The Analysis of Social Process

Florian Znaniecki
University of Poznan, Poland


The subject-matter of sociology is not the social process (progress, evolution, etc.), but concrete social processes analogous to chemical processes. Sociologists have, however, as yet no adequate method of studying and controlling these processes. The old method of enumeration used by statistics merely expresses more accurately and in more technical terms what common sense already knows. To state, even in hierarchical order, the factors involved in a social process in no way help us to control this process, for an effort to deal with any one factor leads us to a mass of other factors involved in that one, and so on indefinitely.

What we need is analytic knowledge which, given a definite set of conditions, can state that a given cause will produce a certain effect with complete certainty, if no other cause interferes; and can with equal certainty determine how this effect will be altered, if some specific cause does interfere. This kind of knowledge can be obtained only if the facts investigated are reducible to relatively simple processes occurring in systems of definite elements, interconnected within but relatively isolated without. We actually find in the active life of every human collectivity innumerable real systems as definitely limited and structurally coherent as those of physics or biology. By confining our study to these and taking a sufficient number of typical cases, we should be able to discover exact causal laws of the same type as physical laws. It will be for the practical leader to make use of these laws in connection with the results of the other social sciences – economics, religion, technique, etc.– in order to control social situations.

The term "social process" is used here in the abstract generic sense (like "chemical process") to indicate a certain general class of facts. I do not mean by it the total historical becoming of human civilization, as those do who would like to substitute it for the discredited "progress" of Comte or the "evolution" of Spencer. Since sociology has broken with the tradition of its early founders and, leaving the grandiose historical problems of the past and future of mankind to philosophy, has settled modestly down to dealing as a special science with a multiplicity of indefinitely repeatable, specific data, it seems more profitable to discuss "social processes" rather than "the social process."

At the present stage of scientific development the processes which sociologists actually investigate as "social" may be defined (without trying to be entirely exact) as prolonged occurrences which affect and modify concrete human collectivities, big or small, such as nations, races, classes, urban and rural communities, neighborhoods, etc. The industrialization of an agricultural community; the growth

(38) or decrease of criminality, poverty, or divorce among the population of a certain territory; a change of political standards in the public opinion of a nation; a war; a strike; a racial struggle; the assimilation of an immigrant collectivity; the spread of communistic organizations among the working class; the segregation of a city population: these are a few random instances of such occurrences.

Now, the problem which I should like to subject to discussion is how these social processes ought to be approached for the purpose of obtaining the best possible scientific results from their study. This problem is particularly important nowadays. We all know that during the life-time of our generation, within the collectivities to which we belong, many social processes have increased in complexity, rapidity, and vastness of the changes involved, at a rate unknown in history; and the prediction seems to be justified that this increase will continue. In the face of this, efforts made by the leaders of social activities to control various specific processes are more and more unsuccessful; and as their impotency becomes obvious, the danger of a possible overthrow of our entire civilization looms ahead. The sciences of man, particularly sociology, economics, and political science, have been challenged to justify their pretensions, loudly voiced for the last century, and to give practical leaders such theoretic laws as could be applied for the control of the future. We, the sociologists, are of course accepting the challenge. But we must be fully conscious what this acceptance implies.

It means that we can no longer be satisfied with merely formulating in more technical terms what everybody already knows more or less from common-sense reflection. Why is it that "discovery" in sociology consists primarily in finding new concrete facts, as in archaeological collecting, not new relations between facts, new laws, as in physical science? Simply because the prevalent type of sociological generalization does not stimulate the kind of research which leads to the discovery of laws. It is the lowest kind of induction, the one that proceeds by mere enumeration; and though sociology has lately improved its reliability as compared with popular judgment, it has not changed its logical foundation.

The unscientific observer reasons: "In nearly all the cases I know of criminals, they come from bad homes; consequently, defective

(39) home education is the main cause of crime." The older sociologist used to reason: "In most of the cases which have been observed, criminals came from bad homes; consequently, defective home education is a potent factor of crime." The present-day sociologist is reasoning: "Out of 1,000 criminals investigated, 600 come from bad homes; whereas out of 1,000 non-criminals in the same population, only 300 do. There is a positive correlation between defective home education and crime; consequently, the former is a factor in the latter." A similar reasoning will be pursued with other apparent factors of criminality, and perhaps the factors will be hierarchized in importance. The results will be, of course, more probably right, if reached by the statistical method, than by untrained observation. But two fundamental points will be common to all these reasonings: ignorance of the precise conditions under which any one of these factors, becomes, if ever, a real cause determining inevitably its effect; and a corresponding neglect of exceptions, lack of any serious attempt to explain cases when the result does not occur, though the given factor is active, or when the result occurs, though this factor is not active.[1]

Very different from this is the kind of analytic inductive reasoning predominant in physical science where, on the basis of a thorough study of a few cases, a conclusion is reached which can be most adequately expressed in the formula: "Under conditions A, if B, then C, unless X"; the final reservation provides a stimulus for research which, starting with apparent exceptions to the law, discovers a new law.

It is commonly assumed that the knowledge given by the kind of induction which culminates in the statistical method of research is sufficient for practical purposes. If we know the most important factors in a social process, surely by influencing these factors we can influence the process! That sounds all right in the abstract—but how dues it work out? Suppose a government has instituted a research into the spread of revolutionary tendencies among the work-

(40) -ing class of its country. Let us assume that this research has reached by statistical methods the view that unemployment, the decay of religion, and foreign propaganda are the most important factors in this process. All the government has to do then is to stop unemployment, revive religion, and check foreign propaganda. But how is it to be done? Obviously by influencing in turn whatever will prove to be the main factors in each of these three processes, and so on, until we find that to influence appreciably any concrete social process with the kind of causal knowledge obtained by the statistical method we must modify the entire cultural life of the collectivity in which this process is going on. That is why some leaders still use the primitive ordering-and-forbidding technique, while others limit themselves to "case work," to counteract in concrete instances undesirable effects of causes beyond their control.

What social leadership needs is theoretic knowledge of the same kind as technical engineering is based upon—analytic knowledge which, given a definite cause under a definite set of conditions, can predict its effect with complete certainty if no other cause interferes; and can with equal certainty determine how this effect will be altered if some specific cause does interfere.

The only way such knowledge can be reached is by methodical analysis of empirical complexities. When sociology passed from the study of the "progress" or "evolution" of mankind to the investigation of specific social processes going on within particular collectivities, it made a step in the right direction. But it did not go far enough. The kind of social process instanced above is still much too complex to be treated as a cause or an effect and subjected to a law: it involves a veritable maze of multiple interacting causes and effects. The problem of its analysis, however, has nothing to do with so-called sociological atomism, viz., the substitution of a mass of individual psychological processes for a social process. Even if such a substitution were valid—which it is not it would not be helpful; since psychology has similar difficulties in subjecting its processes to exact laws. The solution must be sought in an entirely different direction.

In modern science, every process is referred to some static or dynamic system of interconnected elements. It is this system which furnishes the set of conditions under which a cause produces its effect,

(41) and the causal law applies to the process in the very measure in which the system is closed, which means that no appreciable changes in its composition and structure are produced by outside influences while the process is going on, except those implied by the law. Thus, a physical process is viewed as occurring within a kinetic system of moving bodies, or a thermo-dynamic system, or a magnetic field; a biological process goes on within a cell, or a tissue, or an organism.

But the human collectivity to which sociology refers social processes is not this kind of system. It lacks a definite structure, i.e., a combination of forces binding a limited number of specific elements together into a relatively closed whole. Consequently, it cannot be described in scientific terms, nor can the processes which affect it be ever strictly defined. The objective and impartial observer finds there an inexhaustible wealth and chaos of data. Most observers, though, are not objective, or at least not impartial. Some are practically interested in certain data, pick these out, and organize them with reference to their subjective purposes. Such is the attitude of social workers surveying a community. Others approach a collectivity with the preconceived notion that there must be a certain kind of objective structure—usually the kind they specialize in. Thus, the human geographer looks in a community for an ecological structure, the anthropologist for a racial one; the economist sees its whole life determined by the "business of getting a living"; the political scientist identifies its structure with that of the territorial group – federation, state, township – which rules the population; etc. Each one of these finds some data to support his claim, and it will depend on his exclusiveness in ignoring other data or arbitrariness in interpreting them, how far this claim will seem substantiated with regard to the community as a whole.

The facts are that in the active life of every human collectivity we find not one but innumerable real systems, as definitely limited and structurally coherent as those of physics or biology, though built mid maintained by human activity. They belong to all domains of culture. Take a city. There are specifically social systems: human relationships regulated by the mores, standardized personal positions, multiple and various groups, such as the township with its administrative subdivisions and offices, political parties, churches, schools,

(42) trade unions, clubs, associations, and families. There are technical systems—the street, sewer, water, gas, electricity, telephone and railway systems, factories, workshops, machines, cars, etc.; economic systems—banks, commercial and industrial enterprises; aesthetic systems—drama, music, architecture, painting, sculpture, poetry, linguistic systems—conversations, speeches, letters, articles, newspapers, books; scientific and philosophic theories taught and learned, applied, and sometimes created; systems of religious rites, ceremonies, beliefs, and aspirations; hedonistic systems—food, drink sex, and play.

Every one of the complex processes which modern sociologists have been studying can and should be analyzed into a multiplicity of elementary processes, of specific changes going on within these particular limited systems, and specific influences exercised by one such system upon another. There is a chance—I believe there is a certainty—that if we do this, we shall be able to discover exact causal laws of these elementary processes, of the same type as physical laws, and thus open the way to a practical control of cultural reality of the same type as the engineer's control of material reality.

At first glance, the task seems appalling, but on closer inspection it proves workable. Obviously, if we find a process affecting at once many similar systems—say, millions of family groups—we do not need to study all of them. A small number of typical cases is sufficient, if well analyzed; and how to select cases that are typical is a minor problem of scientific method. Less obvious, and therefore more important, is the emphasis this situation forces us to put upon scientific specialization, mainly upon the division of labor between sociology and other sciences. Let us map out our field, our specifically social systems, and stick to them. We cannot afford nowadays to duplicate the work of other specialists. Let us leave business systems and processes to economists, religious ones to religionists, theoretic ones to students of knowledge, technical ones to ergologists. We may co-operate with them, but need not feel responsible for their work.

Nor should we be tempted back into assuming, as of yore, the responsibility for providing a synthetic knowledge of human collectivities in contrast to the other more special social sciences. In the hu-

(43) manistic field, as in the field of nature, the only possible kind of synthesis of the special studies bearing on some concrete fragment of reality is practical—the synthesis a civil engineer makes when, in order to plan and build a particular bridge, he utilizes the results of mathematics, physics, chemistry, mineralogy, and geology. Let us, therefore, leave this task to practical leaders in the domain of human culture.



  1. The statistical method of research as a modern development of the induction per enumerationem must be carefully distinguished from the statistical technique of measurement by which quantities are made numerically comparable. Objections raised against the former do not touch the latter.

Valid HTML 4.01 Strict Valid CSS2