Macrosociology and Microsociology Research Paper

Sociologists thus identify two scales of organization—the microscopic and the macroscopic— which they understand to correspond to two types of phenomena and explanation. Recognizing the existence of levels of reality, however, should not lead us to think that there are ontologically distinct entities that correspond to different types of objects. For the time being, no specific dominant point of view has managed to make the basic structures of these two types entirely homogeneous or unifiable (see Wippler and Lindenberg 1987). Our purpose here is to clarify the debate over the links between macroscopic and microscopic, over whether and how it is possible to move from one to the other, and over the idea that the first is founded on the second.

1. Links Between Micro and Macro and the Problem of Explanation

Although the problems raised by the relations between macroand microsociology are highly relevant today (to judge from the number of studies done on them in recent years) (see for example Collins 1981, Alexander et al. 1987, Huber 1990), few of them are new. They first arose with the institutionalization of sociology as an autonomous scientific discipline positioned at the center of the debates opposing utilitarianism-inspired theorists like Spencer against sociologists who, in search of a foundation for the specificity of the object and the explanations of social phenomena, rejected the hypotheses of classical economy.

Far from being purely rhetorical, these problems determine researchers’ choice of instruments, units of observation, and explanatory procedures. A hypothesis tested at the level of the individual may very well prove false at the systemic level; another, verified at the global or context level may prove invalid for the individual case (see Robinson 1950, Goodman 1953, Duncan and Davis 1953, Goodman 1959). The error of inferring the social from the individual corresponds to that of inferring the individual from the social (‘contextual fallacy’): though it is sometimes a mistake to deduce (without any supplementary hypothesis) the voting behavior of individuals from information on how ecological units voted, it is just as wrongheaded to affirm, on the basis of a relation of positive dependence between an individual’s degree of education and his or her salary, that when the collective level of education rises, the average salary does too.

Let us suppose, as did the founding fathers, that the aim of social sciences consists of explaining the statics and dynamics of social systems (though this definition is neither necessary nor sufficient). Tocqueville (1856) tried to account for political change, mainly the revolutionary one; Durkheim (1893) for the increasing differentiation of functions and the consequences of such differentiation; Weber (1905) for the birth of a new type of socioeconomic organization; Sombart (1906) for the absence of socialism in the USA. Clearly, all these questions pertain to the systemic level. The size of the system may vary, but whatever its components and their number, it constitutes a unit that must not be broken down or dissolved.

We have three competing procedures for conducting an explanation of such phenomena. The first and the most current one is also the least satisfactory. It consists of analyzing the relations between a given social phenomenon and independent social factors; explaining the social by means of the social is the leitmoti of the Durkheimian sociological program. But in order for that first macrological procedure to be legitimate, one’s observations must pertain to the same level. We readily affirm that revolution is the result of improvement in economic, political, and social conditions; that an increase in the volume and density of groups or societies leads to the division of labor; that religious ethics facilitates the birth of capitalism; that a high degree of social mobility makes the development of strong left-wing parties less likely. It seems, however, that this condition of consistency between the content of a sociological proposition and the unit of observation and analysis is not often met. In most studies conducted by interview or questionnaire, the unit remains the individual, and a hiatus is created between the objective of the theory, which is to formulate macrosociological propositions, and empirical research, which remains focused on explaining individual behavior. This holds true despite the fact that Lazarsfeld and Menzel (1961) drew sociologists’ attention to the necessity of consistency. Without consistency, we are caught in an impasse: many hypotheses deduced from macrosociological theories cannot be adequately tested.

The second procedure consists of collecting observations at an infrasystemic level and developing hypotheses on the behavior of such units (individuals, groups, institutions), with the purpose of explaining systemic relations through an appropriate synthesis of these observations. In this case micrological variables positioned in an intermediary position between dependent and independent macroscopic factors are introduced. One affirms that improvement in conditions (macro) generates a level of frustration (micro) such that it increases the probability of protest at the sociopolitical level (macro); that an increase in volume with growing densification aggravates competition between system units confronting each other on the same market, and that this competition is resolved by division of labor; and that social mobility, whether real or perceived as real by individuals, reduces the search for a collective solution by making people believe in the utility of personal effort, thus making the emergence of strong socialist parties less likely.

An explanation operating at this infrasystemic level is sometimes considered more satisfying, stable, and general than one expressed entirely in macroscopic terms. This is Popper’s point of view (1934, 1963), as well as Harsany’s (1968), Simon’s (1977), and Boudon’s (1973, 1977). The argument holds that because the behavior and dynamics of a social system are in fact the result of the actions of its components (individuals, subgroups, institutions), it is these components that should be analyzed if we are to better explain and predict systemic characteristics.

To illustrate this case, let us borrow Tocqueville’s (1856) explanation of the stagnation of French agriculture at the end of the Old Regime compared to the English one as Boudon (1998) stylized it according to his cognitivist model. Agriculture stagnation is the effect of administrative centralization in France. Both phenomena are macroscopic but their relation has to be explained by bringing to the fore the reasons of classes of persons. More prestigious civil servants’ positions in France than in England cause landlords’ absenteeism that in turn provokes a low rate of innovation. By buying positions, French landlords think they can increase their power, prestige, and income. They have good reasons to leave their land, rent it, and serve their King. On the other hand, French farmers cultivate the land in a traditional fashion because they do not have the capacity to innovate. In England, the civil servants’ positions are less numerous and less rewarding while local political life offers them good social rewards. What Boudon constructs is a social mechanism to explain the relation between the two macroscopic phenomena, i.e., a set of ultimate causes that have the character of being individual decisions we perceive as understandable.

The third approach of conducting an explanation is to analyze the effects due to the nature of the positions and distributions of certain variables on the behavior of the system’s component units without formulating hypotheses about individuals (see Blau 1977, Blau and Schwartz 1984). Here we are no longer interested, for example, in the influence of religion (or any other variable) on social relations, but rather in the distribution of religious affiliation and the structural parameters of that distribution, understood to determine the dependent variables. It is important to carefully distinguish this macrostructural research, which works not with variables but positions in the social structure, from studies of the first type like Moore’s (1966), Lenski’s (1966), Skocpol’s (1979), and Turner’s (1984), in which collective properties and aggregate variables are subjected to classic treatment in terms of factors, relations, correlations, and influences.

Whereas microsociology seeks to explain social relations in terms of psychological traits, or processes such as exchange, communication, competition, and cooperation that shape relations between individuals, structural macrosociology accounts for the relations between the different components of a given society in terms of differentiation and integration. For this last approach, explaining certain aspects of relations between religious or ethnic groups—intermarriage or friendship, for example—means mechanically deducing them from structural properties of the groups— namely, their respective sizes—without having recourse to supplementary hypotheses about individuals, such as how liberal they are.

By way of illustration, let us consider the following example. Suppose we have to study the relation between religion and intermarriage between religious groups, and that we have established a correlation between the two variables, namely that the percentage of Protestants marrying non-Protestants is greater than the percentage of Catholics marrying nonCatholics. How can this be explained? We could develop hypotheses concerning the behavior of Catholics and Protestants: Protestants are less ‘practicing,’ less sexually inhibited, more tolerant than Catholics, for example. We would then look for indicators of religious practice, tolerance, and so forth, to account for the observed correlation.

Macrosociology proceeds differently. First, it makes no hypotheses about individual behavior. Rather it analyzes the distribution of individuals among the various positions and the effects of those positions on behavior. Take two groups of different sizes A and B, with A B. Individuals in group A can come into contact with individuals in group B and their members can intermarry. The first theorem deduced from these hypotheses is that the rate of intergroup association for group B must be greater than that for group A. If A and B are composed respectively of 10,000 and 100 individuals, and if there are 10 mixed marriages (10 pairs of spouses), the respective intermarriage rates are A: 10 10,000 1 1,000 and B: 10 100 1 10. If A represents the Catholic group and B the Protestant group, the preceding theorem enables us to affirm that B’s intermarriage rate is higher than A’s—the correlation has been explained in terms of group size.

But no matter how important the analysis of social systems may be, we cannot say that it is the one and only aim of social sciences. Sociological theories also work to explain individual behavior and may study voting behavior, household consumption, and more generally the attitudes and choices of individuals. With the exception of explanations proposed by psychological theories, sociologists explain these behaviors by analyzing characteristics of individuals in their relation to their environment.

It is not hard to show that such problems run through the entire history of sociology; here, however, I will be proceeding logically rather than chronologically. To this effect, I will be proposing a typology whose categories subsume the main sociological theories and bring out relations of inclusion or exclusion concerning models deduced from those theories. I will examine the objectives, problems, and the solutions of macrosociological and microsociological theories, while asking whether they are intractably antinomic or whether there are potential means of passing from one to the other.

2. Outline of a Typology of Micro and Macrotheories

The first definition of micrology and macrology, the simplest but also the least adequate, as indicated above, is etymological: the terms refer to the size of the units of observation used. Micrology is understood to be limited to small units, whereas macrology analyzes groups, collectivities, communities, organizations, institutions, even entire societies. This first, unidimensional definition, is valid for neither sociology nor economics; the observation units used with certain theories do not fall into either of the two categories. The ultimate observation unit in symbolic interactionism or ethnomethodology, for example, is not the individual but the situation in which individuals interact. Nor is size a more relevant criterion, because the unit of observation for certain macrosociological theories, for example, is a limited number of small groups.

Similar definition problems are encountered in economics. The classical distinction between microand macroeconomics also refers to several levels. While microeconomics is the study of the behavior of individual decision-makers (households, firms) or, more exactly, choice as limited by different conditions, macroeconomics seeks to analyze the economy as a whole by structuring its hypotheses around aggregate variables (e.g., national income or consumption). But what is true of sociology is also true of economics: it is not so much the unit of observation that matters as the way in which problems are simplified.

The second definition proposed by Cherkaoui (1997) is multidimensional; it has at least three dimensions: (a) the nature of the patterns to be explained; (b) the type of hypotheses—rationalist or a-rationalist—used in conducting the explanation; (c) the unit of observation and analysis used in empirical testing of propositions. These units may be individuals, situations, or structures; the point is that they must be consistent with the hypothesis demonstrated, as discussed above.

The first dimension pertains to the particular problem under study, which may involve either individual behavior regularities or social and group regularities. Voting for a particular political party, consuming a product, adopting a particular attitude, making an economic or social choice are examples of regularities in individual behavior. On the other hand, accounting for the degree of heterogeneity in a given society (the frequency of intermarriage between ethnic or religious groups, for example); for the existence and intensity of cooperation or conflict within an organization; for integration and differentiation; for macroscopic balances andmbalances, involves studying social patterns.

Let us consider the following problem taken from social stratification studies, one of the major fields of macrosociology. When we conduct empirical research on the effects of the father’s occupation or the individual’s level of education on the social status attained by that individual, we are positioning ourselves at the microsociological level (though we may not be aware of it)—despite the fact that our project is macrosociological. As long as the propositions demonstrated are micrological, we are not violating logical rules; we can legitimately make assertions of the following type: ‘When, for individuals, the level of education increases, so does occupational status.’ Such affirmations derive from empirical research conducted in accordance with known, standardized rules: construction of a random sample, use of analytical statistical methods in which the individual is the relevant unit. But in proceeding this way, we are working from several implicit hypotheses, the simplest of which affirms that individuals’ decisions are entirely independent of each other. Nothing precludes us from applying such simplifications as long as our hypotheses or empirically tested propositions do not go beyond the microsociological level. But it is clear that (a) individual choices are in fact interdependent; (b) there is not an infinite number of occupational positions; (c) the set of such positions constitutes a structure whose parameters function as constraints on individual choices. Obviously, when I decide to apply for a post in sociology in a given institution, (a) I am competing with other applicants (interdependence of individuals); (b) I know that the number of positions in my discipline for which applicants are competing is limited; (c) the job openings in sociology that applicants are competing to fill are the direct result of what other jobs are available both in other disciplines of other departments of the same institution (they also depend, though indirectly, on the number of places available in other institutions, such as all institutions of higher learning); and (d) when making its decision, the institution that has activated competition for the jobs will take into account the interest that other institutions of the same type may have in sociology.

The second dimension of the definition of microand macrosociology proposed here refers to theoretical simplifications and the objectives of such simplifications. I will summarily distinguish between two types of hypothesis about the behavior of units of observation and analysis: a theory requires either rationalist or nonrationalist hypotheses (this last type may be divided into a-rational and nonrational).The rational individual is not necessarily to be reduced to that of Homo æconomicus, whose particularity is to maximize the expected value of his utility function— monetary profit, for example. Following Boudon (1996), we can speak of three different types of rationality: expected, cognitive, and axiological.These theoretical simplifications of course only concern individual actors or units, not structures. If we preclude Hegelian-type visions, which reify universals and endow them with intentionality or rationality, it is obvious that under certain conditions, groups, organizations, and institutions may be assimilated to actors.

The third and last dimension is the level of observation and analysis deemed most relevant to a given theory. That level may be individuals, households, organizations, a context-situation, or population distributions. The individual and the county are the units most commonly used in electoral sociology. Whereas for economic and sociological research, the household is considered the most appropriate unit for studying consumption, in the sociology of organizations and microeconomic theory of the firm, the preferred basic unit is the single firm. For some correlation analysis—that between literacy and the proportion of the state budget allocated for education, for example—rates are required. In the example from social stratification research, we saw why observations made at the individual level were not sufficient to give answers to macrosociological questions; they may even be an inappropriate place to look for such answers as demonstrated by Coleman (1990). The empirical tradition founded at the beginning of the 1950s and popularized by Blau and Duncan’s American Occupational Structure (1967) and Featherman and Hauser’s research (1978), among others, does not allow for a move from the micro to the macro.

With these few taxonomic principles, we can construct the typology of theories (numbers 1–10 correspond to theory types to be identified) as in Fig. 1. This typology is neither exhaustive nor inalterable, and it can be simplified or complexified, namely by specifying hypotheses here designated ‘other than rational.’ It does, however, enable us to observe (a) the fundamental distinction between macrological and micrological theories; (b) the paradigmatic principles underlying these theories; (c) the relations of opposition, distance, or inclusion that a given theory has with the others. Finally, it identifies ‘at a glance’ the historical stages of sociological thought.

To each of the 10 ‘pure’ theory types corresponds one or more theories. Type 1 comprises all theories that share the axioms of the market model of pure and perfect competition, some behaviorist theories, and Homans’ (1961) and Blau’s (1964) theories of social exchange, among others. For Homans, macrosociological structures are no more than repeated individual behaviors: the interaction between individuals is what generates structures; structures do not determine individual behavior. Type 2 is made up of theories in which the market model is applied to particular sociological fields. Coleman (1990, p. 21 et seq.) presents a much-discussed topical example similar in some points to economic markets: the marriage market or, more exactly, the phenomenon described as ‘marriage squeeze.’ Psychoanalysis is a type 3 theory, though certain variants of it are closer to type 9. Type 4 comprises stochastic models and some theories of collective behavior: stochastic theories, like the econometrician explanation of the income distribution, do not hypothesize about individuals, while collective behavior theories make use of a-rationalist postulates about agents’ behavior, such as those implied in notions of ‘suggestion’ or ‘hypnotic effects’ as used by Gustave Le Bon (1895).

Type 5 encompasses the models of adaptive behavior particular to certain behaviorist explanations, together with oligopolistic models like Cournot’s, studied in detail by Simon (1982), while type 6 refers to rational choice theory, including actionist theories that are more complex than type 5. It is no longer simple market models but rather rational behavior models that are called upon in the attempt to build bridges between the micro and the macro. Actionists rediscovered the idea dear to Weber and Popper that all social phenomena (macroscopic level), particularly the structure and function of institutions, should be explained and understood as the result of individual actions, not in terms of universals or collective properties. But in affirming this epistemological principle, actionist theorists did not reduce social phenomena to psychological ones, because in their understanding, action can be influenced by institutions, and institutions are to be explained not only by the actions of individual agents but also by those of other institutions. Here we should cite works by Olson (1966), Schelling (1978), Boudon (1973, 1977, 1990), Hechter (1983), and Coleman (1990).

Type 7 encompasses relativist theories such as ethnomethodology, symbolic interactionism, and phenomenological sociology; primary representatives here are Blumer and Garfinkel. For Blumer, social behavior is negotiated case by case, not imposed by a hidden structure. Macrostructure is unreal, and the empirical foundation of a sociological theory can only be individual behavior in real situations. Whereas Homans left macrosociology intact, interpreting it as a series of individual reinforcements, Blumer denied almost all we claim to know of macrosociological structures. For him, the ultimate basic unit of all research is not the individual, but the situation; that is, interaction between individuals. Psychological principles such as reward and punishment are themselves subject to interpretation in terms of situation (this is why we may categorize this approach as microsociology rather than psychology). Type 8 is made up of role, reference group, and frustration theories. All variants of functionalism and social control theories, from Malinowski through Parsons to Bourdieu, are type 9 theories which explain regularities in individual behavior in terms of variables strictly external to individuals such as values and social rules. Type 9 theories, which stand in opposition to type 1 (classical political economy and Spencerian sociology), dominated sociology until the early 1960s. Finally, macrostructural and network theories make up the 10th and last type (see Fararo 1989).

In fact, only type 2, 4, 6, 8, and 9 theories take on the problem of the connection between microand macrolevels. The first four of these ask macrosociological or macroeconomic questions, but are based on the individual as unit of observation. The assumption—that hypotheses about such units enable us to get from the micro to the macro—is operative particularly in normative theories, which try to explain individual behavior in terms of social norms. All the other theories do not go beyond a single level—microsociological for types 1, 3, 5, and 7 and macrosociological for type 10.

The macrostructuralist project is limited to certain aspects of social reality. It cannot, any more than any other theory, offer a solution to the problem of the links between the micro and the macro. While the rational choice theory presents an undeniable advantage over other theories, it cannot serve as a universal solution: presenting it as unconditionally valid makes it vulnerable to the same dangers other theories have encountered. As for functionalism, it’s error was to yield to the temptation of hegemony; it claimed the title of general theory of social systems, when some of its principles are only valid for particular, tightly circumscribed domains. This means that there is a right way of using functionalism and normative theories, just as there is a wrong way of using such strong, all-encompassing theories. There can be no single solution to the problem of the links between microand macrosociology, any more than there can be a single mode for explaining all phenomena (the first of these problems being only an aspect of the second).

Bibliography:

1. Alexander J, Giesen B, Munch R, Smelser N J (eds.) 1987 The Micro–Macro Link. University of California Press, Berkeley, CA
2. Blau P 1964 Exchange and Power in Social Life. John Wiley, New York
3. Blau P 1977 Inequality and Heterogeneity. The Free Press, New York
4. Blau P, Duncan O D 1967 The American Occupational Structure. John Wiley, New York
5. Blau P, Schwartz J 1984 Crosscuting Social Circles. Testing a Macrostructural Theory of Intergroup Relations. Academic Press, Orlando, FL
6. Boudon R 1973 L’inegalite des chances. Colin, Paris [Education, Opportunity and Social Inequality. Wiley, New York]
7. Boudon R 1977 Effets per ers et ordre social. Presses Universitaires de France, Paris [The Unintended Consequences of Social Action. MacMillan, London, 1982]
8. Boudon R 1990 L’art de se persuader. Fayard, Paris [Art of Self-Persuation. Polity Press, London, 1994]
9. Boudon R 1996 A cognitivist model—generalized rationalchoice. Rationality and Society 8: 123–50
10. Boudon R 1998 Social mechanisms without black boxes. In: Hedstrom P, Swedberg R (eds.) Social Mechanisms. An Analytical Approach to Social Theory. Cambridge University Press, Cambridge, UK
11. Cherkaoui M 1997 Le reel et ses niveaux. Peut-on toujours fonder la macrologie sur la micrologie? [Reality and levels of reality: Can macrology always be based on micrology?] Re ue Francaise de Sociologie 28: 497–524
12. Coleman J S 1990 Foundations of Social Theory. Belknap Press of Harvard University, Cambridge, MA
13. Collins R 1981 On the microfoundations of macrosociology. American Journal of Sociology 86: 984–1014
14. Duncan O D, Davis B 1953 An alternative to ecological correlation. American Sociological Review 15: 351–7
15. Durkheim E 1893 De la di ision du tra ail social. Alcan, Paris [The Di ision of Labor in Society. Free Press, New York, 1947] Etzioni A (ed.) 1961 Complex Organizations. Holt Rienhart, New York
16. Fararo T 1989 The Meaning of General Theoretical Sociology. Tradition and Formalization. Cambridge University Press, Cambridge, UK
17. Featherman D L, Hauser R 1978 Opportunity and Change. Free Press, New York
18. Goodman L A 1953 Ecological regression and behavior of individuals. American Sociological Review 18: 663–4
19. Goodman L A 1959 Some alternative to ecological correlation. American Journal of Sociology 64: 610–25
20. Harsany J C 1968 Individualistic and functionalistic explanations in the light of game theory. In: Lakatos I, Musgrave A (eds.) Problems in the Philosophy of Science. North Holland, Amsterdam
21. Hechter M (ed.) 1983 The Microfoundations of Macrosociology. Temple University Press, Philadelphia, PA
22. Hedstrom P Swedberg R (eds.) 1998 Social Mechanisms. An Analytical Approach to Social Theory. Cambridge University Press, Cambridge, UK
23. Homans G C 1961 Social Behavior. Its Elementary Forms. Harcourt and Brace, New York
24. Huber J 1990 Macro–micro links in gender stratification. American Sociological Review 55: 1–10
25. Lakatos I Musgrave A (eds.) 1968 Problems in the Philosophy of Science. North Holland, Amsterdam
26. Lazarsfeld P, Menzel H 1961 On the relations between individual and collective properties. In: Etzioni A (ed.) Complex Organizations. Holt Rienhart, New York
27. LeBon G 1895 1975 Psychologie des Foules. Retz, Paris
28. Lenski G 1966 Power and Pri ilege. A Theory of Social Stratification. McGraw-Hill, New York
29. Moore B 1966 Social Origins of Dictatorship and Democracy. Beacon, Boston
30. Olson M 1966 Theory of Collecti e Action. Harvard University Press, Cambridge, MA
31. Popper K 1934 Die Logik der Forschung [The Logic of Scientific
32. Routledge & Kegan Paul, London, 1959]
33. Popper K 1963 Conjectures and Refutations. Routledge & Kegan Paul, London
34. Robinson W S 1950 Ecological correlations and the behavior of individuals. American Sociological Review 15: 351–7
35. Schelling T C 1978 Micromoti es and Macrobehavior. Norton Co., New York
36. Skocpol T 1979 States and Social Revolutions. Cambridge University Press, New York
37. Simon H 1977 On judging the plausibility of theories. In: Simon (1977) Models of Disco ery. Reidel Publishing Co., Dordrecht, The Netherlands
38. Simon H (ed.) 1977 Models of Disco ery. Reidel Publishing Co., Dordrecht, The Netherlands
39. Simon H 1982 Models of Bounded Rationality. MIT Press, Cambridge, MA
40. Sombart W 1906 Warum gibt es in den Vereinigten Staaten keinen Sozialismus? [Why is there no Socialism in the United States. M. E. Sharpe, New York, 1976]
41. Tocqueville A de 1856 L’Ancien regime et la Revolution, in Œu res Completes. Edition Gallimard, Paris [The Old Regime and The Revolution. Doubleday, New York, 1955]
42. Turner JH 1984 Societal Stratification. A Theoretical Analysis. Columbia University Press, New York
43. Weber M 1905 Die protestantische Ethik und der ‘Geist’ des Kapitalismus [The Protestant Ethic and the Spirit of Capitalism. Scribner, New York, 1950]
44. Wippler R, Lindenberg S 1987 Collective phenomena and rational choice. In Alexander et al. (eds.) The Micro–Macro Link. University of California Press, Berkeley, CA