On the Role of Values in Social Research: Weber Vindicated?

Introduction

1.1 Recurrent debates over the past century about the role of values in social inquiry have not led to much progress in resolving this issue. Instead, there have been swings in prevailing views, frequently based on shifting taken-for-granted assumptions. In Anglo-American sociology during the late nineteenth and early twentieth centuries the predominant commitment was to a normative approach, whereas in the middle part of the twentieth century there was widespread appeal to the principle of value neutrality. However, over the past few decades, in many quarters, there has been a shift back towards advocacy, or at least to tacit acceptance, of normative approaches. Yet, as I will show, this was not based upon sound argument.^[1]

1.2 The failure to make much progress in resolving this matter, and the shifts in opinion that have taken place, probably derive in part from the conflicting demands associated with the awkward position of social science at the interface between academic forms of inquiry, traditionally emphasizing the need for 'detachment', and the realms of policy and practice, which demand or prompt 'engagement'. And, in relation to teaching, there is the tension between this as initiation into a discipline, on the one hand, and as vocational preparation or citizenship education, on the other. Moreover, because of its boundary position, sociology has been highly susceptible to the changes that have taken place in society: to shifts in the character and role of universities, and in patterns of political governance and economic organization more generally. Particularly significant has been the emergence of systems of mass higher education and the increasing subordination of universities to governments and to commercial interests, even in countries where previously those institutions had had a more autonomous role, including the UK (Collini 2012). In addition, there has been a trend in many areas of natural science and technological research towards industrial and commercial models (seeZiman 2000). During the same period, the growing influence on some parts of social science of political ideas and movements, of a broadly Leftist kind, advocating 'critical' and activist approaches has also been important. Against this background, more conventional kinds of research have been criticized either for irrelevance or for serving dominant interests. Active commitment to the cause of progressive social change, of particular kinds, has become widely regarded as obligatory. All these changes have led to increased demands for, as well as more allowance of, value-laden approaches to social research. At the same time, social scientists have frequently still had to appeal to some notion of detachment or objectivity in order to distinguish themselves from the many other voices claiming expertise about social issues, in what have become crowded and noisy public policy domains.

1.3 Aside from these external factors, intellectual obstacles have also played a significant role in hindering progress towards determining the proper role of values in social research. Indeed, the issue is a complex one, not least because it is bound up with several other equally difficult philosophical and practical matters (see Proctor 1991; Kincaid et al 2007). These include views about truth, and specifically about what sorts of knowledge of social phenomena are possible; as well as questions about whether academic research should be concerned with building up disciplinary knowledge or providing quantitative facts and/or illuminating 'thick' descriptions of particular cases; and about whether it ought to be designed to serve the needs or wants of 'users' – policymakers, political activists, occupational practitioners of various kinds, or citizens generally. Furthermore, what are involved here are not just matters of principle but also practical questions of feasibility and motivation. For example, to what extent can social scientists focus exclusively on contributing to disciplinary knowledge or providing ethnographic descriptions, given that they need resources, including time, to carry out this work, and that the availability of these resources has become increasingly tied to demands for demonstrable 'impact' on policy or practice?^[2] Equally important, many people become social scientists because they want to bring about some sort of practical or socio-political change; for them, making a contribution to academic knowledge, of whatever kind, may seem insufficient.

1.4 A further reason for lack of progress in addressing the issue of values is that confusion has been generated by some of the terms used to identify particular positions: for example, taken at face value the phrases 'value freedom' and 'value neutrality' are systematically misleading, while 'objectivity' has several senses that can easily be confused.^[3] Another problem has been a tendency to caricature opposing views. This is especially true of Max Weber's position, which is my central concern here (Bruun 2007; Bruun and Whimster 2012). For example, Gouldner (1973:4) famously portrayed him as a 'magnificent minotaur' whose lair can only be reached 'by labyrinthine logic' and is 'visited only by a few who never return', but 'is still regarded by many as a holy place' (p3).^[4] Of course, there are features of Weber's writings that have made accurate understanding difficult for those engaging in later discussion of his work – the fact that he was located in what quickly became a largely forgotten intellectual context (in particular, that of German historical economics and neo-Kantian philosophy), that his writing style is often convoluted, and that many commentators have had to rely upon translations from the original German, these frequently consisting of fragments from a large and disorderly corpus. Nevertheless, Gouldner's mythical analogy is tendentious.

1.5 Given all this, considerable care is required if we are to make any headway in determining whether the terms 'value-neutrality', 'value-freedom', and 'objectivity', as conceptualized by Weber, refer to an appropriate principle, or set of principles, for social scientists to adopt. My aim in this paper is to show that there is an important set of principles here, and that Weber had, in fact, resolved most of the intellectual difficulties, if not all the practical ones.^[5]

Clearing away some misconceptions

2.1 An important starting point, against the background of longstanding misinterpretation of Weber's work, is to emphasise that he cannot legitimately be treated as a positivist, in any productive sense of that term; a tendency that is not uncommon in discussions of his concept of value neutrality.^[6] The word 'positivism' has, of course, come to be used in an extremely loose and almost always condemnatory way, being deployed against a variety of positions. However, in the nineteenth and early twentieth century it referred to a specific set of ideas, of varying degrees of credibility.^[7] The concept of value neutrality – that social research should be focused exclusively on the goal of producing sound factual knowledge – was one of these ideas, but it is important to recognise that there was disagreement about it among writers who labelled themselves positivists at the time, as well as among those to whom the label can reasonably be applied. Thus, while John Stuart Mill put forward a version of the notion of value neutrality, in his writings on political economy, in this respect (and some others) he was deviating from the position of the founder of positivism, Auguste Comte. Comte believed that social science, or an ethics based upon it, could authoritatively resolve value issues, for instance via ideas of normal and pathological societal development, an idea subsequently taken up by Durkheim, who is also sometimes labelled a positivist.^[8]

2.2 Furthermore, the main influence upon Weber was not positivism, even though he drew on as well as reacted against Mill's work. Indeed, some of the main sets of philosophical ideas that shaped Weber's thinking – neo-Kantianism and the writings of Nietzsche – were fundamentally opposed to positivism, just as much as they were to the sort of historicism represented by Hegel and Marx. Thus, Weber rejected the idea that science can diagnose social pathology, in the manner assumed by Comte and Durkheim, and also by many of his colleagues in the German historical school of economics. Nor did he believe, like Hegel and Marx, that there is a set of values built into human nature whose realisation can therefore provide an immanent basis for value judgments (see Wood 1990, 2004). In this respect, Weber was a post-Enlightenment thinker, representing a radical liberal tradition that emphasizes the plurality of values – not just across different cultures but also within them – with these values often being seen as 'incommensurable' and therefore as leading to 'undecidable' choices. He also differed from positivists in believing that social science must be idiographic, rather than nomothetic, in character, and that it necessarily relies upon interpretive understanding. What all this should make clear is that if we try to understand Weber's views as if they were a form of positivism we will be misled.

2.3 There are several more specific misconceptions about Weber's notion of value neutrality that are common. One of these stems from the terms 'value neutral' and 'value-free' themselves: these may be taken to imply that what is being proposed is that research can and must involve no reliance upon values, or be neutral towards all values. Yet Weber emphasised that science is necessarily committed to epistemic values. He also recognized that social science operates within the social world, so that social scientists are subject to intellectual and emotional commitments, and material interests, that potentially lead to biases in their conclusions. Indeed, he regarded commitment to value neutrality as the struggle to minimize such biases, viewing this as an ethical responsibility on the part of scientists because of the nature of their distinctive vocation (Weber 1919). He also believed that social inquiry must depend upon the adoption of particular value-relevance frameworks, a point I will discuss later.^[9]

2.4 Another important misconception is the idea that a commitment to value neutrality assumes or demands that we make value judgments quite independently of facts about the world; in other words that the concept implies a metaphysical dichotomy between facts and values (Putnam 2002), or that they represent 'homologous spheres' (Burawoy 2013:749). This is true of Kant, on some readings, but it was not true of Weber. The principle of value neutrality does not require this assumption, and there is no good reason for adopting it. Weber recognised that factual and value assumptions are necessarily intertwined in all human activity, including science.

2.5 It is also sometimes assumed that Weber and other advocates of value neutrality treat value judgments as irrational leaps of subjective faith, or that they must take over the emotive theory of ethics, according to which value judgments are simply expressions of felt emotion (see MacIntyre 1981:24-5). Weber did believe that we necessarily commit ourselves to particular value principles in a non-rational manner, but even this is not essential to the concept of value neutrality.^[10]

Clarifying the nature of value neutrality

3.1 As already noted, part of the difficulty in resolving the issue of whether social inquiry should be normative or objective and value neutral lies in uncertainty and confusion over the meaning of key terms. There are at least three definitional matters that need attention here:

1. What do we mean by 'values'?
2. What do we mean by 'objective'?
3. What do we mean by 'neutral'?

3.2 There are two key senses given to the term 'values':

• As referring to principles that specify good/bad contrasts or polarities. And it is worth pointing out that, in the case of many of the practical values appealed to by those advocating normative sociology today, there are multiple versions of the same principle; so that, for example, justice/injustice can take retributive or distributive forms – concerned, respectively, with reward/punishment or with the equitable allocation of benefits and costs (see Foster et al. 1996: ch3).
• As referring to value-judgments, in other words to evaluations of some person, situation, course of action, etc. as good or bad, right or wrong, in some specific sense. Value judgments will usually draw on a range of value principles; and on various factual assumptions as well.

As already noted, a central pillar of Weber's argument is that we are faced with plural, and incommensurable, value principles. And that for this reason, and others, there are usually conflicting yet equally reasonable (as well as many unreasonable) value judgments about any object, action, situation, institution, etc.

3.3 That we face a plurality of potentially conflicting value principles has been widely recognised, although formulated in diverse ways (see, for instance, Berlin 1981 and 1990, Larmore 1987,Crowder 2002). Weber argued that how we choose among or assign priority to these principles cannot be decided by Reason. In this I think he was wrong, unless we adopt a very narrow interpretation of this term, limiting it to strict logical inference employing apodictic premises; in other words grounds that are beyond all possible doubt. Nevertheless, his position is salutary because it counters the common, but mistaken, tendency to believe that there is a single valid judgment to be made on most value issues, about which there ought to be consensus. Moreover, it is worth pointing out that Weber held that value judgments were open to rational scrutiny to some degree: as regards their derivation from principles, and also in terms of their 'realism' (Bruun 2007). And I believe he was correct about this, even if his model of rational derivation from foundational principles ought perhaps to be replaced by a web or network model (see Hammersley 2014a: ch 3).

3.4 So, one implication of Weber's position is that there is a set of epistemic value principles to which social scientists should be committed, and in terms of which they must make value judgments in carrying out their work. After truth itself, perhaps the most important of these is objectivity. Weber emphasises the fact that Western society involves institutional specialisation, and argues that it has become subject to a process of rationalisation that requires us to pursue any vocation single-mindedly, in other words objectively. At the same time, he insists that 'there is no absolutely "objective" scientific analysis of culture […] All knowledge of cultural reality… is always knowledge from specific and particular points of view' (Weber 1904 in Bruun and Whimster 2012: 113 and 119).

3.5 As this makes clear, the term 'objectivity' is open to divergent interpretations: there are at least five meanings commonly given to it. The first two apply to knowledge claims, the others to research strategies. So, 'objective' can mean:

1. Corresponding to the essential nature of the objects being studied.
2. Relating to features of the world that exist independently of our knowledge of them, and that are scientifically knowable.
3. Restricted to reliance upon evidence that is intersubjectively available.
4. Following procedural rules rather than making idiosyncratic decisions.
5. Taking proper account of what is relevant, and only of what is relevant.^[11]

3.6 The fifth of these senses is the one that is most relevant here. Its opposite is bias. And it is worth noting that this latter concept has general application, referring to distortion of the pursuit of any goal as a result of the influence of values or interests that are not relevant to it. Examples would include bias in job selection or in legal trials. In the case of empirical research, 'bias' refers to deviation from the most effective path towards finding true answers to specific factual questions (see Hammersley 2000: ch6). This makes clear that what constitutes objectivity in pursuing an activity depends upon – in other words, is relative to – its goal(s). And, for Weber, the only legitimate goal for social research is the production of factual knowledge.

3.7 Turning, finally, to the meaning of 'value neutrality', it is important to reiterate that this does not imply that research should be neutral in relation to, or free from influence by, all values. As already noted, Weber insisted that social inquiry necessarily relies upon a commitment to epistemic values: to discovering the truth about social reality, rather than constructing fictions. The values that researchers ought to try to be neutral towards are practical, that is non-epistemic, values: those to do with what is good or bad, right or wrong, about the actions, practices, institutional arrangements, etc being studied. However, this does not mean that researchers must eliminate all their commitments to such practical values, only that their goal in doing research should not be to serve those values, and that they should try to minimise any negative effects (bias) these commitments may have on the effective pursuit of factual knowledge.^[12]

3.8 It must be emphasised that the implication of value neutrality is not that publication of the findings of a study will be politically neutral in its effects, in the sense of being equally favourable or unfavourable to all political or practical commitments. All that can be claimed is that the research was not designed to favour, or to damage, particular interests; and, given this, it is unlikely to do this systematically.

3.9 A number of important assumptions underpin this commitment to value neutrality. These include an insistence that:

• Factual conclusions are analytically distinct from value conclusions; even though value judgments necessarily involve factual assumptions, and to be worthwhile factual conclusions must usually be value-relevant (on which see below).
• Value conclusions cannot be derived solely from factual assumptions: one or more value principles must also serve as premises. So, while in pragmatic terms we may sometimes draw value conclusions from a set of facts, without making the value principles on which we are relying explicit, we are always open to potential challenge as regards the value assumptions involved.
• Social science can only authoritatively validate factual claims, not value conclusions. It can legitimately claim expertise, though not infallibility, about some kinds of factual matter. It cannot claim distinctive expertise as regards value-judgments (Hammersley 2014a: ch3).^[13]

3.10 Along these lines, Weber stresses the limits that properly apply to the contribution social research can make to politics, policy and practice; in other words, there are quite narrow constraints on what it can legitimately claim to produce. He believed that if social scientists put forward value conclusions as if these had been validated by their research they exceed their authority, thereby committing a form of scientism. Moreover, aiming to produce value conclusions increases the danger that researchers' value commitments, beliefs, and preferences will distort the factual conclusions they reach; though, of course, it does not make this inevitable, nor does adherence to the principle of value neutrality ensure the absence of bias. But variation in the degree of this threat is nevertheless important.

Value relevance

4.1 The other, equally important, component of Weber's argument, alongside value neutrality, is that social research must be value-relevant. Following neo-Kantian philosophy, particularly that of Rickert, Weber argued that, unlike natural science, social research focuses on describing and explaining individual phenomena, rather than on producing knowledge of universal laws. And the key point that follows from this is that these phenomena must be selected for investigation on the basis of practical (i.e. non-epistemic) values, and studied from within a particular value-relevance framework (Ciaffa 1998; Hammersley 2014a). In these terms, all social science accounts are perspectival, even though their validity is not socio-culturally relative.

4.2 There are two related reasons put forward for why value-relevance is necessary. One is that it is questionable whether types of social phenomena can be characterised on the basis of sets of essential characteristics that mark phenomena out as belonging to one category, representing a natural kind, rather than another. Instead, the criteria of category membership are constituted in and through the perceptions and actions of the people engaged in the practical activities being studied, and the value frameworks they adopt in the course of these. Moreover, Weber recognised the 'double hermeneutic' (Giddens 1976: 79,162) involved here: that the constitutive work required in identifying and differentiating social phenomena is not restricted to the people social scientists study but is also carried out by social scientists themselves.

4.3 The second reason for reliance upon a value-relevance framework is that the point in studying individual phenomena, for Weber, is to contribute to knowledge that can be used in gaining practical understanding of particular cases (situations, issues, people, etc), and can therefore inform action. Weber argues that even if there were underlying laws governing social phenomena, these would not serve this function because practical actors are interested in the distinctive character of particular events, situations, etc, not in the generic features that these phenomena may share with other instances of some abstract category. For this reason, too, practical – that is, non-epistemic – values must serve as a value-relevance framework for social research: identifying particular questions as important, and framing understanding of them in such a way as to ensure that answers have practical relevance from at least some point of view. These values may be ones to which the researcher is strongly committed personally, but this need not be the case.^[14]

4.4 It is important to emphasise that a value-relevance framework defines what is at issue, but is not partisan in the sense of favouring one view rather than another about that issue. Nor should the framework be employed to draw unconditional value conclusions – evaluations or recommendations – as part of the research, since this would breach the principle of value neutrality.^[15] As an illustration, levels of poverty or of social mobility in particular places at particular times, and reasons for changes in these, are factual issues that are value-relevant from a variety of political perspectives (though not from all). Contrasting views can be taken about them, and research findings are likely to be relevant to many of these views. Furthermore, while how we define 'poverty' or 'social mobility' can vary, depending partly upon value assumptions, the principle of value-neutrality requires us to explore the implications of different definitions for our findings, rather than to promote one of them as the only valid one, or as superior to all others.^[16]

Has social science ever practised value neutrality?

5.1 One of the arguments deployed against commitment to the principle of value neutrality is that it has never actually been practised, and indeed that it cannot be; that, in effect, it is an ideological smokescreen for the necessarily partisan character of all research (see, for example, Horkheimer 1972).^[17] It is true that while, in the middle part of the twentieth century, commitment to something like the principle of value neutrality was often proclaimed by social scientists, they did not always apply this commitment very rigorously (Smith 1994; Storing 1962). In other words, they frequently lapsed into putting forward evaluations and recommendations as deriving directly from their research. Moreover, even when they did not do this, they sometimes failed to make explicit the particular value-relevance framework on which they were relying, thereby tending to imply that it was the only possible, or the only legitimate, one.

5.2 There are at least two plausible reasons for this deviance from the ideal of value-neutrality:

1. The continual demands on social science for value conclusions, not just from outside but also on the part of social scientists themselves. Even Weber did not always avoid indicating his feelings about what he was describing. Evaluation is central to all human activities, and one of the main functions of social science is to provide factual knowledge that can be used in making practical value judgments. As a result, there is frequently a felt need immediately to go beyond producing knowledge and to supply evaluations that seem to derive from them.
2. While, in the mid-twentieth century, there was widespread appeal to the principle of value-neutrality, in practice researchers often seem to have assumed a value consensus, believing that, on this basis, value conclusions could be taken to follow directly from the facts. This was sometimes done by relying upon a misleading analogy with medicine where, often, there is considerable consensus about what would be beneficial and what detrimental.^[18] In other words, the argument was that social science could not determine ends, only means, but that a particular set of ends supplied by this general consensus could be taken for granted, the focus being on obstacles to achieving those ends and how these obstacles could best be overcome.^[19]

5.3 What these points indicate, however, is not that it is impossible to realise the ideal of value neutrality, or that it is undesirable to attempt it, but rather that it is often difficult to pursue effectively. Furthermore, while we can never know with absolute certainty whether the non-epistemic values to which we are committed have biased our findings, we can take precautions to minimise this. A genuine attempt to adhere to value neutrality will usually reduce the likelihood and degree of bias from practical value commitments even when it is not fully successful.

5.4 In summary, then, while from the 1960s onwards explicit commitment to value neutrality was increasingly abandoned, being dismissed as impossible or undesirable, this was largely done on the basis of misinterpretation and caricature. And false conclusions were also drawn from failure to live up to this principle. Equally important, little thought seems to have been given to how normative forms of social science could be justified.

Arguments for normative sociology

6.1 Any commitment to a normative approach needs either to be able to show how, on its own, research can justify particular normative conclusions and disqualify others, or to indicate how research studies investigating the same phenomena committed to conflicting normative ideals can be tolerated, and why they should be.^[20] While some social scientists have put forward arguments claiming to demonstrate that factual premises can lead directly to value conclusions, most of those advocating or adopting a normative approach today appear to treat value judgments as matters of personal commitment.^[21] Furthermore, while many social scientists now reject value neutrality and put value conclusions into their research reports, they rarely provide a clear indication of the value principles upon which they are relying: appeals are made to equality, justice, the common good, for instance, but usually without much specification of which interpretation of these contentious terms is being employed.

6.2 However, there have been some recent attempts to conceptualise and justify the pluralism that is likely to result if social scientists adopt a normative approach. Here, the problem identified is that researchers will come to be committed to competing sets of practical values, and perhaps also to communities, interest groups and/or political organisations that are in conflict with one another.

6.3 One rationale for such pluralism could, of course, be the contractual or commercial model, where researchers offer their services on the market, taking on and doing research that supports clients' value positions. Justification for this might be offered in terms of neo-classical economics. However, this has not been the model that most advocates of normative or partisan research have adopted. One of the main rationales proposed instead has been similar to that for publicly funded provision of lawyers. Thus, Shaver and Staines (1971: 177-9) write that 'Advocate social scientists could be trained, like lawyers, to take up a client's point of view and to defend it as convincingly as possible – gathering supportive evidence and challenging the sources and interpretation of evidence by their clients' opponents' (p. 177). Presumably, the argument here would be that through the process of adversarial discussion the true or best conclusions about policy options would be reached, a kind of legal parallel to the notion of the 'invisible hand' assumed by some economists.

6.4 More recently, Root (1993), Burawoy (2005) and Turner (2007b: 785) have made proposals along somewhat similar lines, but with interesting variations. For example, elaborating on Burawoy's notion of public sociology, Turner claims that this:

provides a novel justification for advocacy scholarship in sociology. Public sociology is intended to have political effects, but also to be funded by the politically neutral state. [Burawoy] argues that public sociology is institutionally neutral, but that committing to an organic relation with a social movement is legitimate as a matter of the sociologist's personal value choice. Although this produces side-taking sociology, by improving the case for particular standpoints it serves to improve democratic discussion generally, which is an appropriately neutral public aim.

Whereas Shaver and Staines envisaged a process in which researchers would be hired out as necessary, Turner, following Burawoy, seems to propose a scheme in which sociologists would choose whom they would (and would not) assist.

6.5 In my view, none of these rationales for partisanship is effective, either as regards the public role of social science or in ensuring the conditions necessary for its flourishing. Beginning with the first of these issues, the market solution will result in those who are able to pay obtaining social science representation, while those without funds will not be able to draw on this resource. Many social researchers would regard this as undesirable, and with good reason. And indeed it would undermine any fanciful hope that this process of contestation would select the best policies.

6.6 The proposal for state funding of representation by public sociologists avoids this problem, in principle at least, though there are still questions about which groups would and would not, should and should not, be provided with representation, how social scientists ought to be recruited to different 'sides', whether they would or should advocate for positions with which they disagree, etc.

6.7 Meanwhile, the Burawoy/Turner proposal reintroduces the problem of uneven representation, since social scientists' political stances are probably not representative of the wider society on many issues. Thus, while what is proposed may result in some counterbalancing of current inequalities in power and influence, there may well still be groups that are left without social science 'representation'. Furthermore, it cannot be taken-for-granted that support from partisan researchers for the various sides in a public debate will 'improve democratic discussion' (Turner 2007b: 785). This seems to assume that such discussion is best conducted on an adversarial basis, analogous to what happens in a court of law, or even worse as in much current public debate. But, even if the adversarial system works well when what is to be decided is a matter of 'guilt' or 'innocence', policy issues are not usually of this kind: they are more complex. Here, I suggest, there needs to be active exploration of different arguments, testing of relevant evidence in a judicious fashion, and so on, if deliberative democracy is to operate effectively. It seems to me that what is proposed would be unlikely to lead to this.^[22]

6.8 These proposals for partisan research are also likely to have serious consequences for social science itself:

1. It will have less internal coherence and lose any clear boundaries, merging into the activity not just of think tanks but also of pressure groups of various kinds. Moreover, this will reinforce public perceptions of it as biased, and undercut whatever intellectual authority it currently enjoys (thereby also weakening its capacity to play an effective public role).
2. Given a plurality of commitments on the part of social scientists, social science will become fragmented into warring factions committed to conflicting value principles. There is already evidence of this in some fields. In others it is only avoided because, in effect, an ideological consensus is enforced; as, for example, in the case of research on ethnic inequalities in education in some countries (see Hammersley 1995: ch 4).
3. Bias in pursuit of factual conclusions will be more likely to result because an important barrier to it has been removed: there is no longer the requirement to consider factual issues in their own terms. Instead, they will be analysed in the process of drawing evaluative or prescriptive conclusions, and judgments about them are likely to be shaped and protected by the value commitments involved. There may be little encouragement to assess whether distortion is taking place, indeed there may be active discouragement of this: in the context of politics, ideas frequently become little more than weapons (Mannheim 1936: 34-5). So too does methodological critique, this being applied selectively – much more rigorously to the arguments of opponents than to the assumptions on one's own side, a tendency that may be conscious or unconscious (Foster et al. 1996). Indeed, many ideas will be treated as 'too good to be false' (Hammersley 2011: ch 5), and therefore as beyond criticism; while others will be viewed as 'beyond the pale'.^[23]
4. As Weber argued, value conclusions will tend to be presented, and be treated, as if they had been validated by research, and this amounts to an abuse of the authority of science. Indeed, it undermines democracy.
5. Despite Turner's claim to the contrary, public funding becomes harder to justify, since it is particular pieces of research that are to be funded and these will be explicitly serving partial interests. Furthermore, even if it is agreed that an overall system of public funding of social science advocacy is desirable there could be still be disagreements about who should and should not be provided with public support.

It seems to me that these are serious and damaging problems, and they count strongly against adopting an adversarial or partisan model of research.

Conclusion

7.1 The principles of value neutrality and objectivity are widely rejected today, in name at least; and social scientists frequently present value conclusions as if these had been validated by research. Yet there is little clarity about the basis for these value conclusions, or about why they are believed to be legitimate. Most of the justificatory strategies employed in the past to validate such conclusions are now rarely adopted explicitly, with good reason; and new ones, such as that developed by critical realists, are unconvincing. Instead it is usually implied that researchers can and should do research on the basis of some notion of the common good or a purported consensus, or (more honestly) on the basis of their own or others' value commitments. I have argued in this paper that acting in these ways undermines the integrity of social science, and increases the danger of bias and scientism.

7.2 In my view, Weber provided us with a sound understanding of the proper roles that values can and should play in social research, centred on his concepts of value neutrality and value relevance. I have shown here that rejection of his position has often been based on misinterpretation and caricature, and on a failure to recognise the indefensible character of normative social science. A revival of commitment to value neutrality on the part of social scientists is overdue, even if unlikely.

---

Notes

¹ On the early period, see Smith 1994. For influential versions of the argument for value neutrality in the middle of the twentieth century, see Parsons 1965 and Bendix 1970.

² Referring to the United States in the first half of the twentieth century, Turner (2007a: 123) writes: 'the idea that sociology could have even gained a foothold in universities without the support of reformers, or by the rejection of them in favour of purity or an ideal of science, is illusory'. Once it had gained a position in the academy more detachment was possible, and indeed required in order to build the status of the discipline there. However, in recent times, the pressure has reversed, arguably universities have increasingly been transformed not just into teaching factories but also into 'knowledge production' facilities whose funding depends upon some demonstration of public 'contribution', however spurious the measures involved.

³ This is a point that will be discussed later, but see also Hammersley 2011. Some of the terms used to characterize other positions, such as 'perfectionism' (Root 1993) and 'partisanship' (Gouldner 1973), are subject to similar problems.

⁴ For a discussion of Gouldner's critique, and of his proposed solution to the problem of values, see Hammersley 2000: ch 4. Even some of those who defend the idea of value freedom manage to generate confusion through careless discussion. See, for example, Black 2013, and my comment on this: Hammersley 2014b.

⁵ I am not offering a new interpretation of Weber's writing on this topic, I rely heavily on the detailed analysis by Bruun (2007), which seems to have been widely accepted among Weber scholars, but does not appear to have had much influence more widely amongst sociologists. The three terms 'value-neutrality', 'value-freedom', and 'objectivity' often tend to be treated as synonyms. I shall present the principle of objectivity as a component of value neutrality, and largely abandon the term 'value-freedom' as particularly problematic, even though it was used by Weber.

⁶ For relatively sophisticated arguments to the effect that his basic orientation is positivist see Kronman 1983, and Doody's (1984) review of this book. Examining Weber's methodology as a whole, Runciman (1972:17 and passim) uses this label, though its meaning is defined in opposition to that of 'idealism'.

⁷ Elsewhere, I have argued that positivist ideas in philosophy still have considerable force and value: there is much to be learned from them (Hammersley 1995: ch 1). See also Cartwright et al 2008. And the same is true of those bodies of work frequently dismissed as positivist in sociology (see Halfpenny 1982).

⁸ On Mill, see Hutchison 1974. For Comte see Wernick 2001. On Durkheim see Watts Miller 1996. Twentieth-century positivism, in both philosophy and social science, generally adopted the principle of value neutrality in some form or another: in particular it denied that science can legitimately, or should aim to, produce value conclusions. See, for example, Lundberg 1947.

⁹ We must recognize too, of course, that some practical values should operate as ethical constraints on how research is carried out, see Hammersley and Traianou 2012.

¹⁰ Even so, his position is different from emotivism, which treats value judgments as expressions of attitude rather than as propositional statements, so that argument is restricted to the factual assumptions they involve (Ayer 1936: 110), and from most kinds of subjectivism, where value judgments are statements about individual preferences (see Frondizi 1971: ch 3). For further discussion of these common criticisms of Weber's position, see Hammersley 2014a: ch 3.

¹¹ There is, of course, room for uncertainty and dispute about what is relevant to any particular task.

¹² This leaves the way open for recognition that practical value commitments can and often do make a positive contribution to research, and for this reason Weber emphasized that research communities should tolerate researchers with diverse political commitments.

¹³ Weber does also identify a need to check that value judgments have been consistently derived from the relevant set of values; though it is not clear that social researchers have any distinctive expertise in doing this. For a discussion of the challenging problems surrounding the concept of expertise as it relates to values, see Rasmussen 2005.

¹⁴ A criticism that has sometimes been made of the principle of value neutrality is that it has frequently been used by 'service intellectuals' (Smith 1994: 27) to legitimate carrying out research that serves governments or commercial organizations. However, what is open to criticism here is not adherence to value neutrality but the tendency for these researchers to adopt some value-relevance frameworks and to neglect others. It is the case, though, that the issue of how value-relevance frameworks should be selected is one that Weber did not address very effectively, and that it is not easy to resolve in practical terms.

¹⁵ Conditional value conclusions are legitimate, taking the form: if we adopt this set of values then, given the factual knowledge available, this is likely to be the evaluation or recommendation reached. However, their conditional character needs to be underlined, along with the scope for drawing a variety of value conclusions from the same set of facts.

¹⁶ For an exploration of the implications of this for the study of social mobility, see Hammersley 2014a: ch 5.

¹⁷ Even in the mid-twentieth century, there were many social scientists who did not accept the principle of value neutrality, and some who challenged it explicitly. It is worth underlining that criticism of this principle came from across the whole political spectrum: from Marxists and advocates of Critical Theory and from Nazi social scientists (Derman 2012). Indeed, there has been sustained criticism from various positions on the Right (see Strauss 1953; Midgely 1983), as well as from the Left.

¹⁸ However, the extent of value consensus even in relation to medical matters is easily exaggerated. Indeed, there are some fields, such as mental health, where there is as little consensus as in social policy issues.

¹⁹ Weber sometimes formulates a commitment to value neutrality on the part of the social scientist as implying a focus on the identification of means, rather than the determination of ends. However, in my view this is misleading. Factual considerations are usually involved, to one degree or another, in setting goals; and value considerations ought to frame the selection of means, for example as regards what is and is not ethical.

²⁰ It also needs to draw distinctions among types of normative activity: from open-minded attempts to reach justifiable value judgments, through social and political commentary, to advocacy and then, at the far end of the spectrum, propagandizing. Also relevant is the role that the concept of objectivity can play in relation to these.

²¹ A recent example of an attempt to justify deriving normative conclusions solely from factual premises is to be found in Bhaskar's critical realism, on which see Hammersley 2014a: ch 4. Of course, in the past Marxism provided a rationale for reaching value conclusions on the basis of facts about society (Wood 2004), but this depended upon a teleological conception of societal development that has come to be widely rejected, along with other 'meta-narratives'. Interestingly, much the same approach was taken by Durkheim: see Watts Miller 1996.

²² For discussion of the requirements of deliberative democracy, and of the 'public use of reason' more generally, see Bohman and Rehg 1997, Gutmann and Thompson 2002.

²³ As already noted, the argument I am putting forward here is not that research by those committed to value-neutrality is automatically free from bias, whereas that of those engaged in 'value-committed' research always reaches false conclusions. The point is simply that research aimed at producing value conclusions is more likely to suffer from bias as regards the pursuit of factual knowledge, and this will be true both when a research community shares much the same value commitments and when it contains warring factions committed to quite different values. The position that Turner and Burawoy take is complex because public sociology is only one of four complementary forms of the discipline, the others being 'professional', 'critical', and 'policy' sociology. I cannot argue the point here but in my view the other three forms erode professional sociology's capacity to resist powerful sources of bias. Moreover, in many fields, in practice, it is very difficult to distinguish amongst these different types of sociological work.

---

References