Abstract
This article is a defence of objectivity in sociology, not as is usually conceived as ‘value freedom’ or ‘procedural objectivity’, but rather as a socially constructed value that can nevertheless assist us in accessing social reality. It is argued that objectivity should not be seen as the opposite to subjectivity, but rather arising from particular intersubjectively held values (both methodological and societal) held in particular times and places. The objectivity defended here is socially situated in the beliefs and values of communities. This on its own would imply an epistemological relativism, but in the final section of the article I make a plea for realism as a regulatory ideal underpinning objectivity and one which can lead us to novel truths about social reality.
In discussions of objectivity it is common to begin by either defining what the author means by the concept, or to describe its various definitions. The latter can be historically illuminating (see, for example, Daston, 1992; Daston and Galison, 1992; Janack, 2002), but usually consists of opposing objectivity against subjectivity in a series of antinomies, for example, private versus public, particular versus universal, erroneous versus true, mental versus physical, etc. (Hammersley, 2011: 90). Any, or all, of these result in setting objectivity outside of its social context and indeed opposed to social values which are seen as parochial or capricious. As Martyn Hammersley (2011: 91) notes, these antimonies become combined into a composite idea of objectivity, which is opposed to subjectivity and then enshrined in methodological procedures which, it is claimed, if followed, produce ‘objective’ research. Those who oppose objectivity take it on, on its own territory, that is either a critique of its impossibility, or a defence of its (apparent) opposites.
In this article, I will claim that this critique is a legitimate strategy when objectivity is conceived as ‘value freedom’ or a series of methodological procedures, because value freedom is unachievable and methodological procedures, on their own, do not capture the totality of objectivity. Yet to abandon any framework that can produce publicly verifiable knowledge, is to embrace an epistemological relativism that cripples sociology as an investigative discipline .
This article is a defence of objectivity, but not objectivity conceived as freedom from or beyond social values. Rather it is an argument that objectivity – the practice of being objective – is socially situated. The values objectivity serve and the conditions for objectivity will vary through time and place. Yet this claim would seem to imply a relativism that would render objectivity as mere sociological description of socially situated practices. However, this is not my intention and the argument of this article proposes: (1) that objectivity is a characteristic that may be found in our own subjectivities, so, far from being opposed, subjectivity might be seen as an antecedent to an intersubjective value of objectivity as a regulatory ideal; (2) that intersubjective agreement is a regulatory ideal. However, an intersubjective regulatory ideal on its own guarantees only agreement about practice and thus my argument has one final component. (3) That in, in sociology as other sciences, this regulatory ideal as a search for the ‘fact of the matter’ will often lead us to a clash with reality, because the facts themselves are autonomous of our theories about them. This does not mean that the search will always reveal such facts, but rather that it is capable of doing so.
I begin, however, with a brief review of challenges to objectivity and some of the responses, in order to contextualize the approach I advocate.
Challenging objectivity
Rejection
Objectivity, in any form, has few current champions in sociology (a notable exception is Martyn Hammersley, 2011). The critics fall into two camps, those who would abandon objectivity altogether and those who seek to reform it.
An example of a contemporary anti-objectivity view is that of Barbara Katz-Rothman. Her rejection of objectivity, in sociology, is a rejection of ‘a scientific world view of a nature independent of our emotions and feelings’ in which scientists as human beings, must ‘restrain themselves from imposing their hopes, expectations, generalizations, aesthetics, even ordinary language on the image of nature’ (Daston and Galison, 1992: 81). Katz-Rothman claims that this ‘ethic of objectivity’ has been replaced (in methodological thinking in the social sciences) by an ethic of involvement (Katz-Rothman, 1996: 50). Her core research has been in the area of birth and midwifery. Hers is not simply a dry investigation of the social context of births, but centres on a particular model of childbirth which places the mother and midwife, rather than the process (as in the medical model) at its centre. For her, sociology is a dialogue and a public activity. The introduction to her blog gives a flavour of this: I’ve been a sociologist for a long time, and ventured into a number of different fields over the years: birth and midwifery (which I still think of as my home base); the new genetics and reproductive technologies; medical sociology; bioethics; issues in disability; adoption; race; and now I’m exploring food studies too … What would be really interesting would be to have people talk, with each other and with me, across areas. I’ve tried, with some success over the years, to talk to midwives about genetics; to encourage people who do new reproductive technologies to think about home birth; to have bioethicists pay more attention to what medical sociology can offer; to get people in Food Studies thinking about midwives and where midwifery issues overlap with their concerns. (Katz-Rothman, 2013)
The dilemma for Katz-Rothman, is that, on one hand, it is clear the value of her work lies precisely in involvement, a ‘value-free’ disengagement would render her work pointless. On the other hand, it is a knowledge claim that the world is a particular way, that the experiences of those she researched (and indeed her own) were not invented, but were actual experiences, from which others may learn. This much must be established if she is to effectively persuade. Her work has a purpose to it, as sociology must, yet her rejection of objectivity becomes paradoxical.
A rejection of objectivity tout court leads to contradiction or relativism. This can be found in postmodernism and poststructualism where the point of sociology (or social knowledge generally) is to subvert the epistemological authority of traditional forms of knowledge, thus dissolving the boundaries between truth claims about the world and fiction (see, for example, Clifford and Marcus, 1986). For many, this task has a political purpose, to bring about ‘progressive cultural change’ (Meacham, 1998: 404), but what counts as progressive cultural change is rarely specified and, if it were, might well be contested. These kinds of positions, though admittedly many and nuanced, render objectivity, not just impossible, but irrelevant. Conceivably what remains is not sociology, but at best socially aware activism.
Most sociology, for most of its history, has been an investigative discipline. What is investigated and how it is investigated change through time and place and indeed may be contested in particular times and places, but what unites most sociological enterprise is that it is investigation for a purpose. As Helen Longino recounts, the philosopher Richard Popkin is reputed to have said, ‘If all one wants is knowledge or truth, why not count the number of bottle caps one can lay down between Los Angeles and San Diego?’ (Longino, 2002: 176). Nevertheless if one is drawing investigative conclusions in respect of some purpose, it that sociologists must claim some epistemological authority, even if it is accepted that their claims to knowledge may be contested. All forms of objectivity aim to provide a means to adjudicate between contested knowledge claims. Objectivity cannot embrace epistemological relativism because its value position is such an adjudication. Thus, reformers of objectivity have defended rational and publicly verifiable rules, definitions or procedures of investigation.
Reformation
In the middle of the twentieth century sociologists did try to grapple with the problem of the social located-ness of sociologists’ theories and definitions. Robert Merton’s view of ‘social problems’ is one such example. He maintained that a social problem is only such when the practices it embodies are dysfunctional for a given society. Michael Root (2007: 43) gives the example of child beating. Some parents in the United States beat their children, but this is only a problem for sociology, in the Mertonian view, if it unsettles a community by violating its norms. One can thus imagine a society where child beating takes place and this reinforces the norms of that society. Merton allowed a moral relativism between societies, but aimed to retain a transcendent epistemological value of ‘function’ for sociologists that can transcend societies. Merton proposed two kinds of function, those that were manifest to society and those that were latent. A sociologist could then elect to investigate a ‘problem’ that, in the view of the sociologist, existed but not in the view of a society.
Herbert Blumer opposed this view, but set a different epistemological standard that retained sociologists’ distance. Blumer maintained that something could only be labelled a problem by sociologists, if people in a society themselves labelled something as a problem. This, as Michael Root notes, presents difficulties. For example, between 1960 and 1990 births out of wedlock, in the United States, rose to around a third of all births, yet in those early decades there was more concern about this in the public mind that there is today. It follows from this, in Blumer’s reasoning, that this was more of a problem when the numbers were lower than it is was later when they were higher.
In their different way, each of these positions echo Weber’s value-free sociology, variations of which dominated the sociological view up until the late 1960s–early 1970s. Weber maintained that social concepts were subjectively constructed and therefore moral and political views would differ between time and place (1974: 64). There were no social concepts free of human subjectivities to discover. However, Weber did not embrace the epistemological relativism this would seem to imply (Weber, 1974: 110). He wanted to show how, under these circumstances, a scientific sociology was possible. His starting point can be summarized as saying that in matters of policy there will be always be a debate about ‘ends’, about what should be achieved and therefore what investigation should be pursued. Investigation is value-driven and in sociology the subject matter of investigation is values. In Weber’s view, it does not follow from this that the moral and political values of commitment should bias investigation. However, in investigating the issues arising from such commitment, the social scientist should examine his/her value positions for their logical coherence and their relationship to other concepts and principles. Indeed, Weber proposes two levels of analysis (1974: 77), the first that of the cultural significance of a phenomenon and the second an investigation of the causal factors that lead to the mass significance of such a phenomena. The existence of (his example) the money economy is a concrete historical fact, thus, he implies, existing outside of any given subjectivity, but nevertheless a product of subjectivities. What is often described as Weber’s ‘value-free sociology’ was not a sociology without values, but rather a sociology that began with values, yet was neutral in the conduct and means of its subsequent investigation.
Blumer and Merton can be seen as variations on Weber, but also embodying (perhaps rather unknowingly) the natural science orthodoxy of logical positivism and its emphasis on testing through observation and operationalization of proxy measurement. Objections to this orthodoxy from Hanson (Hanson, 1965) and Kuhn (1970) implicitly underwrote a doctrine of ‘theory ladenness’ at the heart of post-positivist rejections of objectivity. In philosophy of natural science theory, ladenness emphasized the psychological factors of subjectivity and more generic social motivations (such as the training of the scientist, theory development, etc.), whereas in sociology the challenges were more usually on political or ideological grounds. While many challenges came later, especially from feminism, the first are to be found in the critiques of Howard Becker (1967) and Alvin Gouldner ([1962] 1973).
Gouldner ([1962] 1973) blamed the Weberian influences for the state of US sociology, seen as staid and conservative, in the topics it tackled and the stance it took. ‘Value-free sociology’, he maintained, had produced a group myth that sociology was value-free. The myth had become an excuse for complacency among US sociologists, a self serving narrow professionalism that did not contribute to the public interest (Gouldner, 1968: 109). His alternative to this, of ‘objective partisanship’, is that objectivity should be directed to particular goals, some of which are universal, for example, the alleviation of suffering.
In a paper published in 1967, Howard Becker seems to go further than Gouldner by suggesting that sociologists take sides and cannot help but do so. As Gouldner (1968) notes, though he skirts around the question, he never fully answers it and his meaning, or what it should mean has been the subject of debate since. 1 His question is, should sociologists be value-neutral or should they express a commitment to a particular cause?
Becker maintained that whatever perspective one takes in sociological investigation, it will always be one that is from the standpoint of ‘superordinates’ or ‘subordinates’. The ‘complacent’ US sociology Gouldner complained of is in the first category, because its agenda was shaped by the priorities of a powerful establishment. Yet by the time Becker was writing, in 1967, this criticism was less true and Becker himself had become a leading sociologist of deviance. Becker’s argument leads to the conclusion that sociologists of deviance, presumably because they must try to understand or empathize with the deviant(s), will be accused of bias by superordinates. Superordinates are the voices, the power, of official and approved morality. They may be, for example, the officials of institutions such as asylums. The subordinate parties are those, who it is alleged, have violated that morality and may be the inmates of institutions, such as asylums.
Becker maintained that there is a hierarchy of credibility where the superordinates determine and legitimate definitions, practices, etc., The further up the superordinate hierarchy, the greater the credibility. If researchers conduct their research within the terms of these normative definitions and practices, etc. then they will not be accused of bias by superordinates, though their standpoint will have been that of the normative position. Conversely if they adopt the definitions, or investigate, say, the complaints, or situations of subordinates, they will be accused of bias by those higher up the hierarchy of credibility.
In Becker, there is a recognition of the value-laden nature of enquiry and the role of power and rhetoric, but there is also some justification in Gouldner’s criticism of this position as replacing the myth of value-free social science by a ‘new but no less glib rejection of it’ (Gouldner, 1968: 103).
If we are to take Becker at his (apparent) word, objectivity is unachievable. The sociologist must simply decide whose side she/he is on and conduct research from that ‘side’. Does it then follow that she/he only accepts or even seeks evidence that is convenient to her/his ‘side’? A major problem with such partiality is seeking to convince those who do not support such a position. Sociology becomes little more than an exercise in political persuasion.
Value freedom
As I noted above, for many (critics and advocates alike) value freedom and objectivity are synonymous. Value freedom is short-hand for the absence of moral or personal values that would wish things to be a particular way, rather than them being the way they are. The emergence of this doctrine was much encouraged by the results of science, whereby the physical world could be manipulated to show that the conjecture of the scientist could be proven by outcomes (for example, in Faraday’s experiments, Faraday, [1850] 1995), thus demonstrating the independence of nature from the scientist. At the level of the experiment, this was perhaps a reasonable assumption, if it is meant that experimenting should be free of wishful thinking or indeed the charlatanism associated with practices such as alchemy. In the development of science in the twentieth century, value freedom was justified on the basis of a separation between the context of discovery, in which no epistemic values were acknowledged, and a context of justification within the method of science. Thus, value freedom became the absence of non-epistemic values.
As I have indicated above, the basis of the claim to value freedom in natural science was that of a separation of the knower from the known. 2 Similarly the operation of value freedom in social policy and social science rested on a similar principle of detachment. For example, one can find this in the notion of ‘an objective political dialogue as a procedurally fair dialogue in which no one actively hinders any citizen from expressing his or her views and voting his or her conscience about what justice demands of society’ (Graham, 2002: 297). John Stuart Mill defended this concept of liberal political justice because it increases the likelihood of arriving at the truth (Mill, [1872] 1987: 16–23). Other expressions of it can be found in Rawls (1971) or more recently in a concept of public reason, ‘in which a claim about what justice requires in a given social situation is objective if it can be justified by appeal to a conception of justice that all members of society can be reasonably expected to accept for their own reasons’ (Graham, 2002).
However, Kevin Graham (2002) argues that these concepts of objectivity do not go far enough in accounting for epistemic inequalities that arise from a range of injustices that afflict many members of society. What counts as a fair and open dialogue is determined by those who acclaim it. At the time Mill was writing, the dialogue excluded the poor and women. 3 Even under universal suffrage, people are excluded because they do not have the cultural capital to be part of that dialogue. This argument is well developed in feminist critiques of social science (see, for example, Stanley and Wise, 1979, Oakley, 1981, Smith, 1991). The concerns of sociology in its ‘golden age’ of the 1940s to the 1960s were around industrial and technological development (Letherby, 2003: 20) and when the family was a topic for discussion or research, it was in respect of its function in reproducing industrial society. Women were either invisible or adjuncts and were certainly defined by their relationship to men and their reproductive role in the family. Such exclusion, from the questions of sociology, let alone its investigation, extended to people of colour, disabled and gay people.
The response from feminism, both epistemological and methodological, has been influential on social science. Standpoint theorists, in particular Sandra Harding (1986, 1991, 1993), maintained that by privileging the voices of the oppressed a more valid or authentic knowledge is likely to ensue. As Graham (2002: 306) notes, that does not mean there is a guarantee such better knowledge will always come about, but rather it is more likely to. Moreover, it places the onus on those in positions of power, political and /or epistemic, to try to achieve a ‘more reflexive and critical understanding of others’ viewpoint and their own’ (Graham, 2002: 307).
This honesty about values, beginning from values, taking values into account has taken a number of different forms in feminism. Standpoint feminists, such as Sandra Harding, maintain that better science can be done if it begins from explicit values, in this case, those of the oppressed. Their needs and priorities will be different from those of the oppressor. Helen Longino has argued for the incorporation of non epistemic values in theory choice (Longino, 1990), while others, especially in social science, have emphasized reflexivity in method, or beginning from subjective knowledge and values. Gayle Letherby has described this approach as ‘theorised subjectivity’ (Letherby, 2003) and she maintains that such value explicitness produces better insights. I will return to Letherby’s concept of theorized subjectivity below.
Though they are not wholly compatible with each other, there is much merit in each of these positions. Longino’s introduction of non-epistemic values is similar to my own in essence, though she does not embrace realism as I do. Harding’s standpoint theory brings into relief how the legitimation of knowledge is underpinned by power. Finally Letherby’s insistence on beginning with subjectivity has a long history in sociology, notably of course in the work of Alfred Schutz. In the next section I will support this position and say there can be no objectivity without a reflexive audit of one’s subjectivity, but while reflexivity, or theorized subjectivity, may be a necessary condition for objectivity, neither are a sufficient condition.
Subjectivity, intersubjectivity and values
The subjectivities of others have long been a focus of sociology through the external gaze of the ‘objective’ sociologist. However, more recently the subjective position of the sociologist, as investigator, has received attention The critics of subjectivity as a basis for sociology will often characterize it as, not just as individual insight or knowledge, but as partial, emotional, biased or irrational. Quite possibly an individual might be all of these things some of the time and some of these things all of the time, but it does not follow that to be subjective is always all of these things. Indeed, as social actors, yet alone investigators, it is important to make judgements about how the world is, as opposed to how we would wish it in order to successfully negotiate it. As individuals and social actors we constantly reflect upon and assess the world with varying degrees of success.
Subjectivity is more than just feeling, likes, dislikes or views. Most of us are capable of reasoning why things are the way they are and why we feel 4 a particular way, even if that is not convincing to another. We do not just feel, we are capable of reflecting upon how we feel.
Our mental capacity to reflect upon ourselves is subjective in that it looks inwards, but it is to view ourselves as an object. Indeed, as Hammersley (2011: 90) reminds us, the use of the terms ‘subject’ and ‘object’ have been reversed only in Kantian and post-Kantian philosophy. Previously ‘objects’ referred to the objects of thought and ‘subjects’ things as they are in themselves. Thought of in that way, an audit of our consciousness inevitably requires us to separate out the objects of consciousness from its examination. As an existing subject I feel happy at this moment, why do I feel happy? Why do I not like Jack? Why did I enjoy that novel? It can be to adopt a critical attitude, questioning of self-memory, our recollections or those things we perceive through sense data. It is a peculiarly human trait. To be a subject, is to be aware of being that subject, thus seen as an object of thought and is only something humans can do. So subjectivity and objectivity (in this primordial sense) are human devices, they are of our own making. This is not to imply a Cartesian dualism, but rather that at a practical level such a duality undoubtedly exists.
That we can look at ourselves and question the truth of our own accounts of ourselves or others, or what we ourselves perceive, is to be in possession of a simple form of objectivity – an ability to set aside an experience as differentiated from the flux of other experiences. Being ‘objective’, in this sense, is not a social act, but a mental one. Whether this ability to examine ourselves in such a phenomenological way could ever have developed without social interaction, (thus the reduction may only be partially psychological) is a question beyond the scope of the current article.
Perhaps the ability to reflect, to examine the contents of our memory conferred some evolutionary advantage on our ancestors? Certainly, much later, such an ability played its part in scientific reasoning (Williams, 2000: 9–10). Whether or not the apple really fell on Newton’s head does not matter much, what matters is that he had the ability to reason about objects such as apples and their earthward trajectory. Helen Longino (1996: 264–8) describes how the biologist Barbara McClintock was able to develop a feel for an organism under the microscope. This was not just to describe it, but to have a much deeper understanding. Feminists have often used this kind of example to illustrate the importance of subjectivity in observation, but notwithstanding that, it is an act of radical differentiation, whereby the organism is first identified as an object of consciousness, but then becomes an object of empathy. But it is still an object. I’m not sure if McClintock went on to make generalizing statements about biological behaviour (though it is hard to see a biologist avoiding such statements like, ‘if organism A in circumstances B behaves in such and such a way, then this may be predicted to hold in contexts C….n’, but certainly the thought objects of Newton, Faraday, Bohr, Einstein, etc. became the basis of mathematical or logical expressions that were capable of empirical verification or falsification.
Objectivity might be said to begin, in a somewhat Schutzian way, with a phenomenological act of examining the objects of one’s consciousness (Schutz, [1932] 1967). This is almost a trivial point, but one that is a good starting point for moving on to a socially situated concept of objectivity. A consequence of this phenomenological act – that of reflexivity – has been widely explored in recent decades (e.g. Okely, 1983; Clifford and Marcus, 1986; May with Perry, 2011), where one acknowledges that our observations of the world (and subsequent reasoning about them) are filtered through our experiences of it. In Karen O’Reilly’s words, causing us ‘[to] subvert the idea of the observer as a detached, impersonal research tool; but rather than undermining the scientific enterprise, it means we are being increasingly rigorous, increasingly sceptical’ (O’Reilly, 2009: 191). In this sense autobiography can be objective, indeed, the autobiographer can be her/his harshest critic or interrogator. But however harsh the autobiographer (or the reflexive sociologist) is in her/his judgements, the accounts lack any kind of arbitration or grounding in something beyond the self and the reflection upon the objects of consciousness.
Theorizing subjectivity
Much of sociology requires not just intersubjective agreement and validation by the sociological community, but because it is grounded in the life experiences of those studied, it must be authentic to them. In recent years there has been much discussion of the role of reflexivity in these relationships (see, for example, May with Perry, 2011). Though it takes many forms, at the heart of it is the kind of auditing I described above. Gayle Letherby (2003, 2012: Chapter 5) coined the term ‘theorised subjectivity’ to indicate a wider ‘theorisation of the subjective (which includes the researcher’s motivation and research practice and the respondent’s expectations and behaviour) and its significance to knowledge production (Letherby, 2012: 89). According to Letherby, theorized subjectivity is not a substitute for objectivity, but a necessary precondition. In order to be objective (however the term is defined), we need the authentification that is derived from such theorizing. She goes on to say: ‘I do not believe I am in a position to generate the “true story” of any experience I research but I do believe that “my story” can stand in opposition to and as a criticism of “other stories”’ (Letherby, 2012: 91).
Letherby’s theorized subjectivity moves beyond the sociologist’s own identity to require her/him to exercise a sociological imagination about the world. Nevertheless Letherby offers her story for scrutiny by the world though it remains that she and others wish to influence the world in some way (otherwise why bother conducting an investigation?). In offering her story to scrutiny, she moves from subjective to intersubjective scrutiny.
Intersubjectivity
In much of our everyday activity we make assessments about the world and we act upon them and mostly those assessments are correct enough to permit us to survive and reproduce social life. We put money in the bank and (mostly) we can expect to withdraw it. We put enormous trust in other car drivers, in aeroplane pilots and even politicians. We get it wrong, but getting it wrong leads to self-correction (assuming we survive the breach of trust). To say (as some do) that the social scientist does not have the epistemological authority to claim beyond an acknowledged and self-mediated description is to deny her/him knowledge claims that we take for granted, even in everyday life.
So we differentiate at an individual level, aware that we are doing this, and we necessarily move to an interaction with the physical and social world that requires us to have trust that they exist in the way they seem, or people say they seem. Finally, these everyday interactions of trust are not just validated by our own direct experience, but by the stories of others. I do not need to go to a particular quarter of a city to know it is dangerous, I have the testimony of others which I trust. I trust the claim of food manufacturers not to adulterate my food. I trust (to an extent, but I have the ability to evaluate) a review of a restaurant, a book or a piece of music. Social life and what we might call the socialized aspects of the physical world are social constructions, but they are also things we act upon and trust (often with our lives) and have real consequences (Williams, 2011).
All of this is trivially true, but it is a necessary precursor to thinking about the intersubjective dimension of objectivity. Intersubjective agreement takes place at many levels and is a social validation of the purpose of an activity, the differentiation between objects, or states of affairs, and the truth of the matter. While, after Hanson, and other post-positivists, we can agree that the objects, states of affairs and the methods to know these are theory- or value-laden, in the sense that they come with beliefs about how the world is, or the way it should be, particular values can be challenged on moral or evidence grounds. Values, or the claims driven by values do not have epistemological equivalence. In everyday life we would give more credence to a statement ‘a hand placed in fire will be burnt’ than its counterfactual, thus it is also in the intersubjectivity of investigative communities, such as sociologists, that value statements about objects, or states of affairs, will have different epistemological status.
Values and intersubjectivity
The values that objectivity (as indeed the values of investigation and method) serves will be different in different social contexts. For example, in UK sociology, between the mid-1990s and the present, medical sociology has been both the predominant subject area for sociology and the target of a great amount of funding, to the point where the British Sociological Association (BSA) health and illness section conference (MedSoc) has in some years been bigger than the BSA Conference itself. More recently the Economic and Social Research Council (ESRC) has committed to investing over £60 million in ‘big data’ research (admittedly, much of this is not sociology), and similarly this area of research is itself growing rapidly (see http://www.esrc.ac.uk/news-and-events/announcements/25683/Big_Data_Investment_Capital_funding_.aspx). In both cases one might argue that the agendas are ‘ideological’, they are certainly social in that they reflect societal priorities and interests. In the first case, for example, are there links between illness and poverty, and what are they?; are gender differences in morbidity mediated by social factors?; do people who live alone endure worse health? These research questions embody the value of good health to society as a priority, a value not always present in every society or in every era. In the second case there is also a substantial methodological interest, but it is argued that the analysis of large, often linked, datasets can bring knowledge that will, for example, increase economic prosperity or reduce crime.
Sociology, like other investigative disciplines, is publicly validated through the willingness of societies to fund its research, adopt its results and to sanction its reproduction through training people to be sociologists. Because sociologists (in order to continue to be employed) must research those aspects of societies that are considered important to the society in which they operate, then there is a social contract between the discipline and those people in society who are able to nominate what should be researched.
That this is the case is underlined by the counterfactual of the occasional refusal of sociologists and anthropologists to research certain topics, for example, in the case of Project Camelot, in the 1960s and the later Minerva Consortium. 5 Instances such as these, or more frequently the presentation of research findings inconvenient to funders, indicate that sociologists are not simply the dupes of whoever will fund them, but are capable of agreeing, or disagreeing with, what should or should not be researched.
The social contract between sociology and the wider society is embodied in an intersubjective agreement about values in both directions. While sociology must, to some extent, adopt the investigative priorities of the society it is located in (if it is to enjoy the legitimacy of that society), it often passes on its values in the other direction, by suggesting topics for research and how they shall be researched. In fact, much of what is researched arises from sociologists’ prior research that indicates new or fruitful areas of investigation. The growth of the sociology of health and illness is not solely to do with funding, but more to do with the development of expertise and fruitful research programmes, which in turn attract further funding. The values of society permeate sociology, but sociological investigation is capable of influencing social values through its research findings.
With Helen Longino I accept that not only is the permeation of social values into investigation unavoidable, it is often desirable. Sociology does not simply investigate society in order to explain or understand that society, but mostly it is called upon to investigate that society with a view to resolving a problem, or problems in that society. This is not a peculiar characteristic of sociology, or social science, but is equally so in medicine or natural science (e.g. investigations of cancer, climate change, or genome sequencing).
The values objectivity serves will then vary between time and place and indeed the sociologist is not simply a ‘gun for hire’, but will have her own ‘citizen values’, as exemplified above in the example from Katz-Rothman’s work. But what is it that keeps the sociologist honest and not simply gathering ‘evidence’ to support her/his own ideological position? Of course, some sociologists are guilty of just that, but most are guided by a search for the truth of the matter, with a respect for methods and evidence. There is, then, agreement on what constitutes good methodological practice and indeed that this is the case can be found in methodological critiques of sociology that does not reach those standards (see Devine and Heath, 1999, for examples of such critiques). This is not to say that methodological standards or choices of methods are universal, but rather that publicly verifiable standards will exist in communities of investigation in disciplines or sub-disciplines.
For some reformers of objectivity, a critical intersubjective encounter, of the kind I describe above, in the ‘scientific community’ is enough of an epistemological guarantee. Here I am thinking in particular of Longino’s ‘transformative interrogation’ which commits the community to reject theories based on ‘bad science’. Yet, if we ground objectivity as a value only in some form of intersubjective scrutiny, there is the risk that findings or theories considered deviant to that community will be rejected (Williams, 2005). ‘Bad science’ is not an epistemological given, but an intersubjective judgement that is opposed to the methods and procedures that constitute ‘good science’ in a community of investigation. The ’truth’ of the matter and how to get to it, become a social convention, or, in Longino’s examples, a pragmatic assessment of the value of competing theories to an outcome (Longino, 1990).
To sum up thus far: though a subjective audit is possible and possibly effective, the problem is that the individual is the arbiter of all things. Intersubjectivity at least provides some kind of agreement about what is, or is not the case, (or ought to be the case) and indeed the grounds on which such agreement may be reached and such judgements made. However, intersubjective agreement is about what we collectively believe is the case about the values driving investigation (that constitute its purpose) and the methodological values embodied in investigation. It is an expression of what an investigative community believes to be the case, but not what actually may be the case.
Realism
At the beginning of this article, I argued that objectivity has three characteristics that transcend time and place: purpose, differentiation and truth. Thus far, I have tried to show that purpose is unavoidable, but will be manifested through different values in different times and places (the values that objectivity serves). What counts as objectivity, the conditions of its existence, will be the epistemological and methodological values of an investigative community. The community will therefore hold intersubjective agreement about differentiation (naming, classification, theories, etc.) and what is or is not true (which is a function of the methods of investigation). Yet these things are no more than social conventions. Now while this is understandable and inevitable in moral or aesthetic matters, what is it in an investigative discipline, such as sociology, that allows us to move beyond convention? The decision to investigate X and not Y might be a moral one, a convention shared by a society and the investigators, in this case, sociologists, but is there anything that can underwrite the epistemological or methodological standards, that will move beyond convention?
In this final section I will argue that a realist approach to objectivity provides an ontological grounding beyond intersubjectivity. Realism is a metaphysical doctrine, which proposes that there is an actually existing reality beyond sense perception, and its scientific analogue holds that theory and method can lead us to at least a partial view or explanation of that underlying reality.
As Roy Bhaskar puts it: Things exist and act independently of our descriptions, but we can only know them under particular descriptions. Descriptions belong to the world of society and of men; objects belong to the world of nature … Science, then, is the systematic attempt to express in thought the structures and ways of acting of things that exist and act independently of thought. (Bhaskar, 2008: 250)
An important methodological characteristic of longitudinal analysis is that, unlike cross-sectional analysis, it can show process over time, because it links the records of the same individuals (or other sampling units) observed at different time points. In recent decades much has been made of the increase in the numbers of people who live alone and in some cases unfounded sociological speculation (e.g. Beck-Gernsheim, 2002) that this indicates a rise in ‘non-family’ living, individualization and lifestyle choices. Cross-sectional analysis can provide plenty of aggregate data on the characteristics of those who live alone, but only longitudinal analysis can produce aggregate data capable of demonstrating the transitions from one kind of family/ household structure at time t1, to living alone at t2 and the transition to a further family/ household structure at t3. Research by Ware et al. (2007), using longitudinal census data from England and Wales, indicated that the kinds of transitions experienced were mediated by age, sex and the presence of children. Thus, many younger people spend time living alone, before moving into ‘nuclear family’ arrangements, men who were in ‘nuclear family’-type arrangements exited them to live alone earlier than women (because they are more likely to live with children after marriage dissolution), but once living alone, they were more likely to live alone longer than men. Much living alone is shaped by family origins and we can readily infer only some is by ‘choice’. Indeed, even among young people who live alone, after exiting the parental household, the social class of the originating household is a predictor of the likelihood of living alone (those in non-manual classes being more likely to live alone). While other forms of analysis might have provided some evidence for the presence of these kinds of household change, longitudinal data (in this case a very large sample national dataset) could provide irrefutable evidence of such changes over time through following individual life courses. These data do not provide an entire picture of social ‘reality’, but they do provide a partial picture of a reality, but crucially one that was not ‘visible’ without such analyses and indeed one that ran counter to what was believed to be the case.
The ontological claim of realism is that there is an actually existing world that exists in particular ways that may or may not reveal itself to investigation (Bhaskar, 1978, 1998; Pawson, 1989, 2000). Implicit in this is the metaphysical claim that there is a ‘truth of the matter’, but this claim is not verifiable, in so far we cannot know whether we have found the ‘whole truth’. However, as the brief ‘household change’ example demonstrates, we can uncover partial truths about the world and we can uncover error. We can obtain empirical ‘closure’ in particular domains, which serves to act as motivation of a search for truth as an epistemological and methodological goal.
The example I give is a relatively straightforward one of the relationship between prior beliefs, or theories, investigation and results which are novel or confounding. But a sociologist might reply that the social world is ontologically unstable and complex and moreover the sociologist might be said to be investigating social constructions with socially constructed procedures. The sociologist is (unlike the natural scientist), as the old cliché goes, part of what is being studied.
The first of these (instability, complexity) is certainly the case. However, this does not imply constant flux, devoid of any stable characteristics that permit differentiation. There is, to use the term of Robert Nozick (who was admittedly referring to the physical world), partial invariance (Nozick, 2001). This is not to claim that the world is unchanging, or stable, but rather that there is stability enough for us to make inductive generalizations from one time or place to another and when these do not hold, we also have the ability to identify the translation mechanisms of change.
The second of these implies that the sociologist cannot escape ideology, so according to the interpretation of Becker, should therefore embrace it. The kinds of critiques advanced by Graham (2002) or standpoint theorists (above) make a direct link between ideological motivation (which may not be realized or apparent) and the choice of topic, method, etc. While it is certainly true that sociology and other disciplines will take normative constructions as their subject matter, it does not follow that the assumptions about these constructions are derived from the constructions themselves. As I said above, the choice of topic to study may arise from the personal interests or commitments of the sociologist, societal (and perhaps funding) priorities and these may be set in a wider ideological milieu. But it does not follow that sociology or sociologists are pursuing their own ideological agenda or are (in Becker’s words) on the side of the ‘superordinates’.
Professional sociologists also operate in a cultural milieu, but one that transcends current normative priorities both through its history and through its current battery of theory and method (these of course are also the product of its history). Taking the above example from the Merton–Blumer debate, suppose that a sociologist of the family is to investigate births outside of marriage, because this is perceived by wider society as a ‘problem’. Such an investigation can discover the norms, values and structures of extra-marital births, without pronouncing on whether or not this is a social problem. Asking the sociologist to pronounce on this is a red herring. The data themselves both on the historic trends in such births and the decision-making or pressures that led to extra-marital births can be made apparent, and, as with data on living alone, exist independently of the sociologists’ theories about them, or interpretations of them. Description alone can uncover hitherto unknown ‘truths’ and indeed, description is an important, though not limiting role, for sociology. Nevertheless it would be a legitimate sociological strategy to then ask how this came to be framed as a problem in the current situation. This suggests two lines of investigation: why the numbers of extra-marital births rose until the 1990s and why they then began to fall and, second, how did the social construction of extra-marital births as a problem arise, when the births themselves were falling (the translation mechanism of change)? The sociologist can provide both description and explanation that are not context-free, but equally are not context-dependent. Indeed, it is precisely the ability of sociologists to move beyond the normative that has made the kind of feminist-inspired research of Katz-Rothman possible. Nevertheless, the question for ‘public sociologists’ such as Katz-Rothman, whose investigation begins from particular ideological positions is, what kind of contrary evidence would it take for them (or her) to abandon such positions? For example, her own work is wholly qualitative and draws on the experiences of mothers during childbirth and this leads her to be a passionate advocate of home birth, but what if quantitative data indicated that, in particular jurisdictions, infant mortality rates among children born to mothers at home were considerably higher than that of children born to mothers in hospital?
Conclusion: situated objectivity
In this article, contrary to a widely held view that objectivity stands outside of social context, I have argued that it is itself a value situated in context and moreover I have suggested that in a simple and primordial sense it can be said to begin with an individual differentiation of objects or states of affairs, that may be tested intersubjectively. In everyday life, such ‘intersubjective testing’ is informal and simply fit for purpose. Sociology too, like any science, is driven by purpose. As an investigative community (or more properly communities), sociology will decide intersubjectively what the purpose(s) shall be and these will differ through time and place, as will sociologists’ theories and methods, but these too will be scrutinized intersubjectively. Whatever the purposes, the differentiations and the ‘truth of the matter’ are what a community believes to be the case, but not what might actually be the case. Thus, subjective differentiation and intersubjective testing (or scrutiny) may be necessary conditions, but they not be sufficient to guarantee an objectivity that can transcend the particular time and place. In the second half of the article, I have made a plea for a moderate realist approach, which begins from an ontological assumption of an actually existing world, that may or may not reveal itself to the investigator. This assumption can be borne out empirically, because the theories and methods of sociology can potentially lead us to a confrontation with reality that may uncover novel facts, possibly inconvenient to our theories or purpose.