Dynasty Formation in Eight Imaginary Societies*

L 0 R N E T E P P E R M A N / University of Toronto

B A R R Y T E P P E R M A N / University of Toronto

De quelle manihre les aspirations populaires et les mkcanismes de contrble de l’klite permettent-ils l’ktablissement de dynasties ? Quelles sont les diffkrences qui existent entre les sociktks dynastiques et les sociktks non-dynastiques dans les possibilitks rkelles de mobilitC sociale intergknkrationnelle ? Cet article tente d’apporter une rkponse a ces questions par une analyse mkcanographique simulke de huit sociCtks fictives B dont les effectifs dkmographiques sont modelks a l’image de l’Ontario.

Nos analyses rCv5lent que les sociCtks dynastiques posddent les caractCrktiques suivantes : ( a ) une tr&s forte probabilitk d’hCriter un statut d’appartenance a l’klite; ( b ) une faible chance d’entrer dans l’klite en provenance d u n e classe infkrieure; (c) les individus mobiles se retrouvent dans un nombre l i d 3 de classes; ( d ) une probabilitk mCdiane relativement klevke d’une mobilitC ascen- dante modkrke (une i deux classes) ; et ( e ) un faible risque mkdian d’une mobilitk descendante. Ces rksultats sont expliques et vCrifiks, par aprts, en utilisant des donnkes provenant de sociCtks contemporaines. On discute, en dernier lieu, 5 la lumi5re de ces rksultats, les avantages des sociktks dynastiques et non-dynastiques.

How do popular aspirations and elite controls contribute to the formation of dynasties? How do the opportunities for intergenerational social mobility differ in dynastic and non-dynastic societies? This paper attempts to answer these ques- tions through a computer simulation of eight reasonably realistic “imaginary societies” modelled deinographically after that of Ontario.

The simulations reveal that dynastic societies are characterized by ( a ) a high probability of inheritance of elite status; ( b ) a low probability of entering the elite from below; (c ) the restriction of mobile persons to a limited number of classes; ( d ) a relatively high median probability of moderate (one or two class) upward mobility; and (e) a relatively low median risk of downward mobility. These findings are explained and then verified by data from actual societies.

The relative advantages of dynastic and non-dynastic societies are discussed in view of these findings.

John Porter has written of the Canadian elite, “To allocate these power roles on the basis of ability serves both the principle of equality, which is a norma- tive value, as well as the principle of efficiency, which is an instrumental value in social development and survival.” (Porter, 1965 : 21 8 . )

In this statement, Porter has enunciated several of the basic assumptions of a functional theory of stratification, to which the reader is asked to grant * The authors gratefully acknowledge a grant-in-aid of research from the Department

of University Affairs, which enabled the junior author to give his time to this and related research. This paper reflects a diffuse intellectual debt to Professors George Hornans and Harrison White of Harvard University. 121

Rev. canad. SOC. & Anth./Canad. Rev. Soc. & Anth. 8(3 ) 1971

validity: notably, ( 1 ) power resides in roles, to which people must be allo- cated; ( 2 ) people may be allocated to power roles justly and unjustly, given our normative standards, but also efficientIy (or usefully) as well as ineffi- ciently (or uselessly), given the criterion of contribution to social develop- ment; and by extension, ( 3 ) the utilization of talented or able people in power roles may require, in the absence of structural expansion (i.e., increase in the number of such roles), the death, downward mobility, or non-inheri- tance of these roles by less talented people. Thus, the upward mobility of talented people is limited to some unspecified degree by the inter-generational inheritance of power roles, and arguing back to the beginning, social develop ment and survival are also dependent on the extent of such inheritance.

This general argument can be similarly applied to all scarce and rewarding roles, which will be referred to below as elite positions. The unbroken accession of sons to such positions may be generalized from the case of power roles in the term dynmty, which the dictionary defines as a “succession of rulers from the same family or line; a family or group that maintains power for several generations.” In the discussion below, however, the term dynasty will specifically denote a five-generation succession of fathers and sons through elite positions.

Porter has also argued that “control over (elite) recruitment is an exten- sion of elite power” (1965 :218). If it is true that an incumbent elite can control the entry of people into elite positions, then by continuously doing so, it can largely determine the formation of dynasties, if elite fathers are able to replace themselves with sons and the number of elite positions does not diminish. In the discussion below, “replacement” and “inheritance” within the elite will imply the transmission of elite status from any elite father to any elite son. While kinship particularism is a major component of such transmission, it is not held to be the only one.

Yet one must not neglect the importance of motivation or aspiration to elite status both in the elite class and in the sub-elite population. Even if the elite has perfect control over recruitment, dynasties may fail to form if elite sons do not aspire to their fathers’ elite position. If an elite position carries obligations of social responsibility - such as Porter’s roles of power and authority, or Baltzell’s roles of communal leadership (Baltzell, 1964:7O-86 et parsim) - the sons of elite fathers who are unwilling to assume such responsibility may leave the available elite roles vacant. If, on the other hand, elite control over recruitment is less than perfect, or if perfect control is exercised but with the aim of encouraging competition from below and selecting the best contender for recruitment, then the sons of elite fathers may not be guaranteed a position in the elite. In either event, the formation of dynasties may be uncertain or improbable where aspiration to elite posi- tions is a problematic operating variable.

This raises an interesting practical problem. Assuming as Porter does that ‘‘to allocate these (elite) roles on the basis of ability serves ... the principle of eyu:ility,” a normative value of considerable importance to many, and 122

assuming with Porter again that “control over recruitment is an extension of elite power,” and thus something beyond the control of most, how much leeway is left in the service of the principle of equality (or perhaps more accurately, the principle of rewarded merit) through the operation of the aspirational variable? Assuming for simplicity that all people, regardless of class, could be given the same aspirations, would dynasty formation be increased or diminished by instilling desires to ascend the social ladder?

Most generally, the two purposes of this paper are (1) to measure the relative effects on dynasty formation of (a) elite-held norms of eligibility or recruitment, and (b) commonly-held norms of aspiration; and (2) to examine the effects of dynasty formation on mobility and class composition in the elite and sub-elite classes.

M E T H O D : SOME C O M M O N A S S U M P T I O N S

Assume that there are two types of elite-held recruitment norms and four types of commonly-held aspiration norms, which will be specified in some detail below. Through computer simulation we shall “create” eight imagin- ary societies and each of these societies will be characterized by one of eight (i.e. 2 X 4) pairs of assumptions about these normative orientations. More specifically, each of these eight societies will be characterized by identical structures and processes, except those deriving from the recruitment and aspirations norms which are the independent variables and differ in each of the eight societies.

The eight societies begin their hypothetical existence around the year 1800. Each society has seven classes of which the top class or set of positions is called the elite. We shall purposely avoid discussion of whether this struc- ture is most properly designated as a class system, authority system, or prestige hierarchy, and proceed as though it is a perfectly integrated com- bination of all three systems, for such a simplistic assumption will not affect the analysis in any apparent way. Suffice to say this social structure is highly differentiated with an attractive top, unattractive bottom, and of course a middle.

As these societies are to exist from 1800 to 1950, and during this time the eight imaginary societies will not only increase in population but also “modernize,” it will be necessary to build the effects of structural expansion into this model. The most obvious characteristic of a modernizing class struc- ture is a change in shape of the structure from what could be called a pyramid or cone to what could be called a diamond; from a structure that has a large proportion of people in the lowest classes and few in the top classes, to a structure that has a large proportion of people in the middle classes and rela- tively few in the top and bottom c1asses.l This gradual change in shape will

1 This has been the subject of some debate. Goldthorpe, for example, appears to accept the validity of the assertion (1966:453, 456) as a “formula summing up historical experience” of economic and occupational change during industrialization, but 123

124

be built into each of the imaginary societies, following a somewhat compli- cated mathematical formulation.2

These eight societies have been patterned demographically after Ontario (initially Upper Canada) between the years of 1800 and 1950. The popula- tion of these societies, initially 3 1 ,000,3 increases both through immigration and reproduction. Rates of natural increase for the earliest periods were extrapolated from available statistics for those and later years, while the rates for later years are precisely those of Ontario.‘ We shall concern our- selves only with generational rates of natural increase of males as we are chiefly concerned with father-son inter-generational mobility. These rates are as follows: 1800-1830, 7.000; 1830-1860, 3.3382; 1860-1890, 1.7534; 1890-1920, 1.3874; 1920-1950, 1.3571.6 The meaning of these

rejects the reasoning of those who use this “formula” to put forward a theory of convergence of modem societies. Others who appear to accept this generalization implicitly include Kahl (1961:67), who focuses on occupational stratification, and Lenski (1966:308-313 ct passim), who concerns himself with political, as well as economic and occupational, stratification.

2 Lydall‘s work (1968) implies that this change in shape may be more adequately characterized as a frequency distribution initially skewed to the extreme left (like a Pareto distribution), which moves gradually towards a log-normal distribution (like a normal distribution with a slight skew to the left). Thus, a slight preponderance of lower- and lower-middle-class people remains, and the shape of society is never quite a diamond: that is, never quite symmetrically arrayed about a numerically pre- dominant middle class.

Accordingly, in creating five class structures for the simulation - one structure for each of 1950,1920,1890, 1860, and 1830 (and before) - we have worked from this notion of a density function defining class distribution whose mean is moving gradually to the right, stopping short of dead centre by 1950. An examination of the literature showed that beta density functions with selected constants provided such a set of distributions with intuitively satisfactory amounts of skew and kurtosis.

Beta density functions are given by the formula:

where b is a constant put at 6 in our study; a ranges in value from 1 through 5 (increasing by 1 each generation after the first); and p ranges from 0.1 through 0.9. (See Mooteller and Tukey, 1968:173-178). The proportions of beta values asso- ciated with p values of 0.1 through 0.6, and 0.7-0.9, were used to designate the proportions of the population in Classes 7 through 2 and Class 1, respectively.

3 This figure is something of a compromise among varying estimates. Cook (1967:21) has estimated the population of Upper Canada at 14,000 in 1791 and more than 90,000 twenty years later, while Clark (1962:66) has estimated the population at 33,000 in 1812 and 952,000 by 1851.

4 Rates were calculated from the following sources: for Ontario, 1941-1965, Canada Year Book (1969:224); for Ontario, 1881-1931, Keyfitz (1950:52-54); for Canada, 1851-1871, Keyfitz (1950:51). The remaining rates, between 1800 and 1840, were obtained by extrapolation from a logarithmic plot of the afore-mentioned data. The rates of annual natural increase used were: 1800-1830, 1.0670; 1830-1860,

1.0410; 1860-1890, 1.0191; 1890-1920, 1.0113; 1920-1950, 1.0103. 5Statistics provided by Henripin (1968:30) suggest that these are not too far off.

While an annual rate of natural increase of 6.7 per cent, giving a generational rate of 7.000 (18oCr1830) may seem prodigious or even incredible, it was not quite so remarkable in a pioneering society of one-and-a-half centuries ago. If the earliest, extrapolated, rates err on the high side, they probably do so by 1 son or fewer and may, at worst, be considered a liberal estimate benefiting the lower classes propor- tionally more than the elite, as these reproduction rates are eventually adjusted by class. (See n.6.)

rates is straightforward. For example, between 1920-50, the average father in the population “replaced” himself with 1.3571 sons (and an equal number of daughters).

Specific rates of natural increase for designated fathers are determined by class, as well as by generation. The assumption is made, in keeping with statistics presented by Wrigley (1969: 186-1891, for a British population, that in every generation the lowest class “replaces” itself with roughly twice as many children as the elite class, and rates of the fathers of intermediate classes are graduated between these limits. Thus the rates of increase of the seven classes, relative to those of the top class, may be taken as 6/6, 7/6, 8/6, 9/6, 10/6, 11/6, 12/6 from top to bottom respectively. Given the generational average male rate of increase, classdetermined rates relative to Class 1, and the number of fathers in each class at the beginuing of the generation, one can solve an equation to determine the number of sons born into each class in the given generation.6

There is positive net migration throughout the hundred-and-!ifty-year period, and for simplicity we shall assume that the rates of immigration and emigration are constant, such that there are roughly 125,000 more immi- grants arriving every thirty years than there are emigrants leaving during the same time period.‘ All immigrants enter the bottom four classes in each societyS in equal numbers, but are thereafter indistinguishable from native born members of that class in their chances of mobility. Class of origin then,

6 The equation is

7 7

re C ( k i d = f 1 (md I m 1 1 = 1

where re is the (unknown) reproduction rate adjusted by class, 7 is the average generational rate of natural increase, k, is the class-determined weight relative to Class 1, and mi is the number of fathers in Class i.

The number of sons born into Class i in the given generation is, then, krmtrs. 7 This sum is derived by assuming a constant rate of immigration, zero emigration,

and identical reproduction rates for immigrant and native-born fathers, such that in 1950, the entire population, totalling 4,597,532 (Ontario, 1951) and composed of recent immigrants and native-born descendants is given by

P1951 = [(rAnt 1920-1950) immig.] + [(children of I + I+

((1.3571)(1.3874)(1)]

{(1.3571)(1.3874)(1.7534)(1)) (grt. gr. children of immig.

immig. 1800-1830)

1 183C1860)

{(1.3571)(1.3874)(1.7534)(3.3382)(7.OOO)(31,OOO)} (grt. grt. grt. gr. children of original settlers)

= 4,597,542.

Solving the equation gives I = 124,910. The Ontario populations predicted by this value of 1 and the rates of natural increase specified always fall between 0.85 and 1.15 of the acrual Ontario populations in 1860, 1890, 1920, and 1950. This provides some basis for confidence in our assumptions about immigration and reproduction.

8 This assumption is conveniently simple and probably more valid than any other simple assumption about immigrants. See, for example, Porter (1965:6&103) on “entrance status.” 125

rather than ethnic origin or immigrant status, is held to be of main importance to the opportunity for m~bili ty.~

M E T H O D : S O M E V A R I A B L E ASSUMPTIONS

As noted above, these eight societies are dserentiated by popular aspirations and elite controls over recruitment.

One of four types of aspirations is held in each society. In two societies, all sons regardless of their class of origin, are indifferent to their class of destination. They feel equal interest or disdain for the top, bottom and middle class positions, and therefore have, ignoring momentarily interference from controls over recruitment or the limitations of available space, equal (one seventh) probabilities of ending up in any of the seven classes. Where such popular norms prevail, the societies shall be called free flow societies.

In a second pair of societies, all sons prefer to be as little mobile as pos- sible; accordingly classes of destination are preferred to the degree that they “neighbour” upon class of origin. Inheritance of father’s class is desired above all else; mobility to a class one step above or below father’s class is, let us suppose, one-half as desirable as father’s class; the classes two steps away from father’s class are one-quarter as desirable; and so on, such that a class six levels above or below father’s class is ( % 6 ) or 1/64 as desirable as retention of one’s class of origin (i.e. father’s class). Although it may be difficult to imagine such preferences being exercised, the possibility that such preferences exist has been suggested by both Durkheim’s notion of anomic suicide (Durkheim, 1951 :246-258) and Joel Levine’s reanalysis (Levine, 1969: 12 et passim) of data on intergenerational mobility collected by Glass (1954). Whether such preferences follow a geometric function in “real life” is a matter demanding empirical verification not at present at hand. Where the popular norms described above do prevail, as in two of the eight imaginary societies, these shall be called neighbourhood flow societies.

In a third pair of societies, perhaps much like our own, any downward mobility is despised and all sons seek only to retain or rise above their class of origin. The class a step above one’s class of origin is twice as desirable as one’s class of origin; the class two steps up is (29 or four times as desirable; and the class six steps above one’s class of origin is accordingly (26) or sixty-four times as desirable as one’s class of origin. All classes below one’s class of origin are completely undesirable, or valueless. These two societies shall be called ascending flow societies, for only ascent is valued or held to be desirable in the popular culture.

In the final pair of societies, neither positive nor negative value is attached to the “distance travelled” from class of origin to class of destination, but rather value is attached to the various classes of possible destination them-

9 This assertion, indeed, seems to be the main thrust of Kalbach’s monograph. Ethnic differences, and more specifically, differences between native-born and immigrant, are in most cases pretty quickly subsumed by class differences. See Kalbach

126 (1970:241-312); also Baltzell (1964:48-53).

selves. Everyone is agreed, in these societies, that the bottom class is least attractive of all possible destinations; the next level up (Class 6 ) is twice as desirable as the bottom class, and the class above that (Class 5 ) twice as desirable again; thus the elite class is held by all to be (2*) or 64 times as attractive as the bottom class in society. This agreement on the relative excellence of various class positions is reminiscent of the high degrees of consensus indicated in national surveys of occupational prestige (Hodge et al., 1966); for this reason, where such popular norms prevail regarding the over-all ranking system, as in two of the imaginary societies, they shall be designated as consensual flow societies.

As these popular aspiration norms are analytically distinct from, and for the moment unimpeded by, restrictions imposed by the elite, or by the sue or shape of the class structure, they can be thought of as normative tran- sition probabilities for father-son intergenerational mobility. They are p r e sented as such in Tables I through IV below.

Finally, each of the eight societies is characterized by one of two patterns of control over recruitment into the elite. The first recruitment pattern may be called elitist control; under this condition, whenever two sons are seeking the same position, preference is given to the son from the higher class of origin, and the “loser” of this competition must compete for the next highest class position. If, then, fourteen elite positions are available for incumbency,

TABLE I NORMATIVE TRANSITION PROBABILITTES FOR FATHER-SON INTERGENERATIONAL MOBILITY: FREE FLOW

Son’s desired rank Class of origin I 2 3 4 5 6 7 -

1 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 2 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 3 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 4 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 5 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 6 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 7 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857 0.142857

TABLE ll NORhiATIVE TRANSllTON PROBABILITIES FOR FATHER-SON INTERGENERATIONAL MOBILITY : NEIGHBOURHOOD FLOW

Son’s desired rank Class of origin I 2 3 4 5 6 7

1 0.503937 0.251968 0.125984 2 0.202531 0.405063 0.202531 3 0.093023 0.186046 0.372093 4 0.045454 0.090909 0.181818 5 0.023255 0.046511 0.093023 6 0.012658 0.025316 0.050632 7 0.007874 0.015748 0.031496

0.062992 0.03 1496 0.101265 0.050632 0.186046 0.093023 0.363636 0.181818 0.186046 0.372093 0.101265 0.202531 0.062992 0.125984

0.01 5748 0.007874 0.025316 0.012658 0.04651 1 0.023255 0.09Ow9 0.045454

0.405063 0.20253 1 0.251968 0.503937 127

0.186046 0.093023

TABLE III NORMATIVE TRANSITON PROBABILlTlF!3 FOR FATHER-SON INTERGENBRATIONAL MOBILITY: ASCENDING FLOW

Son’s desired rank Chs of orkin 1 2 3 4 5 6 7

~

1 0.999999 0.000000 0.000000 0.000000 0.000000 2 0.666666 0.333333 0.000000 0.000000 0.000000 3 0.571428 0.285714 0.142857 0.000000 0.000000 4 0.533333 0.266666 0.133333 0.066666 0.000000 5 0.516128 0.258064 0.129032 0.064516 0.032258 6 0.507936 0.253968 0.126984 0.063492 0.031746 7 0.503937 0.251968 0.125984 0.062992 0.031496

~~~

0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.015873 0.000000 0.015748 0.007874

TABLE IV NORMATIVE TRANSITION PROBABILITIES FOR FATHER-SON INTERGENERATIONAL MOBILITY : CONSENSUAL FWW

Son’s desired rank C h s of origin I 2 3 4 5 6 7

1 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 2 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 3 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 4 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 5 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 6 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874 7 0.503937 0.251968 0.125984 0.062992 0.031496 0.015748 0.007874

and seventy sons, ten from each class, are aspiring to those fourteen positions, the positions are Wed by ten sons of elite origin and four sons of Class 2 origin; the remaining fifty-six candidates are forced to compete for Class 2 positions (against those who have entered competition specifically for a Class 2 position, as well as against one another). This procedure continues until all sons have been allocated to class positions and all class positions are filled.I0

The second recruitment pattern may be called egalitarian control. Under this condition, whenever two sons are seeking the same position, no prefer- ence is given to the son from the higher class of origin, and the “winner” is decided by chance. The “loser” as before is forced to compete for the next highest available class position. If, as before, fourteen elite positions are available for incumbency, and seventy sons, ten from each class, are aspiring to those fourteen positions, two sons are selected from each of the seven classes. In all other respects this pattern of competition is similar to the one described as elitist control, and continues until all sons have been allocated to positions, and all class positions have been filled.

1OSince preference is given to higher-born sons not only in the recruitment of elite members, but also in the recruitment to sub-elite positions, this process may be more appropriately considered elite-inspired popular preference, or elite influence exer- cised throughout the class structure. 128

B R I N G I N G M E N B A C K IN

In the year 1800, there are 31,000 people in each society. Before 1830, these people all “replace” themselves with sons and daughters and die. This inattention to the daerential longevity of fathers in general, and fathers of different social classes in particular is expedient but possibly unjustifiable from an empirical standpoint, and may have important consequences for the validity of the simulation. In particular this model may tend to underestimate the number of elite men co-existing with their sons, and thus overestimate the number of vacant elite positions in any given time period. As a coll~e quence, there could be less upward mobility in dynastic societies than we discover (and explain) below. However, being far from certain of such con- sequences, we are willing to proceed with no more than a warning to the reader.

During the same period, 125,000 immigrant men and women enter the bottom four classes. Some time before 1830 the native sons come of age and, together with the immigrants, “express” their preferences for a position in the class structure. As the population grows, the number of positions in each class increases also, although the shape of the class structure does not begin to change until after 1830. In 1830 a tally is made of all the class positions vacated by the death of their original incumbents. Then, taking into account the preferences of sons and immigrants, and the type of control exercised over recruitment into the elite (and other classes), all candidates are do- cated to class positions and all the positions are med. The sons and immi- grants then, according to their class positions, begin replacing themselves and dying off. New immigrants enter, new sons and immigrants express their preferences for a class of destination. As the population grows, the number of positions in each class grows in number, with the upper and middle classes growing more rapidly than the lower class positions. In this way, the shape of the class structure changes slightly from a pyramid shape, although it is still far from a diamond in shape. By 1860, it is again time to allocate men to positions, as in 1830.

This system of processes continues seriatim until, by 1950, the eight societies have each grown in population to the size of Ontario, and each is characterized by a diamond-shaped class structure.ll

By 1920, the original bottom two classes have disappeared from all eight societies except as a point of entry for half of the immigrants.

The five remaining classes are somewhat more consistent in size in elitist than in egalitarian societies. In the former, the population is divided into the first through jifth classes in roughly the ratio of 1/2/3/3/1, or 10 per cent, 20 per cent, 30 percent, 30 percent, 10 percent respectively. This fivexlass structure is reminiscent of that constructed by Warner in “Jonesville” through rating by matched agreements. (Warner, 1960:70.) By contrast, in the

I 1 The computer program by which this was accomplished may be obtained from the 129 authors upon request.

egalitarian societies, Class 5 too has almost disappeared and the remaining classes are stiU in the rough ratio of 1/2/3/3. In all cases, then, the elite class comprises about 10 per cent of the population after 1920.12

Table v shows that as the class structure has changed shape and the elite class has expanded both relatively and absolutely in size, the sons of elite fathers have benefitted considerably. In most cases, the chances of inheriting father’s elite position have been increased by this structural expan- sion, and in no case have they decreased. A three-way analysis of variance without replication confums the obvious: sons of elite men are most advan- taged in elitist societies. Elitist recruitment norms are, in the final generation (1920-1950), providing between four and nine times as good a chance of elite inheritance as the egalitarian societies with identical aspirational norms. In the fifth generation of dynastic societies, sons of elite men have never less than one chance in two of inheriting their fathers’ positions. However, the direct effects of aspirational norms are also sigacant; regardless of recruit- ment norms, there is the least likelihood of elite inheritance under popular norms of free flow, and the greatest likelihood under conditions of ascending flow and neighburhood flow.

DYNASTY FORMATION AND MAINTENANCE

What is the probability, under these various simulated conditions, that direct patrilinear descendants of an elite father will retain continuous elite status through five generations? What is the probability, that is, of an elite father of generation 0, giving birth to a dynasty?

In order to derive expectations about the development of dynasties - families maintaining an unbroken male succession of elite positions through five generations - we must be wilIing to accept the basic assumption of h4arkov chain processes: namely, probabilistic events are determined only by immediateZy preceding events. One must assume, for example, that the probability that a family which has retained its elite status for four genera- tions will retain it for a fifth generation is precisely the same as the probability that a family first achieving elite status in its fourth generation will maintain its status for another generation. Markov chains, then, are strings of events, states or conditions whose distant histories are obliterated with each step forward.

By accepting such an assumption, one may use independent probabilities (of elite sons inheriting their father’s position in each of five generations) to provide the joint probability of a five-generation dynasty. The probability of a dynasty, PO, is then .Po = [(PI) ( p d ( p d ( p 4 ) (pall, where PI ... p5 are the independent probabilities of elite sons inheriting their fathers’ positions in each of generations 1 through 5.

12 It is di!€icult to predict the forms of stratification in these eight societies had their populations grown less rapidly, or even declined; this question and its relation to the “convergence theory” must await further simulations. See also n.1 on this issue. 130

TA

BL

E V

THE PROBABILITY

OF

AN

EL

ITE

SON I

NH

ER

lllN

G A

N ELITE P

OSI

TIO

N,

BY R

eCR

UlT

MEN

T NORM,

GE

NE

RA

TIO

N. a

nd A

SPIR

AT

ION

NORM

Rec

ruitm

ent n

orm

Elit

ist

Ega

litar

ian

Gen

erat

ion

Gen

erat

ion

Asp

irat

ion

norm

1

2 3

4 5

1 2

3 4

5

Free

0.14

3 0.

286

0.57

1 0.

571

0.57

1 0.

003

0.00

6 0.

016

0.04

1 0.

086

Nei

ghbo

urho

od

0.50

4 0.

756

0.88

2 0.

882

0.88

2 0.

071

0.09

7 0.

138

0.19

7 0.

227

Asc

endi

ng

1.00

1.

00

1.00

1.

00

1.00

0.

005

0.01

3 0.

028

0.06

0 0.

113

Con

sens

ual

0.50

4 0.

756

0.756

0.75

6 0.

756

0.00

2 0.

008

0.02

1 0.

052

0.10

5

TABLE VI THE PROBABILITY OF D Y N ~ FORMATION*, BY TYPE OF RECRUITMENT NORM A N D ASPIRATION NORM

Recruitment norm

Elitist Egalitarian

Free flow 0.0076159 0.0000000 Neighburhood flow 0.2612687 0.0000425 Ascending flow 0.9999994 0.0000000 Consensual flow 0.1645308 0.0000000

* Data derived from Table v by the formula PD = Cpl) (pJ (pj) (pd) (ps).

Table VI indicates that egalitarian recruitment norms ensure the formation of Virtually no dynasties, while under elitist norms the probability of dynasty formation is very much a function of the aspiration norms in that society, ranging from a probability of just under 0.01 (for free flow) to a probability of just under 1 .OO (for ascending flow) .13

At fkst glance, the data suggest that in elitist societies, a high common motivation to ascend paradoxically maximizes dynasty formation and thereby maximizes common frustration with the opportunities for ascent. Yet a comparison of “current” (fifth generation) characteristics of the eight societies suggests that such frustration may not develop after all. Indeed another paradox is suggested that will need some resolution. As one might expect, the greater the extent of dynasty formation, the greater is the (current) probability of sons inheriting their father’s elite position or an equivalent position. (See Table w.) As one might also expect from the literature on traditional societies characterized by dynasties (e.g. caste or estate societies) the greater the extent of dynasty formation, the fewer are the (median) number of classes open to those sons born below the elite class. The same impemeabilityl4 that keeps low-born men out of the elite keeps all men locked into a limited range of positions and promotes elite succession.

The paradox is as follows. Although one finds that the greater the extent of dynasty formation, the smaller is the (median) probability of entering the elite from below, yet (a) the greater is the (median) probability of sub-elite sons moving up at least one class from their class of origin, and ( b ) the smaller is the (median) probability of sub-elite sons moving down one class or more from their class of origin. Thus, although dynasty formation does tend to thwart outsiders ambitious to enter the elite, it also tends to encourage

13 As an aside to those particularly interested in the straacation system of Ontario, a recently completed paper by these authors, “The Natural Disruption of Dynasties,” has analysed the dynasty formation of prominent Ontario families and concluded that the province has elitist patterns of recruitment, most likely of the free flow or consensual flow type. No more than 5 per cent of sampled Ontario families have retained elite status for five generations. A copy of this paper may be had on request.

14 See Svalastoga (1965:36-70) on this point. His notion of “permeability” as a key dimension of social structure has been of considerable use in the formulation of the

132 present research.

~~~

~ ~

~~

~~

TA

BL

E V

II

CU

RR

EN

T C

HA

RA

CT

ER

IST

ICS

OF

EIG

HT

IM

AG

INA

RY

SO

CIE

TIES

(192

0-19

50)

BY

EX

TEN

T O

F D

YN

AST

Y F

OR

MA

TIO

N

Exte

nt of

dym

sty f

orm

atio

n H

igh +

+L

ow

Mod

el'

-

2 3

4 5

6 7

8 ~

. ~

I

Prob

abili

ty o

f in

herit

ance

Med

ian

prob

abili

ty o

f en

terin

g el

ite fr

om

Med

ian

num

ber

of c

lass

es

of d

estin

atio

n op

en to

Med

ian

prob

abili

ty o

f su

b-el

ites m

ovin

g up

Med

ian

prob

abili

ty o

f su

b-el

ites m

ovin

g do

wn

* Mod

els a

re, r

espe

ctiv

ely:

asce

ndin

g el

itist

; nei

ghbo

urho

od e

litis

t; co

nsen

sual

elit

ist;

free

elit

ist:

neig

hbou

rhoo

d eg

alita

rian

; asc

endi

ng e

galit

aria

n; c

onse

nsua

l

of e

lite

stat

us

1.OOO

0.88

2 0.

756

0.57

1 0.

227

0.11

3 0.

105

0.08

6

sub-

elite

orig

in

0.OOO

0.OOO

0.OOO

0.OOO

0.15

8 0.

117

0.12

0 0.

130

sub-

elite

2.

000

2.50

0 2.

500

3.50

0 5.

000

5.00

0 5.

000

5.00

0

one

clas

s 0.

653

0.68

5 0.

676

0.57

9 0.

289

0.27

3 0.

271

0.26

3

one

clas

s 0.OOO

0.03

5 0.

012

0.14

3 0.

276

0.27

4 0.

275

0.27

1

egal

itaria

n; f

ree

egal

itaria

n. S

ee T

able

s v

and

VI to

exp

lain

this

ord

erin

g.

c.

W

W

and satisfy such smaller-scale ambitions as moving up a class or avoiding a move down. (The probability of maintaining one’s sub-elite class of origin does not appear to be systematically related to extent of dynasty formation.)

Since our societies are no more than the interaction of structures and processes built into them, the positive relationship between dynasty forma- tion and small-scale opportunity for sub-elites must truly be a functional relationship. Generally, it suggests that a closed dynastic elite contributes to the maintenance of sub-elite opportunities (and vice versa). If true, this may provide a functional explanation of the survival of an aristocratic or feudal polity through many centuries.

Table VIII suggests another aspect of this relationship. Although there is little over-all probability of movement into the elite in dynastic societies, there is sufficient upward mobility from the second and third classes to provide the elite with new talent. Indeed sons risen from Class 2 and 3 origins predominate in the elites of dynastic societies. Although dynastic societies draw their talent from fewer classes than non-dynastic or egalitarian so- cieties and draw a smaller proportion of the elite from non-elite sons, yet membership in the elite does “circulate,” perhaps because of ability or the lack of it, and Porter’s principle of efficiency may indeed be. satisfied. This adaptive mechanism, circulation, too may explain the survival of aristocratic, or more generally, dynastic societies of the past.

However, even if the incidence of upward mobility were high, as in our dynastic societies relative to nondynastic societies, one might st i l l find class consciousness and “much disturbance among the workers,” and the dynastic society’s survival endangered.

If the rising Working-class sons were supplanted by the downwardly moving sons of middleclass parents, the working class would be composed in equal part of working-class sons who might be particularly disgruntled because they have not been able to rise, despite the success of others, and middle-class sons dismayed at

~ ~ ~~ ~~~~ ~

TABLE Wr THE CLASS ORIGINS OF THE CURRENT ELITE, BY EXTENT OF DYNASTY FORMATION (PER CENT)

Extent of dynasty formation High 4 + Low

Model*

Class origins I 2 3 4 5 6 7 8

1 40.8 2 59.2 3 0.0 4 0 .0 5 0 .0 6 t 0 .0 7 t 0.0

Total 100.0

35.8 30.7 64.2 69.2 0.0 0 .0 0 .0 0.0 0.0 0.0 0 .0 0.0 0.0 0.0

100.0 100.0

22.5 8 . 0 41.5 15.7 36.0 20.8 0 .0 37.9 0 .0 13.4 0 .0 0.2 0.0 4 .0

100.0 100.0

3 . 9 3 . 6 10.0 9 . 3 19.9 18.9 30.4 32.2 30.6 30.7 0 . 6 0 .6 4 .6 4 .7

100.0 100.0

3.0 7 . 6

15.4 37.5 30.4 0 .7 5 . 4

100.0

* The models are arrayed as in Table VII. t Non-existent after 1920; see text above. 134

their fall when manual sons are rising. Both parts of the working class as then constituted, the stable and the downwardly mobile, might well be especially aware of class and class interests. (Miller, 1960:3).

If Miller’s hypothesis is modified to include declasse’ sons of Classes 1, 2 and 3, taking Classes 4 and 5 to represent the working classes, we find that in dynastic societies rising working-class sons are least supplanted by down- wardly moving middle and upper class sons. This is, of course, related to the observation discussed in reference to Table VII, that downward mobility is least frequent in dynastic societies. Thus, following Miller, one can expect the greatest awareness of class and class interests in nondynastic (or egali- tarian) societies, and, obversely, the greatest indserence to class and class interests in dynastic societies. This lack of class awareness, Marx would continue, predicts the least

likelihood of class revolution in dynastic societies, and along with the common, small-scale opportunities for sub-elites and the mild circulation of elite membership, suggests a multi-faceted explanation for the survival and resilience of dynastic or, more generally, elitist and aristocratic types of society.

~~

TABLE IX THE PERCENTAGE OF “CURRENT” WORKING-CLASS PERSONS ORIGINATING IN THE MIDDLE AND UPPER CLASSES*, BY EXTENT OF DYNASTY FORMATION

Extent of dynasty formation High 4 Low

Model? Working class

persons 1 2 3 4 5 6 7 8

Class 4 0 0 .0 0 .0 18.8 25.6 36.9 37.2 40.6 Class 5 0 8 . 4 4 . 4 0.0 18.5 36.9 37.2 41.2

* Classes 1, 2, and 3. t The models are arrayed as in Table vn.

S O M E CONFIRMATORY EVIDENCE

How much credibility should be attached to generalizations from a survey of eight imaginary societies? Can these imaginary findings be verified by data obtained in real societies?

The most comprehensive study of intergenerational mobility rates is by Miller (1960), and it is from this study that the present confirmatory data are derived.16 MiUer has collected and summarized the findings of 22 mobility studies in 18 nations. Of these, data are presented on movement into elite positions from various strata in 13 different countries (1960:43) ; two such studies are presented for France. Further, of these 14 studies, all but one,

15 Subsequent presentations of these data, as in Fox and Miller (1966), and Blau and Duncan (1967:ch. 12) do not noticeably increase their utility for our purpose. 135

from Italy, consider movement into both Elites I and 11. Since we have taken pains to not limit the meaning of elite - as, for example, to a status attached to a specilk occupation - we shall include only studies in the following analysis which, with the exception of Italy, include both occupational Elites I and 11, where available. That is, under the circumstances the broadest possible dehition of elite will be used.

We noted above that dynastic societies are characterized by ( 1 ) a high probability that elite sons will inherit an elite position (i.e. there is little movement out of elite positions), and (2) a low median probability that sub-elite sons will enter the elite (i.e. there is little movement into elite posi- tions from below). The data from Miller’s 14 societies can be classified by these criteria, using a median split to discriminate between relatively strong and relatively weak probabilities of elite inheritance ( 1960:49) and rela- tively strong and relatively weak median probabilities of entering the elite from below (1960: 43).

By these criteria, the United States, Puerto Rico, Japan and Great Britain are relatively nondynastic; India, France (11), Sweden and West Ger- many are relatively dynastic; and the six remaining countries, Brazil, Netherlands, Italy, France ( I ) , Denmark, and USSR (emigrb) are ambiguous (i.e. not scoring consistently high or low on both measures of movement in and out of the elite).

One finding of the computer simulation is that highly dynastic societies promote higher probabilities of moderate (short-distance) mobility than less dynastic societies. If Miller’s “manual into non-manual” mobility is used to measure moderate mobility, we can examine the eight selected real societies discussed by Miller (1960: 40) to validate the hypothetical relationship between dynasty formation and the opportunity for moderate mobility.

The eight unambiguous real societies were cross-tabulated by amount of moderate mobility, using a median split to separate the societies of relatively high and relatively low probabilities of “manual into non-manual” mobility. We find that of the four relatively dynastic societies, three - Sweden, India and France (n) - have relatively high rates of “manual into non-manual” mobility, and of the four relatively non-dynastic societies, three - Puerto Rico, Japan, and Great Britain - have relatively low rates of “manual into non-manual” mobility. Only West Germany and the United States fail to follow this pattern. This hypothesis is thus roughly confirmed.

h o t h e r generalization from the examination of eight imaginary societies is that highly dynastic societies diminish the likelihood of downward mobility. If we use as a measure of downward mobility the probability of movement from the “middle-class” into the “manual classes,” we can examine data provided by Miller (1960: 44) from the same eight real societies to address this hypothesis. As before, the eight societies are separated into those with relatively high rates of downward mobility and those with relatively low rates of downward mobility. Of the four relatively dynastic societies, West Ger- many, France (11) and India are characterized by relatively low rates of 136

downward mobility; of the relatively non-dynastic societies, Puerto Rico, Japan, and Great Britain are characterized by relatively high rates of down- ward mobility. The United States is, as before, an exception to the rule; the other exception in this instance is Sweden. This hypothesis is also roughly c o n w e d .

Finally, the simulation indicated that in dynastic societies, mobility is con- fined to no more than 2 or 3 classes of destination from each class of origin, while in non-dynastic societies, there are opportunities for mobility from any given class of origin to any other class. Stated otherwise, in a dynastic society, the greater the distance (in intervening classes) between a specified point of origin and point of destination, the smaller is the probability that someone will reach that detsination from the given origin. This is held to be less true of non-dynastic societies. So, for example, in dynastic societies, sons ori- ginating in Class 2 have a much better chance of entering the elite than Class 5 sons, while in non-dynastic societies, the opportunity differential is less marked.

Again referring to Miller’s data (1960:43), if the “middle class” point of origin is taken as closest to the elite, and the “manual class” as the most distant origin, then the ratio of the probabilities of men originating in the “middle class” to the probabilities of men originating in the “manual class” of reaching the elite should be greater in dynastic than in non-dynastic societies.

In the same eight real societies, the “opportunity ratio” in dynastic societies is found to favour middle class over manual sons by between 4.5 and 6.5 to one; while in the non-dynastic societies the opportunity ratio favours middle class over manual men by between only 2.2 and 3.9 to one. Thus a hypothesis derived from the simulation is once again confirmed.

The confirmation of three hypotheses derived from the computer simula- tion suggests that, at the least, additional systematic comparisons of real dynastic and non-dynastic societies should be undertaken; and at the most, conclusions should be drawn from the findings obtained in eight imaginary societies.

W H Y M O B I L I T Y ?

An analysis of variance showed that the direct effects of control over recruit- ment and aspirational norms on the probability of dynasty formation are not significantly in excess of the inter-action effects of these variables. Thus the formation of dynasties is the result of an interaction between elitism and strong popular aspirations to attain (or retain) elite status. But what are the mechanics of this interactive process?

In dynastic societies elite sons, like other sons, want to retain their status or improve it, but unlike other sons, are given preferential treatment in this endeavour. (If the elite sons wished to move down, they would likewise have preferential treatment in that endeavour, but they rarely desire to move 137

down.) As a result of the differential birth rate and structural expansion too few elite sons are born to fill all the available elite positions, so some sons of Class 2 and 3 origins can move into the elite, and this in turn vacates positions in Classes 2 and 3 for lower-born sons aspiring upward. Thus the combina- tion of a general aspiration upward, preferential treatment for higher-born sons, and an inverse relation between class position and the reproduction rate sets off a “chain reaction” in the class structure that enables a majority of sons to move up one or two class levels each generation.la

If there were more elite sons than could be accommodated in the elite, under elitist norms a chain reaction would commence to drive down the chances of upward mobility for sons of all sub-elite classes. However, under the sway of egalitarian norms, or diminished control over recruitment, such an excess of elite sons would not depress the opportunities for upward mobility among sub-elite sons. Elite sons might choose or be forced, under these conditions (as during some epochs of European history) to enter or create innovative or marginal social positions (such as entrepreneur or speculator, cf. Stone, 1967: 160-182).

What then are the relative advantages and disadvantages of dynastic and non-dynastic systems? Dynastic systems offer little prospect of extreme up- ward mobility, but a strong likelihood of moderate upward mobility and the security of little downward mobility, as long as the elite birth-rate remains relatively low. In the most extremely dynastic system, the elitist aspiration- flow society, where the median probability of moving up one class is 0.65 and the median probability of moving down one class is zero, the great- grandson of a Class 5 man has a 0.34 probability of being in Class 2 with a good (p = 0.56) chance of seeing his son enter the elite. By contrast, in the least dynastic system, the egalitarian free-flow society, where the median probabilities of upward and downward mobility are roughly equal, the pro- bability of consistently upward intergenerational mobility is slight, but there is always a slight (median) chance, p = 0.13, of moving directly into the elite. Indeed, the probability of moving directly from Class 5 into the elite in this society is precisely 0.12, or a chance of one in eight. Thus it takes three or four generations as long in a dynastic system to get within striking distance of an elite position as in the non-dynastic system. The egalitarian system, then, offers the possibility of more immediate access to the elite than does the dynastic system. But this accessibility does not increase with time; that is, a family is unlikely to get into better striking distance of the elite with each passing generation. A family’s chances of entering the elite under egalitarian conditions of recruitment are constant and slight.

Under these conditions, moreover, the competition for desirable positions is increased by virtue of the fact that no sons are protected by preference, and all sons are liable to failure as well as success. As Table VII above indicated, the probability of moving upward or downward one or more

16 Cf. White (1969) for a similar though more sophisticated formulation of mobility 138 through “vacancy chains.”

classes is roughly equal. The social “cost” of this type of system is frustration and a sense of relative deprivation for those who are immobile or down- wardly mobile, insecurity for those who are entering the competition, and a loss of self-respect among the “inferior” (Young, 1967: 108).

The “payoff’ of this increased risk is that elite positions are, like all others, unprotected by preference; however in view of the total number of com- petitors for elite positions, the probability of entering the elite from a sub- elite origin is very slight. This probability is roughly equal to the proportion of elite positions in the class structure. Thus, an egalitarian system of any kind is much like a free flow situation in outcome; in either case, the prob- ability of entering the elite is chiefly proportional to the number of elite positions (or size of the elite). But if the small probability of entering the elite in a free-flow situation is acceptable, since inspired by indderence, in an egalitarian system this probability is slight in spite of one’s aspirations. There is some variation in mobility due to the aspirations that are held, but the median probability of entering the elite under egalitarian conditions of any kind is never less than about 0.12 and never more than about 0.16 (see Table vn) .

The unprotectedness of elite positions under egalitarian conditions, then, cannot be valued for the increase in opportunity it offers average men to enter the elite. Is it to be valued for its greater utilization of talent, its adherence to the “principle of efficiency”? We have seen that although a dynastic system assures elite sons of an elite position, it also recruits new talent from below in every generation. The elite of a dynastic society, as in a non-dynastic one, is fluid in its membership. It can conceivably provide as talented and efficient leadership as the egalitarian society, without the risk and insecurity that may be associated with egalitarian systems.

If the dynastic systems provide opportunity for much moderate moblity, the opportunity of eventual access to the elite for many families, and relief from the insecurity, frustration and sense of failure associated with universal competition, then its main disadvantage must lie in the protection it offers the incumbents of elite positions and their sons. Stated conversely, the only apparent advantage of an egalitarian system is that it deprives today’s elite men of the certainty that their sons wilI be tomorrow’s elite. In common terms, this suggests that a fully egalitarian system is favoured mainly because it makes the elite as insecure as the subelite. The egalitarian’s motives are envy and spite, rather squalid motives for favouring one social arrangement over another. If an egalitarian system offers no unequivocally better op portunities or satisfaction than a dynastic system, the choice between the two may as well be haphazard as not; but with this in mind, it is ditEcult to choose to favour egalitarianism over elitism while suspecting that no more than envy and spite may be at the root of our choice.

Many are dissatisfied with the stratification system; many wish to change it, and seek criteria by which to judge how to change it. The examination of eight imaginary societies has suggested that the choice between elitist and 139

egalitarian recruitment or dynastic and non-dynastic stratification is less of a choice than one might have imagined, and the change from one to the other less of a change than we might have wished. Perhaps the only significant choice, if available, can be between stratification and non-stratification.

REFERENCES

Baltzell, E. Digby 1964 The Protestant Establishment: Aristocracy and Caste in America.

Blau, Peter M., and Otis Dudley Duncan 1967 The American Occupational Structure. New York: John Wiley. Clark, S. D. 1962 The Developing Canadian Community. Toronto: University of Toronto

Press. Cook, Ramsay, with J. T. Saywell and J. C. Ricker 1967 Canada: A Modem Study. Toronto: Clarke, Irwin and Company. Dominion Bureau of Statistics 1969 Canada Year Book: 1969. Ottawa: Queen’s Printer. Durkheim, E d l e 1951 Suicide: A Study in Sociology. Translated by I. A. Spaulding and

Fox, Thomas, and S. M. Miller 1966 “Occupational stratification and mobility.” In Richard L. Merritt and

Stein Rokkan (ed.), Comparing Nations. New Haven: Yale University Press.

Glass, D. V., ed. 1954 Social Mobility in Britain. London: Routledge and Kegan Paul. Goldthorpe, John H. 1966 “Social stratification in industrial society.” In R. Bendix and S. M. Lipset

(eds.), Class, Status and Power, 2nd ed. New York: The Free Press. Henripin, Jacques 1968 Tendances et Facteurs de la Fkconditk au Canada. Ottawa: Dominion

Hodge, Robert W., et al. 1966 “Occupational prestige in the United States: 1925-1963.” In Bendix

Kahl, Joseph A. 1961 The American Class Structure. New York: Holt, Rinehart and Winston. Kalbach, Warren E. 1970 The Impact of Immigration on Canada’s Population. Ottawa: Dominion

Keyfitz, Nathan 1950 “The growth of Canadian population.” Population Studies 4 (June) :

Lenski, Gerhard E. 1966 Power and Privilege: A Theory of Social Stratification. New York:

Levine, Joel H. 1969 “A two-parameter model of interaction in father-son status mobility.”

Lydall, Harold 1968 The Structure of Earnings. London: Oxford University Press. Miller, S. M. 1960 “Comparative social mobility: A trend report and bibliography.”

New York: Random House, Vintage Books.

G. Simpson. Glencoe, Ill. : The Free Press.

Bureau of Statistics, Queen’s Printer.

and Lipset (eds.), Class, Status and Power.

Bureau of Statistics, Queen’s Printer.

47-63.

McGraw-Hill.

University of Michigan: unpublished manuscript.

140 Current Sociology 9: 1-89.

Mosteller, Frederick, and J. W. Tukey 1968 “Data analysis, including statistics.” In G. Lindzey and E. Aronson (eds.) ,

Handbook of Social Psychology, rev. ed. Reading, Mass.: Addison-Wesley .

Porter, John 1965 The Vertical Mosaic. Toronto: University of Toronto Press. Stone, Lawrence 1967 The Crisis of the Aristocracy: 1558-1641. London: Oxford University

Press. Svalastoga, Kaare 1965 Social Differentiation. New York: David McKay Company. Warner, W. Lloyd, et al. 1960 Social Class in America. New York: Harper and Row, Harper Torchbooks. White, Harrison C. 1969 “Control and evolution of aggregate personnel: flows of men and jobs.”

Wrigley, E. A. 1969 Population and History. New York: McGraw-Hill, World University

Young, Michael 1967 The Rise of the Meritocracy, 1870-2033. Baltimore: Penguin Books.

Administrative Science Quarterly 14 (March) : 4-1 1.

Library.

141