Positivism and Sociology

The nineteenth-century French philosopher Auguste Comte is generally credited with inventing both of the terms ‘positivism’ and ‘sociology’ (see Andreski 1974; Keat and Urry 1975; Benton 1977, 2015; Halfpenny 1982).

This positive knowledge was, of course, science. However, the problem as Comte saw it was that each branch of knowledge goes through the three stages, but that they don't all reach scientific maturity at the same time. Astronomy, physics, chemistry and biology had all, Comte argued, arrived at the scientific stage, but accounts of human mental and social life were still languishing in the pre-scientific, metaphysical stage. The time was now ripe for setting the study of human social life on scientific foundations, and Comte set out to establish ‘social physics', or ‘sociology, as a scientific discipline. Since Comte's day the term ‘positivism' has been used extensively to characterize (often with derogatory connotations) approaches to social science which have made use of large data sets, quantitative measurement and statistical methods of analysis.

1.The empiricist account of the natural sciences is accepted.

2. Science is valued as the highest or even the only genuine form of knowledge (since this is the view of most modern empiricists, it could conveniently be included under 1).

3. Scientific method, as represented by the empiricists, can and should be extended to the study of human mental and social life, to establish these disciplines as social sciences.

4. Once reliable social scientific knowledge has been established, it will be possible to apply it to control, or regulate the behaviour of individuals or groups in society. Social problems and conflicts can be identified and resolved one by one on the basis of expert knowledge offered by social scientists, in much the same way as natural scientific expertise is involved in solving practical problems in engineering and technology. This approach to the role of social science in projects for social reform is sometimes called ‘social engineering'.

There are several reasons why positivists might want to use the natural sciences as the model for work in the social sciences. The most obvious one is the enormous cultural authority possessed by the natural sciences. Governments routinely take advice on difficult matters of technical policy-making, from food safety to animal welfare and building standards, from committees largely composed of scientific experts. In public debate (until quite recently - see Beck 1992) scientists have had a largely unchallenged role in media discussions of such issues. Social scientists might well want to present their disciplines as sufficiently well established for them to be accorded this sort of authority. Not unconnected with this is the still controversial status of the social sciences within academic institutions. Strong claims made by social scientists about the reliability, objectivity and usefulness of the knowledge they have to offer may be used to support their claim to be well represented in university staffing and research council funding for their research.

Table 2.1 Suicides in different countries per million persons of each confession

Source: Durkheim (1896, 1952: 154).

That positivists should have accepted the empiricist account of science is not surprising, given the pre-eminence of this view of science until relatively recent times, and given its clear justification for science's superiority over other forms of belief-system. However, the positivist commitment to extending scientific method to the human sciences is more obviously contestable. In later chapters (particularly 5, 6 and 7) we will consider in detail some of the most powerful arguments against this positivist doctrine, but for now we will just consider the case for it. We will use some of the work of Durkheim as our example here, but it is important to note that we are not claiming that Durkheim was himself a positivist (see Lukes 1973; Pearce 1989; Craib 1997). For our purposes in this book, he shares some important features in common with positivists, and this is what we will focus on.

In his classic work on suicide (Durkheim 1896, 1952), Durkheim drew on a vast array of statistical sources to show that there were consistent patterns in suicide rates. He showed that these patterns could not be accounted for in terms of a series of non­social factors, such as race, heredity, psychological disorder, climate, season and so on. He then went on to show that they could be accounted for in terms of variations in religious faith, marital status, employment in civilian or military occupations, sudden changes in income (in either direction) and so on. Table 2.1 shows the pattern for religious faith.

Although there are variations in suicide rates over time in each country, comparison between countries shows remarkable constancy - some countries having consistently higher or lower rates than others. Similarly with religious confession: though absolute rates vary a great deal for the same faith in different countries, there is constancy in that in each country Protestants have higher rates than Catholics, and Catholics higher rates than Jews. Durkheim argues that this pattern cannot be explained in terms of doctrinal differences between the religions, but, rather, is a consequence of the different ways the churches relate to individual followers:

If religion protects a man from the desire for self-destruction, it is not that it preaches the respect for his own person to him...

(Durkheim 1896, 1952: 170)

In his book on suicide, and his methodological classic The Rules of Sociological Method (1895, 1982), Durkheim uses a series of arguments to establish that society is a reality in its own right. The facts, ‘social facts', of which this reality is made up exist independently of each individual, and exert what he calls a ‘coercive power' over us. For example, each individual is born into a society whose institutions and practices are already in existence. Each of us, if we are to participate in our society, communicate with others and so on, must learn the necessary skills, including those involved in speaking and understanding the local language. In this sense, as well as in more obvious respects, we are coerced into following the established rules of our ‘social environment', or ‘milieu'. There is a particularly powerful statement of this towards the end of Suicide:

[I]t is not true that society is made up only of individuals; it also includes material things, which play an essential role in the common life. The social fact is sometimes so far materialized as to become an element of the external world. For instance, a definite type of architecture is a social phenomenon; but it is partially embodied in houses and buildings of all sorts which, once constructed, become autonomous realities, independent of individuals. It is the same with avenues of communication and transportation, with instruments and machines used in industry or private life which express the state of technology at any moment in history, of written language, and so on.

(Durkheim 1896, 1952: 314)

This is enough for Durkheim to show that there is an order of facts, social facts, which are distinct from facts about individual people and their mental states, or biological characteristics. This class of facts, most obviously detected through the analysis of statistical patterns, justifies the existence of a distinct science - sociology - which takes it for its subject-matter. This science, having its own distinct subject-matter, will not be reducible to biology, or to psychology.

However, a further step in the argument is required. As practising participants in social life, it could be argued that all of us possess knowledge of it - this seems to be implied in Durkheims own argument. If this is so, why do we need a specialist science to tell us what we already know? In answer to this Durkheim could point out that his analysis of statistical patterns in the occurrence of suicide had come up with results which most people would find surprising. This apparently most individual and lonely of acts, when studied sociologically, turns out to be determined by variable features of the social environment. In The Rules of Sociological Method he offers us a more general argument. As the facts of social life exist prior to each individual, are independent of their will, and exert a coercive power, they resemble facts of nature. We all interact with natural materials and objects, and we do so through ‘lay’ or common-sense understandings of their properties, but just because of this we would not generally claim that there was no need for natural science. The history of the natural sciences shows innumerable instances of common-sense beliefs being corrected in the face of new scientific evidence and theory. So why should we assume that common-sense assumptions and prejudices give us reliable knowledge of the social world? If, in general, science progresses by increasingly distancing itself from common-sense assumptions, and gaining deeper understanding of its subject-matter, we should expect this to be true of the social sciences too.

Finally, some brief comments are due on the fourth doctrine of positivism - the proposal to apply social scientific knowledge in social policy-making. This view of the public role of social science has continued to be very widely held, and it provides yet another justification for extending the methods of the natural sciences into the study of society. Only on the basis of the sorts of claims to quantitative reliability, objectivity and general applicability already made by the natural sciences could the social sciences expect to be taken seriously by policy-makers. Today in most countries official statistics are collected on virtually all aspects of social and economic life - on patterns of ill-health and death, on marriage and divorce, on unemployment, income differentials, attitudes and values, consumption patterns and so on - and social scientists are employed to collect and interpret these, as well as to give advice on policy implications (in the UK, such publications as Social Trends and British Social Attitudes contain selections from such statistical surveys).

The logical form of a scientific explanation as represented in the empiricist ‘covering law’ model shows how the link between such knowledge and policy might be made. To oversimplify considerably, the statistics might show that criminal behaviour by juveniles was more common among the children of divorced parents. This is not a universal law, but a statistical generalization (though the required element of universality might be present, if it is held that this statistical generalization holds across different cultures and historical periods). However, the basic structure of a scientific explanation can be maintained:

If there are high divorce rates then there will be high rates of juvenile crime. Divorce rates are high.

Therefore: There are high rates of juvenile crime.

If policy-makers are convinced by public opinion that high rates of juvenile crime are a bad thing, and are charged with coming up with policies to reduce them, then this piece of scientific explanation will yield the policy recommendation to take action to reduce divorce rates. Of course, there are some obvious complications here. One is that a mere statistical association between divorce rates and juvenile crime does not show that one causes the other. It could be that some third social fact, such as unemployment rates, causes both high rates of divorce and juvenile crime. A policy of dealing with unemployment therefore might be more effective than trying to do something about divorce. But there might be more subtle problems with the statistical association. It might, for example, be that the association of juvenile crime with divorce holds only where divorce is stigmatized by prevailing values. If this were so, then the appropriate policy might be to work for a cultural shift in favour of more liberal social values. However, none of this counts against the positivist notion of ‘social engineering' as such. Each of these possibilities can in principle be addressed by more exact data-gathering, and more sophisticated analytical methods. There are, however, other lines of criticism, which we will explore in the next chapter.