1 Introduction¹

One way in which social thinkers have grappled with questions of justification, evaluation and political orientation in social research is with reference to ideas of objectivity and subjectivity. These concepts, and the debates around them, are long-standing in the field of sociology with important contributions being made by founding figures such as Emile Durkheim ([1895]1964) and Max Weber (1949). Nevertheless, sociologists continue to debate whether objectivity is an appropriate goal (see, e.g., Romero 2020; Williams 2015; Hammersley 2011). One of the difficult issues raised in debates about objectivity is what the words objective and subjective actually mean, and it is often noted that these terms have been used in a range of different, sometimes incompatible, ways (see, e.g., Megill 1994). In this article, I focus on critically evaluating one particular conception of objectivity: what I call the ‘object-focused’ conception, which associates objectivity with an accurate grasp of the object of analysis. I do so by considering the way that this conception is used in critical social science, exploring Michael Burawoy’s use of the objective/subjective divide in his analysis of domination (Burawoy 1979, 1998, 2012a, 2017).

Although the focus here is on objectivity and critical social science, this is not the orientation most commonly linked to the idea of objectivity. Indeed, probably the most common usage over the course of sociology’s history is by thinkers who reject politically committed, critical orientations to social science, and associate objectivity with value-freedom. I would like, therefore, to briefly review this approach to draw out the contrast with usages in critical social science. For those who associate objectivity with value-freedom, an objective social scientist is one who appropriately restricts the influence of subjective values on research, whereas a social scientist who fails to be objective allows subjective values to shape or, more accurately from this point of view, mis-shape research findings. Max Weber is the most famous sociological defender of a notion of value-freedom, although, as is commonly noted, Weber does not exclude values entirely from the process of research. Rather, he argues that values will shape the topics that social scientists investigate and the questions that they ask, but need not (and should not) determine the answers to those questions (for discussion, see Hammersley 2000). While there are undoubtedly variations in the positions of those who link objectivity to value-freedom, one of the key over-arching beliefs is that the influence of values on research findings (as opposed to questions) is problematic because values are not subject to empirical and rational assessment in the same way as claims about facts. For these writers it is this feature that makes values subjective; they are not in the realm of the objective, the realm of that which is empirically verifiable and/or rationally arguable. Thus, for some, there is a need to uphold the objective/subjective division to counter the risk of factual claims being influenced by non-rational elements and, at the extreme, knowledge claims becoming merely projections of subjectively based political views rather than having genuine empirical and rational justification.

The value-free conception of objectivity has been subjected to serious critical scrutiny, with many thinkers who advocate critical orientations to social science rejecting it as an ideal. Some have done so by arguing that value-free objectivity is not possible—that values always impinge upon investigation and cannot be meaningfully excluded from inquiry, only concealed (e.g., Lather 1992). Others have made the related, but not identical, argument that certain kinds of value-commitments may improve accounts within a particular domain, rather than undermine them, by opening up alternative or more comprehensive avenues of inquiry (see, e.g., Wylie and Nelson 2007; Harding 1991). Thus, both the possibility and benefits of value-free objectivity have been questioned.

Nevertheless, arguments against value-freedom do not necessarily undermine other views of objectivity based on quite different grounds. The key alternative we are concerned with here is an ‘object-focused’ conception of objectivity. This is based on the idea that an objective analysis gets in touch with, and reflects, the characteristics and features of the object of investigation, whereas a subjective analysis is one that reflects the assumptions and/or imaginings of the knowing subject. Aspects of this approach to objectivity can be found in Durkheim’s The Rules of Sociological Method ([1895] 1964, 31–36), but the contemporary approach to social theory most closely associated with it is critical realism (e.g., Collier 2003). For critical realist thinkers like Collier, an objective claim is one which is oriented to the properties of objects themselves and has been generated by a responsiveness to those properties. By contrast, a subjective claim is one which has been generated out of the knowing subject and their assumptions, biases and wishful thinking; the limitations of their historical context and so on (Collier 2003, 134–137). As proponents of this view believe that what is required for the knowing subject to be successful is openness to the properties of the object, the effects of subjective bias and limitations are problematic because they obstruct the accurate apprehension of the object. However, for at least some exponents of this approach, social science itself need not be ‘value-free’ to achieve object-focused objectivity (cf. Sayer 2000, 58–59). Further, this approach does not entail a view about which knowers are best placed to generate objective claims. But amongst proponents of this view, it is not unusual for scientists to be seen as knowers who can step outside of their historical context, and their subjective blinkers, in order to generate objective claims, both in the natural sciences (e.g., Collier 2003, 219) and the social sciences (e.g., Durkheim [1895] 1964, 31–36).²

Although not entailed in subscribing to the object-focused approach to objectivity, the approach undoubtedly has an affinity with a certain kind of critical orientation to social science. The object-focused approach opens up a gap between subjective and objective accounts, and this gap can be utilised by social scientists who wish to offer a robust objective critique of the subjective viewpoints and understandings of lay members of society. In this mode, lay actors’ subjective perspectives are understood to be potentially compromised by oppressive social conditions or other everyday influences, and objectivity should be strived for by the social scientist as a means to step outside of those conditions and achieve a more convincing account of social life that is critical of those oppressive conditions. Two prominent thinkers who have defended this approach to objectivity and subjectivity are Steven Lukes and Michael Burawoy. In the second edition of Power: A Radical View (2005), Lukes argues that the subjective perceptions that actors have of their interests are potentially compromised by forms of domination. For Lukes, the critical social scientist can legitimately move beyond these perceptions and try to formulate objective views grounded elsewhere, for example in a theory of human health and welfare which can be used to assess what is, and is not, good for human beings (see Lukes 2005, 82). I have evaluated Lukes’ approach elsewhere (see Kemp 2012), arguing that although his use of the objective/subjective distinction is somewhat tentative, it runs the risk of simply assuming, rather than demonstrating, the superiority of the social scientific account of lay actors’ interests over the views of lay actors themselves. In this article, I explore the critical social science perspective by focusing on the arguments of Burawoy, extending this argument to evaluate his bolder defence of a hierarchical distinction between objective and subjective perspectives.

Burawoy is known both for his defence of public sociology, issued during his time as president of the American Sociological Association (Burawoy 2005), and for his theoretically minded ethnographies of the manufacturing process. A key example of the latter is his analysis of the exploitation of workers in Manufacturing Consent (1979), which arose out of a ten-month period of fieldwork in which he worked on the shop floor of an industrial plant in Chicago during the mid-1970s. In this discussion, I will be largely focusing on Burawoy’s 2012 Sociology article that reflects on this classic work. As implied by its title, Burawoy’s article ‘The Roots of Domination: Beyond Bourdieu and Gramsci’ is a theoretically engaged analysis of domination in the workplace.³ As part of a critical dialogue with Bourdieu and Gramsci, Burawoy demonstrates his use of the objective/subjective distinction to set up and justify his critical perspective on labour processes. Burawoy’s analysis is important for the aforementioned debates on the relation between objectivity and subjectivity in the production of social scientific knowledge because he gives us a very clear account of how the objective/subjective distinction operates in practice and argues that objectivity confers authority on the sociological perspective.

The article proceeds as follows. In Section 2, I outline Burawoy’s approach to domination, moving, in Section 3, to critically evaluate his use of the objective/subjective distinction. I agree with Burawoy that it is legitimate for the accounts of social scientists and lay actors to diverge, and, thus, it is legitimate for a social scientist to have a critical perspective on lay understandings. However, I argue that he prejudges the outcome of this disagreement by associating social science with an objective analysis of social relations and lay perspectives with the (strong) potential for flawed subjectivity. As an alternative, I suggest that such disagreements should be investigated in a dialogue between social scientists and lay actors that does not prejudge whose account is more valid but is oriented towards evaluating these potentially different understandings of the social world. I then move on, in Section 4, to consider whether other writings by Burawoy, which address similar issues, offer an alternative, preferable perspective on these issues. In the penultimate section of the article, I acknowledge that productive dialogue is not a simple matter and consider some of the challenges faced by participants in dialogue.

2 Burawoy: Objective and Subjective Views of the Labour Process

One of Burawoy’s central concerns, in both 1979 and 2012, is to explain why workers within capitalism do not accept a Marxist account of the labour process. More specifically, he is interested in why what he sees as the domination of the labourer, through the expropriation of surplus value, is not seen as such by labourers themselves, with the consequence that these labourers do not challenge capitalism to stop this form of domination. In order to both explain the discrepant points of view and justify his own understanding, Burawoy differentiates the ‘objective’ and the ‘subjective’ truths of the situation and explicitly associates them with the different perspectives of the researcher and the socially located participant. Burawoy refers to the ‘objective truth of the sociologist’, a truth that is generated from the perspective of ‘the outsider studying the game’ (Burawoy 2012a, 189). By contrast, it is the participant in the workforce who experiences the subjective truth ‘of the insider playing the game’ (Burawoy 2012a, 189). The idea of multiple ‘truths’ is often associated with relativist modes of analysis, in which the beliefs based on each perspective are adequate on their own terms and cannot be shown to be problematic from another perspective. However, Burawoy’s use of the subjective/objective division signals that he is not treating these ‘truths’ as equally adequate. This is apparent in the way in which Burawoy analyses competing understandings of labour and exploitation. For Burawoy, there can be a ‘fundamental gap between the objective and the subjective truth of labour’, one which is ‘achieved through mystification rooted in the social relations into which men and women enter’ (Burawoy 2012a, 198). This mystification is a ‘social process that produces the gap between experience and reality for all who enter a specific set of social relations’ (Burawoy 2012a, 191, his emphasis). Thus, the gap between the objective and the subjective, between the sociologist’s view and the participant’s view, is explained by reference to the mystification of the latter. Burawoy does attend to the subjective truth of participants, but this subjective truth is only deemed to be adequate when it converges with the objective view of the observer. Furthermore, what interests Burawoy about cases where subjective consciousness and the objective perspective of the observer diverge is the way in which subjectively based misunderstandings help to reproduce objective exploitation by mystifying it.

Burawoy is particularly interested in the way in which capitalist relations of production are reproduced, despite what he analyses as the exploitation of workers within capitalism. For Burawoy, the objective situation of the labourer is that identified by Marx. Workers in capitalism are exploited in that they are only paid for some of their labour time, the unpaid work creating the value that results in profit (Burawoy 2012a, 188). However, many workers do not see themselves as exploited in the way that Marxist theory suggests, including those at the Chicago plant. Thus, Burawoy is interested in why these workers ‘consent’ to capitalist relations and the ways in which capitalism has generated such consent, and he discusses three aspects of this (Burawoy 2012a, 192–198). The first of these is the way in which factory organization produced ‘the concrete coordination of the interests of capital and labour through collective bargaining’ (Burawoy 2012a, 193). This involved some concessions being granted to workers that helped to gain their consent without altering the fundamental structure of capitalist relations. The second of these was the ‘internal labour market’ which gave those workers higher up the hierarchy a reason to stay at a particular firm because their seniority would not be recognized at another firm (Burawoy 2012a, 193). But it is the third factor which Burawoy discusses most extensively in his 2012 article, ‘the constitution of work as a game’ (Burawoy 2012a, 193, his emphasis), so let us consider that further here.

Burawoy’s treatment of work as a game invokes the separation of the perspective of participants and the perspective of the sociological analyst that I have discussed already. Participants were engaging in a game of ‘making out’. This game related to the quantity of piecework completed, and workers ‘made out’ and earned bonuses by producing a percentage of output that ‘was not higher than 140 percent and not lower than 125 percent’ of a target amount specified by the company (Burawoy 2012a, 193). From the perspective of the worker, the game countered the boredom of repetitive work and provided a feeling of satisfaction when they achieved the required percentages. Workers who made out felt that they were getting one over on the management, both by limiting their possible output and gaining ‘small monetary concessions’ (Burawoy 2012a, 195). From Burawoy’s perspective as an observer, the game looks rather different. For Burawoy, those who willingly participate in a process that allows the extraction of surplus labour for profit are misunderstanding the conditions of their production (Burawoy 2012a, 195). Despite workers believing that they were outwitting the management, the game of making out encouraged workers to apply themselves to the job, such that management were able to gain profit from their labour (Burawoy 2012a, 193). The game produced ‘symbolic rewards and emotional satisfaction’, but this enchantment simply masked the alienating character of the work (Burawoy 2012a, 198, 193). Capturing the difference in perspectives on this game, Burawoy states that ‘[t]o the outsider “making out” appears as absurd, to the insider it is what gives meaning to life’ (Burawoy 2012a, 195).

Burawoy is not saying that participants themselves are ‘absurd’ for getting caught up in the making out game. The game was socially sanctioned such that workers were subject to the evaluation of others and could be ostracized for not participating in it. Burawoy also acknowledges that, as a worker, he found it absorbing, and this is a key point about the social relations of labour which, Burawoy argues, produce ‘mystification’ insofar as participation in the game makes it hard to see it for what it is. Firstly, it encourages workers to focus attention on the game rather than on ‘the conditions of its existence’, that is, the mode of production in which capital and labour are opposed to one another (Burawoy 2012a, 195). Secondly, playing the game makes it hard to challenge its rules. Burawoy states that ‘one cannot play the game of “making out” on the shop floor and at the same time question its rules—rules that are socially sanctioned by workers and shop floor management alike’ (Burawoy 2012a, 189).

Thus, for Burawoy, the social relations of production mystify the exploitation that they involve, such that the subjective experiences of the workers do not match the objective analysis of the sociological observer.

In setting out his analysis of domination, Burawoy does not specify precisely how he is using the terms objectivity and subjectivity. Furthermore, as I shall discuss later, his usage has varied a little over time (Burawoy 1998). Nevertheless, paying attention to Burawoy’s wider theoretical commitments, his usage of the terms in the 2012 article makes his view fairly clear. Beginning from his theoretical commitments, we can note Burawoy’s association with Marxist ideas and his (not uncritical) use of Gramsci’s ideas. Given this association, it is highly unlikely that Burawoy intends the social scientist’s objectivity to involve a commitment to ‘value-freedom’. Although there is an issue within Marxism about whether it is best described as oriented to ‘values’ rather than concerned with the real movement of society, there is undoubtedly an evaluation of capitalism present within Marxist analysis. Relatedly, Burawoy’s critical analyses, which use terms like exploitation and domination, are clearly infused with values.⁴

A more plausible reading of Burawoy’s article is that he sees an objective perspective as one that captures the real features of the research area (object): in this case, the social relations of labour. As we have seen, Burawoy makes a distinction between the ‘experience’ of a set of social relations and their ‘reality’, the experience potentially being mystified (Burawoy 2012a, 191). It is clear that, when analysing labour, Burawoy associates the ‘objective truth of capitalist work’ with the real relations in which surplus value is extracted, as compared to the ‘subjective truth’, which is the potentially mistaken beliefs that actors have. Such beliefs are seen as being produced by the obscuring features of the actors’ social situation, which helps the objective situation to be perpetuated (Burawoy 2012a, 188). Thus, I would argue that, in his more recent work on domination in the labour process, Burawoy’s conception of objectivity is object-focused, such that the objective truth of the situation is understood to be a correct characterisation of the object of analysis. It is apparent that, for Burawoy, the objective/subjective distinction allows him to legitimise having a different view of the social relations that he is studying from that of the participants. Furthermore, his account of the subjective participant experience allows him to explain why it is different to that of the sociologist—because the social relations the participant is located within produce a mystified sense of those relations.

3 Evaluating Burawoy’s Approach

Having explained key aspects of Burawoy’s treatment of objectivity and subjectivity, I would now like to turn to evaluating them. One key issue is as follows: is Burawoy justified in associating the correct characterisation of social relations with the sociologist’s perspective? Undoubtedly, we can reasonably hope that those sociologists who study particular social relations have something insightful to say about them. These sociologists are likely to have studied the social relations carefully, generating evidence about them through procedures that have benefitted from methodological reflection, as well as drawing on theoretical tools that have been sharpened through social scientific debate. Burawoy’s own study Manufacturing Consent is a good example of this. But to acknowledge this is not to say that the sociologist, therefore, has an objective perspective which grasps the real characteristics of the social relations in question. This is an unjustified leap.

For one thing, we are rarely dealing with a sociological perspective so much as the perspectives of sociologists. Sociologists can, and often do, dispute with one another about the character of social relations and other features of the social world. Burawoy hints at this himself when he characterizes his activity after conducting the Chicago ethnography as seeking to ‘persuade my fellow academics of the superiority of my theory of the labour process and of manufacturing consent’ (Burawoy 2012a, 202). Burawoy’s book was undoubtedly valuable and influential, but it would be inaccurate to say that every sociologist dealing with labour relations has been persuaded to take the view that surplus labour is extracted in the way that Burawoy describes. Even in a symposium about Manufacturing Consent in Contemporary Sociology, in which most of the contributors are former students of Burawoy (Gottfried 2001, 435), there is meaningful disagreement about his analysis. For example, Michael Freeland criticizes Burawoy for failing to recognize the role of ‘non-rational’ commitments that originate outside the factory but contribute to sustaining relations in the workplace (Freeland 2001, 447–448). Setting aside the question of whether Freeland’s approach is more convincing than Burawoy’s, the key point here is that there is not a single consensual sociological approach to analysing the social relations in the Chicago plant. And, of course, Burawoy is himself disagreeing with other sociologists, as in his dispute with Bourdieu about whether institutional forms of internalized dispositions explain consent to the labour process, and indeed about whether the habitus can be said to exist at all (Burawoy 2012a). The existence of competing sociological views would suggest that we cannot simply consider the sociologist’s perspective to capture the truth about social reality, unless we are willing to admit the possibility of multiple truths.

Another way to explore the limitations of Burawoy’s approach to objectivity and subjectivity is to question the idea that a sociologist’s perspective is intrinsically superior to that of participants in social relations. When Burawoy analyses domination, he argues that a sociological perspective is objective, insofar as it correctly captures the features of social relations, whereas the subjective actors’ perspective is either in line with this sociological perspective and, thus, adequate, or out of line and, thus, at least confused, and possibly mystified. But is it not possible that, at least in some cases, the participating actor has a better understanding than the sociologist? If we admit this possibility, it is surely questionable to use a conceptual division that intrinsically associates a sociologist’s perspective with reality and, thus, sees any divergence from this on the part of participating subjects as a failure to grasp reality (Holmwood 2014). Rather, we should admit that there are competing accounts of the social relations in question, giving rise to issues of the relative validity of these accounts. Sociologists should, and many do, avoid assuming from the beginning that a sociological account accurately represents reality, and rather see its validity as subject to assessment and consideration. As noted above, we would hope that sociologists have insights to bring to a discussion about social relations that they have systematically studied. But participants in those social relations may also have insights, partly because they have lived those social relations and this may be a source of understanding, as many feminist thinkers have argued (see, e.g., Harding 1991). Alongside—and, indeed, surely as part of—these lived experiences, actors may gain insights about their world from culturally available resources, from education to cultural commentary that appears in the media, to insights from documentaries, novels, films and so on. The vast majority of actors live mental lives that are not simply determined by their immediate context. Rather, how they think about the social world is influenced by a wide range of cultural elements—which are part of their lived experience insofar as they are reflected on within their experience.⁵

Many sociologists do acknowledge that participants have insights that need to be attended to. And much sociological and social theoretical writing has also reflected on the potential limitations of a sociologist’s/analyst’s account (e.g., Haraway 1988; Megill 1994). One way of doing so is to extend Burawoy’s point about the situatedness of actors in social relations and apply it to sociologists as well. For Burawoy, individuals act and think within social relations, and, following Marx, Burawoy argues that these social relations can impact on how the social world is experienced (Burawoy 2012a, 190). Again, drawing on feminist arguments, this point should surely prompt sociologists to consider how their own socio-cultural backgrounds and positions in social relations shape what they see (see Haraway 1988). After all, a sociologist is a situated individual, involved in various kinds of social relations and potentially ‘games’, both within the academy and in wider social life. Thus, feminists have argued that occupying dominant positions in society can make it harder to see and understand aspects of the lives of disadvantaged groups (Smith 1987; Harding 1991). This point can be linked to a criticism of Manufacturing Consent which argues that Burawoy did not attend enough to the gendered character of social relations in the plant that he studied, nor acknowledge that it is a site of the ‘production of masculinity’ (Leidner 2001, 441). One of the implications of a sociologist’s positioning in social relations is a caution against over-confidence about a sociological perspective or the status of sociological accounts. In making this point, I am not arguing that sociologists and other actors are trapped in perspectives determined by their backgrounds. If this was so, criticism of the limitations of a particular perspective would be pointless, as the actor in question could not change what s/he was seeing. Rather, it is to argue that reflection on a perspective can help to explain why an actor saw certain things, but also help them to a better way of seeing.

It is perhaps worth pointing out here that aspects of Burawoy’s specific account of how social relations mystified workers at the manufacturing plant, particularly those relating to ‘making out’, are rather unconvincing. Undoubtedly, Burawoy’s account of the activity of making out and the commitments it involves is an exemplary piece of ethnographic work. But his wider theory of the significance of making out seems problematic. For Burawoy, participating in a game is crucial to the concealment of important aspects of social relations:

Games obscure the conditions of their own playing through the very process of securing participation. Just as one cannot play chess and at the same time question its rules, so one cannot play the game of “making out” on the shop floor and at the same time question its rules—rules that are socially sanctioned by workers and shop floor management alike.

Burawoy 2012a, 189

Here, Burawoy is explaining the inability of workers to see their domination and their disinterest in removing it partly by reference to the blinkers that the game of making out imposes. Certainly, it is apparent from Burawoy’s account that playing the game helps workers to pass the time, challenge themselves and claim status. But the manufacturing plant is not a ‘total institution’, and workers are outside of the game for substantial portions of the day. Thus, although Burawoy reports finding himself very focused on the game during working hours, he is able to think critically about it outside of those hours. Similarly, those who play football a great deal can still believe that the off-side rule should be changed, and campaign for this to happen outside of time on the pitch. I would argue that Burawoy overstates the significance of the absorption of workers within the game of ‘making out’ for their understanding of social relations in the plant and society. An overemphasis on mystification risks failing to take participants’ understandings seriously as an alternative perspective on the social situation to one’s own.

As Burawoy assigns a special privilege to a sociological observer’s perspective, it is perhaps surprising to find that he implicitly criticizes Bourdieu for adopting an elitist conception of sociologists as privileged knowers.⁶ For example, Burawoy writes that for Bourdieu, ‘We are all fish in water unable to comprehend the environment in which we swim—except, of course, Bourdieu and his fellow sociologists.’ (Burawoy 2012a, 192). Burawoy takes Bourdieu to be arguing that all social orders necessarily produce a misrecognition of their hierarchies and forms of domination through the dovetailing of habitus and the wider social world. Burawoy sees his position as different to Bourdieu’s, insofar as Burawoy argues that in some social orders, such as that of state socialism in Hungary, the domination of workers may be apparent to them within their subjective experience (Burawoy 2012a, 198–200). Certainly, Burawoy’s position is less elitist than this strand of Bourdieu’s thought.⁷ But Burawoy does not seem to see that his own approach retains a degree of elitism, insofar as it implies that the validity of a participant’s perspective is evaluated against the sociologist’s position. Where accounts diverge, Burawoy gives no indication that this disagreement is to be taken seriously, or that it should spark reflection about whether or not the social scientist has the better account.

It is also worth pointing out that the social positioning of the knower is not the only limiting factor when it comes to their understanding of social relations. Simply put, producing knowledge is a difficult task. The subject matter does not present itself straightforwardly or directly. Our knowledge of one aspect of the social world is intertwined with our knowledge of other aspects and is part of a complex ensemble of understandings and practices (for discussions of the intertwined aspects of elements of knowledge see Duhem [1914] 1962 and Kuhn 1970). For this reason, knowledge claims should always be considered provisional, and further investigation may require them to be importantly modified or overturned. As such, it is illegitimate to characterise any particular account as one which captures reality and states the truth about it. Clearly, we should reject the idea that a sociological account captures the objective truth of social reality. But this also means that, as an alternative to Burawoy’s privileging of the social scientist, what is required is not so much a decision about whether it is the sociologist or the participant who has correctly grasped reality in any particular case, but an investigation of the comparative merits of accounts in order to see which is currently the most convincing.

4 Do Burawoy’s Other Writings Offer an Alternative Perspective?

So far, I have been focusing largely on the ideas Burawoy expresses in his article ‘The Roots of Domination: Beyond Bourdieu and Gramsci’ and their relationship to one of his classic earlier ethnographies, Manufacturing Consent. What I want to consider now is whether two of Burawoy’s other articles, one preceding and one post-dating the Bourdieu/Gramsci piece, could be taken as offering an alternative perspective.

The first of these is ‘Critical Sociology: A Dialogue Between Two Sciences’ (1998), in which Burawoy argues that ‘positive science’ is not the only appropriate model for social science, contending that another model is also of value, which he calls ‘reflexive science’. Whereas positive science attempts to produce detached knowledge of the social world, reflexive science acknowledges its locatedness as a starting point for knowledge. There are various other differences between the models, including contrasting views of objectivity. Burawoy argues that positive science believes in ‘procedural objectivity’, where following the right procedures ‘leads to ever more accurate representations of the world’ (Burawoy 1998, 14). By contrast, reflexive science is committed to ‘embedded objectivity’, in which the requirement placed on researchers is to recognize their biases and commitments, whilst also recognizing the intersubjective character of knowledge production (Burawoy 1998, 14). As an example of this in action, Burawoy discusses the reflexive interview, in which dialogue is important and the research method can be empowering for the subjects being interviewed (Burawoy 1998, 16).

As such, we can see that Burawoy’s earlier article outlines a concept of objectivity which does not seem to invoke a hierarchical division between the objective social scientist and the subjective lay actor. However, we can also see that there are seeds of Burawoy’s later approach in this earlier discussion. For one thing, Burawoy is not arguing for ‘reflexive’ social science instead of ‘positive’ social science in this article, but for sociology to embrace both (Burawoy 1998, 19). As such, it is not at all clear that he sees the embedded conception of objectivity as superseding the procedural one and its representational orientation. For another thing, Burawoy already expresses a doubt in this earlier article about doing away with the objective/subjective division. For example, when discussing critics of science, he states:

The denunciation of social science as “objectivism” (the standpoint of the outsider) turns into the embrace of “subjectivism” (the standpoint of the participant). Instead of replacing one one-sided standpoint with another, I propose an alternative model of social science that takes context as point of departure but not point of conclusion.

Burawoy 1998, 13

Although different interpretations of this remark are possible, in my view it is plausible to see Burawoy already moving towards his later viewpoint here. As ‘context’ is being associated with the understandings of lay actors, Burawoy appears to be advocating a social science that starts from an analysis of the “subjective” perspectives of lay actors but goes beyond this to also incorporate the “objective” standpoint of the social scientific observer. This raises the issue of the relation between the two standpoints, and it is this which is resolved, in his 2012 treatment of domination, in favour of the non-dialogical division between social scientific objectivity and lay subjectivity. Although Burawoy’s earlier ideas seem to have dialogical potential, they are later developed in a direction which undermines these possibilities.

The other, later, article of Burawoy’s that I want to discuss is entitled ‘On Desmond: The Limits of Spontaneous Sociology’ (2017). A little bit of explanation is required to situate this piece, but it is worthwhile doing so, because in this piece Burawoy usefully expands on his views about the relationship between lay actor and sociological accounts. Burawoy’s article is a response to Mathew Desmond’s ‘Relational Ethnography’ (2014), and these impressive exponents of ethnography disagree about the theoretical basis and framing of ethnography. Desmond argues for the value of relational ethnography over substantialist ethnography, his discussion strongly implying that Burawoy’s studies fit into the latter category. According to Desmond, substantialist ethnography tends to isolate its object and treat a place or group as a meaningfully distinct object in itself; by contrast, relational ethnography focuses on the way in which what might seem to be a distinct object is crucially shaped by its relations with others, with, for example, group identity being crucially shaped by inter-group interactions. Burawoy counters that both he and Desmond conduct relational ethnographies, the difference is that his own are structural in character, whereas Desmond’s are transactional. That is to say, whereas Burawoy insists on putting the case being studied into a wider macro/structural context, Desmond tends to assume that the truth of the matter resides within the case/locale being studied.

It is this latter distinction that takes us closer to the concerns of this article, because Burawoy argues that the division between transactional and structural approaches connects to different orientations to the analysis of the relationship between lay/participant and sociological/researcher accounts. Although Burawoy makes only limited use of the subjective/objective division in this article, nevertheless there is a clear link to the issues we have been discussing, insofar as Burawoy is assessing the value of lay/participant accounts compared to sociological/researcher accounts. This emerges through Burawoy’s argument that, despite Desmond explicitly taking inspiration from Bourdieu’s claim that sociological researchers need to break with lay understandings, Desmond fails to do this in crucial respects. This is important to our considerations because, in criticising Desmond for failing to break with lay/participant perspectives, Burawoy indicates his ongoing support for doing so.

One instance of this is to be found when Burawoy criticises Desmond for failing to ‘break with common sense’ (Burawoy 2017, 264). Although Burawoy suggests that the common sense one might wish to break from can be found in the dominant ideology, in the media and even in a discipline like sociology, his criticism of Desmond is that the latter does not break with the ‘common sense’ of lay actors/participants (see, e.g., Burawoy 2017, 266–268). According to Burawoy, Desmond tends to rely on the ‘folk theory’ of participants, with the result being that he does not develop an appropriately ‘critical’ sociological perspective (Burawoy 2017, 264) and often fails to get to grips with ‘wider institutional forces’ (Burawoy 2017, 268). Alongside this, Burawoy sees Desmond as staying too close to the ‘lived experiences’ of participants (Burawoy 2017, 277), decisively limiting Desmond’s analysis. What is needed instead, suggests Burawoy, is to use ‘comparative analysis or theoretical labor’ to ‘go beyond lived realities to their determinations’ (Burawoy 2017, 281). Similarly, Burawoy suggests that a comparative and theoretically minded approach is needed to ‘move beyond common sense to science’ (Burawoy 2017, pp. 272–273).

In my view, these arguments reflect Burawoy’s ongoing commitment to the superiority of the (implicitly single) sociological perspective over the views of lay actors. The idea of ‘lived experience’ is not helpful when it involves the implication that the lived experience of lay actors/participants is intrinsically local and bounded, and that such actors are incurious about the wider influences shaping their lives—that reflection on this is not part of their ‘experience’. Assuming a clear-cut contrast between ‘lived experience’ on the one hand and reflection on wider influences on the other allows lay actor/participant perspectives to be assumed to be unreflective and not of value compared to those of social scientists.⁸

A similar issue can be raised with Burawoy’s broad invocation of “common sense”. To a large extent, Burawoy’s approach seems vulnerable to a critique of standard forms of sociological analysis offered by Anderson and Sharrock (1983, 566–569), who contend that the notion of ‘common sense’ is not well-explicated by sociologists but that they nevertheless assume that it is a problematic form of reasoning characterised by ‘procedural faults’ and systematic errors.⁹ Burawoy’s analysis largely fits into this mould, insofar as he generally treats it as an obviously problematic way of thinking shared by non-sociological actors. The fact that the concept is not explicated or made more precise suggests that Burawoy thinks ‘common sense’ ways of thinking are straightforwardly wide-spread and problematic—the latter clearly applying insofar as he suggests that common sense is something that an analyst can be ‘trapped’ in (Burawoy 2017, 277). But it is surely legitimate to request that assumptions about the wide-spread existence of a common sense and its doubtful character should be supported by further argument and evidence.

I should admit that, in his 2012 article, Burawoy does sympathetically cite the Gramscian idea that, in certain circumstances, ‘good sense’ can be found in ‘common sense’ (Burawoy 2012a, 196). But two remarks can be made about this. Firstly, if Burawoy thinks this is the case, why is he so sweeping in his critical characterisation of common sense at other points? Secondly, although this link with ‘good sense’ is an improvement on Burawoy’s critical approach, it only returns us to the position in which lay actors’ views may be considered to be insightful if they concur with the views of sociological structural, scientific analysts. As such, the invocation of common sense does not seem to open up or encourage a sense of genuine dialogue, or encourage one to reflect on the possibility that, in certain cases, sociologists could be less insightful than lay actors/participants in terms of their views of the world.¹⁰ And this suggests to me that Burawoy’s 2017 position has not rectified the issues I identified in his earlier work.

5 Relating Social Scientific and Lay Perspectives: The Challenges of Dialogue

So far in this discussion, I have argued against a position which associates social scientists’ perspectives with an objective orientation in order to privilege them. I have suggested that lay social actors can have sources of insight and that their perspectives should not be placed in the distinct category of the subjective. Where there is divergence between accounts, what is required is a dialogue between social scientists and lay actors, where we understand this as a discussion about the value of the different viewpoints offered in which the epistemic superiority of social scientific accounts is not assumed. I should emphasize that dialogue in this context is not necessarily a conversation between just two perspectives. There can be multiple social scientific perspectives on any particular issue, with both intra- and inter-disciplinary disagreements possible (think of clashes between, say, sociologists and neuroscientists). Likewise, there will often be a range of lay perspectives on an issue.¹¹ As such, the dialogue will involve defenders of a range of competing viewpoints. However, I do not suggest that dialogue between advocates of different positions will be straightforward. In this section, I want to acknowledge some of the challenges and complexities of the process by briefly discussing three aspects of what is involved in dialogue: the conditions for productive dialogue, the significance of emotion and, finally, the uncertainty of meaning in dialogic exchanges.

Firstly, there is the question of the conditions required for productive dialogue. These have been extensively debated, with a key focal point of discussion being Habermas’s (1990) arguments, which have both enthusiastic defenders and strong critics. Although it is impossible to meaningfully summarize and respond to these debates here, I can flag up a range of issues that it would be important for those engaged in dialogue to take seriously. Those positioned on the pro-Habermasian side of the debate have established that actors engaged in dialogue should be mindful of the value of giving all participants an equal chance to contribute to the dialogue, of all participants having the right to throw the terms of debate into question and, indeed, of all participants being permitted to raise questions about the procedures for debating in the first place (see, for discussion, Benhabib 1994). Drawing on the work of those who are critical of the Habermasian position I would suggest that those engaged in dialogue need not consider consensus to be the ideal outcome or the only measure of success (see, e.g., Mouffe 1999). Even setting aside theoretical questions about the value of consensus, there are productive and reasonable outcomes of dialogue that do not require consensus. For example, I would argue that a dialogue where different participants challenge one another, such that all participants have to defend and develop their views in relation to those challenges, is one that has been productive. Furthermore, while those hoping that social scientists and lay actors could work together to promote mutually agreeable social projects and goals might fear the consequences of a lack of consensus, it is worth remembering that different groups do not have to agree fundamentally on everything in order to be positively motivated to engage in coordinated action on particular issues. (For feminist considerations about coalitional politics, see Carastathis 2013 and Cole and Luna 2010).

Secondly, it is worth emphasizing that dialogue between social scientists and social actors is likely to involve emotions. Feminist writers like Susan Strickland have argued that in cases where there are tangible power and status inequalities between parties in a discussion there are likely to be conflicts and tensions. She suggests such dialogues may be uncomfortable, being

more like an argument where you go away angry and hurt and defensive of your own point of view, but can’t forget theirs, which you keep mulling over and in the process, gradually altering your own point of view, so that at the next encounter, it has changed in interaction with theirs and so on.

Strickland 1994, 271

Even where power inequalities do not obviously shape the discussion, dialogue and disagreement are likely to involve emotions within the interaction. Insofar as parties are involved in and care about the consequences of a dialogue, this will have an emotional component to it, potentially including emotions such as engagement, satisfaction and irritation, among others. However, to say parties to a dialogue are not emotionally ‘neutral’ is not to say that such debates will, therefore, be irrational, or that emotions introduce a problematically ‘subjective’ component. As feminists and others have argued, emotions are not intrinsically unreasonable (Jaggar 1989; Barbalet 2008). In dialogue, emotions exist in a relation with the intellectual or practical commitments that participants bring to the matter under discussion.

The final issue that I want to bring out here is less discussed in the literature. This is that the actual substance of a dialogue may not be clear-cut and straightforward. In dialogues, what is at stake is not just the validity of the arguments of others, but also what it is that the other participants are arguing in the first place, and what the implications of those arguments are.¹² To use the example of a two-party dialogue, a social scientist will have an idea about what s/he is arguing and what the lay actor’s perspective is, and the lay actor will have an idea about what s/he is arguing and what the social scientist’s perspective is. But the justification of all these characterisations is up-for-grabs in a dialogue, rather than fixed. It is, of course, possible to misunderstand what others are arguing. But, it is also possible, in some sense, to misunderstand elements of one’s own argument. This is because the assumptions and implications of what one is arguing may not be fully apparent to oneself. To give an example from social theoretical dialogue, Peter Winch did not intend in The Idea of a Social Science and its Relation to Philosophy to assume the correctness of a Humean account of causality, but came to accept, in response to critics, that this was in fact assumed in his initial statement of his position (Winch 1990, xi–xii).

I have taken the time to emphasize the potential to misunderstand others and oneself because I want to suggest that this dimension of uncertainty about what is being argued is actually crucial in many dialogues. It only takes a brief exposure to dialogues between academics, for example, to see that one of the most common responses to a critique of a theory is for its originator to declare, ‘My argument has been misinterpreted!’¹³ Thus, dialogue as a process is in part about moving towards understanding what is being argued, as well as about evaluating the validity of those arguments. Acknowledging this dimension of dialogue does not make discussion any less worthwhile, and it does not reduce the possibility of arguments influencing the views of the parties to the discussion. But it does require us to acknowledge the complex layering of dialogues, rather than seeing them as straightforward clashes between clear-cut viewpoints.

6 Conclusion

The overarching concern of this article has been to contribute to an evaluation of the mode of critical social science which makes a clear-cut separation between subjective and objective perspectives on the social world. Such an approach separates the subjective, potentially compromised perspective of the lay social actor from the objective perspective of the social scientist, the latter conceptualized as being in touch with the real characteristics of the social world. This evaluation has been carried out by closely examining Burawoy’s analysis of domination, which applies the objective/subjective distinction to differing perspectives on labour relations. Criticising Burawoy’s approach, I argued that it mistakenly homogenizes social science into a single perspective and is also based on an unjustified prioritisation of social scientific perspectives over and above those of lay actors. I suggested that an important consequence of this is that where there are differences between the viewpoints of social scientists and lay actors, Burawoy’s analysis of domination undermines the conditions for a dialogue between competing perspectives, giving an ‘in-principle’ priority to the sociological perspective. I also argued that, although there are hints of alternatives in other aspects of Burawoy’s work, ultimately these cannot be used as a robust basis for a dialogic conception of social science.

What I would like to emphasize in this Conclusion is that, in raising doubts about the distinction between ‘objective’ social science and the ‘subjective’ perspective of lay actors, I am not denying the capacity of sociology to contribute to knowledge. Nor am I arguing against the value of calling on and further developing social scientific methods in order to enhance our understanding of the social world. The strengths of sociological research tools and modes of reasoning may reasonably be invoked in debates (although their limitations should also be admitted). Rather, my argument involves a criticism of the position that one can make a judgement about the superiority of sociologists’ accounts ‘in principle’ where divergent accounts and explanations are offered about the character of the social world. As many sociological researchers recognise, neither the use of particular methods nor other aspects of social scientific insight and expertise take sociological understanding into a special, distinct, objective realm which guarantees justification. Where the knowledge claims of social scientists clash with those of lay actors, what is required is genuine engagement between the different parties, where we try to convince others of the value of our insights whilst retaining an openness to learning from others and revising our claims accordingly.