Problems of Definition

One of the basic aspects of the structure of a country is stratification, meaning the grouping of its inhabitants into social classes. Perhaps because of this strategic position itself, not only from a theoretical perspective but also considering its practical and ideological implications, the topic of social stratification has been the object of endless discussion within the human sciences2. We will not be referring to them here: we will limit ourselves to fixing the criteria that have been employed in this work for the determination of the concept of class, its number and characteristics in relation with the available empirical material for the statistical analysis of social stratification in our country.

Indeed, the problem that faces the researcher is not only the formulation of definitions and theoretical frameworks that are as accurate and refined as possible in terms of their logical coherence, empirical scope and harmonious linkage with the body of sociological theories, but also to formulate them in a way that allows for the optimal use of the available data. This last requirement always sets rigid limits to the labour of research, especially when it takes as its object a national society and does not make use of special and intensive studies carried out in restricted areas or sectors. Most recent research on social classes has been of this type: their empirical material, even if always subject to the limitations inherent in the social object, is gathered considering the theoretical objectives of the research, thereby reducing the divorce between objectives and data.

The present case provides census data and other information whose compilation obviously did not aim to answer specific theoretical requirements. It mostly amounts to the distribution of active population in occupational categories and other data that can be used as complementary. In addition, only the last national census and the one of the City of Buenos Aires (1936) consider the minimum indispensable details to formulate estimations relatively well-grounded, whereas previous ones lack important attributes that make their usage even more uncertain. As it is patent, it is always about estimations, of varying validity, reached via data from the most variegated sources. The preceding warnings allow us to circumscribe now the aforementioned theoretical problem in its actual scope. The item on which inference will be based refers to the “occupation” and consists more precisely in estimations about the occupational structure of the population. Therefore, the following question is formulated: What is the relationship between such occupational structure and social classes?

There is now what can be considered unanimous agreement on the central role played by occupation in the determination of class3. Even when no one identifies these two concepts, it is commonly accepted that social classes are made up of certain occupations or groups of occupations. But this in turn raises a new question: what principles govern the classification of occupations into classes? On this point, agreement is far from complete: numerous different definitions and criteria have been proposed, sometimes contradicting one another.

In this controversy we must bear in mind that class is an object with a real sociological existence; that is, it is not a mere classificatory name but rather refers to a set of individuals having certain common elements that are manifested in their forms of thinking and acting.

On the other hand, this community of psychological and behavioural attitudes — these mental facts — are not dis-embedded but are necessarily rooted in facts of an extra-psychological order. We do not raise here the problem of causal relations between the psychological order and the objective (structural) order, but we do insist that both are necessary components of any social phenomenon, and specifically of class. The “determinants” of class must therefore be sought in these two orders of phenomena: namely, in structural criteria and in psychosocial criteria:

1. Among the former we find first and foremost value judgment, according to which occupations are ordered in a series of layers that are superimposed and fix solutions of continuity that denote class boundaries. These “points of rupture” in the hierarchically ordered series of occupations manifest themselves as greater “social distance” between the members of the different occupational groups (difficulties in “social” frequentation, marriage, etc.). We assign a structural or objective character to this value judgment insofar as it manifests itself as a socially established — though not codified — norm whose existence is acknowledged by the members of society even if only to oppose it. This pattern of social inequality — which contrasts with the egalitarian ideologies that underlie our modern societies — is naturally in turn a social-historical product whose origins it is not appropriate to discuss here. It is only necessary to remember that this value criterion is linked to the distribution of real power among the different social groups, and that such distribution — at least in our Western society — expresses in each historical phase the existing balance among these groups.

The type of existence, which characterizes the different social classes, also bears an objective character; in fact, the groups of occupations that make up each social class present certain common ways of living — clothing, dwelling and many other elements of “material culture” — that are the result not only of their similar position within the social structure, but also of the traditions that the community of position itself forms through a typical process of “institutionalization” with the passing of time. The type of existence is also linked to other objective criteria: to the economic level, which refers to the minimum and maximum limits between which the incomes of the different occupations that make up the classes should oscillate, and to the personal characteristic — first and foremost the type and degree of education and personal culture — which are considered peculiar to each social class.

2. Among the psychosocial criteria we find mainly two: the self-identification of the members of each occupation with a given class, and the system of attitudes, norms and values that characterize the individuals of each class and distinguish them from the others. These psychosocial criteria are currently synthesized in the concept of status social personality, an expression that denotes the typical mental configuration that, as a result of the community of life and similarity of position and outlook within society, most individuals of a class are assumed to possess.

All these criteria — structural and psycho-social — do not mechanically add up to the classification or ordering of the different occupational groups: on the contrary, their reciprocal linkage results from the concrete historical and social processes that have shaped each national or regional society. At any given moment the class structure of a country bears the imprint of its history;4 sometimes a history that is already remote, and always that of the economic and social development of two or three generations. In the highly dynamic societies of modernity, the rapid changes that occur in the structure of occupations, in the value judgments that rank them, in the type of existence, in the system of attitudes, etc. are far from being synchronized: because of the universal phenomenon of cultural “lag”, the different criteria no longer correspond with each other in the perfect way that a — theoretically — static society should have. Occupational groups may see their real position of power within society modified while all or part of the two elements — value judgment, type of existence, etc. — remain temporarily lagged behind in the positions which correspond to the previous class structure. The system of attitudes and expectations of the members of an occupational group may prove impossible to satisfy in a new situation which has displaced the traditional environment too quickly. Sometimes the “lag” is also due to the coexistence of anachronistic elements within the social structure. For example, feudal remains can explain to some extent the survival of the hereditary principle in the determination of classes. This occurs especially in certain countries of the old world, but there are plenty of examples in new nations where the existence of an economic-social organization closer to the feudal than to the capitalist type gives rise to such survival. However, even where this does not occur, the hereditary principle may retain a certain limited validity for the upper class, in contrast to the principle of the dominant occupation. This phenomenon of “cultural lag” is mainly responsible for the difficulties surrounding the concept of class: precisely for this reason only a dynamic perspective, one that perceives the class structure (at least in our societies) as in perpetual motion, can provide accurate theoretical frameworks for the social reality of the present.

For these reasons it would be futile to seek a clear-cut distinction of social classes: the reality we face presents us with a variety of groups characterized by different combinations of the structural and psycho-social criteria indicated above. Classes represent only, so to speak, areas of the social structure in which a certain combination of criteria occurs with greater statistical frequency. According to the brief analysis carried out, the classification of a country’s inhabitants into social classes would require data on the following topics:

a. Occupational structure of the population: inhabitants classified by occupational category;
b. Hierarchy assigned to the different occupations according to the dominant socio-cultural patterns, and ways in which occupations are grouped into classes according to these patterns.;
c. Type of existence, economic status and personal characteristics — particularly education — that characterize in average the different occupations or groups of occupations.;
d. “Self-identification” of the members of the different occupations with one or another social class.;
e. Characteristics of different systems of attitudes, norms, values (status, social personalities) that should characterize the different occupations or groups of occupations and distinguish them (to justify their inclusion in different classes).

With these elements it would be possible to establish the number and characteristics of social classes in our country, as real sociological entities, and not as mere classificatory names [categories]. This analysis, to be complete, would then have to be extended to the study of the socio-economic roots of the real groups discovered, a study that would necessarily also have to trace the historical origins of present social stratification.

Unfortunately, we find ourselves very far from this ideal position, and moreover the nature of this project assigns much more modest limits to the development that can be given to the subject of social classes. We have some estimates of the occupational structure of the country, but we lack concrete data about the other elements. For these we can only rely on common-sense knowledge and some analogies with other countries.

It must be emphasized first of all that the class system in our country is close to that of industrialized Western societies: whatever its typical features, as well as regional peculiarities that can be observed in its different areas, it seems that as a first approximation we can accept the schemes whose relative validity has been proven in other countries of the same type.5 On the other hand, the very scarcity of basic data simplifies the problem, since it is only a matter of ordering the broad categories of occupations detailed in Table 1, and eventually of attempting finer distinctions — always on a conjectural basis — with the help of other data.

Since it is not possible to detail here the background and the bibliography reviewed, we will limit ourselves to some brief indications referring exclusively to empirical research.

a. The social judgment of value, according to which different occupations are ordered and grouped into classes, has been studied with different methods. In some countries —the US and the UK— empirical research has been carried out regarding the prestige of occupations in an attempt to determine “prestige scales” valid in certain communities or in the country as a whole.6;
b. Other studies have employed sometimes very elaborate techniques to determine through observation of behaviour the number and content (in terms of occupational groups, etc.) of social classes.7;
c. The problem of “self-identification” and, in general, the study of the psychology of social classes are the subject of many other works.8 Despite the remarkable discrepancies in the results, some general conclusions can be drawn:.

1. There is a clear notion of the social hierarchy of occupations in the collective consciousness, i.e., a scale of prestige that orders them in a certain sequence. There is considerable agreement on this, and in particular it can be said that Western industrialized societies show marked similarity in this respect. Recall, for example, the similarity of objective and subjective (self-identification) stratification recorded in France and the United States; not only is there in both countries a very close relationship between occupation and class self-identification, but also the structure assumed by the hierarchization of occupations is very similar. The most marked divergences are observed in those sectors most strongly influenced by historical factors, especially in the countryside.9 Some uncertainty is apparent with respect to intermediate categories (for instance, clerks, specialized workers, etc.), both from the point of view of their assignment to different classes and their sense of belonging (or self-identification) research results are ambiguous. This ambiguity reflects a real situation: these are groups which, although numerically important, occupy “marginal” positions and possess contradictory attributes (for example, being salaried and having an occupation and type of existence oriented to the model of the upper classes).

2. There is no agreement as to the number of social classes, which usually ranges from two to six or seven. These discrepancies, which also occur with respect to the same community, seem to detract from the consistency of class as a real sociological group, transforming the problem of its number into a matter of mere convention. Some authors base their denial of the reality of classes on this disagreement: for them only a continuous series of social positions exists, without fractures.10 However, that classes are not mere “statistical sets” is demonstrated, along with common consciousness, by an impressive body of research attesting to their differential behaviour. The existence of a big number of ambiguous situations and marginal groups tends of course to conceal the “points of fracture” between the different strata of the population; these, however, reveal themselves quite dramatically in the significant spheres of collective behaviour. What happens is that often groups of occupations assigned to the same class reveal different attitudes towards a certain issue. This fact shows that, from the point of view of collective consciousness and concrete behaviour, occupational groups have perhaps a greater reality than “class”.

Such a phenomenon does not, however, contradict its real character; it only means that in any analysis the character of relative heterogeneity that it possesses should not be forgotten. The occupational group represents the link between the class structure and the economic-technical structure and, in short, its greater consistency and internal coherence is a confirmation of the conditioning and determining character that such a structure has on the classes. For this reason, the appropriate method for overcoming the disadvantages of mere conventionalism in the definition of social classes and their number is to maintain a close connection with the categories deriving from the technical-economic structure of society, i.e., with the main groups of occupation, differentiated above all on the basis of their position within the economic organization (legal status: owners, wage-earners, independent workers, etc., and type of activity: branch of activity, branch of industry, trading, services, etc.) and the meaning that this position bears with respect to the functioning of the economic system itself.

d. Occupations and classes in Argentina: For this paper we will use the conventional tripartite classification into “upper”, “middle” and “popular” classes, which also agrees with the one adopted in the very few studies on the subject that exist in our country.11 We will speak, however, of “middle classes” and “popular classes” in the plural, to emphasize their “composite” character, insofar as they are the result of the conjunction of occupational groups endowed with certain dynamics of their own, despite the common characteristics that justify their inclusion in a same class category. A more detailed analysis of the economic structure of each of the major occupational groups will later allow us to distinguish not only the “upper” class from the “middle” classes, which in a first classification can remain indiscriminate due to the low numerical importance of the former, but also to subdivide them into two groups (“upper” and “lower”).;

Based on all these considerations and on the Argentinian and foreign antecedents cited, we adopted the following scale in order to classify the different occupational groups into social classes:

a. Popular classes. Urban sector: “workers” and “apprentices” in secondary activities, trading and services. In this denomination have been summarized all the occupations that by their functions, form of remuneration, prevalent instruction and type of existence, correspond to the popular classes — distinction based on notions of common sense. “Self-employed”: all branches of industry, commerce and services have been included, with the exceptions indicated below. The workers included are largely home-servicing workers or people who, although they exercise their occupation without being legally in a position of dependence (self-employed), correspond by their characteristics to the popular classes.
b. Popular classes. Rural sector: “workers” and “apprentices” in the primary sector, almost entirely agricultural workers. “Self-employed” of the same sector.
c. Middle classes (including upper middle classes). Urban sector: “employers, businessmen” in industry, trading, services. Family “helpers” working at the company for the head of the family. “Self-employed” corresponding to the branch of “graphic-press and paper”. “Wholesale trade”. “Trade”. “At-desk”. “Public shows”. “Hotels”. “Healthcare services”. “Land transportation”. “Liberal professionals” — it is assumed that in this branch the majority or in some cases the totality of the people correspond to the social type of “middle class”). “Employees” and “Cadets” of the secondary and tertiary sectors. We also added “Rentiers” and “Retired pensioners” not included in the active population, whose characteristics correspond to the middle class.
d. Middle classes (and upper middle classes). Rural sector: “bosses, businessmen, employers” in the primary sector. “Helpers” and “employees” of the same sector (including owners and tenants).
e. Autonomous middle classes and dependent middle classes: within the middle classes, these two groups have been distinguished, differing in terms of their legal-economic position, with important consequences in terms of the remaining characteristics.

Table 1: Economically active population by categories of occupation and major branches of activity. 1947.*
(thousands of persons)
* A.E.R.A. (1948: volume 10) and details from I.C.N..
Table 1: Economically active population by categories of occupation and major branches of activity. 1947.* (thousands of persons) * A.E.R.A. (1948: volume 10) and details from I.C.N..

Each of these occupational categories brings together a heterogeneous set of people. The distinction between “workers” and “employees” is provided by the same censuses we used, and they were made on the basis of a social class criterion. It is not possible to arrive with these data only at finer discriminations; for example, between skilled and unskilled workers; between subordinate and managerial employees. The same is true of “bosses”,12 which evidently include owners of small artisanal businesses and owners of large factories, etc. For these further distinctions, one should seek the help of special censuses which, although not directly comparable with the results of the population census, can shed some light on the composition of these groups.

Before continuing with this analysis, however, it is necessary to ask to what extent it is possible to assign sociological reality in our country — in the sense indicated above — to “classes” thus defined and to the subgroups of which they consist. As already noted, many scholars have pointed out the dangers of taking as groups that are mere aggregates obtained by a manipulation of the statistics.13 Therefore, even if it is a classification formulated on a conjectural basis and only as a hypothesis, it is indispensable to count on some objective proof or indication — beyond the common sense that has been used in the classification and the analogies and aprioristic reasoning that the criteria have provided us with — that these figures somehow reflect real existing differences in the population. However, as a preliminary verification, we can cite some facts; first, it has been shown that the demographic characteristics of the two large groups differ substantially — vital rates, family composition and size, and the recent evolution of the birth rate are different in the popular classes14 and in the middle classes, and also in their urban and rural sectors. In addition, Bunge was able to observe that there are probably other differential characteristics in demographic terms for a certain part of the upper class; in numerous families of the upper class (the “aristocracy”) the pattern of reduction in family size typical of the middle class is absent (Bunge 1917: 342). (See Table 2 on the following page).

Moreover, it is possible to show that these occupational groups and classes differ in their political orientation as revealed at election time. The correlations which have been computed in Chapter 16 and in which the proportion of the different groups in question has been taken as an independent variable, is quite conclusive in this respect. This fact is of some importance since it indirectly demonstrates that the “self-identification” of the occupational groups with the different classes approximates, on average, the resulting one from the classification adopted. As Centers has shown, in fact, political orientation is closely correlated not only with the category of occupation — e.g., the vote of workers differs substantially on average from that of employers — but also with self-identification — e.g., those workers who identify themselves as members of the middle class tend to vote similarly to middle class, etc.).15

Table 2: Table 2. Labour force, distinguishing classes and branches of occupation. 1947
Estimates according to text. These figures differ considerably from another estimate made on the basis of
data from the III School Census (1943); Cf. Germani (1950). It should be noted that this Census, due to its
nature, was not suitable for such estimates.
** Includes “helpers”.
**** Includes “helpers” and some “own account” groups (see text).
Table 2: Table 2. Labour force, distinguishing classes and branches of occupation. 1947 Estimates according to text. These figures differ considerably from another estimate made on the basis of data from the III School Census (1943); Cf. Germani (1950). It should be noted that this Census, due to its nature, was not suitable for such estimates. ** Includes “helpers”. **** Includes “helpers” and some “own account” groups (see text).

Data are available on other “cultural” differences —in the anthropological sense— between classes and between different occupational groups. For example, the distribution of expenditures in a family budget is not only a function of the size of the family and its purchasing power, but also reveals marked differences in terms of its needs: two families that are comparable in terms of their composition, and with an almost equal or very similar level of income, will distribute their expenditures differently if they belong to the middle class (employees) or to the popular class (workers).16

And finally, we will cite another example: in our country, too, there are differences in the average intellectual capacity (measured by tests) of the different social classes. As is made clear in the chapter dedicated to this subject, such differences are of sociocultural value and, rather than reflecting supposed biological conditions, they are the expression of different environments; that is, the different types of “subcultures” correspond to the different social classes.17

It can thus be stated with some foundation that the two major groups in our classification, “middle” and “upper” classes on the one hand, and “popular classes” on the other, are not mere statistical sets but represent relatively differentiated groups, with respect to several characteristics, two of which — at least — are verified by the correlations just indicated.

It is important to emphasize the possibility of establishing uniformities of this type, since precisely in this lies the usefulness that can be derived from a classification of the population into different socio-economic levels. To prove that 40% of the economically active inhabitants can be assigned to the “middle classes” and 60% to the “popular classes”, and that these groups are in turn made up of certain proportions of employees, employers, professionals, etc., will only make sense to the extent that it allows us to formulate valid hypotheses about other static and dynamic aspects of social reality. Class constitutes a fundamental category within modern society and can therefore be assumed as an independent variable in the study of several social facts; its quantitative determination thus represents the only means of grounding such studies on more concrete and precise foundations.

The reading of Table 2, rather than giving us an answer, suggests questions: In what way do the different groups distribute the national profit? What are the social origins of the members of each of them? To what extent do heritage or personal characteristics determine the assignment to each class? How do they differ in their standard of living, type of existence, psychological traits? What is the effective relation of power in which they find themselves at this moment? In what way do they contribute to the formation of intellectual and political elites? It is beyond the nature and possibilities of this research to answer these and other questions; we merely formulate them, as they represent a vast research plan offered to scholars of the country’s social reality.

The fact that the estimated figures refer to the working population rather than to the totality of the inhabitants undoubtedly places a limitation on their use. It would be of greater utility to provide a classification of the families in the different socio-economic levels, but the data necessary to make such an estimate is lacking. Even within these limits, however, the estimated figures offer a good expression of the social structure of the country as it is at present, and their meaning will become clearer when the characteristics of the economic structure in which the different groups are rooted are examined.


I would like to express my gratitude to the Gino Germani Research Institute of the University of Buenos Aires, Argentina, for having shortened the distance between the scientific work of Gino Germani and the students of the Licentiate Degree in Sociology of the same university, which I was part of. The legacy of Gino Germani has marked the history of sociology in Latin America, and this has been transmitted by the professorship’s community to the students across generations till today.

Especially, I would like to thank Ignacio Mancini, from the Documentation Centre of the Gino Germani Research Institute, who kindly offered me the connection with the son of Gino Germani. Finally, my special thanks for Sergio Germani, who allowed me to translate this piece of research and made the transnational knowledge of this article possible.


2 Cf. Pfautz (1953). This author gives a list of 333 papers in 8 years, only in the United States.

3 Quantitative research also confirms this hypothesis. See Cattell (1942: 293-308); occupation has the highest correlation with social class estimated based on other criteria. Warner et al. (1949: 168 and 177) have found correlations above 0.90.

4 On this often-forgotten point, see the interesting article by Cole (1950: 275- 290).

5 Among the Latin American countries, Argentina and Chile are the closest to the type of class structure of Western Europe and the United States. Cf. Beals (1953: 338).

6 See especially Davies (1952: 134-147), which contains a general review of the subject, and Smith (1943), one of the most complete investigations. Especially from the point of view of the use of census data, the study by Hall and Caradog Jones (1950: 31-55) is of importance.

7 Perhaps the most important research of this type is that carried out by Warner et al. (1949); this work examines different objective and subjective criteria for the determination of social class; among the former are occupation, amount and source of income, education, type of housing and area of residence.

8 We particularly recall the book by R. Centers, The Psychology of Social Classes (1949): a quantitative study on a national scale. This author was able to prove that three quarters of industrial, commercial and agricultural owners, professionals and employees identified themselves with the “middle” or “upper” class, while 80% of workers identified themselves with the “working” and “lower” class.

9 Cf. Rogoff (1953: 327- 339).

10 Typical is the recent study by Lenski (1952: 139-144).

11 See Bagú (1950), Poviña (1950), Barcos (1936: 243- 319) and Bunge (1917). Bunge classifies the economically active population into three groups: workers, non-workers and the wealthy, the first corresponding to the “popular classes” and the third to the “middle classes” of our classification.

12 [Translator’s note]: boss is used here as the translation of the word patrón.

13 Also Carr Saunders and Caradog Jones (1927: 67), after having analyzed the census statistics, prove on the basis of these that “there is no measurable characteristic at the disposal of the statistician which would allow the classification of the population into social classes”. And finally, although he admits that these methodological difficulties do not in themselves imply the non-existence of social classes, he asks: “Are we not wrongly interpreting the social structure of this country by using the concept of social class, when with respect to clothing, language, recreation, all the members of the community look alike?”

14 [Translator’s note]: in this translation the term “popular classes” has been used to refer to what Germani called clases populares, which is usually known as “working classes” in the Anglo-saxon sociological literature.

15 See Centers (1949: 119, 126 and 127).

16 National Labour Department: Cost of Living (1935: 20-21). Here are two groups of families of workers and employees with a very similar level of income (230 and 250 respectively) and identical consumption capacity.

17 See Chapter XV.


  • Bagú, S. (1950). “La clase media en la Argentina”, in Crevenna, Th. (ed.) (1950). Materiales para el estudio de la clase media en la América Latina. Unión Panamericana.
  • Barcos, J. R. (1936). "El trágico destino de la clase media". Hechos e Ideas.
  • Beals, R. L. (1953). "Social Stratification in Latin-America". American Journal of Sociology.
  • Bunge, A. E. (1917). Riqueza y Renta de la Argentina. Ag. General de Librería y Publicaciones.
  • Carr Saunders, T. M. y Caradog Jones, D. (1927). A Survey of the Social Structure of England & Wales as illustrated by Statistics. Oxford University Press.
  • Cattell, R. B. (1942). "The concept of social status". The Journal of Social Psychology, 15(2), 293-308.
  • Centers, R. (1949). The Psychology of Social Classes. Princeton University Press.
  • Cole, S. D. H. (1950). "The Conception of the Middle Classes". The British Journal of Sociology, I.
  • Davies, A. F. (1952). "Prestige of Occupations". British Journal of Sociology, 3, 134-147.
  • Departamento Nacional del Trabajo. (1935). Costo de la vida. Series C, 1.
  • Germani, G. (1955). “Clases sociales. Introducción”, in Germani, G. Estructura social de la Argentina. Análisis estadístico. Raigal.
  • Germani, G. (1950). “La clase media en la Argentina”, in Crevenna, Th. (ed.) (1950). Materiales para el estudio de la clase media en la América Latina. Unión Panamericana.
  • Hall, J. y Caradog Jones, D. (1950). "Social Grading of Occupations". British Journal of Sociology, 1(1), 31–55.
  • Lenski, G. E. (1952). "American Social Classes: Statistical Strata of Social Groups". American Journal of Sociology, 58(2), 139-144.
  • Pfautz, H. P. (1953). "The current literature on social stratification: critique and bibliography". American Journal of Sociology, 58(4), 391-418.
  • Poviña, A. (1950). “Concepto de la clase media y su proyección argentina”, in Crevenna, Th. (ed.) (1950). Materiales para el estudio de la clase media en la América Latina. Unión Panamericana.
  • Rogoff, N. (1953). "Social Stratification in France and in the United States". American Journal of Sociology, 58(4), 347-357.
  • Smith, M. (1943). "An Empirical Scale of Prestige Status of Occupations". American Sociological Review, 8(2), 185–192.
  • Warner, W. Ll., Meeker, M., & Eells, K. (1949). Social Class in America; a manual of procedure for the measurement of social status.