On reductions – examining a British-Bourdieusian sociology of education

Abstract In this article I explore a disposition towards a critique of ‘reductionism’ and ‘determinism’ that seems to me to be very much prevalent within British sociology. I take a qualitative educational sociology that uses Bourdieusian concepts as one body of research where this disposition is expressed with particular fervour. A close examination of this work reveals its tacit acceptance of the limits of specific, and mostly statistical, classifications. It also reveals a distorted approach to reflexivity and the view of rival approaches and critiques. This points towards a specific and systematic disposition that, through the abundant warnings of ‘reductionism’ and ‘determinism’, itself reduces the development of a particular theory to a specific epistemology, thereby endangering the potential of rupture and epistemological break advocated by Bourdieu. This prompts us to pose fresh Bourdieusian questions about the state and practice of Bourdieusian reflexivity.


Introduction -a Bourdieusian view on Bourdieusians
Some time ago, Matt Dawson, one of my colleagues at Glasgow who specialises in sociological theory, wrote a few blog entries about the British Sociological Association's annual conference (Dawson 2015(Dawson , 2016. He looked at the frequency with which various theorists were used in the presentations (based on a name search in the abstracts). Barely surprising, the name 'Pierre Bourdieu' always comfortably came up as first place. No doubt, sociology, perhaps particularly in the UK, has seen a kind of revolution in the last 30 or so years, transforming the 'hidden gem' Pierre Bourdieu into a towering figure whose ideas have become a basic staple of the sociological diet for the next generation of sociologists (Thatcher et al. 2018).
If one listens to these researchers at the said conferences (or its offshoots), the critical outside observer may be quite amazed at the amount of change that Bourdieu's ideas have experienced in this process of 'canonisation' and 'import' . Or so it feels, for he or she may, at first, barely have something to point towards when it comes to the adaptation of Bourdieusian ideas to British lands. All the basic concepts -above all capital, habitus, and field -are there, and so is the theoretical eclecticism and political engagement in the form of siding with the weak and powerless for which Bourdieu (also) became famous. Still, there may remain a queasy feeling in the stomach of that critical outside observer. But it is one not easily spelled out, nor made specific. One may engage, as a reaction, in an unfair polemic. But this leads nowhere as it only serves to compound one's own learned prejudices. The point in this article, however, will be to subject part of this new generation of Bourdieusians to a Bourdieusian-inspired critique in order to substantiate, within the vernacular it itself uses, what may be a 'mere' feeling otherwise. For I confess that I do feel that, in the process of 'canonisation' , Bourdieusian thought, not unlike that of other thinkers before him (Mills 1959, 3-131), has been subjected to a constriction from which it needs to be freed if it is to have the impact that its founder hoped it would have.
This article, then, is a critical, yet close objectification and clarification. It will be about uncovering and exploring the blind-spots of a particular sociology of education, usually associated with Cambridge Professor Dianne Reay as the main representative (Bathmaker et al. 2016;Burke 2011Burke , 2017Burke, Emmerich, and Ingram 2013;Mellor et al. 2014;Reay 1998Reay , 2004Reay , 2017Reay, David, and Ball 2005;Reay, Crozier, and James 2011;Wallace 2017). It will do so in a Bourdieusian and rather 'inductive' way. That means it assumes, on the one hand, the existence of the usual Bourdieusian concepts -field, capital, habitus (Bourdieu [1980] 1992a) -while, on the other hand, attempting to fill only a part of these conceptual vacancies with content. This filling-procedure will concern the expression of a scientific habitus of a group of researchers active in what may be called the contemporary British field of sociology. After a brief historical positioning of the sociology under investigation here (Section 2) I will uncover a systematicity in the research practice of these Bourdieusians that points towards specific tendencies of vision and di-vision. In this I will concentrate on two main aspects of their scientific practice -the executing of empirical projects, and the conduct of Bourdieusian reflexivity, which includes 'softer' aspects of deference and demeanour (Section 3). What can be seen are systematic divergences in the meaning of practices in relation to the original Bourdieusian method and sociology. Remaining within a Bourdieusian framework myself, I claim that this investigation allows us to draw certain conclusions about the texture of the habitus of those who carry out this research, and about the field in which they are active. It furthermore pushes us to formulate fresh and more precise questions towards a sociology of sociology of this field (Section 4).

Expressing 'anti-determinism' in the field -a British-Bourdieusian sociology of education
It may be said that the literature discussed here represents and specifies an attitude of 'not impressing one's ideas onto the data' , a phrase often heard at sociological conferences these days. However, it may also be acknowledged that it is one possible implementation in academic sociology among others. The point here is simply to outline the main attitudes in the field into which the original disposition is congealed and channelled.
By way of a brief, and certainly more conventional, contextualising, it is probably fair to say that this educational sociology emerged out of a context of what might be called the 'class and feminism' debates within the sociology of the 1980s and 1990s. The rise of feminist social research (Skeggs 2008) was directed first against objectivist-quantitative forms of social science (see Acker 1973;Goldthorpe 1983;Goldthorpe, Llewellyn, and Payne 1987;Skeggs 2008;Stanworth 1984), but also against grand theories and narratives. These included, especially during the 1990s, claims of the vanishing of social classes (Beck and Beck-Gernsheim 2002;Pakulski and Waters 1996). In contrast to these, Diane Reay and others developed an approach which, in their view, '[…] understand[s] social class in terms of relationships, not just economic relationships […]' (Reay 2017, 2), but which at the same time acknowledges that '[d]iscourses of classnessness are in effect class discourses insofar as they operate in class interests' (Reay 1998, 261 (Reay 1998, 259). Therefore, the original disposition 'not to impress one's ideas onto the data' is interpreted and minted in such a way that it is associated with an attitude towards 'qualitative' social research, above all in-depth biographical interviews, and an open political partisanship with dominated groups such as the 'working classes' , women and non-white ethnicities. It is decidedly not interpreted in a way in which this 'anti-determinism' would come to mean something closer, say, to a position that is more cautious of taking a political side that openly, or a more quantitative approach.
The question is now two-fold: first, how are Bourdieusian ideas transformed and deformed in this kind of social and intellectual context? And second, why is this so? I will attempt an answer to the first question in the pages that follow, engaging in a close analytic comparison of Bourdieu's works and those of Diane Reay and her associates. I wish to map out to which methodological decisions and which argument structure this 'anti-deterministic' disposition leads. This examination aims to show that at the bottom of the 'anti-determinism' of this type of educational sociology there is a determinism and a reduction of the comprehensive and open Bourdieusian epistemology (Bourdieu, Chamboredon, andPasseron [1968] 1991;Bourdieu 1992b). This reduction leads to an empiricist-qualitative sociology that misrecognises itself as such, a sociology which systematically privileges specific forms of evidence and specific interpretations of evidence over others. Moreover, this determinism is misrecognised by the skilful and creative transformation and rededication of Bourdieusian concepts. I will show this recurring duality of a specific reduction and misrecognition of this specific reduction, of what is done and what is represented to be done.

A peculiar disposition -the reductions in 'anti-reductionism'
The 'anti-reductionist' disposition can be found in many scientific operations of this British-Bourdieusian sociology of education. I am here singling out two areas for further examination -the empirical projects that follow from it (3.1), and the understanding of reflexivity and theory (3.2).

Empirical projects
How do these researchers structure their qualitative-empirical endeavours? Following the rejection of both quantitative-objectivist and grand theoretical approaches as 'reductionist' , the thrust of these studies is to provide a closer and more intimate look on education. It is in this thrust that the application of Bourdieusian concepts is embedded. These operations are supposed to balance out the theoretical and economistic determinisms of the other approaches by focussing on 'agency' , 'consciousness' , 'reflexivity' , but also 'fluidity' and 'complexity' , making room for the 'indeterminacy' , 'change' and 'messiness' (Reay 2004, 432) of the world.
However, even though there is, again and again, a repeated insistence and critique of the reductionist character of the usual statistical categories, like the British national-statistics socio-economic classification (NS-SEC), which need to be replaced with a narrative that is 'more than just' objectified category, it is astonishing to what limited extent the original categories themselves are questioned. This is the case with Reay, David, and Ball's (2005) study on educational choice in which they, on the one hand, rebuke the 'simplistic divisions' of quasi-official classifications of middle and working classes because they '[…] convey only a fraction of the story of social class' (Reay, David, and Ball 2005, 16) Instead, they hold, these divisions need to be 'overlaid' by the study of '[…] affective responses to the higher education choices process, such as sense of security or insecurity, familiarity or unfamiliarity, and attitudes and inclination […]' (Reay, David, and Ball 2005, 16). This further suggests, and resonates with, a distinction between the 'rational' and the 'affective' or 'irrational' . There are the 'quantitative' approaches that take stock of the 'high status activities' on a grander scale, and then there are the 'qualitative' approaches like that of Reay et al. which take care of the 'blind spots' that the original classification has produced, telling '[…] more nuanced-inflected tale[s]' (Reay, David, and Ball 2005, 16), taking into account complexities that 'cannot be bracketed' (Burke, Emmerich, and Ingram 2013, 166). Bourdieusian concepts are transformed to reflect this desire, and that means they are multiplied. Thus, there are specific concepts like 'black cultural capital' developed that conceptualise the educational processes of black Carribean youth in London (Wallace 2017) to account better for racial problems in educational sociology. This is because cultural capital per se is perceived by Wallace as a 'fixed, static resource' linked to 'whiteness' , which therefore must be 'de-essentialised' in order to '[…] challenge durable deficit narratives about black and ethnic minority young people and to produce asset-based perspectives […]' (Wallace 2017, 912f.). There is, furthermore, the concept of the 'institutional habitus' (Reay, David, and Ball 2005, 35-60) to account for specific 'school cultures' which are neglected in earlier studies. There is also the notion of 'familial habitus' (Reay, Crozier, and James 2011, 23-43) to show in what ways parents influence the educational choices of their children. In terms of groups there is a clear focus on hitherto neglected categories such as the working class (Reay 2017, 57-74), gender or race. Most of the added concepts regard what is called 'the median level' , looking at '[…] positive, deliberate and active processes […]' (Burke, Emmerich, and Ingram 2013, 166) such as 'conformity' , 'agreement' or 'cooperation' within it.
But the theoretical concepts that follow from the empirical studies produced still exhibit a strong logical link to the originally criticised 'crude' classifications. For some of the added concepts, such as 'conformity' , this rootedness in common sense has been shown a long time ago (Jahoda 1959). The newer concepts of 'institutional habitus' , 'familial habitus' or 'black cultural capital' -even if intended to counter conventional statistical categories like 'middle class' or 'working class' , are still outgrowths of the acceptance of the logic that underlies them, in the sense that they define and structure the reality to be studied, in the sense of an epistemological acceptance. In the same vein, Reay, Crozier, and James's (2011) study on 'middle-class identities and urban schooling' -despite its critique of the 'homogeneity' of the usual quantitative studies which are in need of a portrayal of 'cross-cutting differences' , recognising dimensions like geographical dispersion, political persuasion or education -assumes a surprisingly conventional definition of these classes. They are to be distinguished by '[…] the ability to wield power over others […]' (Reay, Crozier, and James 2011, 12), of having 'individualism' , 'citizenship' , 'responsibility' as their core values. Other examples would include the study of 'social mobility' (Reay 2017, 101-130) which is of course a long-standing concept of quantitative educational sociology, at least in Britain (Goldthorpe, Llewellyn, and Payne 1987). What kind of logic is this? It is a logic that orients itself on the real and already existing -existing either in statistical-quantitative or state classifications or in popular discourse. If these classifications constitute a relational point, it is not shown in a more precise manner how they are relational, despite protestations to the contrary. However, this already implies the acceptance, despite or even through the critique levelled at them, of the initial and mostly statistical classifications. In short, this approach, which wishes to combat 'reductionism' in terms of empirical topics and theoretical concepts, is itself quite content with a reductionism in epistemological terms, the reductionism laid out by 'orthodox' social science and even 'common sense' .
We might attain a clearer view of this epistemological reductionism by utilising an analysis of Popperian epistemology, which is the basis of much of the 'quantitative' and 'orthodox' sociology that Reay and her colleagues position themselves against. This analysis is provided by Norbert Elias, who shares some fundamental epistemological principles with the Bourdieusian project (Bourdieu and Wacquant 1992, 92f.). In his critique (Elias [1985] 2009) of Popper's Logic of Scientific Discovery' ([1934] 2002) Elias points towards the isolated position that, for the Popperians, epistemology and theory have in relation to scientific objects or 'reality' . In their quest to avoid the seeping in of 'inductive' and therefore nonuniversal criteria of theory-production, the Popperians favour a 'deductive' approach based on Mathematics and theoretical Physics. But by doing so, Elias suggests, they neglect recent advances in the empirical sciences, such as sociology, general physics or biology, that stress the necessary interaction of theory and 'inductive' empirical evidence. In doing so, Elias holds, they create a kind of 'pure science' , hermetically closed to all outside influences, a phenomenon he calls 'one-level sciences' . Against this, 'two-level sciences' have as a reference point not themselves but an empirical plane with which they have a dialogue, an exchange of suggestions, giving rise to a history of concepts and logics, relations and viewpoints, with eventual attempts to synthesise them (Elias [1985] 2009, 165f.). What emerges from these thoughts is the realisation that scientific activity has essentially three layers that consist in 1. actual social and/or natural events to be explained, 2. specific theories constructed to make sense of these events, and 3. theories of these theories, i.e. epistemology (Elias [1985(Elias [ ] 2009. This assumption is shared by Bourdieu (Bourdieu, Chamboredon, andPasseron [1968] 1991, 30f.) and other philosophers of science (Polanyi [1958(Polanyi [ ] 1974. The central issue now evolves around the relationship between these three layers, especially between social events and specific theories on the one side and epistemology on the other. Should the latter be isolated from the former, in some kind of metaphysical way? Or should it be open to a kind of dialogue with specific theories and social events that it produces? Popper clearly advocates for the first option, justifying this move with reference to the unexamined and empirically unproven 'value-judgments' that would seep into epistemology by doing otherwise. But by doing so, Elias suggests, he fixes and universalises the empiricist epistemology as it was practiced during Newton's and Kant's time as the one and only model to arrive at theories and a properly scientific interpretation of social acts. Elias clearly sees the ethnocentrism involved in this. He uses a metaphor to explicate what this 'deductivism' amounts to: '[…] one proceeds more or less in the same manner as one would if one were to work out a model of the common properties of mammals only on the basis of one's knowledge of lions, perhaps because the lion corresponds best to one's ideal of a mammal or maybe because the prestige and status of lions as mammals is particularly high' (Elias [1985(Elias [ ] 2009. Now, it is obvious that the 'anti-reductionist' sociology discussed here would tend to agree with this argument wholeheartedly. After all, its work is presented as exactly the opposite of the 'fixating' and 'reductive' approaches to education that characterise the 'orthodox' approaches like that of John Goldthorpe, for whom rational action is a central part of the model of science that he deems universal, his 'lion' , so to speak: '[…] the tendency of actors to act rationally in the circumstances that prevail is the common factor influencing them -even if relatively weak -while propensities to depart from rationality operate randomly in many different ways' (Goldthorpe [1996] 2010, 316, emphasis in the original). 1 No doubt, this would be felt to be 'reductive' by scholars like Diane Reay: 'Given the dominance of ideas like 'choice' , 'diversity' and 'the market' in educational policy and implementation, it is perhaps tempting to see secondary school choice through the most apparently simplistic of economic perspectives, as if it was best understood through choices made by rational, calculating individuals in an increasingly information-rich environment. Yet even the most casual of conversations with parents making such choices, or with school head teachers, or with young people, will quickly expose other kinds of consideration, and indicate that school choice is much more than a sum of intelligent use of market data and the odd pragmatic consideration. For this reason, we cannot begin to understand it unless we bring to bear some appropriate tools' (Reay, Crozier, and James 2011, 23). Thus, the justification to use Bourdieusian concepts in the way they are used. Nevertheless, the very fact of aiming to add to this picture already implies the acceptance of the logic that undergirds it, in the name of a 'personal integrity' and 'passionate partiality' that compels these thinkers to put '[…] a strong focus on social justice and an emphatic focus on working-class experience […]' (Reay 2017, 2).
Thus, political engagement is directly infused into epistemological-Bourdieusian principles which are therefore rendered very much flexible and nuanced in their relationship to theory and data. In complete contrast to much of quantitative sociology, this relationship is not fixed but essentially merged. By doing so it tends to enclose empirical research on education in an exclusive alternative between fixed vs. fluid relationships of epistemology and theory/data. Of course, this enclosure itself represents a reduction. 2 This 'fluidity' , which is quite distinctive for this adaptation of Bourdieusian ideas vis-à-vis other national contexts, can be seen at every step of the analysis. If it is assumed from the outset that all fixating of meanings of concepts like capital, habitus or field is 'reductive' then one is forced to operate with either vaguely defined and/or multiplied versions of these concepts (see also Atkinson 2011 for this point) and/or to accept what 'reality' offers. These concepts can then in fact follow the impressions one gains from the field, and can be linked with the concepts that arise from these impressions. This is rather easily done precisely because of the vagueness of the initial epistemological-theoretical concepts. One example may show this particularly clearly. The concept of the 'familial habitus' applied to a group of 250 pairs of middle class families is discussed (2011,. These were interviewed for a project on educational choice. At the outset habitus is defined, in 'anti-reductionist' terms, as 'durable, transposable dispositions' and as a kind of 'active past' . A very high percentage of this sample compared with the average of society went to private grammar schools, and an even higher percentage went on to university (Reay, Crozier, and James 2011, 24), often as the first in their families. It is therefore inferred that this fact is '[…] a highly visible marker of the role education played in their class mobility' (Reay, Crozier, and James 2011, 25). The authors have therefore linked (and effectively reduced) habitus dispositions to a specific kind of schooling history of this class, to a specific kind of expression of this habitus. Then they go on to examine the views of various of these parents on their children's schooling. They find that quite a few of these parents do not want to send their children to private grammar schools, the very same kind that these parents themselves attended, but to 'normal' state grammars. The parents interviewed give various reasons for this, such as the bad experiences of another family member, or their own professional reflections (see Reay, Crozier, and James 2011, 31-36). Because this is, at face value, different from the original habitus Reay and her colleagues see in it '[…] a great departure from the family habitus […]' (Reay, Crozier, and James 2011, 39), and even an attempt to '[…] avoid history repeating itself […]' (Reay, Crozier, and James 2011, 38). In conclusion, they see a new, 'collective, moral vision' of 'reciprocity, care, empathy' on the rise (Reay, Crozier, and James 2011, 167), after these parents have become 'sociologically conscious' of their roots. This, too, can be framed in Bourdieusian terms: 'Whilst habitus reflects the social position in which it was constructed, it also carries within it the genesis of new creative responses which are capable of transcending the social conditions in which it was produced […]' (Reay, Crozier, and James 2011, 32). Notice how the epistemological concept (here the middle class habitus) is merged with an expression of this habitus (private schooling), and how this merged concept-expression is then conserved into the present time. 3 At the root of the 'freedom' and 'anti-determinism' that is exhibited here is therefore a double reduction of Bourdieusian ideas -one that merges an epistemological concept (familial habitus) with an alleged expression of it (attendance of private schools) based on quantitative state classifications, and another one that conserves the meaning of this expression over a generation. The specific 'anti-reductionist' disposition discussed above therefore leads to the relinquishing of the 'primacy of consciously constructed theory' (Bourdieu, Chamboredon, andPasseron [1968] 1991, 35f.). Unlike in Bourdieu's epistemology, it is not the idea that makes the fact but the fact that makes the idea.

Reflexivity
This double reduction tends to lead to a very empathic and sympathetic (two phrases essentially merged in this sociology of education, as epistemology is merged with values, theory and fitting data) depiction of some groups (mostly working-class and female and non-white ethnicity students). It also leads to a more distanced, irritated and unfavourable depiction of others (esp. middle-class students but also working-class students critical of working-class practices, see for instance Abrahams 2017, but also Mellor et al. 2014). The former are often accorded subjective agency while the latter are shown to be linked to their class habitus. 4 But if this is so, how does this sociology nevertheless claim to be faithful to Bourdieusian ideas? How, for example, is reflexivity, as a central feature of Bourdieusianism, dealt with? This is supposed to be a tool to free oneself from the (self-)limitations of one's own conditions of existence and of the socially positioned and dispositioned habitus. It means to construct the specific, academic space with its specific forms of capital and specific positions to find out where the 'blind spots' of one's view are, what is taken as doxa (Bourdieu and Wacquant 1992, 104f.). It is to be accomplished by rigorous self-application of the same concepts to oneself that are applied to others. Would this not mean that this kind of sociology would eventually become conscious of its de facto empiricism and change course? When observing it in the practice of reflexivity we, too, can see the disposition towards 'anti-reductionism' at work. We can see how reflexivity is interpreted and executed in a specific way.
In fact, reflexivity itself is already rendered in 'anti-reductionist' terms. This, in the view of Diane Reay, originates in her working-class turned parvenu trajectory: 'My working-class background influences everything I research and write. That passionate partiality is helpful in ensuring a strong focus on social justice and an emphatic focus on working-class experience, but it can also result in an over-simplification of that experience that is particularly problematic at a period when class has become such a confusing, contradictory and slippery concept' (Reay 2017, 2, emphasis mine). There are three claims congealed in this sentence. First, that there is some direct link from social origin to scientific sensibility. Second, that Reay knows of the doxa of either her field of sociology and/or the people that come from her origin and that are positioned in that field, and that this doxa is 'over-simplification' . And third, that Reay has overcome this 'over-simplification' by '[…] thinking against the person that class and community made […]' (Reay 2017, 2) her.
Her colleagues, too, seem to share these claims. They often, for example, refer to their own 'working class origin' and the resultant 'fish out of water' state in which they presumably find themselves in contemporary British sociology (see for example Bathmaker et al. 2016, 93-125). This, they argue, tends to give them the lucidity of a similarly 'cleft habitus' as Bourdieu had (Bourdieu [2004(Bourdieu [ ] 2008(Bourdieu [ , 2006. This assumes a kind of shared 'working-class' origin -understood in a rather comprehensively statistical and fuzzy way -that is assumed to be obvious to everyone involved as well as an equally 'clear' , and contrasting, 'orthodox' state of the sociological field. The alleged clash between both then produces the 'cleft habitus' that this sociology claims for itself. At the same time, however, it is difficult within this view to enquire what exactly working-class entails since this, once again, would constitute a 'reduction' that is incompatible with the disposition shown here: '[…] class experiences are intersected by positions such as ethnicity, age and gender, meaning that there is no unitary class experience. We argue that this allusion of class homogeneity is a limitation that could influence that data analysis and questioning process […]' (Mellor et al. 2014, 146). We are therefore thrown back to our primary '[…] capacities for reflexivity, empathy, communication, curiosity, analysis and respect together with skilful use of life experience […]' (Mellor et al. 2014, 147), and thus complete a circular movement by which the refusal to 'reduce' is justified by the taste for the 'refusal' to 'reduce' . By contrast, alternative Bourdieusian approaches critical to this stance, like that of Will Atkinson (2011) who develops a critique of it similar to the one developed here, are labelled as 'objectivist' , as not '[…] embracing [their] own intellectual context […]' (Burke, Emmerich, and Ingram 2013, 168).
But there are other issues with Bourdieusian reflexivity that pose potential problems for this disposition. What about the insistence of theory-building in Bourdieusian writings in order to circumnavigate empiricism? How can one end up with empiricism while at the same time nominally promoting an epistemology and theory that eschews and warns of empiricism? One strategy seems to be the reliance on and rededication of dichotomies. Sometimes, for instance, there are complaints about the narrow 'empirical focus' that many (presumably quantitative) empirical Bourdieusian studies often have (Reay 2004). This, in the view of this sociology, shows once more the 'latent determinism' of this use of Bourdieu's concepts against which a seemingly radical relationalism has to be put: '[…] habitus becomes active in relation to a field, and the same habitus can lead to very different practices and stories depending on the state of the field' (Reay 2004, 432). There is a theoretical aura borrowed when criticising what must obviously be quantitative-positivist empiricism as exhibiting '[…] the danger [of] habitus becoming whatever the data reveal' (Reay 2004, 438). And since it is also clear that habitus must be used as a 'tool' , i.e. not in any non-empirical, 'theoreticist' way, it therefore follows for this sociology that it is best to merge epistemology, theory and 'reality' into one: '[…] the difficulties, inconsistencies, risks of determinism, and aspects of circularity inherent in habitus can be viewed as far less problematic if habitus is viewed more fluidly as both method and theory, a way of understanding the world' (Reay 2004, 439). Thus, through the skilful rededication of labels and meanings -empiricism associated with only quantitative empiricism, the primacy of the theoretical movement in the research process associated with theoreticism -it looks as if the specific empiricism of this sociology which accords theory a submissive role is the only viable theoretical alternative in the Bourdieusian spirit, rather than, say, the possibility of combining theoretical primacy with qualitative and qualitative elements as it is done not only by Bourdieu but also works like that of Goffman (Goffman [1961a or Willis ([1978Willis ([ ] 2014. This is further compounded by the relegation of rival Bourdieusian accounts into that dichotomy, so that Atkinson's (2011) critique is labelled as having theoreticist 'neutral pretensions' and as being 'prescriptive' and 'inflexible' (Burke, Emmerich, and Ingram 2013, 169f.), in short as being aloof and 'arrogant' .
A similar operation by this sociological school can be observed with work on Bourdieusian epistemology (for example Burke 2011) which it both interprets and manipulates. Theory, it is held, 'cannot be avoided' for the Bourdieusian. This is so because we need to 'abandon' the 'autonomous role' of the researcher and of value-free sociology. Again, the warning is of a specific kind of positivism, i.e. its quantitative side. There are a few remarks that formally criticise qualitative research, but in such a way that it is not about the lack of theory of it but about not being empathetic enough due to the '[…] feeling of superiority from the supposed ability to understand or appreciate an individual's perspective' (Burke 2011, 1). In other words, if there is a problem with qualitative research it is that it is too 'arrogant' . What is done here is to effectively juxtapose epistemological concepts and moral points by associating arrogance with theory and value-freedom. Against this, open 'partisanship' is positioned as 'a healthy disposition' , necessary to figure out one's biases and to avoid 'purely inductive research' . Thus, 'spontaneous sociology' as understood in Bourdieu (Bourdieu, Chamboredon, andPasseron [1968] 1991, 20-24) is shifted and narrowed down in its meaning to mean the blind acceptance of the 'neutral researcher' , now ignoring Bourdieu's warning that too much blind trust in humanism is another dead-end, that '[…] good sentiments [also] make bad sociology' (Bourdieu, Chamboredon, andPasseron [1968] 1991, 251). Following this, there are two options for empirical research that are claimed to be taken out of Bourdieu's writings -one that is merely the '[…] recital of a transcript […]' (Burke 2011, 2), thus 'inductive' , and one where 'theoretical consideration' has to be spent. Again, the impression is created as if this sociology is in line with Bourdieu's avoidance of empiricism. In reality, however, Bourdieu's original meaning of empiricism is once again shifted in a way that it now only encompasses a rather caricatural depiction of it (transcript recital), whereas other and maybe more 'realistic' kinds of empiricism, e.g. ones that more openly involve theory, now take and replace the place of (initially) 'data-independent' theory in Bourdieu. This sociology allows only for a specific kind of theory, that is, theory dependent on 'data' and impressions that spring, as we have seen above, from integrated and common-sense-rooted classifications. 'Rich data' is to be embraced a la grounded theory which merely has to be 'theorised' . 'Common sense understandings' are due because '[…] social researchers […] are also social actors' (Burke 2011, 2). Because of this dualism there should be a 'working partnership' between 'theory' and 'empiricism' so that the former 'appreciates grounded conditions' while the latter 'looks outward' . 5 Thus, by 'appreciating' 'empiricism' , 'theory' is subsumed under data and also under a corresponding inductivist epistemology which nevertheless understands and misrecognises itself to be Bourdieusian. The original Bourdieusian epistemological break with 'common sense' , then, is surely not enough since there is the danger of simply replacing 'common sense' with 'scholarly common sense' , which is of course 'theory' (that also includes theory independent from the given and accepted reality, i.e. non-empiricist theory), so that, again to avoid a 'reduction' , a 'second break' needs to be accomplished.
But how do we know what common sense is? Would this not have to be shown a bit more precisely in a Bourdieusian perspective? After all, it is not at all clear that 'theory' independent from given data as such is 'common sense' . But then, surely, it would be 'reductionist' to try to specify this 'common sense' which is also full of 'nuances' and 'complexities' . Here, too, the disposition under observation knows a way out of this impasse, which is by pleading in moral terms '[…] that all theory up until now has been a theory created by, and for the advancement of, men' (Burke 2011, 3) -a sweeping, quasi-philosophical statement that seems to serve as a placeholder for an empirical and reflexive investigation of what 'common sense' in a specific time and place really means. The use of authority figures like W.E.B. DuBois, C.W. Mills, Alvin Gouldner or Anne Oakley as well as the reference to past states of the field therefore serve to symbolically reinforce and to justify the disposition of 'anti-reductionism' and to evade a properly empirical investigation into what 'common sense' actually is. The price paid for this specific 'anti-reductionism' is therefore an epistemological mutilation of the Bourdieusian understandings of 'reflexivity' , 'theory' , 'empiricism' and 'common sense' .
One more time, therefore, we see an epistemological reduction of everything non-empiricist and qualitative into the camp of theoreticism, 6 objectivism and quantitative empiricism.

Discussion and conclusion -towards a more nuanced and disciplined reflexive practice
This article set out to clarify what was initially a 'mere' feeling, i.e. the constriction of Bourdieusian thinking in a specific intellectual context (British sociology). By examining a prominent part of this thinking, I hope to have fulfilled this task.
I tried to do this by looking at two different analytical levels: there is the epistemological or 'logical' level of comparing Bourdieu's works with those of the sociology of education examined here. On this level, I have shown the more or less tacitly empiricist and inductivist epistemology inherent in the latter approach, arrived at by implicit or explicit acceptance of received classifications, statistical categories and everyday concepts. Essentially, this leads to a reduction of epistemological vigilance, to a reduction of the scope of theoretical variety and thus ultimately to an impoverishment of the sociological imagination and practice. This is the 'logical' level of analysis. But there is another, equally important level that is intimately wedded to the first, though it is often neglected even by critical observers like Atkinson (2011Atkinson ( , 2013. That is the level of legitimation by classifying symbols. Truncated and modified Bourdieusian ideas such as 'practice' , 'habitus' or 'capital' are one group of symbols used. But there are also prominent other sociologists' names, specific labels and socio-morally loaded terms such as 'orthodox' , 'value-freedom' and 'theoreticist' which persuade more than they convince. What is surprising is the systematicity with which this transformation and deformation of Bourdieusian thought is carried out and defended. What follows from this analysis for the practice of doing Bourdieusian analysis? I see three areas for improvement in order to arrive at a more reflexive, less reductive Bourdieusianism. First, more care has to be taken to distinguish epistemology from theory and data. This means, above all, to move beyond the largely metaphorical use of concepts. It needs to be specified in each case what, and what not, is meant by 'field' , 'capital' , 'working-class habitus' or other concepts. This also involves to fruitfully relate these specified notions to each other so that they make up a particular and revisable theoretical model that provides concrete, justified leads and boundaries for empirical enquiry as well as focus, rather than a mere repository, to be used at free will.
This would also involve, second, a more thorough consideration of the epistemological break that goes well beyond a vaguely defined 'scholarly common sense' . Rather than to simply and conveniently equating habitus, capital or field with content easily found in 'reality' -in institutions, or families, or skills -they would need to be more consciously constructed. The same goes for concepts such as 'mobility' , 'working class' or 'choice' . This would imply a more thorough break with strands of research such as quantitative sociology whose limitations are presently more embroidered than overcome.
The necessity of these two steps is especially clear with the use of the term 'field' by these authors (as in 'the field of higher education' , see for instance Bathmaker, Ingram, and Waller 2013;Bathmaker 2015). It de facto covers both arenas of everyday encounters and more abstract compositions of institutions, and thus is, in practical use, more of a metaphor than a working tool. A more epistemologically rigorous Bourdieusianism would force itself to specify what, in terms of persons, institutions and types of actions, is affected by the 'field' , and built a clearer model to explain why the habitus of certain groups (working-class, male or female, ethnic minority students) do what they do in it. Even though, of course, this would constitute the dreaded 'reduction' , it would be the basis to actually initiate (rather than to forestall) the close dialectic of specific theory and empirical investigations that was the hallmark of the original Bourdieusianism. For example, it would force us to rethink the very applicability of the term 'field of higher education' for students, since they, by definition, do not have field-relevant capital to speak of, which is the entering condition for existence in the field. Overcoming this metaphorical use of 'field' , so common in the social sciences (Wacquant 2014, 133, FN9), would also mean to think about a more selective, yet 'epistemologically disciplined' (Wacquant 2014) use of Bourdieusian concepts. Rather than to imply canonically and mechanically some 'inseparability' of field-habitus-capital, it would encourage us to think of creative and selective mergers with those of other theoretical strands. For example, for the 'students in the field of higher education' this would mean to search for conceptual alternatives for 'field' that take into account the, as of yet, unfamiliar/ unsettled state of students in the environment in which they will be active, and which therefore requires an explanation of their socialisation and acquisition of capital within it, something that is completely put out of view otherwise. Thus, the creative, yet selective fusion of habitus or capital with more interactional theoretical tools, such as 'self ' , 'encounters' , 'game' or 'instrumental formal organisation' , as they are supplied by thinkers like Goffman ([1961Goffman ([ ] 2013Goffman ([ , [1961b, would seem helpful. This, in a way, would be to change priorities -rather than to be close to specific Bourdieusian quotes or concepts and to relax the consequential following of Bourdieusian epistemology in an attempt to go 'beyond' or 'think after' (Bathmaker 2015) Bourdieu, 7 one should rather do the reverse.
This would also increase the attractiveness of this sociology of education for broader strata of the profession, including foreign academics, because a clearer conceptual realisation of Bourdieusian concepts will allow for much more precise comparisons by way of homologies, and thus fertilise a desperately needed informed internationalisation of the field which is, still, in its infancy (Mosbah-Natanson and Gingras 2014).
Thirdly and lastly, this practice, if applied to the researchers conducting it, would also have a freeing effect on the way reflexivity is carried out. It would force these thinkers to bring into much closer explanatory interaction their personal trajectories in social and academic space, and their concomitant position-taking in it as a 'result' of this. It would mean to go beyond vague categorisations such as 'working class' , 'cleft habitus' or metaphors like 'fish out of water' in order to specify what these terms really mean in the field in question. A very fleeting glimpse at the long-standing distaste of British sociology for statistical or 'analytical' methods by both substantial numbers of staff and students (Halsey 2004, 187-194;Williams, Sloan, and Brookfield 2017;Panayotova 2019) and the overall drive of class sociology towards subjectivism and microscopism (Carrier 2012) would already suggest that this particular Bourdieusian sociology is more integrated into broader field traditions and developments than it is prepared to admit at present. This seems all the more clear if one looks at the very different ways in which Bourdieu's sociology is interpreted in other national contexts -just look at the much more statistics-affine, almost structural-functionalist interpretation that currently succeeds in German sociology (Bernhard and Schmidt-Wellenburg 2011). These broader developments must be brought into much closer explanatory interaction with the trajectory and perspectives of these thinkers in order to achieve a Bourdieusian reflexivity that is not content with (positive or negative) labels, but for which understanding is central. This, then, also implies a fresh look at the sympathy-empathy (not) shown for various groups of students. A more nuanced and precise reflexivity built in this way promises to contribute substantially to the sociology of sociology, sharpening the efficiency of sociology for real social change that is so badly needed in our neoliberal times. Again, this will also foster sociology's internationalisation as it facilitates comparison and homologies across very different national traditions. Then perhaps, it will at last be possible to achieve a more decisive break with both the personal and sociological past which, at the moment, seems more imaginary than real. Notes 1. I am not suggesting a complete and simple overlap of Popper's metaphysical philosophy of science and that of John Goldthorpe. After all, the latter arrived at his conclusions about metatheory after decades of empirical research (see Goldthorpe et al. 1968 for an early example of this empirical orientation). On the other hand, even in these early works one may see an implicit philosophy of science at work which fairly well resembles the one criticised by Elias above, and which therefore is, in my view, applicable to the same criticism. This would suggest a specific disposition towards this which has been compounded and congealed over the years and through empirical research. It would also push us to enquire into the differences between an Eliasian and Bourdieusian social theory beyond the common bases that I assume here (see for a start Bourdieu and Wacquant 1992, 92f.), but this is a topic beyond this paper. 2. 'To confuse the theory of sociological knowledge, which is at the level of meta-science, with the partial theories of the social that implement the principles of sociological meta-science in the systematic organization of a set of relations and principles explaining those relations, is to condemn oneself either to renounce the practice of science by expecting a science of metascience to stand in for science, or to treat a necessarily empty synthesis of general theories (or even partial theories) of the social as the meta-science that is the precondition for any possible scientific knowledge' (Bourdieu, Chamboredon, and Passeron [1968] 1991, 30f.) 3. Atkinson also notes the the mix-up of expressions of habitus with the habitus itself: '[…] although we may read off the existence of a habitus from the practices it generates, one would not call a taste for fish an 'aspect' of the 'dominant' class habitus but a practice generated by its qua complex of dispositions toward privileging form over function based on relative distance from necessity' (Atkinson 2011, 337). 4. This points us again to the deep links this sociology continues to entertain with more economy-centred approaches. It reminds us of the 'resurfacing' of the opposite epistemological position (here objectivism) in a mutilated form when repressed otherwise: 'The historical analogy helps us to see that the theory of action and, more precisely, of the relations between the agents and objective conditions (or structures) that is implemented by economics constantly oscillates, from one text to another and sometimes from one page to another, between an objectivist vision that subjects freedoms and wills to external, mechanical determinism or an internal, intellectual determinism and a subjectivist, finalist vision that substitutes the future ends of the project and of intentional action, or, to put it another way, the expectation of future profits, for the antecedents of causal explanation' (Bourdieu [1980] 1992, 46). 5. 'The definition of the process of science as a dialogue between hypothesis and experience can, however, easily degenerate into the anthropomorphic image of an exchange in which the two partners take perfectly symmetrical and interchangeable roles. But it should not be forgotten that reality never has the initiative in this exchange, since it cannot reply unless it is questioned' (Bourdieu, Chamboredon, and Passeron [1968] 1991, 35f.) 6. This rejection of most theory can also be seen on another level. Matt Dawson, in one of his blog articles about the BSA annual conference's use of theorists, notes the peculiar fact that although Bourdieu is dominant overall in terms of abstract mentions, he is far less so within the theory stream of the conference (2015). 7. In fact, it is astonishing to see the uniform lack of agreement on these basic, epistemological issues with so many recognised Bourdieusians of all variations (see for instance Alexander 1995, 128ff;Atkinson 2016;Burawoy and von Holdt 2012).