After reading Pertti Laulajainens article in SPS No. 1, 1984, I would like to comment on some of his conclusions. I find it necessary to do so both from a theoretical and an empirical point of view. In my opinion Laulajainen still fails to look at the interrelationship between electoral and organizational strength, political tradition, and socioeconomic structure. He fails to put the existence of organization in the Finnish case into a wider causal context (Selle 1982, 191, 214). When reconsidering my own approach to the Norwegian Communist Party in the immediate postwar period, I am even more convinced that this is the case (Selle 1983).

To me the most important weakness of Laulajainen's approach is that it is almost exclusively an empirical one. Correlations that are in themselves important and informative are interpreted (at least indirectly) into a causal language without any sufficient discussion of the relationship between the theoretical and empirical levels. Both of us have found a close relationship between political tradition, organizational strength and communist voting. However, such a finding is a long way from proclaiming that these variables explain the communist support, because it would mean that political tradition and organizational strength are the main causes for that support. I will argue that such a statement is more than problematic.

I will begin with some methodological remarks. When Laulajainen claims the influence of socioeconomic factors to be secondary, this is mainly an a priori starting point. The data — numbers in 'big' socioeconomic sectors — and the method used — correlation analysis — are too broad to reveal anything finite about the relationship between socioeconomic factors and communist voting behaviour. More detailed microstudies might have given quite different results.

Further, the main objective of Laulajainen is to analyse the communist defeat in 1948. The quarrelsome question is if a decline from 23.5 to 20% really is a great defeat, electorally speaking. Is it at all relevant to argue against a sociostructural approach analysing such relatively small fluctations? The Finnish election of 1948 was held in a politically very cold and special period of European history. In almost every country in Western Europe the communist parties had a dramatic decline in the second election after the war. In the next Finnish election the party was back to 21.6% and stayed above the 20% level for the whole period up to 1970.

I also find problematic the way Laulajainen uses the variables of electoral and industrial change. Electoral change is always very difficult to explain by correlation analysis — much more so than when explaining general strength. Even if one correlates electoral change with general industrial strength, one should for statistical reasons not be surprised by the lack of a strong relationship between the variables. This does not mean that such factors are without any influence in the real world. In Laulajainen's article it is even difficult to find out what kind of change one is talking about. Is it change in real numbers, percentage differences or some kind of standardized measurement? The answer to this question will have consequences for the correlations found.

To claim that some of the change in communist support between 1945 and 1948 can be explained by organizational factors together with the impact of political tradition seems plausible, but to make this into a general model (or theory) of communist voting in Finland is something quite different. Since the same model is underlying other publications of Laulajainen (1979a, b; 1983), articles trying to explain communist support over a longer period of time, I will discuss more generally the limitation of such a model.

My perhaps most important objection has to do with the way political tradition and organizational strength are used. First of all, I do not think one should use the same strategy to explain short time fluctations versus real change over time. Next, the electoral support between two elections is never random. In real life one will always find some kind of stability at an aggregate level. The important but very difficult problem is to decide how strong a relevant correlation must be. We don't have that many theories at an interval level in the social sciences. That one finds statistically significant correlations should not surprise anybody. Much more interesting are of course differences between regions and change in such correlations over time. This argument also implies that if one operationalizes political tradition as previous voting strength, for instance using the first election year of the party, one should not be surprised at finding a substantial covariation. In my opinion there is already at this stage a touch of tautology in the approach. More about that later on.

Even more complicated is the use of the organizational variable. If one

wants a high explained variance, this variable is, together with political tradition, an ideal one. However, to build an organization as strong as possible is an integrated part of communist ideology and organizational theory. To me this means that the pure existence of such an organization can be seen as an important indicator of general communist strength. In the Norwegian context, for example, there was an enormous difference in support for the Communists between communes with and without an organization, and this even increased over time (Selle 1982, 203). Organizational strength and behaviour do have consequences for electoral behaviour. The main problem empirically is how to separate from each other what we can call the reflection of the general strength and the direct mobilizing effect of organizational characteristics.

Laulajainen does not discuss the fundamental theoretical and methodological questions which arise. These problems may be one of the main reasons for the ambivalent attitude towards the meaning of party organization in the extensive literature on voting behaviour. In the 'cleavage literature' (Rokkan & Lipset 1967) the reflection perspective dominates. To understand the political parties and politics more generally you have to understand the social cleavage structure in a country. Not even the very interesting approach on political mobilization by Cameron (1974), making the organization itself into the fundamental agent of mobilization, takes into consideration the question of how to separate reflection from the mobilizing effect. To really grasp the process of mobilization and thereby the development of social movements, I find such a separation of the greatest importance. However, I do not see any operationalization of organizational characteristics that can solve this problem empirically in a satisfactory way. In addition, a large part of the relevant organizational data on organizational strength and behaviour is seldom available, or if it is, only rarely reliable.

I agree with Laulajainen that party activists play a central role in the mobilization process and therefore that the often used member/voter ratio for a party is a very limited measurement if one wants to analyse the importance of organization. However, the relationship between activists and: voting results does not have to be unambiguous. Internal quarrels and different political views on important political questions can move a strong (in members) and active organization into passivity with time for little external activity. Furthermore, research from Great Britain has questioned whether the activity level and the activists always have a mobilizing effect. Research on the relationship between Labour's activists and potential voters shows no 'symmetry' in attitudes (Kavanagh 1982). This means that we here refer to situations where activists actually have a demobilizing effect.

To sum up, using organizational behaviour as an explanation of communist
voting can to a large extent be a tautology. It is the mobilizing effect of

organizational behaviour that is of interest. However, we do not have any
good method of separating reflection from mobilization.

In my opinion this discussion questions the interpretations made by Laulajainen. Furthermore, using mainly organizational strength and political tradition as explanatory variables, Laulajainen does not take into account the trivial truth that not all kinds of people in all areas have the same probability or potential for voting communist. A good model of communist voting behaviour must discuss the opportunity structure more thoroughly (Tilly 1978). One must discuss the impact of ecological characteristics and how these are 'filtrated' through organizational structure and behaviour and through individual status and attitudes. This opportunity structure will of course change over time and set up the borderline for communist mobilization in different periods.

A period of communist mobilization that shows the importance of this perspective and at the same time clarifies some of the limitations of Laulajainen's approach is the period just after World War 11. We have already said that in most cases one will find a strong connection between organizational strength and communist voting. However, in what we can call general mobilization periods of communism, such as just after World War 11, one will in many countries find strong communist support also in areas without any communist tradition or communist organization. In the Norwegian case this communist expansion took place in areas of previous social-democratic tradition. This means that one can talk about a gradual radicalization. In this particular period of Norwegian communism, organizational strength was not at all a necessary condition for communist strength. However, it was always a sufficient condition (Selle 1983). Further, in these new areas of communism the party quickly tried to build up an organization to stabilize its support and make communism into a political tradition. In order to understand processes like this, Laulajainen's approach is too limited. There is still a lot to learn from the traditional communist research of Allardt (1970) and Lipset (1968), and the rational main perspective on communism maintained by Korpi (1971).

Altogether what we have found is a blurred and dubious relationship between Laulajainen's main variables — organizational strength, political tradition and communist voting behaviour. In another publication Laulajainen argues as follows: 'It is particularly organizations that have maintained the political tradition and electoral support of the communist party' (1979b, 178). In another passage he calls the relationship between political tradition and organization Siamese twins (1983, 221). It is unclear what explains what and this fact must have consequences when using a causal language. In my opinion this has consequences for causal modelling on a more general level as well. We have a situation where at least part of the political tradition variable and in most cases an extensive part of the organization variable belong to the same

phenomenon which we try to explain as our dependent variable. As long as the possibility of separating the mobilizing effects from what we have called the reflection aspect of the same variable is lacking, it becomes very difficult to interpret the results from such a causal model. We actually do not know how much is tautology and how much is explanation.

This is also the case in the relationship between political tradition ?nd organization. Generally one finds a very strong correlation between political tradition and organizational strength. This very close and dynamic relationship means that in most cases they measure partly the same phenomenon. Political tradition is of course more than previous voting behaviour. It has to do with both the organizational tradition and the voting tradition as well as with political structure and political integration at a more general level. When operationalized as previous voting strength, this means that only one of several indicators measures the phenomenon. These are, however, often the only reliable data we have.

I for one see political tradition as a much wider concept than organization, and perhaps even more problematic to handle empirically. However, I will strongly argue against Sten Sparre Nilsson (1983a, b), who wants to reduce political tradition into organizational activity. Both Laulajainen and myself have found strong correlations between political tradition and voting behaviour in the immediate postwar period. We also uncovered that this connection vas strongest in the period of decline. The communist decline in the late 1940s was greatest in the 'new' communist areas without communist tradition. After World War II the parties were much stronger in their traditional strongholds than anywhere else. This was so in spite of the fact that the communist party was outlawed in Finland in the interwar period and that the communist party in Norway was without any substantial support in most communes in the 1930'5. In my opinion it is impossible to explain the communist resurrection at an aggregate level in 1945 by the daily work of communist organizations during the 1930'5, as Nilsson seems to think (1983b, 195).

Political tradition goes far deeper and is also a broader concept than organizational behaviour, and there is nothing miraculous about the fact that the new rise of the party was strongest in areas of communist tradition. That some organizations continued to exist throughout the 1930's cannot explain satisfactorily the general increase. To me it makes sense to say that communist political traditions survived in the 1930's and stayed latent in spite of the absence of organizations to maintain them (Selle 1982, 210). After World War II this tradition was a very important part of the opportunity structure of the party.

Using the Laulajainen approach, one will always get a high explained
variance but a far from sufficiently comprehensive model of communist
voting behaviour. This high amount of explained variance is of course in

itself substantially important. However, this faet should not stop one from innovative theory building concerning communist voting. What we need is a real clarification of the relationship between the variables used, and the relationship between these variables and sociostructural and individual characteristics. The connection between causality (the theoretical level) and association (the empirical level) has until now not been discussed sufficiently. In my opinion what we should do is to try out a combination of extensive non-recursive models and detailed time series analysis. However, as far as I can see, we will still be in an area of research where the relationship between our theoretical model — showing causation — and our measurement model — showing association — remains problematic.

REFERENCES

Allardt, E. 1970. 'Types of Protest and Alienation', in Allardt, E. and Rokkan, S., ed., Mass
Politics. N.Y.: Free Press.

Cameron, D. 1974. 'Toward a Theory of Political Mobilization', Journal of Politics, vol. 36.

Kavanagh, D., ed. 1982. The Politics of the Labour Party. London: Allen & Unwin.

Korpi, W. 1971. 'Working Class Communism in Western Europe', ASR, vol. 36.

Laulajainen, P. 1979a. 'Some aspects of the Division of the Finnish Working Class Movement
after the Civil War: Research note', SPS, vol. 2, New Series, no. 1.

Laulajainen, P. 1979b. Socialidemokraatti vai Kommunisti (English Summary). Mikkeli: Itä
Instituutti.

Laulajainen, P. 1983. 'Vänsterradikalismen i finska Lappland', in Lindstrøm, U. and Karvonen,
L., ed., Nordkalotten i Politikken. Lund: CWK Gleerup.

Laulajainen, P. 1984. 'Organizational Factors in the Defeat of the Finnish Communist Party',
SPS, vol. 7, New Series, no. 1.

Lipset, S.M. 1968. Det politiske menneske. Oslo: Gyldendal.

Lipset, S.M. & Rokkan, S. 1967. 'Cleavage Structure, Party Systems, and Voter Alignments:
An Introduction', in Lipset, S.M. and Rokkan, S., ed., Party Systems and Voter Alignments.
N.Y.: Free Press.

Nilsson, S.S. 1983a. 'Kommunismens store tid i Norge', Aftenposten, April 6.

Nilsson, S.S. 1983b. 'Sunt vett, avansert analyse', Statsvetenskaplig tidsskrift, no. 3, 1983.

Selle, P. 1982. 'The Norwegian Communist Party in the Immediate Postwar Period', SPS,
vol. 5, New Series, no. 3.

Selle, P. 1983. Norges Kommunistiske Parti 1945-1950. Bergen: Universitetsforlaget.

Tilly, C. 1978. From Mobilization to Revolution. Massachusetts: Addison — Wesley.