such as
descriptions of the organization and membership of specific factions,
or circumstantial evidence of their existence, such as incidents which
display the modes of conflict I have described. On the other hand, we
would have to abandon the model as an explanation for Chinese
political behaviour where we found evidence of other conflict structures
or of modes of conflict not consistent with the model. The Cultural
Revolution period is a good place to begin the search because of the
relative openness of political conflict at all levels. The available
evidence, I will argue, permits the view that Chinese politics at the
central government level has been structured largely by factions,
especially if the Cultural Revolution is regarded not as a first outbreak
of factionalism but as an episode in which a long-standing factional
system attempted to defend its existence against an attack based on
outside social forces; that political conflict at the intermediate (regional
and provincial) levels has been non-factional or, at most, intermittently
factional; and that the local-level mass movements of the Cultural
Revolution were essentially non-factional.
The centre is an ill-defined elite of several hundred high officials
of the party, army and state bureaucracies. All the members of this
elite have devoted their lives to the revolutionary movement, are Party
members and have served extensively with the PLA. Despite differing
current affiliations, therefore, it is a single elite whose political ties
and concerns cross institutional boundaries.50While Central Committee
membership is prestigious, it does not define the central elite, whose
composition shifts more rapidly than Central Committees can be recon-
stituted.51 Informal organs such as the work conference,52with flexible
membership, are the locus of political conflict within this central elite.
There is some direct evidence of the existence of personal factions
50. Cf. in particular Stuart R. Schram, "The Party," p. 171; Howard L. Boor-
man, "Teng Hsiao-p'ing: a political profile," CQ 21 (January-March 1965),
p. 119; Whitson, "The Field Army," passim.
51. " In fact, CC membership has been in recent years a badge of recognition
or a reward for things done." Donald W. Klein and Lois B. Hager, "The ninth
Central Committee," CQ 45 (January-March 1971), p. 37.
52. Parris H. Chang, " Research notes on the changing loci of decision in the
Chinese Communist party," CQ 44 (October-December 1970), pp. 170-3.
in this top elite. Scholars have been able to identify certain figures as
followers of, among others, T'ao Chu,53 Lin Piao,54 Liu Shao-ch'i,55
and, of course, Mao. Analyses of the backgrounds of purged leaders
have suggested that certain shared experiences may have led to the
formation of personal ties which help to explain who was purged
during the Cultural Revolution.56 Many purged leaders have been
charged with crimes like "fostering personal forces, establishing their
personal authority, winning over defectors and renegades, forming
private parties. . . ." 57 However, none of this evidence is really con-
clusive and it remains impossible to delineate with any confidence
who the major factional leaders have been, much less their followers.
Turning to circumstantial evidence, the central political arena quite
clearly displayed most of the characteristic modes of conflict of fac-
tional systems during the Cultural Revolution. Almost all commentators
noted the absence of extreme sanctions against defeated politicians.58
Most defeated leaders were punished only by re-education, while even
the most severely sanctioned political leaders suffered only the humili-
ation of mass denunciations; few, if any, were executed. At the time
of writing some purged leaders are even returning to office.59 Also
noteworthy was the effort to provide some legal validation, however
flimsy, for the deposition of leaders like Liu Shao-ch'i and P'eng Chen,
suggesting a concern for the constitutional continuity of the regime.
Defensive tactics predominated during the Cultural Revolution. One
thinks of P'eng Chen's effort through his 12 February directive to turn
the brunt of Mao's attack onto the "academic authorities "; the counter-
mobilization of loyalist Red Guards or revolutionary rebels by local
officials under attack; the disregarding or distortion of directives from
the centre; the "sham" confessions which were actually self-justifica-
tions.co By contrast, political initiatives such as Yao Wen-yuan's attack
53. Ezra F. Vogel, Canton under Communism: Programs and Politics in a Pro-
vincial Capital, 1949-1968 (Cambridge, Mass.: Harvard University Press, 1969),
p. 325; Michel Oksenberg, " Occupational groups in Chinese society and the
Cultural Revolution," in The Cultural Revolution: 1967 in Review, Michigan
Papers in Chinese Studies, No. 2 (Ann Arbor, Mich., 1968), p. 26.
54. " Lin Piao: a historical profile," Current Scene, Vol. VII, No. 5 (10 March
1969), p. 3; Bridgham, "Mao's Cultural Revolution: the struggle to consolidate
power," p. 16; Ralph L. Powell, "The increasing power of Lin Piao and the
Party-soldiers, 1959-1966," CQ 34 (April-June 1968), pp. 62-3.
55. Schapiro and Lewis, "The roles of the monolithic Party," p. 131.
56. Parris H. Chang, " Mao's Great Purge: a political balance sheet," Problems
of Communism, Vol. XVIII, No. 2 (March-April 1969), p. 8: Whitson, "The
Field Army," pp. 13-21.
57. Cited from Hung-ch'i (Red Flag), No. 12, 1967, in " Quarterly chronicle
and documentation," CQ 32 (October-December 1967), p. 198.
58. E.g., see Charles Neuhauser, " The Chinese Communist Party in the 1960s:
prelude to the Cultural Revolution," CQ 32 (October-December 1967), p. 3.
59. Among these the most prominent so far are Chao Tzu-yang and Ch'en
Tsai-tao, but many less prominent cadres are also said to have been reinstated.
60. Cf. Ezra F. Vogel, "The structure of conflict: China in 1967," in The
Cultural Revolution: 1967 in Review, pp. 106-7.
on Wu Han, or the launching of the Red Guards, were relatively rare,

and were apparently prepared in secret and sprung by surprise. Opposi-
tion to the emergence of a strong leader was shown by the resistance
of other leaders to the accretion of power in Mao's hands. The shifting
of factional alliances is reflected in the rise and fall of various pro-
paganda directors, chiefs of staff, and the like; in the purging by 1969
of three-fifths of the pre-Cultural Revolution Central Committee; in the
fall of 23 of 28 provincial leaders, of four out of five members of
the original Group of Five directing the Cultural Revolution and of 13
of 18 members of the Central Cultural Revolution Group.61 The
irrelevance of doctrine in determining factional alliances is suggested
by the fact that many attacked as anti-Mao during this period had
records which were utterly pro-Mao,62 while Chou En-lai, for example,
who has not always been particularly pro-Mao, has so far managed to
retain the title of a Maoist. The purges of Ch'en Po-ta and Lin Piao
in 1971 gave further validation of this point.
That decision-making within the central elite was made by consensus
and compromise is suggested by the contradictory orders which were
often issued simultaneously or even within the same document, such as
the famous Sixteen Points; and by the oscillations noted by many analysts
between radical and moderate policies emanating from the centre.63
A closer look at these oscillations suggests that they followed the
typical factional cycle in which a crisis leads to a policy decision, which
itself becomes the focus of renewed conflict leading ultimately to another
crisis. For example, the events of July-September 1967 began with
a crisis, the Wuhan incident, which temporarily unified the factions at
the centre against rule-breaking violence by local military commanders
against central officials. The call was issued to "pluck out a small
handful" in the Army. During August, the Red Guards responded by
stealing weapons and attacking the PLA. Thus, the July decision
which originally represented the joint interests of all the factions at the
centre in suppressing armed action by military dissidents began to
redound too strongly to the political advantage of the pro-Red Guard
factions at the centre. This led to renewed policy debate at the centre
and presumably to a showdown which issued in a new decision - this
time repudiating the concept of attacking a handful in the Army, and
purging three members of the Central Cultural Revolution Group.14
The "comic opera" conflict of personal vilification and imaginative
61. See Richard Baum, "China: year of the mangoes," AS IX: 1 (January
1969), pp. 3-4; Gordon A. Bennet, "China's continuing revolution: will it be
permanent?"AS X:1 (January 1970), p. 3.
62. Cf. Merle Goldman, "The fall of Chou Yang," CQ 27 (July-September
1966), pp. 132-48.
63. For example, see Bridgman's three articles cited above, note 3. Also cf.
Baum, " China: year of the mangoes," pp. 4-5.
64. See "Quarterly chronicle and documentation," CQ 32 (October-December
1967), pp. 184-221.
attack was certainly a feature of the Cultural Revolution, and, as I have

shown, such tactics are serious and meaningful in factional systems,
where the personal prestige, position and contacts of the politician
are his prime resources. Equally evident was doctrinalism, as seen, for
example, in the practice of "waving the Red Flag to oppose the Red
Flag."
Of the 15 characteristics of factional politics, then, 10 are visible in,
or consistent with, the events of the Cultural Revolution. The eleventh -
the relatively small number of factions in the system - cannot be
confirmed or denied on the basis of our present inadequate information.
It is perhaps too early to reach a judgment with respect to two further
characteristics, immobilism and stability, although one could certainly
argue that no matter how much the Chinese political system seemed
to be breaking down during the Cultural Revolution, it did not actually
break down; and no matter how much it seemed to be making definite
policy decisions, it did not opt for any single consistent foreign policy,
development strategy, defence policy, policy on the role of intellectuals,
or the like, but wavered between contradictory poles very much like
a system composed of constantly shifting factional alliances.65
This leaves two features to look for - a taboo on discussion of the
principle of legitimacy, and solidarity of the factions in the face of
outside threats. The episode of the Red Guards would seem to contra-
dict these predictions, since members of the elite went outside the elite
itself to mobilize a social force, the students (which had previously
threatened many elites in modem Chinese history), and encouraged
this force to attack the leaders of the Party itself. This contradiction
can be resolved by arguing that the politics of the central elite was
already factional in the early 1960s and that Mao, for reasons discussed
below, decided to break the rules of factional politics and directly
mobilize a social force to destroy the factional system; that the non-
Maoist members of the factional elite responded, as the model predicts,
with solidarity against the student threat and took steps to demobilize
the Red Guards; and that in doing so, they scrupulously avoided
criticizing Mao, since they had in recent years increasingly based the
legitimacy of the regime upon Mao's personal charisma rather than on
the claimed historical mission of the Party as an institution. I will
expand upon this interpretation of the Cultural Revolution, after survey-
ing the evidence on intermediate and local-level conflict.
Data on political conflict at the intermediate levels of the Chinese
political system have been scarce. Perhaps the most confident assertion
we can make is that relatively independent political sub-systems (" inde-
pendent kingdoms") have existed in the Chinese political system. These
include provinces - two famous examples being Inner Mongolia under
65. Of course at any given moment there seemed to be a definite policy line,
but the question is whether the line was consistently sustained or whether it gave
way to a contradictory line.
Ulanfu and Szechwan under Li Ching-ch'tian66- and bureaucracies such

as the five Field Army systems of the PLA 67 and the state bureaucracy
under Chou En-lai.
There was political conflict within many or all of these "independent
kingdoms" during the Cultural Revolution. There were struggles in
every province and region over the composition of the Revolutionary
Committees,68 and conflict was reported in such bureaucratic units as
the PLA Air Force 69 and the Ministry of Foreign Affairs.70Although
such episodes often displayed earmarks of factional conflict - they were
long, complex, and indecisive, with a succession of contradictory policy
outputs; there seemed to be limits on sanctions imposed on defeated
leaders; they were characterized by personal vituperation and by "wav-
ing the Red Flag to oppose the Red Flag" - we know too little about
events in most of these arenas to characterize them with confidence.
Even if we tentatively conclude that the Cultural Revolution witnessed
factional conflict in many of the intermediate-level arenas of the Chinese
political system, a number of important points remain obscure. Was
political conflict factional in these arenas in the years before the Cultural
Revolution, or did the Cultural Revolution bring out factions by causing
the disintegration of institutions like Party committees and PLA hierar-
chies? How did intermediate-level factions during the Cultural Revo-
lution relate to factions at the centre? Perhaps the most satisfactory
guess is that before the Cultural Revolution such "independent king-
doms" were institutional power bases dominated by factions whose
leaders were participants at the central level. Factions within these local
arenas were merely latent, and intra-province or intra-bureaucracy
conflict was primarily "institutional" (resolved by authoritative institu-
tional decisions arrived at by following legitimate, regular procedures
of decision-making). The Cultural Revolution, initiated at the centre
and necessarily involving the fates of the "independent kings" around
the country, weakened these institutions and gave intermediate-level
factions the chance to emerge and fight for power.
66. On Ulanfu, see Paul Hyer and William Heaton, "The Cultural Revolution
in Inner Mongolia," CQ 36 (October-December 1968), pp. 114-28; on Li Ching-
ch'tian, see "Stalemate in Szechuan," Current Scene, Vol. VI, No. 11 (1 July
1968), pp. 1-13. Frederick Teiwes, in his Provincial Party Personnel in Mainland
China, 1955-1966 (New York: East Asian Institute, Columbia University, 1967),
has shown that provincial leadership was mainly stable during the 1956-66 period
(p. 62), which seems consistent with the " independent kingdoms " view advanced
here, although Teiwes himself concludes that the data do not support the idea of
provincial-level factions (pp. 27-8).
67. Whitson, " The Field Army."
68. Cf. Jiirgen Domes, " The role of the military in the formation of revolution-
ary committees 1967-68," CQ 44 (October-December 1970), pp. 112-45.
69. " Quarterly chronicle and documentation," CQ 32 (October-December, 1967)
p. 196.
70. Melvin Gurtov, "The foreign ministry and foreign affairs during the Cul-
tural Revolution," CQ 40 (October-December 1969), pp. 65-102.
As for the local levels, the evidence for factionalism there seems,
at first glance, to be unambiguous. Beginning in January 1967, the
Chinese press periodically launched attacks on " factionalism, anarchism,
and sectarianism" at the local level. But upon closer inspection, it
appears that what the Chinese press called factionalism is not what is
specified in our model. In the first place, the Red Guard units did not
behave like factions in the sense as defined here. 'They engaged in
violent clashes, sometimes killing their enemies, and seemed to stress
aggressive initiatives rather than defensive tactics. Rather than entering
shifting alliances to block the emergence of strong leadership, Red
Guard units tended to amalgamate in each city into two large armed
camps, each under a strong leader. While it is often unclear why a
Red Guard unit chose to ally with one camp rather than another, there
is no evidence that pre-existing clientelist ties structured the choice of
association; rather, shared class background and ideological position
appear to have been the determining factors. Instead of cycles of con-
sensus and conflict, Red Guard organizations seem to have engaged in
a downward spiral of increasingly violent confrontation.71
Indeed, the few concrete studies of the Red Guards and other units
of "mass factionalism" show that these units were not structured as
factions.72 The most detailed study, that of Bennett and Montaperto,
shows that the Red Guards in one Canton school at first organized as
an elite formal organization consisting of "five-red" students, structured
along the same lines as the school itself, so that the lowest unit in the
Red Guard organizational hierarchy was a group of students selected
from a given class (pan).73 As this structure crumbled, the movement
took on the loose structure of a mass of "friendship groups," each of
which was apparently a corporate unit rather than a faction.'" The
Red Guard units then sorted themselves out into two stable alliance
systems.75
The structure and political behaviour of Red Guard and other "mass
factions" bear more resemblance to those of certain loosely structured
mass social movements 76 than to those of factions in our definition.
The coherence of these "mass factions" of the Cultural Revolution was
based not upon the mobilization of exchange-based clientelist ties for
71. This analysis is based on Vogel, Canton, pp. 323-35; Gordon A. Bennett and
Ronald N. Montaperto, Red Guard: The Political Biography of Dai Hsiao-ai
(Garden City, N.Y.: Doubleday, 1971); and John Israel, "The Red Guards in his-
torical perspective: continuity and change in the Chinese youth movement," CQ 30
(April-June 1967), pp. 1-32.
72. See, e.g., " Mass factionalism in Communist China," Current Scene, Vol. VI,
no. 8 (15 May 1968), pp. 1-13.
73. Bennett and Montaperto, Red Guard, pp. 72-7.
74. Ibid. p. 142.
75. Ibid. p. 186; Vogel, Canton, pp. 329-30.
76. See Ralph H. Turner, " Collective behavior," in Robert E. L. Faris, ed.,
Handbook of Modern Sociology (Chicago: Rand McNally, 1964), pp. 382-425; and
Lewis M. Killian, " Social movements," ibid. pp. 426-55.
the pursuit of political gain, but upon the mobilization of social tensions
and resentments which arose because of the differential access to rewards
enjoyed by different social categories. People allied themselves with
those whom they perceived as belonging to the same category, as dis-
advantaged in the same way and as seeking the same type of reparation.
The difference between the Cultural Revolution "mass factions" and
most other mass movements was that they were highly responsive to
signals from "the centre" and thus lent themselves to manipulation
as somewhat clumsy instruments in the factional struggle being waged
there.77
This does not mean that there was not also factionalism in our
sense in villages and enterprises. It seems probable that the launching
of the Red Guards into factories and suburban areas weakened institu-
tional control and led in at least some places to the mobilization of
factions to pay off old scores or otherwise take advantage of the
opportunity presented. But although this sort of thing very probably
happened, we lack the evidence to prove that it did.
CCP Political History Interpreted in Terms of the Factionalism

Model
This preliminary survey of the evidence suggests that we are justified
in delving further into the implications of the model for Chinese politics.
Let us accept as a working hypothesis that during the Cultural Revolution
the politics of "the centre" were structured largely by factions, that
the politics of the Red Guards were non-factional, and that the
intermediate-level "independent kingdoms" were at first internally
organized as stable institutions, but suffered outbreaks of factionalism.
At this point, let us also take the leap of assuming that if factionalism
was present in the Cultural Revolution it had also been present in
earlier periods. It would be tedious to defend this assumption for each
stage of party history, and unnecessary in an article whose purpose is
to establish the plausibility and not the necessary correctness of a
line of analysis. Although the factionalism argument is easiest to make
for the 1959-72 period - which is indeed stressed in what follows -
our exploratory purpose is best served by sketching the interpretation
that would flow from the maximum assumption: factionalism since
the earliest days of the party.
Let us begin with two well-established facts: that the Chinese Com-
munist Party and its Army have formed throughout their history a
single institutional system with a single elite performing simultaneously
the functions of political and military leadership 78; and that while the
institution as a whole was striving towards power in the Chinese revolu-
77. Cf. Israel, " The Red Guards,"p. 19.

78. Schram, " The Party "; Whitson, "The Field Army."
tion, its elite was from the beginning frequently racked with internal
conflict.79 Studies of intra-Party conflict have focused heavily on the
explicit content of the ideological and policy issues under debate and
have not analysed the way in which the Party/Army elite organized
itself for intra-Party conflict or the way in which such conflict was
carried out. The doctrinalism of these debates, the lack of extreme
intra-Party sanctions, the inability of Mao to exclude his opponents
and their views from the Party, all hint at the operation of factions; this
suspicion is encouraged by the charges aired during the Cultural Revolu-
tion that Mao's enemies had built personal followings and opposed him
throughout the history of the Party. As the Party/Army grew, I would
argue, each major leader played upon his personal network of clientelist
ties to cultivate a personal faction, seeding his followers in important
Party and Army posts 80 and using them and their resources of office,
prestige, and personal connections to further his own standing in the
movement and the adoption of the policies he thought it should follow.
As it happened, the organization of intra-Party debate by factions boded
better for the future of the movement than would some other types of
intra-Party dissension. Through its limitation of sanctions and its shifting
alliances, factionalism bound the movement together and prevented its
fragmentation into permanently hostile rival movements.81 Since the
CCP as a whole existed in a condition of strong threat from the Kuomin-
tang and the Japanese, it could not have survived if it had not limited
its internal rivalries to factional forms.
With liberation in 1949, the organizational sub-structure beneath the
elite increased in size and complexity, and the resources of power avail-
able for control by elite factions increased. However, the central elite it-
self, I would hypothesize, remained factional in its internal divisions. I am
not able to identify the factional leaders, but the obvious starting points
for further research are such names as Mao Tse-tung, Teng Hsiao-p'ing,
Liu Shao-ch'i, Lin Piao, P'eng Chen, P'eng Teh-huai, and Kao Kang.
Factionalism would have been superseded (i) if intra-elite conflict had
ceased, which it clearly did not; (ii) if a member of the elite had mobil-
ized the institutional means (as did Stalin) or the personal charisma to
earn the sole power to resolve conflict, which also clearly did not occur;
(iii) if the Party had split into permanently hostile segments which
moved beyond the rules and resources of factional conflict towards the
goal of mutual extirpation, which obviously did not occur, or (iv) if the
Party had established procedural rules (e.g., voting in the Central Com-
mittee) which all its elite members accepted as the sole legitimate way
79. See, for example, Schwartz, Chinese Communism and the Rise of Mao; John
E. Rue, Mao Tse-tung in Opposition, 1927-1935 (Stanford: Stanford University
Press, 1966).
80. Cf. Whitson, "The Field Army," pp. 3-8 and 21-6.
81. This generalization applies best to the history of the Party after about 1938.
Before that there were several incidents of intra-Party military conflict and of
permanentsplits.
to resolve controversy, and whose decisions were accepted as binding

by all. This final alternative - the institutionalization of the Party as a
conflict-resolving mechanism - may have seemed for a time to have
occurred, but as the 1950s wore on the evidence accumulated that, on
the contrary, factionalism continued.
The factional conflicts were for the most part deeply hidden. Only
occasionally, as in the Kao-Jao purge of 1954 and the P'eng Teh-huai
purge of 1959, did factionalism erupt into public view.82 But the 1950s
and early 1960s were full of circumstantial evidence of factional conflict.
This evidence included the regularly shifting policy lines (the alternation
of radical and moderate policies in agriculture, foreign policy, cultural
affairs and the like),83 the often simultaneous issuing of contradictory
policy signals, and the resulting failure of the system to move con-
sistently in any single policy direction after the initial establishment of
order and the destruction of the landlord class and other opposition
elements. If, despite factional conflict, the allocation of most of the top
posts in the polity remained essentially stable for most of the period,
it was perhaps because many positions, such as Central Committee
seats, were more honorary than substantive. Real power shifted more
rapidly than formal position. Moreover, the struggle for position in the
first decade or so after liberation probably centred on control of newly
liberated regional and provincial "independent kingdoms." 84 It was
a phase in which factional efforts were heavily directed at competing
to consolidate power within the broad polity. The process which William
Whitson has discerned in the PLA, in which each Field Army became
the consolidated power base of one or more factions, while "observers"
from other factions were present in each Field Army (also serving as
"hostages "),s5 was probably widely repeated in other institutions
82. On the Kao purge, see Bridgham, " Factionalism in the central committee,"
pp. 205-211; Teiwes, " A review article," pp. 122-6. On the P'eng Teh-huai purge,
see both of these articles and David A. Charles, "The dismissal of Marshal P'eng
Teh-huai," in Roderick MacFarquhar, ed., China Under Mao: Politics Takes Com-
mand, A Selection of Articles from The China Quarterly (Cambridge, Mass.; The
M.I.T. Press, 1966), pp. 20-33. What is known about these episodes is consistent
with the view that they represent instances of factional conflict.
83. The phenomenon of policy oscillation in the CPR has been noted by many
analysts, and various efforts have been made to explain it. What is striking about
many - not all - of these analyses is that they are non-political, avoiding any hint
that elite conflicts might underlie the oscillations. For some noteworthy attempts
to deal with this problem, see Schurmann, Ideology and Organization, pp. 73-104;
Shinkichi Etb, " Communist China: moderation and radicalism in the Chinese
revolution," in James B. Crowley, ed., Modern East Asia: Essays in Interpretation
(N.Y.: Harcourt, Brace and World, Inc., 1970), pp. 337-73; G. William Skinner
and Edwin A. Winckler, "Compliance succession in rural Communist Cihna: a
cyclical theory," in Amitai Etzioni, ed., A Sociological Reader in Complex Organ-
izations, 2nd ed. (N.Y.: Holt, Rinehart and Winston, 1969), pp. 410-38; and
Alexander Eckstein, "Economic fluctuations in Communist China's domestic de-
velopment," in Ho and Tsou, eds., China in Crisis, Vol. I, Bk. 2, pp. 691-729.
84. Cf. Vogel, Canton, chs. 2 and 3.
85. Whitson, " The Field Army," pp. 15, 23-4.
throughout the system. As control of these bases became stabilized, the

struggle shifted to new posts higher up in the system.
There is no reason to suppose that Liu Shao-ch'i and Mao were
consistently on opposite sides of the issues during these years; quite
possibly in the shifting alliances of elite factions they sometimes
found themselves on the same side of an issue. But in 1959 in the
aftermath of the Great Leap they found themselves opposing one
another. Factional rivalries perhaps at this point reached a new stage
of intensity, with the "independent kingdoms" stabilized under factional
control and conflict now focused on positions at the very top of the
system. Most factional leaders were unified in the wish to subtract from
Mao's too dominant power. The bargain which was struck called for
greater power to pass into the hands of Liu Shao-ch'i and his followers
and allies, and for Mao to move into a weaker position.
The period of 1959-62 was one of contradictory developments of the
sort to be expected in the aftermath of such a factional consensus.
Mao's actual power was weak, and within the small world of the
elite he was subjected to persistent, damaging attack by followers of
one of his rivals, P'eng Chen. At the same time, his charisma was
exploited by the regime to strengthen its legitimacy, harmed by the
Great Leap; the cult of Mao was launched even while - and perhaps
all the more willingly because - Mao was himself in eclipse. The new
Minister of Defence, Lin Piao, in an effort to solidify the control over
the PLA of his own faction, initiated a programme to study the
thoughts of Mao within the Army.86 Whether he was already then
allied with Mao or simply turned to the Mao-study campaign as a
convenient means of consolidating his personal influence against rival
factions in the PLA is unclear. In any case, the campaign made Lin
a natural ally for Mao in the latter's counterattack on his rivals which
began in 1962.
Mao's counterattack of 1962-4, carried out with the help of Lin Piao,
was still within the rules of factional strife. It included the socialist
education campaign, the learn-from-the-PLA campaign, the extension of
PLA influence throughout the country, the escalation of the cult of
Mao, and the pressure for attacks on intellectual cadres in the Party."'
Mao's strategy was to use his national prestige, which his opponents
had been obliged to enhance for the sake of the regime itself, and his
alliance with Lin Piao, to undermine the personal power of his rivals,
founded in the Party and cultural apparatuses. The rivals responded in
86. The developments of this period are reviewed in Bridgham, " Mao's 'Cul-
tural Revolution': origin and development," pp. 2-7. Also see Powell, "The
Increasing power," pp. 45-6 and 62-; Joffe, " The Chinese Army under Lin Piao,"
in Lindbeck, ed., China; and Merle Goldman, " Party policies toward the intel-
lectuals: the unique blooming and contending of 1961-2" in Lewis, ed., Party
Leadership, pp. 268-303.
87. Bridgham, " Mao's 'Cultural Revolution': origin and development," pp.
8-12.
typical factional fashion by stymying, blocking, and evading Mao's

initiatives.88
If Mao had continued to attack his rivals in the factional tradition
of the Party, he would eventually have precipitated another of the
periodic elite crises, a realignment of factions and a redistribution of
power which might have improved his position. But Mao decided
to break the rules, to mobilize new sources of power from outside the
elite. We may speculate that his goal was no longer merely an improved
position for his faction in the elite, but an end to factionalism and its
associated policy oscillations, and an institutionalization of the Party
as an instrument of Maoist will, capable of outliving Mao himself.
Perhaps a sense of impending death aroused at this moment a new
sense of urgency.89 Lacking the organized intra-Party power base of
Stalin, Mao prepared to mobilize student power to drive his rivals out
of office. The students were an attractive choice because of their
historical role as official-topplers in modem China, because of Mao's
faith in the revolutionary qualities of youth, and because of their
suppressed tensions which were capable of exploding into a passionate
political movement upon orders from a respected authority. "Perhaps
for the first time in the history of a Communist state," as John W.
Lewis has written, "a leadership faction [was to go] outside the party
for its main support." 90
If the essence of the Cultural Revolution was, as I have argued, a
student-based Maoist attack on the factional Party elite, then its decisive
period was the year in which the battle was fought over mobilizing
versus demobilizing the students. This year (August 1966 to September
1967), saw a series of efforts by the factional elite to resist Mao's extra-
Party offensive, demobilize the Red Guards, and restore the factional
conflict system, if not the factional alignments, of the first 15 years of
the regime. The ultimate reason for the victory of Mao's rivals was
that Mao had no direct organizational link to the Red Guards, and was
obliged to employ means under the control of the Party centre (the
authority of the Party centre, the hierarchy of Party committees, the
official press) to call the Red Guards into existence and to direct their
activities. Even as he reached outside the factional system for his
instrument of power, he remained dependent upon the Party for the
means of wielding the instrument. In order to achieve his aims, Mao
would have had to retain a shifting majority within the Party centre
as he isolated his enemies one by one for attack by the students. This
88. Neuhauser, " The Chinese Communist Party," pp. 19-25.

89. As Robert Jay Lifton has argued, in Revolutionary Immortality.
90. John W. Lewis, " Leader, commissar, and bureaucrat: the Chinese political
system in the last days of the revolution," in Ho and Tsou, eds., China in Crisis,
Vol. 1, Bk. Two, p. 474. Note that if one views the Cultural Revolution not as a
factional attack on Mao, but as a Maoist attack on factions, one disposes of most
of the objections to a factional interpretation raised by Teiwes, " A review article,"
pp. 126-35.
ultimately proved impossible. His efforts to achieve this shifting majority

account for the rapid series of typical factional cycles of consensus,
conflict and crisis, each finding its response in the alternately rabid and
restrained behaviour of the Red Guards, which characterized the central
year of the Cultural Revolution. Although the shifting alignments within
the Party centre are invisible to us, the rhythms of the struggle can be
perceived.
The first cycle began with the launching of the Red Guards in August
1966, under a compromise document, the Sixteen Points, which ordered
the Red Guards to be both radical and restrained; proceeded through
the counter-mobilization of loyalist student and worker-peasant groups
by party officials; and moved to a crisis of disorder in September which
caused the factions at the centre to agree on the Red Guards' returning
home and on the focusing of attacks on a small number of high Party
officials who had already suffered factional reverses in the preceding
months. In the second cycle, the Maoists launched new attacks centred in
Peking and Shanghai on officials in government organs, and denuncia-
tions were published of some high non-Maoist officers of the PLA;
a crisis of disorder in factories, docks and railways caused a new
decision from the centre in January 1967, for the intervention of PLA
units to restore order. In the third cycle, factions resisting Mao seized
the initiative and, in the "February adverse current," took a number
of steps further restricting the Red Guard movement; the Maoists
countered by publishing new attacks on the "top Party person," signal-
ling a revival of Red Guard activism throughout the country; the result-
ing disorder finally reached such serious proportions that the centre
once again reached a decision in late April or early May 1967, for the
re-imposition of order by the PLA on the local scene. In the fourth
cycle, the Wuhan incident (July 1967) provided the Maoists with a
pretext for demanding an attack on non-Maoist officers in the PLA;
the resulting upsurge of the Red Guards led to the stealing of weapons
and the outbreak of serious fighting; in response, the centre reached
the decision once again to clamp down on the masses; in September
Chiang Ch'ing gave a speech attacking mass factionalism and three
Maoist members of the Central Cultural Revolution Group were
purged.
The defeat of the fourth launching of the Red Guards, I would
argue, settled the fate of Mao's effort to end factionalism at the Party
centre. Key factions at the centre - perhaps an example was the faction
of Chou En-lai - had supported Mao long enought to enable him to
launch the successive Red Guard outbursts; we may speculate that such
factions advanced their own positions as members of other factions
became discredited by Red Guard attacks. Now such swing factions
feared further Red Guard activism, either because they were ready
to consolidate their own positions or because they felt that future attacks
would be directed at them. Mao could no longer garner the support
he needed within the central elite to give him access to the tools neces-
sary to direct the Red Guards. He could no longer divide and conquer
in the Party centre; on this issue, the other factions of the centre were
now united against him and his allies. Although the Cultural Revolution
was to continue in name and with occasional outbreaks of violence, and
although the press was to carry word of struggles between the left and
the right until 1969, after September 1967 there was no longer any hope
of using the Red Guards to cleanse the Party centre of factionalism.
But this did not mean that disorder came to a halt. Conflicts at the
intermediate and local levels of the system had been called into being
by the factional struggle at the top. They now had a momentum of
their own. At the mass organization level, there was the problem of
putting a stop to the downward spiral of vengeance, resentment and
violence, and of channelling aroused student activism into some controll-
able organizational framework. At the intermediate levels, many of the
formerly secure "independent kingdoms" had seen outbreaks of internal
conflict, possibly of a factional nature, and most of these struggles were
still under way. Thus the period from the autumn of 1967 to the autumn
of 1968 was spent in trying to organize the masses in "great alliances"
and the provincial and regional officials in revolutionary committees.
The job of resolving local controversies and restoring order must have
been rendered all the more complex by the fact that the interests of
the factions of the Party centre were intimately bound up in each
decision made. The shape of the future would be determined by the
allocation of power in the regional and provincial revolutionary com-
mittees. These would be the new "independent kingdoms" upon which
future factions would base their power.
The decision on who would control the contested intermediate levels
of 1968 was not solely in the hands of the Party centre (as it had been
in the 1950s), but also depended upon the balance of strength of the
locally contending groups (whether factions or not). The support of
a central faction was not adequate to assure the dominance of an other-
wise weak local figure over the revolutionary committee, while a strong
figure without central backing could bargain for such legitimation with
central factions. The implications of this fact for the personnel makeup
of the factional centre of the Party was enormous. Many of the old fact-
ional leaders had already been humiliated and pushed from power. Now
leaders on the regional and provincial levels found themselves in a
position to bargain with the remaining factions of the centre for
exchanges of support. The result has been called an expansion of the
central-level arena of conflict 91 to include the intermediate-level leaders;
I would interpret it as a partial changing of the guard in which a new
generation of factional leaders 92 was drawn into the central elite to
91. Chang, " Research notes," p. 177.

92. Cf. Klein and Hager, "The ninth Central Committee," pp. 55-6. This is a
" new generation" in the sense that these men had not previously been operating
at the central level. Most or all of them had previously been high officials of the
replace those of the old elite who had fallen during the preceding two
years. The system of factional conflict, however, remained unchanged.
Since 1968 there have been conflicts over whether the Cultural Revo-
lution was good or bad, conflicts over the role of students and conflicts
over the constitution of the Party. These conflicts have been carried out
with many echoes of the Cultural Revolution - attacks on Liu Shao-ch'i
"and other political swindlers," exhortations to rebel, occasional out-
breaks of local violence. There have also continued to be changes
in personnel at the higher levels of the system - e.g., the purges of
Yang Ch'eng-wu, Ch'en Po-ta, Lin Piao and his followers. But I would
argue that, although these conflicts have been intricate and significant,
they indicate not that the Cultural Revolution remains unsettled, but
that China remains in the pre-Cultural Revolution pattern of factional
conflict at the centre.
Some Problems and Implications of the Model

The style of argument I have adopted - development of an abstract
model, and its speculative application to a concrete case - has advant-
ages and drawbacks. Two of the advantages seem especially important.
First, since the Cultural Revolution, our analysis of Chinese politics
has been undergoing a re-evaluation; various explicit and implicit
models are being advanced, and it would be useful if these could be
discussed and compared more systematically than has so far been the
case. Second, since data on Chinese elite politics are scarce, an explicit
model has the heuristic and exploratory function of suggesting one
reconstruction of what the unknown facts may be.
The drawbacks must not be minimized. A model's ability to provide
hypothetical bridges across gaps in the data is earned at the price of
rigorous assumptions that may be only imperfectly fulfilled in complex
reality. The factionalism model posited here assumes a situation where
politicians rely exclusively upon clientelist ties to structure political
action - a condition which must be set if a relatively coherent and
comprehensible model is to be constructed, but which is not likely to be
fully satisfied by many actual cases. A given political system is more
likely to be a mixed than a pure type, both at a given moment in time,
and, even more emphatically, through time, as forms of political
organization change. Systems are usually mixed, but models are pure:
the correspondence is seldom perfect. A model remains useful to the
extent that it provides a relatively accurate diagram of the predominant
dynamics of a system at a given point in time.
Party or the Army or both. They had been followers in factions, we assume, but
now they emerged as leaders in their own right, in the aftermath of the deteriora-
tion of many of the pre-Cultural Revolution factions.
The model simplifies, too, by considering only one of the many sets
of constraints that mould behaviour. It explains the effects on political
behaviour of a given organizational framework, but it does not explain
why men adopt that framework, how long they will stick to it, or
indeed, why they insist on disagreeing with one another at all; it leaves
unanswered such questions as why a certain government office is import-
ant for a certain faction to control, why a certain constitutional or
charismatic legitimation is accepted, and why some revelations discredit
a politician and others do not. A more complete explanation of behaviour
in any political system would have to take into account not only what
might be called the " organizational-structure" constraints on behaviour
which have been discussed in this paper, but also cultural constraints,
institutional constraints, and ideological constraints.
Particularly important in any fuller explanation of CCP elite politics
would be the interplay between organizational and ideological con-
straints. There is certainly no contradiction betwen commitment to Maoist
ideology and participation in factions. Although "faction" as such is
deplored, factions may simply be conceived as ways of co-ordinating
the efforts of like-minded colleagues to achieve goals in which they
believe. As the model argues, factional struggle occurs within the
context of a broad ideological consensus on goals and methods. It would
therefore be a mistake to identify the factionalism model with a crude
power struggle theory, if the latter assumes that leaders are cynical in
their ideological statements. But it would be equally foolish to believe
that in China alone men's perspectives on ideological and policy issues
are not influenced by their individual political vantage points. The
occasional impression that this is so may be the result of our knowing
so much more about the issues than about the vantage points.

A Factionalism Model For CCP Politics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Factionalism Model For CCP Politics

Uploaded by

Copyright:

Available Formats

A Factionalism Model for CCP Politics

Author(s): Andrew J. Nathan

Until the Cultural Revolution, the predominant western view of con-

from the "unorganized opinion group" model are interpretations which

and "Mao's Cultural Revolution: the struggle to consolidate power," CQ 41

which at various times may have encompassed disagreements about the

6. See, for example, Richard M. Pfeffer, "The pursuit of purity," Problems of

The Structure of Factions

Such ties include patron-client relations, godfather-parent relations,

17. Cf. Foster, " The dyadic contract."

The third possibility occurs in an oligarchic or relatively small-scale

L Simple factions and

forgo formal definitions and systematic comparative analysis. What

and followers. Structurally, the faction is articulated through one or

The Structural Characteristics of Factions

24. Each structuralcharacteristicdiscussed is not necessarily unique to factions

it is likely to become.28 This is one of several structural characteristics

number of subordinate leaders it contains and the farther removed

The Characteristics of Factional Politics

31. Cf. Amitai Etzioni, Modern Organizations(Englewood Cliffs, N.J.: Prentice-

telist tie, their one-to-one communications structure, and their tendency

3. When a faction does take a political initiative (which it does only

the factions. To attempt to take action without first achieving such a

A second set of propositions is based upon the fact that factions

38. Cf. Leites, On the Game, pp. 48-9.

A third set of propositions concerns the size and shape of the

to the system comes from within, from a factional leader attempting to

factions for other forms of organization (often clientelist political

Does China Have Factional Politics?

on Wu Han, or the launching of the Red Guards, were relatively rare,

attack was certainly a feature of the Cultural Revolution, and, as I have

Ulanfu and Szechwan under Li Ching-ch'tian66- and bureaucracies such

CCP Political History Interpreted in Terms of the Factionalism

77. Cf. Israel, " The Red Guards,"p. 19.

to resolve controversy, and whose decisions were accepted as binding

throughout the system. As control of these bases became stabilized, the

typical factional fashion by stymying, blocking, and evading Mao's

88. Neuhauser, " The Chinese Communist Party," pp. 19-25.

ultimately proved impossible. His efforts to achieve this shifting majority

91. Chang, " Research notes," p. 177.

Some Problems and Implications of the Model

You might also like