Copenhagen School of Economics and Social Science and
The University of Texas

Recent years have witnessed an upsurge in research on the structure and interaction of elites at national levels in a number of societies (Mills, 1956; Guttsman, 1965; Porter, 1965; Zapf, 1966; Lijphart, 1968). In general, however, the analysis of labor union elites has been a somewhat neglected area in the study of national elite systems, and insofar as researchers have at all concerned themselves with labor union elites, they have reached conclusions that have differed rather remarkably. In the United States, Mills studied labor leaders and concluded some years later, "Well below the top councils, they are of the middle levels of power" (1948; 1956: 263). Keller, echoing Mills, classified them as a "segmental" rather than a "strategic" elite (1963: 306). In Canada, Porter placed them "on the fringe of the confraternity of power" (1965: 540). Somewhat by contrast, researchers in Europe have tended to view labor leaders as being more central to elite processes. In Great Britain, Guttsman saw them as comprising one of sixteen elite groups but was unable to locate them on a hierarchy of elites (1963: 328, 369). In the Netherlands, Lijphart identified labor leaders as key members of three of the four elite groups dominating the pillars of Dutch society (1968: 59—68). On the other hand again, Zapf, in constructing an index of "social distance" between various institutional elites in Germany, concluded that labor leaders were farthest removed from the other elite groups and only related to the socialist party elite (1965: 199).

At least one feature common to the Scandinavian societies makes them particularly interesting from this research perspective. We refer to the strategic positions of labor union leaders in the elite structures of these countries. On the face of things, it appears that labor union leaders in Scandinavia occupy posi

* This is the initial report of a research project supported by NAVF (Norwegian Research Council for the Sciences and Humanities), the US Educational Foundation in Norway, the University of Texas at Austin, the American-Scandinavian Foundation, and the Institute for Social Research in Oslo. Director Erik Rinde and his staff at this last-named institute were especially helpful during all phases of the project. In addition, we appreciate the assistance of Erik Arthur, Lisbeth Higley, Else Stuenaes, and Sally Cook. Further, we are deeply indebted to many officials of the LO unions and headquarters for their active cooperation in this research. The authors alone are responsible for the interpretations and conclusions contained in the article. They participated equally in all phases of the research and are therefore listed alphabetically.

tions considerably more powerful than do their counterparts in other Western societies. Summarizing their role in Sweden, Elvander writes, "That LO is the organization which has the greatest influence [of all interest organizations] is beyond all doubt" (1969: 292).

In this article we present data on the leaders of the Norwegian Federation of Trade Unions (Land sorganis asjonen i Norge, or LO) which permit a tentative evaluation of their position in the more general system of elites in Norway. A more definitive treatment of so large a question is beyond our purposes here, depending as it does on making similar evaluations of other elite groups in Norwegian society and of how they jointly interact with the labor elite. In this article we are concerned only with a reconnaissance of the ground that must be covered as regards the labor elite.

There is good reason to believe that Norwegian labor leaders have great influence not only on economic affairs, but also on decisions related to education, social welfare, urbanization, environment, and a host of other questions not strictly economic in nature. Intertwined with the Labor Party, the LO is part of a major power system which has had a far-reaching impact on all Norwegian institutions. From its humble birth in 1899, when it comprised seven small unions with a total membership of about 5,000, the LO has experienced dramatic growth and now includes all major blue-collar and many white-collar unions with a total membership of about 578,000. But membership figures alone do not explain the stature of LO as a power bloc in Norwegian society. The loyalty and solidarity of the membership and the cadres, the centralized organizational network, the linkages to other power systems, the legitimacy of LO's structure and functions, must all be considered in an evaluation of labor power in Norway.

The institutionalized participation of big economic interests such as labor, farming, fishing, industry, and shipping in political decision-making is a very important feature of Norwegian politics. The organizations articulating these interests are not just pressure groups or parapolitical units on the boundaries of the political system. On the contrary, they must be regarded as extremely powerful actors within the political system so that very few "authoritative allocations of values" take place without their involvement in one way or other. In Norway's multi-party politics there is ample opportunity for a meshing of these organizations with the political system. At least three of the five major parties are connected rather directly with specific economic interests. Of these clearly the most elaborate connections are between LO and the Labor Party. To a great extent both units comprise one unified power bloc symbolized by the term Labor Movement. Nevertheless, the very scope of this bloc, the increasingly differentiated functions of union and party, and the fact that the party is subject to the caprices of electoral victory or defeat, while LO's power is stable, makes it possible to view the two as separate systems. Each has a distinct organization, leadership group, set of functions, and distinct internal problems.

In defending his research focus on the lower echelons of American political
parties, Eldersveld argued, "The critical action locus of the party structure is at

its base" (1964: 10). The opposite formulation is the basic rationale for studying leadership groups in Norwegian unions: The critical action locus is at the top. This is so because of the unique centralization of power in these unions, because the very success of Norwegian unionism has created rank-and-file apathy against which the leaders must constantly struggle, and because of a "two-tier" system of decision-making wherein institutionalized bargaining between leaders of major organizations and elected officials takes place in a "constitutional vacuum" quite apart from normal legislative and cabinet processes (Rokkan, 1965: 105 ff.). Thus in LO, goals and strategies are to a large extent set at the top, and policies are implemented through an extended network of full-time and part-time officials. In this situation, the composition and training of the leadership element raises problems of fundamental significance.

In spite of the crucial importance of labor power and leadership in Norwegian society, there has been very little systematic research on the structure and functioning of various parts of the labor movement. True, various surveys have provided valuable information about the backgrounds and attitudes of labor voters and union members (Valen & Katz, 1964; Fivelsdal, 1964; Lysgaard, 1965). But the higher echelons of the labor movement have remained a blank spot on the sociological map of Norwegian society. In what follows we ask a number of questions about the persons who inhabit some of this unknown region. Specifically, we ask, (1) Who are they and what are some of their demographic characteristics? (2) What kinds of social, educational, and occupational backgrounds do they share in common? (3) What is the nature of the career system by which they reached leadership positions? (4) As leaders of LO, how are they linked to the political party system, and especially to the Labor Party? (5) What are some of the ways in which they participate in national elite processes in Norway?

Elite, Apparatus, and Delegates in LO

Successful performance of basic union functions such as bargaining with employers, organizing, socializing and mobilizing members, and administering the formidable economic resources of a giant organization such as LO requires an extensive leadership structure. A complex ordering of cadres comprising elected and appointed, full- and part-time personnel is a necessity. The starting point for a study of those persons who guide and direct the labor union movement is a specification of the size and organizational characteristics of this leadership structure. (To avoid confusion we want to stress that when we hereafter employ the terms "LO leaders", "labor leaders", or "union leaders" without further qualification, we will be referring to all the persons situated in the several categories of this structure which we specify below.)

The basic organizational principle in LO is that of industrial unionism, introduced in 1923. However, even today there are exceptions to this principle. One exception is the continuation of half a dozen small craft unions, (e.g. Electricians Union, Printers Union), organizing about 5 °/o of all LO members. Another

exception is groups of white-collar workers who have been permitted their own unions, (e.g. Policeman's Union, Office Workers Union). These organize approximately 10% of all members, although many more white-collar workers are organized in hybrid, mixed-collar unions. At the moment there is much discussion in LO about a fundamental restructuring of the entire organization, and one can expect major changes in the next few years.

Because some readers may be unfamiliar with the current formal organization of LO, we present in Figure 1 an outline of its principal authoritative bodies and their relations to one another. Due to space limitations, however, our comments on this formal structure must be brief in the extreme.¹

Constitutionally, the basic unit in each union is the union branch or local (avdeling). The locals are composed of clubs (bedriftsklubber) located at work units within each local's area of jurisdiction. Locals are themselves aggregated in district committees (stedlige styrer) set up by each national union. Above these district committees is only the national union (forbund) itself, presided over by a union executive committee (forbundstyre) which is elected by the national convention (landsmøte) which each union holds every third, fourth, or fifth year. The national convention or, in some cases, the membership also elects a national committee (landsstyre) which has the responsibility of overseeing, usually on an annual basis, the actions of the union executive committee during the

years in which there is no national convention. In any case, the union executive
committee, averaging 10.6 members in 1967, is the effective seat of power
within each union. Subordinate to it is the union's full-time administrative staff.

Comprising an important superstructure above and beside the individual unions is the LO network, in effect a federal structure tying together the member unions. The supreme body in this network is the quadrennial LO Congress, consisting of 315 voting members. During the long periods between Congresses the highest authority is the General Council (Representantskapet), consisting of 135 voting members and usually not meeting more than twice a year. Immediately below the Council is the LO Secretariat, the single most important decisionmaking body in the entire system. The Secretariat consists of fifteen members and nine alternates, all of whom are elected by the Congress under a rotation system that attempts to give the large unions regular places without entirely excluding the smaller unions. With but one or two exceptions, the union representatives on the Secretariat are union presidents. They are joined by the president, vice-president, treasurer, and first secretary of the LO network, with the second and third secretaries sitting as alternates. All six LO officers, incidentally, are elected by the Congress. The Secretariat is basically a committee of the leaders of the major unions. Meeting weekly, it is vested with authority to set dues payments from the unions to the LO network, to set the broad framework in which wage demands are made, to speak to the public on behalf of LO, to enter into negotiations with cooperating or competing interest groups as well as with the government, and, not least, to supervise the activities of the network's central and district administrative offices. Thus the Secretariat is the primary mechanism by which the power of the unions is centralized and brought into play.

These then, very briefly described, are the most important components of the LO system in Norway. The system has two features that deserve special attention. The first is that of double centralization. Individual unions in LO are themselves highly centralized at the national level. For example, nearly all bargaining is carried on at the national level and the locals undertake little independent action. Beyond this centralization, the individual unions have delegated considerable power to the LO network, and especially to the Secretariat. This system of double centralization makes the Norwegian trade union movement, as one observer puts it, one of the most centralized in the western world (Dorfman, 1966: 51). Second, it will be seen that the structure is representative throughout and has obviously been established to approximate democratic forms and processes as close as possible. But even our superficial description suggests that the unwieldy sizes of many of the bodies, the infrequent meetings they hold, and the absence of expertise their compositions indicate, make day-to-day reliance on them an invitation to organizational disaster. It is therefore necessary to examine the more informal leadership structure of LO in order to better understand its dynamics. However, let us first state that because there is a more informal leadership structure does not mean necessarily that LO is an oligarchy in disguise. Rather, it means only that the organization is a power unit of great

scope, and as such it must necessarily have recourse to bureaucratic mechanisms
to be effective.

LO's bureaucratic mechanisms are not, however, fully rationalized. This becomes immediately clear when one asks how many persons are actually involved in organizational work on something like a full-time basis. In a well-developed bureaucracy there would be no doubt about the answer, the line between those inside the bureaucracy and those outside it being sharply defined. It is thus a striking reflection of the uneasy coincidence of bureaucratic and representativevoluntary processes in LO that no one, inside or outside the organization, knows with any precision the size of its cadres. In a comparison of American and European union movements, Lipset, using data collected by LO in the early 1960'5, estimated the number of paid officials at 240, or one official for every 2,200 members (1967: 217—223). Lipset went on to view this comparatively small body of paid officials as evidence of the "lay" character of Norwegian unionism. He speculated that Norwegian unions depend heavily on voluntary, rank-andfile participation in order to perform organizational tasks. This in turn lessens the control full-time leaders have, and it implies the presence of many local activists with a "sense of mission".

Whatever the merit of Lipset's speculations, our investigations suggest that the number of paid officials in the LO system is considerably larger than his estimate, even allowing for some expansion in size during the half decade separating his estimate from our own. In 1966, LO compiled figures showing 256 paid officials at national and district levels as well as 27 officials, including three union presidents, paid for part-time work at these levels.² In addition, LO reported employing 326 paid functionaries, at least 50 of whom performed important jobs entailing considerable expertise and responsibility.³ But these 609 persons are only the top of the organizational iceberg. Below them is a large network of paid positions in other district offices, in union locals, and on factory floors, the number of which is known only to the individual unions.⁴ Replies to a questionnaire we circulated in 1969 revealed that in 18 large unions orgainzing mainly the private sector there were 54 full-time, factory-paid officials at the local level. In addition, 3 of the 18 unions surveyed said they employed a total of 10 full-time officials at this level.

To this extensive organization of line and staff personnel must be added an unknown number of paid officials in the Workers Information Association (AOF) and in the LO school system. But our aim is not to delimit precisely the extent of LO's apparatus, which is a subject for independent study. Rather it is merely to show that the leadership and bureaucratic components, including clerical personnel, of the LO system are very sizable indeed. We estimate that they total upwards of 1,000 paid full-time positions at all levels, and probably a corresponding number of paid part-time positions. This easily makes LO the largest of all bureaucratically organized interest groups in Norway, and it makes LO one of the 100 or so largest private employers in the economy.

An organization having the size and diversity of LO is necessarily a fusion

of many hierarchies, and any analysis of the structure and functioning of its leadership must take this into account. Thus if we are to understand LO's leaders as important power wielders in Norwegian unionism and in the society as a whole, we must first try to understand how power is distributed within LO itself. A detailed and comprehensive analysis of this question would, however, require a union-by-union study of electoral and decision-making processes based on intimate research access over time to the "corridors of power". Not having the data for such an analysis, we confine ourselves here to suggesting a broad scheme of power which seems to organize plausibly those limited materials we have obtained.

Sitting at the apex of Norwegian union organization and being primarily responsible for wielding organizational power internally and externally is a relatively small group of persons we shall call the labor elite. At a minimum this elite group consists of the 6 elected officers who oversee the LO network and the presidents of LO's member unions (37 full- and 4 part-time presidents in 1967) plus the union vice-presidents who have full-time status (34 in 1967).⁵ This group of 81 persons, individually and collectively, possesses the general attributes of an elite. A third of them constitute regular and alternate members of the LO Secretariat. In their individual capacities as presidents and vicepresidents, members of this group play the major roles in labor-management negotiating rounds, are frequently appointed to government committees intended to thrash out questions of national importance, in practice have the authority to convene or delay meetings of their union executive committees, direct the administrative staffs of their unions, have a major say in nominating or appointing lower-ranking officials to higher positions, and are the persons representing their unions' points of view to the public at large. They thus meet the first requirement of elite status: the possession of power.

But there are additional reasons for designating this group elite. Their small number, their lengthy, very similar careers as union officials, their cohabitation of the two or three Oslo office buildings in which nearly all unions are headquartered, their frequent travels inland and abroad to conferences and various representative functions, their common membership on innumerable formal and ad hoc committees inside and outside LO, and their rather uniform membership in the Labor Party all suggest very considerable cohesiveness as a social group. There is a fellowship and a sense of belonging among them that transcends mere similarities in positions. As Porter has noted, social cohesion, coupled with the possession of power, transforms a simple statistical category into an elite, a group worthy of study as a unique element in the structure of power (1965: 230).

Situated immediately below the elite in the union hierarchies is a more heterogeneous group we designate the union apparatus. The core of this group in 1966 was the 81 other full-time officials who were elected to positions on the executive committees of the LO unions. Formally distinct from these elected officials, but in practice combining with them, was another group of 94 officials

employed by the unions in full-time administrative positions.⁶ These 175 individuals perform essentially interchangeable jobs extending across the full range of tasks, from publishing union newspapers to applying and interpreting work contracts, that are the core functions of the unions.

While our lumping together elected and appointed officials in the apparatus category might seem a careless mixing of "democracy" and "bureaucracy", our investigations suggest that the distinction is exceedingly vague in practice. For example, the mere fact of holding elective office does not guarantee an official will be concerned with mainly representative functions. Norwegian unions are so small that an official, elected or not, must be a jack-of-all-trades, and he will often be involved in administrative work that in other, larger unions might be delegated to hired experts. This is presumably why in everyday usage the title "tillitsmann" or "representative" tends to be applied not just to those who are elected but to appointed officials as well. Thus we make no distinction between elected and appointed officials at the middle echelons of union leadership, and we see both types as performing basically the same tasks. Indeed, several unions (e.g. the Steelworkers Union, the Office Workers Union) have already abolished the distinction at the apparatus level, and there is every reason to believe that this modification will be adopted by other unions in the near future.

Another component of the union apparatus group is the staff of the LO network. Excluding clerical personnel and those engaged in specialized but routine jobs (e.g. auditing) this part of the apparatus consists of approximately 40 persons. At least 9 of them are experts employed in the legal, economic, press, and rationalization divisions of LO's administration. Seventeen others are employed as full-time district secretaries whose responsibility is to represent LO to union organizations at local levels. The remainder are persons employed in or elected to administrative and leadership positions in various subdivisions of the LO network (e.g. the Government Employees Cartel, the Secretary for Women's Affairs, the Secretary for Organizational Affairs).⁷

A third important group in LO's leadership structure is the persons we have labeled delegates. This group consists, first, of the 284 persons who in 1967 were sitting as executive committee members of the 41 individual unions, exclusive of committee member's already classified in the elite and apparatus. The overwhelming majority of these delegates from local union organizations to the executive committees occupy a "lay" status, that is, they are unsalaried, part-time volunteers. We found only 20 of them holding down full-time union jobs at the local level, although because of incomplete survey response it is possible that there are a few more full-time delegates.

Among students of unions, considerable importance is attached to the presence of lay members on union executive committees. Since they often constitute a voting majority on the boards, it is thought that, at a minimum, they comprise an effective veto group representing membership interests. As Galenson argues, ". . . These rank-and-file men have strong local backing, and are quite capable of standing up to the permanent officers . . . This pattern of lay majority on

executive boards [committees] is quite common in Scandinavia, and is regarded
as an important contributor to union democracy." (1961: 74—75).

Unfortunately there is not available any study of the influence of these delegates, and Galenson is unable to furnish data to support his assertion. Quite possibly in this area we have only "native theory" to rely on. In the absence of data, we can only say that the roles of delegates are likely to vary considerably between unions and between individuals. Some obvious variables are the skills and experience of individual delegates, the size and importance of the locals they represent, whether there is consensus in their locals, whether there is consensus among the delegates themselves, and so on. It seems likely that "palace revolutions" go both ways, sometimes resulting in the displacement of full-time officials by lay persons, and sometimes resulting in the exclusion of delegates who are regarded by constituted leaders as disruptive. It therefore seems prudent to treat as an open question that of delegates' power.

Much the same considerations apply to the other component of the leadership group we call delegates. This is the 66 members of LO's 115-member General Council we have not already classified in some other leadership category. Since the General Council is vested with supreme authority in LO during the four years between sessions of the LO Congress, election to this body is a mark of considerable recognition. In 1967 at least 13 of the 66 delegate members of the Council were salaried local union officials, most of the remaining 53 being lay members. However, because the Council meets but twice a year, it is doubtful that a position on the Council is itself an important leadership position. Rather, we expect that its great constitutional authority, in most cases, means that its elected members are important rank-and-file leaders in other capacities. For this reason we include them in the delegate category of union leaders.

Our analysis leads us, then, to view labor union leaders as situated in three relatively well-defined categories: elite, apparatus, and delegates. The basic criterion used to assign persons to one or another of these categories is a combination of positional rank and membership in formal bodies. Doubtless this classification is somewhat artificial, but from the standpoint of the ability to influence decisions both inside and outside the union movement it is likely to be closer to reality than a classification by formal bodies alone would be.

By our estimate these three groups of leaders, elites, apparatus, and delegates, numbered approximately 600 persons in 1967.⁸ In order to study their backgrounds, careers, and present activities on the basis of uniform data, in May 1967, using a list of names especially compiled by LO, we mailed each leader a lengthy questionnaire accompanied by a letter of endorsement from the president of LO. Over the following twelve month period a total of 482 questionnaires were returned, a response rate of 80 °/o. However, 10 of these were insufficiently complete for use, thereby requiring that our analysis proceed with data for 472 leaders, or 78 °/o of the universe. Response rates differed somewhat for each category with the final rates being 80.2 °/o of the elite, 69.1 °/o of the apparatus, and 81.7 °/o of the delegates.

During the winter and spring of 1969 we supplemented the questionnaire data with considerable material derived by personal inquiries and by fairly systematic interviews with 30 union leaders including all of LO's elected officers and the presidents of the twenty largest unions. Since these interview data relate primarily to a larger study of Norwegian elite structure and of elite ideologies in Norway we make but limited use of them here.

A Statistical Profile of Union Leaders

A statistical profile of LO's leaders is necessary for a more detailed understanding
of who they are and where they come from. The profile presented here concentrates
on four dimensions: age, sex, residence, and income in 1967.

A key assumption underlying the maintenance of an organizational hierarchy is that persons at its top will generally be the most experienced and knowledgeable. As a practical matter this usually means that high leadership position is directly correlated with advanced age. Table 1, trichotommng the elite, apparatus, and delegate groups by age in 1967, shows LO's adherence to this general pattern. More than a third of the elite were born in the first decade of this century or earlier while 6 °/o were born in the decade or so immediately preceding the Second World War. By contrast, persons located further down in LO's

hierarchy, those in the apparatus, are noticeably younger as a group. Roughly equal numbers of them (about a fifth in each case) were born in the first and third decades of the century. The delegate group occupies an intermediate position in age distributions. But looking at the three categories in combination we find that one quarter of all leaders were born in 1911 or earlier. These persons were socialized into the union movement during the tumultuous labor-management conflicts in the 1920's and '30's before the processing of those conflicts became pacific. Further, these persons saw first-hand how the political factionalism of the 1920's paralyzed recruitment of new members. It is plausible to infer that the lessons learned then explain much of the staunchness with which the principle of movement solidarity is defended even today.

Another question about age distributions is in which unions are the most elderly leaders located. The answer to this is interesting for two reasons. First, to the extent that such leaders control the largest and most powerful unions are they better able to make their conceptions prevail, and their conceptions may well differ from those of younger leaders who grew up in changed circumstances.

Second, sociologists have often discussed Michels' hypothesis that the bureaucratization of mass organizations tends to result in an entrenched leadership, one indication of which will be low turnover resulting in disproportionately advanced ages of leaders. In the present context this implies that leaders of the large unions which have the most elaborate bureaucracies will also be the oldest leaders.

In Table 2 the location of leaders who were 51 years old or more in 1967
is depicted as regards the sizes of their unions. In both elite and delegate cate

gories, leaders 51 years old or more comprise majorities in large unions (over 10,000 members) while they are minorities in small unions. The concentration of older leaders is most noticeable at the elite level in large unions where 63 % were 51 years or more in 1967. Only among apparatus personnel is the frequency of younger leaders greater in large unions than in small ones. This reflects the more rapid expansion of the large unions and their policy of deliberately hiring young persons for apparatus positions. This policy is even more in evidence in the apparatus of the LO network where 78 °/o were 50 years or less in 1967.

There is thus substantial support for both propositions about the location of elderly leaders. First, they tend to be concentrated in the larger and more influential unions, thereby enjoying an advantage in pressing whatever distinctive views they may have concerning what the union movement should be and do. Second, the data lend modest support to the hypothesis that the size of union bureaucracies slows down the advance of young leaders to the top and tends to relegate control to those persons who are likely to have been around the longest.

In the late 1960's approximately one quarter of LO's members were women. Yet an examination of the compositions of union leadership bodies hardly reflects this. Of the 300 delegates attending the 1969 LO Congress, for example, only 14 were women. Although the Municipal Workers Union counted 33,800 women in its ranks in 1969 (47%), only one of its 32 delegates was female. Likewise, with 22,800 women members (56 %) the Office Workers Union had but 3 women among its 18 representatives.⁹

The infrequent appearance of women in top union positions is dramatically
shown by our data. In 1967 there were no women to be found among the 65 persons
we studied as elite.10 Of the 121 apparatus personnel, only 7 were women,

and of the 286 delegates to union executive committees, only 24 were women. Still, as far as we can gather there is no evidence that female aspirants to top positions are deliberately excluded. If such were the case then at a maximum one would expect to find women leaders only in those unions where the rank-andfile is predominantly female. In fact, there is but a modest statistical relationship between the number of female rank-and-file members and the presence of women leaders in individual unions. For example, 2 of the 4 female union apparatus leaders, and 10 of the 24 female delegates were located in unions in which less than 25 °/o of the membership was female. Thus women seem to enjoy relatively equal opportunities among themselves for reaching leadership positions. But beyond this, either the equal opportunities women enjoy among themselves are uniformly unequal in competition with men, or women more or less consistently choose not to take advantage of the opportunities available to them.¹¹

A third statistical dimension of union leadership is its residential concentration in Oslo. Three-fifths of all leaders, about triple the corresponding percentage of the work force, resided in Oslo or its surrounding county, Akershus, in 1967. But the proportions of elite and apparatus leaders living in Oslo were greater still: 94 °/o of the elite and 79 °/o of the apparatus. By contrast, somewhat less than half the delegates (47 %) resided in the capital district, and it was only among delegates that leaders were found to be located in significant numbers in all major geographic regions of the country.

This marked concentration of union leaders in and around Oslo is of considerable importance for the structure and effectiveness of LO. Residential proximity facilitates ease of communication and daily interaction. Meetings can be frequent and impromptu, differing points of view can be thrashed out spontaneously, and organizational unity can be cemented in friendship and fellowship. In a word, there is good reason to believe that residential concentration in the capital significantly enhances the cohesiveness of the leadership group. This observation is further supported by our finding that not only do the great majority of top leaders live in Oslo but they also tend to reside in particular areas of the city. Thus, plotting the residential addresses of all full-time leaders on a city map, we found that with few exceptions they resided in the eastern sections of Oslo, close to or east of the Akers River.

But the organizational benefits flowing from residential proximity also entail organizational costs. One of these is likely to be, from the rank-and-file standpoint, the image of a remote, somewhat inaccessible "Oslo Establishment". Of course, the system in which delegates from outlying districts are elected to union executive committees is designed to counter this problem. But the fact that close to a majority of all delegates themselves live in Oslo, plus the fact that another 12% live within commuting distance of the capital in the counties touching the Oslofjord, seem to limit considerably the effectiveness of this system. Whereas the concentration of elite and apparatus personnel in Oslo is explained by the functional need for coordination, the same location of large numbers of delegates is probably due to finances. The frequency of committee meet

ings requires that lay members be able to attend without incurring major tran portation expenses and loss of wages. One other organizational cost of the res dential concentration described here, this time from the standpoint of the leader is the necessity to travel to distant local organizations with wearying frequenc in order to maintain lines of communication. As Table 3 shows, to be a unic official often means to be "on the road" two or more months each year, a tas that no doubt diminishes the attractiveness of full-time positions.

A final dimension in the profile of union leaders sketched here is their in comes. As with any bureaucracy, LO has found it necessary to establish a systen of income differentials to reward incumbents for their burdens and responsibili ties. However, this recognition of the fact of hierarchy contradicts somewha the strong egalitarian traditions and values of the union movement. LO ha sought a compromise in this dilemma by using the graduated salary scale de veloped by the state to pay civil servants. Different union positions are peggec at various points along this scale, and union officials receive salary increment generally in accordance with the rules applied by the state to its own employees Application of the state's salary plan does not appear to be uniform across al unions, however, and there is a measure of variation between similar positions

Table 4 shows the 1967 incomes of the leaders we studied in four brackets. The resulting distributions show that within the ranks of full-time paid officials the positional hierarchy is reflected by an income hierarchy. Among elite leaders 56 % earned more than 40.000 Crowns compared with 42 % of the apparatus

leaders. Further analysis reveals that the advanced ages and seniority of some of the latter apparently make their incomes roughly comparable to elite incomes, although the variable of age operates only to a limited extent. Perhaps the most striking distinction shown in Table 4, however, is that between salaried union officials in the elite and apparatus categories and the part-time delegates whose incomes come from non-union sources. There is an income differential in the range of 10,000 Crowns separating part-time and full-time union leaders. There can be little doubt that the greater monetary rewards of full-time union positions contribute to generating competition for these jobs. This datum is of importance for the analysis of the career system in LO, and we will return to it later. Here we may conclude that the income differentials shown by our data rather effectively distinguish the three categories of leaders on which we are focussing. But they do so by describing an income continuum extending from delegates to elite persons, the upper extreme of which is still sufficiently modest to indicate that monetary rewards obtained from union leadership are quite limited. Our data show that officials earn a middle class income, but there can be no doubt that union leaders' incomes are far below those of the employers with whom they bargain. We may say then that income differentials in LO moderate somewhat the egalitarian values of the movement, but they are far from flatly contradicting those values.

Social Backgrounds of LO Leaders

Research on social backgrounds and recruitment is a major part of the literature on elites. This research is based on the arguable assumption that one can infer from patterns in background and recruitment data a good deal about the attitudes and behavior of elites (Edinger and Searing, 1967: 430; Dahrendorf, 1967: 217—279). As they have concerned the social backgrounds and recruitment of labor leaders much research and discussion have been dominated by the associated phenomena of the "labor aristocrat" and the "affluent worker". Specifically, the role played by that stratum of the working class consisting of craftsmen, foremen, and other skilled workers in shaping the direction of labor movements has been and is still a controversial question (Hobsbawn, 1964: 273 ff.; Porter, 1965: 344-346; Goldthorpe, et al., 1968: 116-117; Butler and Rose, 1960: 15-16).

Enjoying greater security derived from possession of scarce and marketable skills, better conditions of work, a strong sense of occupational identification transmitted from father to son, and prospects for upward social mobility, this privileged stratum of the working class was seen by Marx and Engels as an essentially conservative element (Goldthorpe, et aL, ibid.). Expounding this point of view, Michels saw "a veritable class distinction" between skilled and unskilled workers, and concluded that, "The aristocratic element of the working class, the best paid, those who approximate most closely to the bourgeoisie, pursue tactics of their own" (1962: 272-273).

This focus on the role of labor aristocrats in working class movements provides a perspective from which we may examine the social backgrounds of labor leaders in Norway. For we take it as fact that after a brief revolutionary period in the 1920'5, LO has consistently been reformist (i.e. Social Democratic) in its social and political stance. Whether the persons who brought about this strongly reformist orientation were labor aristocrats is an historical question we cannot take up here. Few if any of the leaders we studied in 1967 were old enough in those years to have significantly influenced this change in orientations. All we can say is that today's leaders inherited a reform-oriented union movement, and they have steadfastly adhered to this orientation. What interests us, rather, is the question of whether there are empirical grounds for the proposition that one reason why today's leaders have so consistently been reformist is that they share a common background in the labor aristocracy from which, according to theory, they obtained a basic disposition toward reformist strategies ¹² To ascertain whether this special stratum of the work force did in fact contribute disproportionately to the ranks of LO leaders, we turn next to some of our data bearing on their social class origins.

Probably the best single indicator of social class origins is the occupation of
a person's father. Table 5 presents this datum for the fathers of all labor leaders
in our study. In the table the social class origins of the three leadership groups

are seen to be quite similar. Slightly more than three fifths of all leaders had fathers who were manual laborers, the component of the work force which today supplies approximately 75 °/o of LO's membership. But within this manual labor group there is a close division between those fathers who were in the craft and skilled stratum (47 °/o) and those in the unskilled or semi-skilled stratum (53%). However, only 18°/o of the former possessed the prestigious "craft certificate" (svenneprøve) indicating attainment of the highest craft status.

Table 5 also shows that a quarter of all leaders had fathers who were employed

in the non-manual component of the work force. Of these only a small proportion were independently employed, all others being salaried white-collar employees. The number of leaders with origins in the non-manual stratum constitutes a significant exception to the expectation that a working class movement will be led by persons from the working class. Nor can this exception be explained away by the recruitment of leaders from white-collar families to white-collar unions. Thirty-seven percent of the leaders whose fathers were in non-manual occupations went on to leadership positions in &/#e-collar unions. Instead, the anomalous movement of white-collar sons into blue-collar leadership positions is explained by the relatively low status rank of the occupational group we have labeled "white-collar". Being, in the main, government employees such as policemen, military personnel, and municipal civil servants of various sorts, fathers classified here were members of the lowest echelon of the then emergent middle class. Their status proximity to skilled and craft categories of the manual labor component was probably considerable and warrants viewing them as an extension of the labor aristocracy.¹³

Taken together, the persons whose fathers were skilled workers or lowranking white-collar employees comprise 54 % of all union leaders. But the fact that only a tenth of all fathers had formal education beyond grammar school and that only slightly more than one-third had specialized occupational training reminds us that the opportunities and higher statuses available to them were limited. Broadly speaking then, LO's leaders have working class origins. These appear to be more or less evenly divided between an upper, "aristocratic" stratum and a lower, less differentiated one. Certainly the contribution of sons to union leadership roles by fathers situated in the upper echelons of the working class is impressive numerically. But the labor aristocracy was far from being the only source of recruitment. Our data suggest that today's labor leaders approximate a cross-section of the working class with the lower, unskilled elements having considerable representation, particularly at the elite level.

If the recruitment base for labor leaders does not lie clearly in a specific social stratum it seems likely that there are influences which cut across strata to propel persons with diverse backgrounds toward union leadership. The extent to which particular family milieux, irrespective of their locations in the status order, contain strong commitments to unionism as a social movement might be one such influence. Our study provides data with which we can examine the strength of this aspect of the socialization process through which labor leaders passed.

We asked union leaders to tell us if members of their families were organized in LO when the respondents were growing up. Three-quarters of all leaders indicated that either their fathers, their mothers, or both parents were members of LO at that time. Nor, apparently, did the unionized fathers comprise a random combination of active and passive union members. When asked to recall the extent to which their fathers were active in the union movement, the 352 leaders to whom the question was applicable classified their fathers' activities as follows:

These distributions do not differ significantly between the three leadership categories, and, bearing in mind the obvious limitations of these recall data, we can conclude that at least three-fifths of today's union leaders had fathers who were themselves active union members.

Family emphasis on unionism as a social movement may also be manifested by the intensity of parents' political participation, allegiances, and discussions. Examining this aspect of labor leaders' family milieux, we have first the finding that fully 22 °/o of their fathers held elective municipal office at some point in their lives. Another 5% were appointed to various official committees and boards in their communes while 3 fl/o held political positions at the national level, three of them as members of Parliament. In addition to these forms of political participation it is reasonable to assume that an unknown but substantial number of parents also held positions in party organizations. More generally, we asked labor leaders to classify for us the degree to which their fathers were interested in politics. Their classifications are as follows:

Subject again to the reliability of recall, these data, in combination with those
on political participation, demonstrate rather convincingly that large numbers
of union leaders grew up in highly politicized family milieux.

If the classifications of fathers' activities in unions and interests in politics are cross-tabulated, the resulting combinations in Table 6 indicate an extraordinarily strong commitment to unionism as a social movement in the family milieux of union leaders. For us it is a fair assumption that much more than 15 °/o of the working class would, in a similar classification, be placed in the lower right-hand corner of the table, and we interpret these distributions as quite unrepresentative of the broader mass of working-class fathers. One question raised is whether the distributions in Table 6 are related to the skilled and unskilled occupational statuses of the fathers. But when a control for occupation is in

troduced there is only a small tendency for skilled fathers to be more consistently active in the union movement and interested in politics. We are therefore inclined to conclude that specific union and political norms which labor leaders probably internalized through socialization in their families are critical in an explanation of why these leaders are where they are, and that such norms have not been the attribute of any one segment of the Norwegian working class.

Labor unions are preeminently urban phenomena. It is the commonly experienced grievances of large numbers of persons situated in close proximity to each other that provide organizing potential for unions and give them their power base. Accordingly, union leaders are in the normal case products of urban environments, persons who are especially sensitive to the harshness of working class life in urban society. It is thus one of the real curiosities of the union movement in Norway that a study of the geographic origins of its leaders uncovers substantial rural influence.

Table 7 shows the urban-rural backgrounds of union leaders in the elite, apparatus, and delegate categories. The data indicate that persons with urban backgrounds have a better chance of reaching elite positions than do those with rural backgrounds. This accords with the foregoing observation that unions are urban organizations and thus, not unexpectedly, more often have leaders with

urban backgrounds. But besides showing a steady relation between urban backgrounds and high leadership position, table 7 also shows that large numbers of LO leaders grew up in rural areas of Norway. Thus if the table is dichotomized so that leaders from other than "thinly populated rural districts" and "rural villages" are classified as growing up in an urban environment, we find that 29% of the elite, 42 °/o of the apparatus, and 51 °/o of the delegates grew up in rural areas.

The combination of urban and rural backgrounds among LO leaders raises the further question of their regional origins. Regional identities have always been strong in Norway, and it is interesting to examine the extent to which various regions are represented among union leaders. Table 8 classifies the counties in which these persons grew up by region. The table shows that slightly less than a majority grew up in Oslo or around the Oslofjord. The regional backgrounds of all other leaders are quite diverse with no single region contributing a clearly disproportionate number. This regional diversity demonstrates again the substantial rural influence in union leadership ranks. More particularly, and remembering that today almost all full-time leaders and half the part-time leaders reside in Oslo, these data on regional origins indicate the substantial geographical mobility that careers leading to top union positions imply. They also reflect the comparatively late industrialization and urbanization of Norwegian society, and they show that immigrants to the cities have not been excluded from the union movement. A final implication of these data is that a substantial number of powerful persons in the unions have had first-hand experience with the problems and needs of the rural sector. This experience probably facilitates appeals to, and negotiations with, agricultural and rural interest groups. Indeed, the Norwegian labor movement has been comparatively successful in maintaining political support in the rural areas of the society. More generally, the rural backgrounds of so many union leaders may contribute to the cultural rapport between town and country Eckstein believes to be a key stabilizing factor in Norwegian political life (1966: 121—128).

The last aspect of labor leaders' social backgrounds we will examine here
is their formal education. Access to formal education is everywhere shaped by

differential class opportunities. The amounts of education various groups are able to obtain, therefore, can serve as indicators of their points of origin in the status order. Moreover, one of the key features of elite groups in modern societies is their possession of superior education. Superior education serves as a prime means of legitimizing power and is often instrumentally necessary to effective wielding of authority within and between the complex institutions of modern societies.

The formal education of Norwegian trade union leaders has, in the past, been a two-step process. The first step was that of general education in the public school system at several levels. During the early decades of this century, the period in which most leaders obtained their general schooling, seven years of attendance at grammar school were required of all children. Table 9 shows that nearly half of all leaders terminated their general education upon completing this requirement. Another quarter continued for one or two years, while those going on to further secondary, 'gymnasium', or university educations comprised a final quarter of the group. It is to be noted that apparatus leaders depart significantly from the general pattern, having attained as a group considerably more education. This is explained by two factors, first, that they are younger and have enjoyed the benefits of greater affluence and a more open education system in the postwar years. Second, this group includes all the full-time professionals and specialists staffing the union bureaucracies. The presence of these persons pushes the group's education profile up.

Although most LO leaders had access to very limited general education, it is important to ask about the extent to which this was a common experience across different types of unions. Table 10 classifies all leaders studied by the type of union to which they belong and the highest education they obtained. Differences in the amounts of education obtained by the several groups are striking. Close to two-thirds of the blue-collar leaders obtained only the minimum

education required by law while this was the case with but one-quarter of the white-collar leaders. Similarly, less than one-tenth of the former had access to further secondary, 'gymnasium', or university education compared with more than half the latter. It is interesting to note that the education profile of the LO network group much more closely approximates that of the white-collar than that of the blue-collar leaders. This probably reflects the greater educational requirements for jobs in the LO network than in the unions.

On their face these data provide a perspective from which to view differing union interests and political activities, as well as differing conceptions about the role of LO in the society, which sometimes distinguish the blue- and white-collar branches of the union movement. But it is easy to exaggerate whatever divisive effects aggregate differences in education may have. That is, other variables such as age, size of union, experience, and, not least, the fact of genuinely divergent interests on some issues, are likely to be of equal or greater importance.¹⁴

In any case, the foregoing analysis shows that with relatively few exceptions labor leaders on the whole have had limited general education. But it has been observed in studies of labor leaders in other societies that although they hardly differ from large numbers of persons where general education is concerned, they may often have highly skilled and technical occupational training in addition. Two measures of occupational training are the extent to which labor leaders are educated in specialized trade and technical schools and the frequency with which they possess official certification of high skill status.

We found that at least 334 of the 472 leaders studied (71 °/o) had some occupational training beyond their general education. One might suspect that it was into this sort of training many of the persons with only grammar school education went, thereby reducing some of the differences noted above in general education. However, the facts are somewhat to the contrary. Two-fifths of the 229 leaders who reported having only grammar school education also reported having had no occupational training. Thus 92 leaders apparently had no formal education of any sort beyond the legal minimum. More than 70 °/o of

these persons were delegates and over half of these delegates were employed privately
as unskilled or semi-skilled laborers.

Be that as it may, substantial numbers of leaders did couple occupational training to their general education. In many cases this training appears to have been extensive. For example, if the longest occupational training program of each leader for whom detailed information is available is measured (107 leaders participated in two such programs and 58 did so in three or more) data show that 8 % attended one such program for at least three years, 16 % for two to three years, 22 °/o for one to two years, and 54 % for a period lasting less than a year. The highest achievement in many of these programs is the receipt of a craft certificate, normally after three to four years of organized study and apprenticeship. In 1967, 20% of all blue-collar leaders and 6 °/o of all mixed-collar leaders held these certificates.

On the basis of these data then, it appears that occupational training is a fairly standard feature of most leaders' educational backgrounds. Keeping in mind the group of 92 leaders to whom they do not apply, the data on occupational training indicate that today's labor leaders have been recruited from several echelons in the generally skilled component of the manual labor force. Evidence of fairly extensive occupational training reinforces the more general observation that this skilled component contributed leaders to the union movement in numbers probably disproportionate to its size. On the initial assumptions that reformist orientations are most characteristic of this stratum of the working class, our description of the substantial anchoring LO's leaders have in the skilled stratum suggests one reason for the maintenance of LO's reformism in the postwar period.

However, we are less impressed by this interpretation of our data than we are by the alternate view that data on social origins really describe a coalition of leaders with very different backgrounds. Rather than concluding, for instance, that LO has been reformist because so many leaders come from the skilled stratum, we think its more significant to say that LO has been reformist even though nearly half its leaders come from lower levels of the working class. We notice again the intensity of fathers' commitments to the union movement across all levels of that class. This datum suggests a uniformity in socialization that seems to counteract intra-class differentiation, substantial urba.n-rual dichotomization, and equally substantial blue-collar versus white-collar differences in education. Thus LO's leadership is broadly representative of a number of divergent interests and background patterns. It is best described as a more or less open coalition rather than a cooptive monopoly by one stratum. We may now ask how this leadership coalition functions and is maintained. This requires structural analysis and we therefore turn to an exploration of the career system in LO.

LO Leaders' Careers

The concept of career is closely linked to the social mobility of individuals
within and between hierarchical structures, and in popular parlance it is linked

to the notion of promotion or success. Social mobility presupposes relatively stable hierarchies as regards selected values such as prestige, power, income, skill, and so on. The more unstable, incongruent, or diffuse hierarchies are, the more difficult it is to apply the concepts of mobility and career in empirical research. This is a major problem in the study of careers within political and union organizations where status systems are strongly crystallized at only a few points, these points being surrounded by rather diffuse networks of elective and appointive positions. This organizational pattern is best viewed as a functional compromise between the conflicting demands of "bureaucracy" and "democracy" in mass organizations, the former necessary to organizational effectiveness and the latter necessary to organizational morale and support. Thus the diffuseness characteristic of union and political organizations gives to them an egalitarian appearance that serves to activate rank-and-file members toward part-time service. Hierarchy is hidden behind this egalitarianism and is of necessity weakly articulated and somewhat informal, crystallizing around personalities as well as positions.

What is required from a research standpoint, therefore, are intensive studies of the stratification systems of unions and parties. Such studies do not exist for Norwegian unions, and without them we cannot present more than a crude analysis of the career patterns of union cadres and top leaders. It is of some consolation to note that we are not alone in this predicament for Porter seems to have come to the same conclusion in his monumental study of Canadian elites. He tried to analyze the career stages of union leaders, but remarked that "such confusing data are difficult to put into any kind of order", and then fell back on the assertion that "obviously the top men select those who are appointed" (Porter, 1965: 357).

To be organizationally significant a career system must be fairly well articulated. At least some people must perceive it as one of several ways in which they could conceivably move. The Norwegian union career system meets this criterion, but only to a limited degree. The complexity and diffuseness of the union system, especially in the larger unions, make it difficult for a young worker to "plan on" or "choose" a union career. Mostly in contact with part-time, unpaid officials, the routes from the periphery to the center of the union system, from voluntary to salaried work, are likely to appear labyrinthine and, possibly, accidental to such a person. Moreover, even if the route to the lower rungs of the career ladder is apparent, the individual will have difficulties seeing any mechanism by which he can be reasonably certain that once he moves ahead he will not quickly fall back again. This is because, at least at the local level, advancement largely depends upon elective and representative processes, an eminently insecure basis on which to chart one's future. Indeed, estimating that there are approximately 500 full-time, non-clerical jobs in the LO system that might attract members, we can quickly see that on an absolute basis only one of every 925 male members can be absorbed into the full-time structure. By contrast, it has been calculated that about one of every 10 Norwegian workers experiences

some upward mobility along one or another occupational route (e.g. promotion to foreman or establishing one's own business) (Aubert, 1968). What Schlesinger has called the "structure of opportunities", the number of offices available and the frequency with which new men attain them, is in Norwegian unions both difficult to perceive and sharply limited in size (1966: 11—21).

In one sense the typical leader's career begins at the time he is first gainfully employed. Given their limited formal education it comes as no surprise to find that two-thirds of all leaders were working full-time by the age of 16. These first jobs were usually of an unskilled or apprentice type, and an examination of them suggests no particular patterns. We do know that approximately one-fifth of those in manual occupations obtained craft certificates at some point, and this may be taken as an indicator of some intra-generational mobility. If we shift our attention to a classification of their occupational statuses at the time they went over to paid union positions, we can make a few inferences about mobility patterns.

Table 11 enables us to compare the highest occupational statuses attained by our respondents and those of their fathers (Table 5). Persons in the elite category had basically the same occupational patterns their fathers had. Insofar as elite members have experienced upward mobility as a group, union jobs have themselves been the operative mechanism. By contrast, a comparison of Table 11 and Table 5 shows that many members of the apparatus category must have experienced upward mobility before they entered a full-time union position. These occupational data require one more comment, namely, that with very great frequency labor leaders are recruited from the skilled and white-collar components of the work-force. Thus, our data describe a labor leadership whose immediate occupational backgrounds are very clearly within the stratum we earlier called the "labor aristocracy".

Turning next to the economic sectors from which paid union officials are recruited, we learn from Table 12 that more than a third were employed by government, 43 °/o by private enterprise, and 16 °/o by various organizations (including co-operatives). It is remarkable that capitalist productive units, historically the well-springs of unionism, now contribute in direct fashion less than half the professional union leaders. The typical union leader in the late 1960'5, in a movement whose rank-and-file is still predominantly employed in the private sector, is recruited from the public or organizational sector. Certainly this recruitment pattern reflects the major inroads and successes LO has achieved in organizing the public sector in recent years. But we suspect there is an additional explanation why so many union leaders have not recently, at least, had much first-hand experience with private enterprise. We refer to the tendency of elite groups to recruit from their "doorsteps" partly because, as Marvick argues, "only in that way can they apply their own judgment of the man or use the judgment of familiar and trusted associates" (1968: 280). The Labor Party's long-lasting majority in national politics and even longer majorities in many municipalities' politics have probably given LO unions unusually good familiarity with and access to governmental personnel. The Labor Movement has to some extent "colonized" the governmental sector, so that government units may tend to become "doorstep organizations" for the recruitment of union leaders.

Table 12 also reveals a certain amount of internal recruitment in the LO. 6 °/o of the full-time union officials earlier held paid clerical or other specialized positions within the union movement. Another 4 °/o were employed in other areas of the Labor Movement, for instance in the Workers Information Office or in the labor newspapers, prior to moving to full-time union jobs.

Parallel with their occupational careers persons who eventually become professional union leaders get involved in part-time, voluntary union activities at an early date. On the average, elite members joined unions at the age of 19 while apparatus members joined at the age of 20. Nearly two-fifths (38 °/o) of the elite reported they were unionized by the age of 17 compared to 30 % of the apparatus and 27 °/o of the delegates.

Although our data do not permit a specification of how and when leaders first became active in their unions, the data show clearly that most of them had extensive experience in elective positions at the local level before becoming fulltime officials. Without exception (and excluding 16% who gave incomplete

answers) leaders in the elite category report they had at least one local position, and 70 % of them were at one time or other President or Vice-Presidents. Further analysis of the data showed that 82 °/o of the elite and 77 % of the apparatus held two or more local elective positions, and more than 50 °/o in both categories held three or more positions. These figures are based on incomplete reporting and are therefore conservative, but they still indicate very strongly the intensive involvement in local union activities which precedes leaders' entries into full-time union work.

But how long do promising leaders stay at the local level before they shift
into full-time work? Tables 13 and 14 supply answers to this question.

The latter shows that a very large majority has been unionized 9 years or more before assuming a full-time position, and that about one-third has been unionized for 21 years and more before achieving a full-time position. This pattern is reflected in Table 13 from which we learn that a majority of the leadership (78 °/o of the elite and 58 °/o of the apparatus) were in their thirties and early forties when they assumed full-time jobs.

To summarize, it is evident that those portions of careers spent in part-time, voluntary union activities at the local level are lengthy. The sudden rise of a person to full-time status is definitely an exception to the rule. The typical leader serves in part-time, usually unpaid positions for many years before shifting to a full-time job. Without doubt Norwegian unions follow the same rules as British unions, which "not only insist that candidates for leadership should be drawn from the ranks of the workers they organize but: also that they should have a detailed knowledge of trade-union machinery and should, therefore, pass through the various levels of trade-union activity" (Allen, 1954: 195—196).

For the young and ambitious worker, union activity does not offer a quick
path to power and prestige in the society. While effectively discouraging and
probably excluding persons with excessively personal and short-term motives,

the requirement of long and fairly unrewarding service prior to a full-time career probably costs the union movement a certain number of talented persons who go elsewhere in search of more promising careers. In addition, it should be mentioned that unions in many cases have helped workers in launching company careers by making talented people visible to management or by giving them training and experience which qualified them for promotion within the enterprise. It is reasonable to assume that those leaders who do persevere through the many years of minor positions required by the system must do so for reasons other than immediate personal reward. In the process, they acquire an intimate knowledge of the problems of union organization at the local level and an understanding of rank-and-file sentiments.

We argue, however, that local involvement and local experience are necessary, but not sufficient conditions for successful work in full-time leadership positions. The local perspective is too narrow a framework for effective roleperformance at the top level where different interests must be aggregated within a framework of national union administration. The competence developed at local levels must be supplemented by formal training and various relevant experiences (e.g. political office-holding).

The union school system, which has expanded significantly during the last twenty years, provides a more formal introduction to union problems and to national economic and political problems. This schooling is an integral part of the typical union career. It is one of several ways in which a prospective leader may increase his visibility within the union and demonstrate his commitment to the goals of Norwegian unions. In addition, the union educational system probably contributes in a central way to the development of social cohesion within the union system by knitting together the cadres and top leaders through the various activities organized for the students.

The union school system gives unions a high degree of control over the supply of recruits to the category of "trained activists", but neither union locals nor national headquarters can guarantee that a candidate they would like to see elected to a local position will gain support among his co-workers. Union members apply for stipends to go to the union boarding school close to Oslo, and we assume that local leaders have considerable influence on the selection of students. These local leaders are in a better position to evaluate the candidates' potential for union work and to know their political attitudes. We take it that at this early point many left-wing extremists meet a closed door.

We may summarize by saying that in general the chances for election or appointment to a full-time, higher-level union position are better the more visible and politically acceptable a man is. Visibility at the rank-and-file level flows from Labor Party activity, election to an inter-union municipal council, participation in union courses, election to delegate status for national conventions, and, perhaps most important, successful work in a big local in Oslo or its vicinity. Presidents of big locals, for example, will often be noticed and respected for the votes they can influence, and this, we think, is an important power basis for an

aspiring leader that also tends to structure his ambitions toward the attainment
of a permanent higher-level position in the national union hierarchy.

So far our analysis has been concentrated on the shift from part-time to fulltime status. Moving into a full-time job, usually in the apparatus category, is for a large number of officials the final promotion: they remain union secretaries, editors, and other sorts of specialists for the rest of their working lives. Elite positions are scarce and demanding, and those apparatus personnel prompted to compete for them must do so not only among themselves but also against part-time officials who may already be lay members of the national executive committee with substantial local support.

But if promotion opportunities within one union are limited, cannot a man compete for top-level positions in other unions? Union constitutions stipulate that at least the top officials (President, Vice-President, and Treasurer or First Secretary) must be elected by the union's national convention. In general there is no explicit rule that they must be recruited from the particular union's membership, but such a rule is implied in practice and puts an effective barrier on cross-union careers at the higher levels. Movements between unions may take place, but at a rather low level in the hierarchy. And as soon as a person has started a full-time career he is generally expected to stay in "his" union. This is clearly demonstrated by Table 15. Unions do not raid each other for personnel, and there are very few career opportunities in other branches of the union system. LO headquarters has some positions at the district level, but yearly vacancies are few.

This implies that the single union constitutes a closed promotion sphere. The mobility potential within the whole system is highly limited, and this makes it extremely important for a newcomer to get along with his superiors and fit into the team. We asked twenty top leaders, for example, what they mean when they describe someone as "an effective union leader" (en dyktig tillitsmann). Responses were varied, but those most frequently voiced were "human qualities" and "ability to understand and get along with others". These eminently subjective attributes were valued above "leadership capabilities" and "organizational capabilities*, the two characteristics which received second most frequent mention. Still less frequently mentioned were functional skills such as bargaining capability and social-political involvement, while strictly instrumental language,

writing, and speaking skills were rarely singled out. Explaining why they value personality characteristics over instrumental abilities, some of our respondents stressed that the circle of officials within each union is so small that it it very important that a man be emotionally stable and display a cooperative attitude.

We have no data bearing directly on turn-over in Norwegian unions, but various indicators suggest that in top positions it is extremely slow and results mostly from natural causes (age, illness, and death). Table 16, for example, shows that among persons occupying elite positions in 1967, 54 °/o had been full-time officials for 13 years or more. The corresponding figure for apparatus personnel

was 28 %. It appears that once admission to full-time status is obtained that status is reliably self-perpetuating. In this respect Norwegian unions conform to a pattern common to union movements in most industrialized countries (Allen, 1954; Porter, 1965; Zapf, 1965). Often such stability in leadership positions is interpreted as a clear sign of "oligarchy", that the leaders are "following a logic of self-interest, of exploiting the masses to maintain or extend their own privilege and power" (Lipset's preface to Michels, 1962: 35). In Norwegian unions, however, several factors militate against abuse of power. First, union constitutions limit the power of the top leaders; second, the union system is pervaded by egalitarian norms; and third, the long period of local training creates in the leader a commitment and responsibility toward the membership.

In interviewing union leaders, we did not feel that the problems of reelection were tabooed issues. Quite a few of the leaders maintained that they felt some discomfort about elections and other informants suggested that before elections there are tendencies toward alliance-building and jockeying for positions. But election results indicate that there is little ground for apprehension. In nearly all cases it seems that the top incumbencies are permanent and that elections, therefore, tend to have a ceremonial character.

There are several reasons for this. The most obvious reason is that there is no general opposition within unions. Without institutionalized competition for top positions incumbents may feel safe so long as they do "a good job" in their office and do not violate elementary rules of the union game. Elections, then, become occasions for delegates to express their approval of the top leaders, an occasion for the affirmation of legitimacy. But there is a sharp edge to all elections, and in our interviews we were told that they may be used as instruments of warning ("skudd for baugen") to incumbents who pursue unpopular policies or who have slackened in their efforts. Nevertheless votes are carefully calculated so that incumbents are not suddenly thrown out of office. Why this

cautiousness? Several of our respondents expressed the view that there are very strong norms against throwing out an elected officer so long as he does a moderately good job. Such an official has a moral claim to re-election. Many union members are familiar with the problems of job insecurity, and this may be a reason why they tend to be relatively lenient toward their leaders. As one union president said to us, "Our National Convention is no worse than private employers in general".

But the analogy of the employment relationship should not be carried too far. There are other important, non-rational elements in leader-member relationships which may account for the longevity of incumbents. In his analysis of British trade unions, Allen stressed that a top leader symbolizes his organization, so that "to admit that its senior official is incompetent is a grave reflection on a trade union which it wants to avoid for as long as possible, even at the expense of keeping the incompetent person in office or creating a figurehead post for him to which he can retire with honour" (1954: 233). Allen further observed that union members may engage in hero-worship which makes them blind to the defects of their leaders. In analyzing political parties, Michels speculated that "the mass is sincerely grateful to its leaders, regarding gratitude as a sacred duty. As a rule, this sentiment of gratitude is displayed in the continual re-election of the leaders who have deserved well of the party, so that leadership commonly becomes perpetual" (1962: 92).

LO Leaders as Political Actors

The Norwegian Labor Movement is often described as a unitary movement with political and union branches. The unitary character of the movement is reflected in ceremonies such as parades and speeches on the First of May, in regular meetings of party and union leaders, in the presentation of joint legislative proposals, and in electoral contests when the two branches tend to fuse into one mobilization system directed toward the single purpose of achieving a labor majority in parliamentary and municipal elections. But despite these unitary features, a clear division of labor separates the two branches. Numerous formal links between them indicate that coordination is not automatic, rather it is planned and controlled at all levels. Unfortunately, there are no systematic studies of the various linkages or of the exchanges which occur between the branches. Such studies are essential to an understanding of the integrative as well as divisive forces within the Labor Movement. In this section we examine some aspects of labor leaders' political activities. It is common knowledge that these leaders generally support the Labor Party, but we are here presenting for the first time some data on their differential involvement at various levels of the political system.

Norwegian political parties are "mass parties" in the sense proposed by Duverger. That is, recruitment of members is a fundamental activity, members pay dues to the party, and there is a distinct organizational apparatus "centralized and firmly knit" (Duverger, 1963: 62—71). This type of party organization has an impact on political culture, and within the Norwegian political culture

party membership is regarded as a role having rather strong claims on a person's energy and loyalty. Ideally, an adult party member is expected to be deeply interested in, and well informed, about political matters so that he identifies strongly with his party and is able to defend its policies on all important issues. There is a feeling that it is a member's duty to devote time and energy to the promotion of the party. Moreover, there are norms opposing changes in party allegiance. Such changes are regarded as signs of ideological instability and a certain lack of moral consistency. Thus there are strong boundary maintenance drives within the political system, and various surveys indicate that party membership is a very reliable predictor of political attitudes (Fivelsdal, 1965; Valen & Katz, 1964). Union leaders in Norway share these norms of the political culture and play an important role in maintaining them.

Table 17 not only confirms the common knowledge that LO leaders support the Labor Party but also shows that the party enjoys a virtual monopoly in their ranks. In the elite and apparatus categories there is not a single member of any other party. Among delegates the Labor Party has very strong support with 72 °/o members and only 5 °/o members of other parties. It is of topical interest to note that the left-wing socialists are weakly represented. SF (The Socialist People's Party) has directed much of its propaganda toward union members, but the table shows that the party has been quite unsuccessful in penetrating the union leadership.

Table 17 accords with other studies demonstrating that the Labor Party is strongly entrenched in LO cadres. A national survey of the lower and middle strata of the white-collar population showed that among males holding elective positions within union locals 28 % were Labor Party members, an additional 41 °/o were classified as strong Labor identifiers, while 9 °/o were classified as weak identifiers. Only 2fl/o2^fl/o were left-wing socialists (SF or NKP) (Fivelsdal, 1965: 116). We may conclude, then, that in moving from lower to higher echelons within LO one finds a rapidly increasing proportion of Labor Party members, while non-socialists, left-wing socialists, and non-members appear as insignificant groups of deviates. Within the elite and apparatus it seems that the only political deviation tolerated is that of being politically uncommitted with regard to actual party membership.

esting when confronted with electoral data showing that Labor Party support among blue- and white-collar workers, although of majority proportions, is far from uniform. It seems likely that leadership uniformity does not result from random political choices, therefore, but is the consequence of selective control and recruitment processes. These processes are probably not supervised in detail from LO headquarters, however, but are carried out by loyal cadres at the local level. Wc have no systematic data bearing on this question, but it is our impression that left-wing socialists are barred from higher-level positions. As far .as non-socialists are concerned, there are no indications that they constitute any serious problem for LO. We may speculate that they are few and are generally less ideologically motivated toward leadership positions than left-wing socialists who see it as their essential mission to "reinvigorate a complacent movement". We also speculate that, given Labor Party majorities among local cadres and rank-and-file members, non-socialist aspirants are perceived as class deviates, a perception sufficient to trigger automatic boundary maintenance mechanisms. Left-wing socialists, on the other hand, can more easily be perceived as articulating working-class interests, and because of this it is necessary for the Labor Party and LO leadership to denounce them as "false prophets", and prevent them from infiltrating the higher echelons.

One problem remains. How can we account for the presence of non-partisans who comprise 16% of all union leaders? Are the non-partisans randomly distributed, or can they be located in certain areas of the union movement where Labor Party dominance is weaker? One hypothesis would be that partisanship is less accentuated in white-collar unions, many of which have a non-partisan tradition and have joined LO only during the last decade or two. In Table 18 we have collapsed the categories of elite and apparatus in order to ensure an adequate basis for calculating percentages. We have also isolated respondents belonging to the LO network. At elite and apparatus levels the table shows a correspondence between type of union and partisanship. Blue-collar unions have very high proportions of Labor Party members while white-collar unions have relatively fewer Labor partisans. Mixed unions are in an intermediate position.

Party membership among the LO headquarters and district officials we studied is, however, 100 %. The distribution of delegates' political affiliations also shows that political nonconformity is related to type of union. In blue-collar unions there are only deviates to the left of the Labor Party (apart from the non-partisans). In mixed unions there are followers of both left-wing and nonsocialist parties. But in pure white-collar unions deviation is exclusively to the right (and to non-partisanship) and tends to reflect inchoately the political affiliations of white-collar rank-and-file members.

We have no data on the extent to which nonconformists and nonpartisans among the delegates act as organized factions at meetings, nor have we any data on the extent to which they form alliances with their brethren in the elite and apparatus groups. But there is every reason to suspect that such high proportions of nonconformists and nonpartisans among the delegates of some unions are a factor influencing their general policies. Other data indicate that whitecollar unions affiliated with LO are generally very cautious not to take open political stands (Fivelsdal and Dahl Jacobsen, 1962). In several of the whitecollar and mixed unions affiliation with LO has been criticized on political grounds, and white-collar leaders within the LO system have tended to describe their unions as nonpartisan. But in spite of this distancing from the Labor Party, recent surveys show that white-collar members still perceive their unions as being heavily involved in party politics and party support (Fivelsdal, 1965, 1969).

Party membership is a ticket to the periphery of the political arena. But how many union leaders use their ticket only to watch the game of politics at a closer distance than is possible for non-partisans, and how many have at some time or other actually participated in local and/or national politics? Our use of a mail questionnaire did not permit an extensive series of questions on past and present political activities so we focused on a few salient areas: the holding of elective political positions, the length of party membership, the holding of elective positions within the party, and having had a position on the party slate.

As mentioned previously, party activity may be important for a man's union career. A position in the local party or government can make a man more visible to the rank-and-file as well as to higher union cadres. Participation in Labor Party politics, especially office-holding, may serve to groom a man for a union job. Some people, then, may climb two ladders in their careers, that is, they have one foot on the union ladder and one on the party ladder.

Our data show that a large majority of union leaders joined the Labor Party

in their early twenties after they had been unionized. Table 19 shows a lag in the average ages at which future union leaders were recruited to union and party membership. This may be interpreted in various ways, but we think the data indicate the more deeply embedded position unions have at the workplace level compared with party organizations. Strong pressures on newcomers to join the union are an integral part of the workplace culture whereas, given the seriousness with which party membership is defined in Norwegian political culture, the individual is permitted more leeway in deciding whether to join the party. Because unions are generally closely linked to the Labor Party, then, they may serve as communication networks and instruments of political socialization for the party. When we recall that the leaders we are studying came out of highly politicized family milieux in a majority of cases, and when we now see that, even so, two to three years elapsed between joining the union and joining the party, the important socializing impact of the unions seems even more firmly established.

Turning to the question of union leaders' political activities, we learn from Table 20 that close to one half of all respondents hold or have held an elective municipal position (kommunestyre). Further, a substantial number in all leadership categories have at one time or other served as delegates to the Labor Party's biennial national conventions and have held elective positions in the party at the local level. Only a few, however, have been or are now representatives (01 alternates) to the Storting.

The fact that approximately half of the union leadership has had more 01 less extensive political experience within local government may be of great significance for understanding not only union-party relations., but also the ethos oJ union leaders. First, we assume that the pattern of overlapping leadership 01 multiple office-holding in union and party is a very strong integrative mechanism

on the structural level. Second, experience in politics may serve to integrate the LO unions on a cultural level by giving prospective leaders common experiences beyond their specific workplaces and locals, and by instilling political values which reach beyond bread-and-butter union concerns. We may also speculate that participation in local politics with its pragmatic problems such as the logistics of municipal administration serves to develop in leaders the incrementalist orientation towards political problem-solving we discussed earlier. Third, we assume that having themselves worked as politicians makes it easier for leaders to understand and cooperate with party people in later stages of their careers.

The occupancy of various elective positions within the Labor Party probably also serves to familiarize union leaders with the complexities of the politician's role, and Table 20 shows that a fair number hold or have held district positions in the party. Such positions are often combined with elective municipal positions.

An interesting aspect of the table is the pattern of political activity among members of the labor elite. Apparently these persons were more active in earlier periods than in 1967, but it would be wrong to interpret this as indicating a withdrawal from politics. We must not forget that the table covers mainly lower levels of the political system while elite activity can be expected to occur at the highest levels.

At the elite apex, for that matter, it is difficult to distinguish between union and political activities. Members of the labor elite assume important political roles as members of the powerful National Executive Committee of the Labor Party, and many elite leaders perform political functions as members of the party's various standing committees. A special so-called Cooperative Committee of the Labor Party and LO is maintained to achieve peak coordination of forces, and in 1967—68 this committee consisted of the party chairman, vicechairman, and general secretary, plus the LO president, vice-president, and one union president, the last three elected to the committee by the LO Secretariat.

Wtihout doubt LO and party connections are well established at the top, btu why do so few union leaders hold elective positions within the Storting or the municipalities? Our interviews with union leaders gave us the impression that there exists an informal rule that union work should have top priority, and a feeling that work as a political representative will take too much time and energy from union responsibilities. If a leader is away from his office frequently his absence tends to interrupt his personal communication network and disrupt his union's administrative processes. Being away too often may also result in a loss of contact with day-to-day union problems and finally endanger a leader's prospects for re-election.

The Labor Elite in the National Elite System

Elite persons are those who hold major power roles in the central institutional
orders of society. A special aspect of such roles is the necessity to make decisions
with an eye to their impact on other, usually competing elites and institutional

orders. "What distinguishes elite decisions from decisions along the 'line'," Porter writes, "is that the former have to take rival bureaucracies into consideration. In other words elite decisions are involved in all the decision-making and coordinating functions that lead to conflicts and accommodations between competing bureaucracies" (1965: 223—24). Both the definition and subsequent analysis of elites from this standpoint entail extensive special research on how single elites participate in the national elite system, research on who is involved with whom, when, where, and how. The assumption is that interaction with leaders in other societal areas is sufficiently frequent and intensive so as to make such activity a distinctive feature of all elite roles. Etzioni formulates this assumption rather sharply when he writes that, "National elites devote most of their time and energy to dealing with organized collectivities and their elites and only a small part of their resources to direct appeals to 'the people'" (1968: 433).

But Etzioni's criterion for defining national elite status (devoting most of one's time and energy to dealing with organized collectivities and their elites) is too restrictive. It implies a boundary between those who do little else than interact with other elites and those who only infrequently or never do this. There is in reality no such boundary, only a complex continuum. Moreover, persons clearly at the "elite" end of this continuum, we think, are so few in number as probably not to comprise a distinct group. Speaking descriptively, the time and energy most persons, classifiable as elite on other plausible dimensions, devote to inter-elite affairs are lesser portions of their total work efforts. At least, this seems to be the situation with regard to the Norwegian labor elite. Whether the labor elite is in this respect representative of other Norwegian elites, and whether the Norwegian elite system as a whole is similar to that of other societies are of course empirical questions about which we can only speculate here.

We supplied all full-time officials with a list of ten groups with which they normally have contact, and we asked them to rank the four groups requiring the most time under normal work conditions. A comparison of elite and apparatus responses is presented in Table 21. It shows first that groups inside the union movement make by far the largest demands on leaders' time. Looking at those groups which receive the first and second greatest amounts of time, we see that elite persons are more involved with other leaders in their own unions and less involved with union members in general than are apparatus officials. This presumably indicates the elite's constant involvement with and direction of lower level cadres in the unions, and, accordingly, a reduced opportunity to give top priorities to problems raised directly by rank-and-file members. The greater involvement of apparatus officials with the latter problems and persons is a part of the service function performed by the bureaucratic echelons. Neither the elite nor the apparatus are seen in Table 21 to be intensively or differentially involved with union leaders outside their own unions, although the elite naturally devotes more time to Secretariat affairs than does the apparatus. This finding reminds us that although the LO system, viewed comparatively, is highly centralized, it is still a federation of unions which individually possess considerable

autonomy. Insofar as leaders' energies are given over to groups internal to
the union movement, clearly the greatest part of these energies go to groups
within their own unions.

Looking now at groups outside the union movement which require the first and second greatest amounts of time, we find that members of the labor union elite devote more time to employers' associations and less time to individual local employers than do members of the apparatus, a reflection of the institutionalized system of collective bargaining between those who control the major economic resources in Norway. Requiring only slightly less time than employers' groups are civil servants. The extensive administrative power held by these persons makes them desirable partners for large power blocs such as LO, and our data indicate that labor leaders at both elite and apparatus levels interact fairly extensively with government officials. More surprising is the relative lack of interaction, insofar as our data measure it, with politicians. We suspect that we see here a consequence of the division of labor between LO and the Labor Party wherein the party's leaders both inside and outside parliament have the responsibility of communicating LO's policies and goals to fellow politicians. Indeed, the data cause us to speculate that LO leaders are probably careful not to establish direct lines of communication with non-socialist political leaders lest they upset the delicate and complicated relations with the Labor Party while simultaneously opening themselves to charges of ideological deviance. So long as the Labor Party is able to protect and further the unions' political goals, LO is content to let direct contacts with other parties lie dormant.

Norwegian union leaders often speak about two "roads" on which issues are carried toward solutions, the road of industrial bargaining (forhandlingsvejen) and the road of legislation (lovveien). Generally, they seem to prefer the former road, but when this leads to a dead end they may turn to the party and the legislative road. But besides these two roads there exists another, less conspicuous but still very important road in the political landscape which carries a continuous traffic of big and small issues. This is the road through the Norwegian system of government created committees, boards, and councils.

The development of the committee system is clearly related to the institutionalization of organized interests in Norwegian society. It gives interest organizations direct access to administration and policy-making at the national level, and it provides higher government officials with good opportunities to sound out the big interests on controversial issues. Generally, the committee system serves as an input channel to the state, as an arena for non-public articulation and aggregation of interests, and as a mechanism for coordinating the major institutional sectors of Norwegian society. Since the system provides direct access to policy- and decision-making in the state administration, organizations regard committee representation as a valuable asset. Therefore, struggles for representation may erupt when new committees are being appointed and when various organizations feel that a given committee's jurisdiction impinges on their interests. How representatives are appointed, has not been studied in detail. So far as we are able to ascertain, the normal procedure is that the Ministry contacts the large organizations which have interests at stake in the area covered by the committee, and asks them to nominate one or more representatives with alternates. To ensure status equilibration within the committee, the Ministry may provide informal guidelines as to the hierarchical level at which the committee is supposed to function. All committees are obviously not of equal importance, and many of them deal with fairly routine problems, while others work on problems of top-level planning and policy-making. Committees, then, draw upon personnel from various levels of the bureaucracies involved, and we expect that the representatives' ranks are related to the "grade" of the committee.

Table 22 confirms our expectation that labor elite members as well as bureaucratic personnel serve as committee representatives. But we also note striking differences in the frequency of appointments to committees (including alternates): 72 % of the elite is represented compared to 39 % of the apparatus group and only 16 % of the delegates. Elite persons hold on the average 1.9 committee positions, apparatus personnel 0.9, and delegates 0.2. A closer analysis of the official listing of all committees and their members in 1967 reveals that the top 6 elected officials of the LO network held an average of 5 positions each, and the professional leaders of the four administrative sections in national headquarters held an identical number. It is evident that these professionals are often appointed to committees of specialists or appear on committees as advisers to elite members. Based on the criterion of participation in the committee system

alone, then, the top experts in the LO system obviously qualify as members c the national elite system, while several of the members of our elite categor (mainly leaders of smaller unions) are disqualified from elite status on this ac coun

Finally, LO leaders may be viewed as interacting with elites and collectivities external to the union movement when they participate as leaders, and, in a few cases, members of certain non-union voluntary associations. In these capacities they may be able to influence such associations, or, where interests are already compatible, their participation can further cooperation and solidarity between the unions and the association. Table 23 presents data showing the extent to which LO leaders participated in a fairly wide variety of voluntary associations in 1967.

The most striking feature of the table, however, is the relatively low frequency with which labor leaders actually are involved in non-union associations. Leaders in the elite category held but five positions of national leadership in such associations and no local positions. The larger apparatus and delegate

groups held proportionately even fewer national positions, although they were somewhat more active at the local level. Moreover, if the comparatively less important youth, sports, and hobby associations are held separate, membership in the remaining associations is sparse and infrequent. If the 317 member and leadership positions shown in Table 23 are viewed as a proportion of all union leaders studied we find 0.67 positions per leader.

There are probably a number of reasons why non-union activities are so infrequent. Certainly limited time is a major factor, particularly at the elite level where working hours can often extend into evenings and over weekends. Then too, it can be surmised that LO officials have less unfettered resources to contribute to voluntary associations than businessmen or politicians. LO's funds are tightly controlled and cannot be committed by any one official, and this may make the deliberate recruitment of union officials to leadership positions in nonunion associations less attractive. Further, it seems probable that the ethos of a solidary, independent labor movement having boundaries beyond which one should not go is a factor limiting cross-cutting memberships for union leaders. In any case, we conclude from these data that LO leader«; do not significantly extend their power roles and their social prominence by participation in interest groups and associations outside the sphere of the labor movement.

The foregoing conclusion lends support to a generalized image of LO leaders as comprising a relatively closed corporate structure in Norwegian society. Their social backgrounds effectively distinguish them from other key elites such as business, civil servants, non-socialist politicians, academicians, and so on. They reach power positions in the LO system through long and complicated career patterns and mechanisms which ensure experienced, politically seasoned "insiders" and which effectively train and screen aspiring "outsiders". Their power is in part wielded through a closely allied Labor Party in which many of them have much experience and considerable influence. Insofar as they exercise power outside the channels of party politics they do so through direct contacts with their counterparts in the business and civil service elites. These contacts occur within more or less formal contexts of job performance and the extensive system of government-sponsored committees to which they are named with conderable frequency at the uppermost levels.

It would be premature to infer from this analysis of the labor elite any hard and fast conclusions about the larger system of elites in Norway. At best our research merely raises questions worthy of further investigation. The material in this article essentially depicts the labor elite as located at the top of one organizational pillar in the society. Its members have virtually unchallenged control of an organizational network whose boundaries are more or less coterminous with those of a major social collectivity, the old working class and its contemporary extensions in the lower echelons of the non-manual component of the work force. Are other elites in Norway similarly situated atop corresponding organizational pillars? If such is the case, as we suspect it is, then what are the

consequences for the society and its adaptability to change? Dahrendorf has suggested two possibilities. The first is a kind of cartel of elites, a situation in which groups in control of the various social pillars act defensively to "reduce risk and maintain the status quo". They in effect enter into "an agreement not to hurt one another, but to administer public affairs together" (1967: 255). The other possibility is that through interaction and shared values, elites are able to transcend their strictly defensive interests to form a cohesive, dynamic stratum able to process problems and proposals that imply structural change. The search for empirical referrents of these and similar models in Norway and other modern societies is a primary task for future research.

---

NOTES

1 For further information on labor union structure and problems in Norway, see Galenson (1949), Dorfman (1966), Bull (1968), Fivelsdal (1964, 1965, 1969), Lysgaard (1965), and Leiserson (1959). There is currently a discussion on the reorganization of the whole LO system, cf. the following LO publications: Fagbevegelsen og framtida: Referat fra LO-offensivens startkonferanse 24. september 1968 and Debattopplegg om organisasjonsformene (1969. See further Labor Party Secretary Ronald Bye's pamphlet, Nytt syn på fagbevegelsen (Oslo: AUF, 1968).

2 These figures were made available to us by the LO administration. So far as we can establish, Lipset used a comparable set of figures in making his estimate.

3 25 specialists and experts in LO's central offices; 17 district secretaries employed by LO; an unknown number of specialists in the headquarters of member unions (e.g. 4 in the Municipal Union). Moreover, the 326 paid functionaries were exclusive of service personnel, of whom 63 were employed in 1966.

4 For example, in the LO figures cited The Building Trades Union is reported as having 11 elected and 3 appointed full-time officials. It is also reported to employ 25 functionaries. However, these figures do not include approximately 8 full-time district officials, an unknown number of functionaries staffing the offices of various district committees and union locals, and a national system of paid "measurers", which together number approximately 200 persons. See Engstad and Harm (1957).

5 The number of unions refers to 1966_—1967, the time our study was initiated. Consolidations of a number of the smaller unions in LO have been reducing the total number of unions rapidly over the past few years. Accordingly, the size of the labor elite as it is positionally defined here has been decreasing and now stands at about 75 persons. Of course, exact numerical specifications of analytical categories such as "elite" are always to be treated skeptically. In our specification, for example, we include presidents of some very small unions who certainly have but limited power.

8 The figure of 94 administrative officials is somewhat inflated due to the inclusion of 40 employees of the Sailors Union who man that union's offices in the ports of the world. These persons are obviously too removed and specialized to play significant leadership roles in the Norwegian union movement per se.

---

7 It seems clear that in practice a number of individuals we have classified in the apparatus category possess power and influence commensurate with an elite position. This is, for example, likely to be the case •with several of the elected officials in the larger unions who are the "third men" in their respective hierarchies after the presidents and vice-presidents whom we have called elite. The same consideration probably applies to a few of LO's central office personnel. But in the absence of a detailed study there is no uniform method for singling out such individuals. We have therefore chosen to make a distinction between elite and apparatus that no doubt cuts artificially across the fluid continuum of power at these upper levels of union leadership.

8 This figure excludes the 40 employees of the Sailors Union whom we discussed earlier.

9 Arbeiderbladet, May 21, 1969, p. 5.

10 However, by what is apparently a special arrangement, the third-ranking official of the Clothing Workers Union in 1967, a woman, held a permanent seat on the LO Secretariat instead of the president of that union.

11 The female President of the Danish Tobacco Worker's Union (with a majority of female members) said in an interview in 1969: "Women lack self-confidence, and therefore they don't have much confidence in each other. When I was nominated for the position as Union President, many of the members felt uncertain about female leadership. They came close to feeling it was a little shameful." Berlingske Tidende, January 19, 1969.

12 Of course there are many reasons for the reformist character of labor movements in Northern Europe. One such reason, particularly applicable to Norway, is the relative openness of the political system to newly mobilized groups. See T. H. Marshall, "Citizenship and Social Class" in Marshall (1964).

13 We may observe here that the exact boundaries of the labor aristocracy must be adjusted to take into account particular features of stratification patterns in different societies. One such feature in Norway was the absence of a sharp discontinuity between skilled laborers and minor functionaries, probably as a consequence of the pervasive egalitarianism of Norwegian culture. This further suggests one reason why LO has been so successful in organizing white-collar workers compared with continental labor movements. Cf. Fivelsdal, 1965: 82^—85.

14 For further analysis of problems related to the blue-collar/white-collar cleavage, cf. Fivelsdal and Dahl Jacobsen, 1962, and Fivelsdal, 1965.

REFERENCES

Allen, V. L. Power in Trade Unions, London, Longmans, 1954.

Aubert, Vilhelm. "Lagdeling", pp. 130—183 in N. R. Ramsøy (ed.), Det norske samfunn, Oslo,
Gyldendal, 1968.

Bull, Edvard. Norsk fagbevegelse, Oslo, Tiden, 1968.

Butler, D. E. and Richard Rose. The British General Election of 1959, London, Macmillan, 1960.

Dahrendorf, Ralf. Society and Democracy in Germany, New York, Doubleday, 1967.

Dorfman, Herbert. Labor Relations in Norway, Oslo, The Norwegian Joint Committee on International
Social Policy, 1966.

Duverger, Maurice. Political Parties, New York, John Wiley 8c Sons, 1963.

Eckstein, Harry. Division and Cohesion in Democracy — A Study of Norway, Princeton, N.J.,
Princeton University Press, 1966.

Edinger, Lewis J. and Donald D. Searing. "Social Background in Elite Analysis: A Methodological
Inquiry", American Political Science Review, vol. LXI, no. 2, 1967.

Eldersveld, Samuel F. Political Parties, Chicago, Rand McNally, 1964.

Engstad, Paul and Henry Harm. Fagorganisasjonens oppbygging, Oslo, Arbeidernes Opplysningsforbund,
1957.

Etzioni, Amitai. The Active Society, New York, The Free Press, 1968.

Fivelsdal, Egil and Knut Dahl Jacobsen. Interesseorganisasjoner og stortingsvalg, Oslo, Institutt
for samfunnsforskning, 1962.

Fivelsdal, Egil. Funksjonaerenes syn på faglige og politiske spørsmål, Oslo, Universitetsforlaget,
1964.

Fivelsdal, Egil. "White-Collar Unions and the Norwegian Labor Movement", Industrial Relations,
vol. 5, 1965: 80—92.

Fivelsdal, Egil. Hva mener medlemmene? Oslo, Statstjenestemannskartellet, 1969.

Galenson, Walter. Trade Union Democracy in Western Europe, Berkeley, University of California
Press, 1961.

Goldthorpe, John H. et al. "The Affluent Worker and the Thesis of Embourgeoisement", in
Joseph A. Kahl (ed.), Comparative Perspectives on Stratification, Boston, Little, Brown and
Company, 1968.

Guttsman, W. L. The British Political Elite, London, MacGibbon & Kee, 1965.

Hobsbawm, Eric J. Labouring Men: Studies in the History of Labour, London, Weidenfeld and
Nicolson, 1964.

Keller, Suzanne. Beyond the Ruling Class, New York, Random House, 1963.

Leiserson, Mark W. Wages and Economic Control in Norway, 194)—1957, Cambridge, Mass.,
Harvard University Press, 1959.

Lijphart, Arend. The Politics of Accomodation: Pluralism and Democracy in the Netherlands,
Berkeley — Los Angeles, University of California Press, 1968.

Lipset, Seymour Martin. The First New Nation, New York, Doubleday, Anchor Books ed., 1967.

Lysgaard, Sverre. Arbeidernes syn på faglige og politiske spørsmål, Oslo, Universitetsforlaget,
1965.

Marshall, T. H. Class, Citizenship and Social Development, New York, Doubleday, 1964.

Marvick, Dwaine. "Political Recruitment and Careers", International Encyclopedia of the Social
Sciences, vol. 12, 1968: 273—281.

Michels, Robert. Political Parties, New York, The Free Press, 1962.

Mills, C. Wright. The New Men of Power: America's Labor Leaders, New York, Harcourt, Brace
and Company, 1948.

Mills, C. Wright. The Power Elite, New York, Oxford University Press, 1956.

Rokkan, Stein. "Norway: Numerical Democracy and Corporate Pluralism", in Robert A. Dahl
(ed.). Political Oppositions in Western Democracies, New Haven, Yale University Press, 1966.

Valen, Henry and Daniel Katz. Political Parties in Norway, Oslo, Universitetsforlaget, 1964.