paramount importance in explaining or predicting changes in ideology. To some extent, the small number of references is also a consequence of the fact that a rigid definition of ideology is used here. Ideology, in this context, refers to any of the various "isms" and "thought." Types of ideology are classified according to the charismatic leader who is associated with the particular "ism" or "thought." For example. Maoist thought is classified as Chinese, while Leninism is classified as Russian. General ideology does not refer to any specific individual. While it is plausible that an ideology may, at some point in its development, transcend the confines of national boundary and become universal, it is maintained that, at least in Communist China, the nationalistic element in different ideologies can still be distinguished from their universal meaning. For example, Marxism-Leninism is still foreign and Russian, while Maoist thought is Chinese, in spite of their presumed universal applicability.

Analysis: Of the eighteen references to ideology in the May, August and December 1957 issues of the People's Daily, eleven referred to Chinese ideology, three to foreign ideology, and four to general ideology. The fact that Chinese ideology predominated supports Schurmann's hypothesis concerning decentralization policies in the latter part of 1957 as well as the fact that these policies coincided with a conscious attempt to develop an economic model which was not merely a mirror image of the Soviet economic model. Change in emphasis on different types of ideology was but one of the many changes which occurred in 1957. Schurmann pointed out that during the first fourteen years of the Chinese People's Republic, the leadership applied two distinct developmental strategies, one for each of the two five-year plan periods. The terminal year of the First Five-Year Plan, 1957, was crucial not only for economists, but also for sociologists and political scientists since many social and political changes occurred. In the present discussion, we are only interested in the ideological dimension. A detailed examination of the hypotheses raised by Schurmann will be treated elsewhere.

In a strict sense, our data on Chinese versus foreign ideology do not enable us to test the hypothesis that there was a shift in ideology accompanying the transition from the First to the Second Five-Year Plan period. Schurmann postulated that China had gone through two radical phases, one during the First Five-Year Plan period when the Chinese Communists tried to repeat the Soviet experience of industrialization, and the second during the Great Leap Forward when they used their own mobilizational techniques to try to achieve an economic breakthrough.8 Our corollary to this general postulate is that the transition from the First to the Second Five-Year Plan period was also accompanied by a change in ideology. To test this corollary, data from the First Five-Year Plan period are needed. But until data for 1952 are available and analyzed, such a comparison is not possible.

Since the Foreign Ideology Index is not made up exclusively of

categories on Soviet Ideology, it would be interesting to look at the statistical variation of Soviet ideology categories in comparison with those for Chinese ideology (see Table I).

This table reveals several striking facts. First, there was only one reference to Soviet ideology in the entire sample. This discovery seems to be relevant in the interpretation of the Sino-Soviet conflict. Although a general break in Sino-Soviet relations did not occur until years later, it has been suggested that changes in Sino-Soviet relations began as early as 1956 in connection with the problems of de-Stalinization and with crises in Poland and Hungary. In the early period, according to one authority, it was very much a conflict over the question of the "correct" ideology, although the fact that economic problems were involved-including the stopping of Soviet loans and the disillusionment with the Soviet economic model-cannot be denied. Another striking fact in Table I is the variation in references to ideology by month; this variation seems to be related to the timing of the "Hundred Flowers" campaign and the anti-rightist campaign which followed. MacFarquhar stated that anything printed before June 8 in the People's Daily can be taken to have been straight reporting, June 8 being the day on which the People's Daily initiated the counter-attack. Our data show that in May there were nine references to ideology, for August there was none, and for December there were 3. To interpret this variation, let us divide the general event period covering the "Flowers" campaign into four sub-periods:

1. February to April-period of anxiety following Mao's speech on contradictions in February 1957.

2. May and part of June-blooming and contending period during which there was a liberalized policy on ideological debates.

3. Part of June, July and August-period of the anti-rightist campaign. 4. September to December-aftermath.

Our data cover sub-periods 2, 3, and 4. During the period of liberalization on communication, the number of references to ideology was largest.

TABLE I

REFERENCES TO CHINESE VERSUS SOVIET IDEOLOGIES, BY MONTH*

*Figures in parentheses represent number of references after adjustment for variation in sample size (adjustment ratios for May, August, December, equal to 1.30, 1.07, 1.00 respectively). Tests of significance are based on adjusted figures. **Kolmogorov-Smirnov 1-sample test for row or column difference.

During the most intensive period, the anti-rightist campaign, there was no reference to ideology at all. Finally, there was a limited reemergence of ideological discussion in December. In May, many individuals openly attacked the regime and its ideology; study groups (Hsiao-tsu) were formed and mass mettings were held to discuss alternative ideologies and practices. But after June 8, intellectuals, with the exception of the students, became more apprehensive and reserved. After July 20, when a government resolution was passed whereby in future every student has to produce proof of ideological reliability before he can obtain a job, the students also became more quiescent. It was only toward the end of the year that the situation began to calm down, and even some of those who had been singled out for special reprimand turned up again in public life, although shorn of all influence. 10

Thus there tends to be a minimization of communication about the source and focal point of failure during a crisis situation, especially the role which the elite played in contributing to this failure. However, while there is a minimization of communication about the elite's role in policy failure, there is also a maximization of statements blaming "others" for this failure. "Others," in this sense, may refer to individuals groups, nations or even abstract ideas." There was a tremendous increase in references to the campaigns against counter-revolutionaries (sufan) and references to antirightist campaigns in August 1957.

The findings on minimization and maximization of communication, in terms of the number of references in the People's Daily, may be expressed in the form of a general hypothesis:

If a policy fails, the system tends to minimize communication about the source and focal point of this failure, minimize communication about the role the elite played in contributing to the failure, but maximize communication in blaming "others" for the failure.

In the case of ideology, the "Flowers" campaign was formally endorsed by Mao; moreover, the effects of the campaign were viewed as extremely unfavorable for the entire system; consequently the restriction on discussion about ideology was so great that there was no reference to ideology at all in the entire sample for August. In December, the threat of the consequences of general discontent became less pronounced, as evidenced by the reappearance of a limited number of references to ideology.

It is by no means assumed that the data on ideology conclusively confirmed the minimization hypothesis mentioned previously; a confirmation of this hypothesis requires testing against a variety of different situations, different types of policies, and different time periods. A comparison of the references to the ideology of Mao, Engels, Stalin, and Lenin reveals three important facts. First, nine of the ten references to ideology of individuals concerned Maoist ideology. Second, there was no reference to ideology of any individual either in August or December. The last fact, when combined with our finding that there was a limited reappearance of references to ideology in December, means that this reap

pearance was limited to general ideology; apparently ideology of specific individuals was still too sensitive an issue to be discussed in the press.

THE POLITICAL PARTY INDEX

The political Party Index consists of three categories: The CCP (Chinese Communist Party), Communist parties of other countries, and workers' parties of other countries (see Table II). First of all, looking at the total column for all months, we find that of the 209 references to parties, 186 referred to the CCP. On the basis of our limited data, we must accept the validity of the postulate that the Chinese Communist spokesmen consider the Chinese Communist movement as part of the world Communist movement. On a long-term basis, this postulate is probably true. It was certainly not accidental that Teng Hsiao-Ping, in an article written for the Pravda in celebration of the 10th anniversary of the People's Republic of China in 1957, stated that:

We stand for proletarian internationalism as opposed to all kinds of bourgeois reactionary ideologies of big-nation chauvinism and narrow nationalism. The modern revisionists, as represented by the Yugoslav ruling clique, use bourgeois nationalism to oppose proletarian internationalism, use the nation as cover to oppose international solidarity and have fully become an echo of imperialism. . . .12

In fact, Teng's article is a reflection of a more general ideological stand which emphasizes the interdependence of the Chinese Communist and world Communist movements on a long-term basis.

While, on the one hand, we can speak of a long-term ideological stand, at the same time, short-run fluctuations may occur due to practical considerations made necessary by such problems as domestic crises. In our sample for the latter part of 1957, for example, there was a preponderant number of references to the CCP as opposed to references to foreign Communist and workers' parties. The nature of such short-run

*Figures in parentheses represent number of references after adjustment for variation in sample size (adjustment ratios for May, August, December, equal to 1.30, 1.07, 1.00 respectively). Tests of significance are based on adjusted figures. **Kolmogorov-Smirnov 1-sample test for row or column difference.

fluctuations is worth exploring in future research, although it will be, undoubtedly, a tedious and costly job. In order to distinguish between short-run fluctuations and sudden changes, one must also have detailed knowledge of the long-term developmental patterns. Methodologically, one should study long-term patterns in detail qualitatively and, at the same time, select a large sample of documentary materials covering a number of years for quantitative analyses. What we have said for timevariation holds, of course, also for regional variation.

Looking at the total number of references to all parties, we found that, in comparing the distribution for the three months, our hypothesis on minimization in communication in periods of stress is again confirmed. There were 122 references in May, but the number was reduced to 8 for August, and finally there was a reemergence of references to political parties in December, the number being 79. Looking at the distribution by month for each party, we found that for the CCP the restriction in communication hypothesis is also confirmed; with 112 references for May, 4 for August, and 70 for December. The number of references to foreign Communist parties or foreign workers' parties during the three months were too few in number to warrant speculation.

ADMINISTRATIVE ORGANIZATION AND DECENTRALIZATION IN 1957

If the proposition that decentralization occurred late in 1957 is valid, then a comparison of the three time periods in our sample with regard to a series of indicators about decentralization in its various manifestations should provide evidence that in the later months of 1957, decentralization was much more widespread than the earlier months of the year. Since our sample consists of three months in 1957-namely, May, August, and December-we will be satisfied if, in fact, there were more indicators of decentralization such as a decrease in emphasis

TABLE III

on cen

REFERENCES TO LOCAL AND NATIONAL AUTHORITATIVE
AND ADMINISTRATIVE ORGANIZATIONS, BY MONTH*

*Figures in parentheses represent number of references after adjustment for variation in sample size (adjustment ratios for May, August, December, equal to 1.30, 1.07, 1.00 respectively). Tests of significance are based on adjusted figures.

**Kolmogorov-Smirnov 1-sample test for row or column difference.