Seeing the point from which you see what you see: An essay on epistemic reflexivity in language research

This essay deals with epistemic issues in language research, focusing particularly on the field of language planning and policy (LPP). It outlines Pierre Bourdieu’s principle of epistemic reflexivity as a device for understanding what the view of the research object owes to the researcher’s past and present position in social space. I hold that developing such an understanding is particularly vital for LPP scholars, by virtue of the ways in which the objects investigated here tend to linger in the borderlands between science and politics. Accordingly, the essay unearths the philosophical roots of epistemic reflexivity and highlights some of its implications in the research practice with examples from Swedish LPP research. It also examines the value of a reflexive stance in interviews as a way of pinpointing the relevance of epistemic reflexivity in every moment of the scholarly investigation. In conclusion, the argument is that since epistemic reflexivity is a useful device for any critical researcher who wishes to grasp the knowledge he or she produces, it is so also for language researchers, and particularly so in relation to the ideologically normative practices of LPP scholarship. Therefore, a reflexive gaze is a pivotal driver for yielding better language research.


LAnguAge, poLiticS, And LAnguAge poLiticS
The title of this essay derives from the work of the French sociologist Pierre Bourdieu, whose writings on academic knowledge production deal at length with the relationship between the view of the research object, on the one hand, and the viewpoint of the researcher, on the other. For Bourdieu, understanding this relationship offers researchers "a chance of seeing the point from which you see what you see" (Bourdieu, 1989, p. 18-19), and thereby grasp the knowledge that is produced. Bourdieu discusses this matter as one of 'epistemic reflexivity.' The present work aims to essay some applications of this idea in the practices of researchers, including language researchers. By using it, I propose, they can avoid being "the toy of social forces" in their research practices (Bourdieu & Wacquant, 1992, p. 183, emphasis removed). Therefore, it is at once a pivotal driver for yielding better research.
People in general -including researchers -have strong sentiments attached to languages and linguistic practice, and this fact seems to be particularly salient when such languages are perceived as being "theirs" -their mother tongue, their heritage language, etc. Language thus embodies all kinds of imaginaries with important bearings on people's investments and senses of selves. Like other intellectual inquiries, then, language research is faced with pivotal questions about the status of academic knowledge, and the fact that when researching language, we "bring our biographies and our subjectivities to every stage of the research process, and this influences the questions we ask and the ways in which we try to find answers" (Cameron et al., 1992, p. 5). It should be stressed therefore that many research topics concerning language are vested in language ideologies, that is, "socially positioned and politically interested constructions of language and communicative processes" (Briggs, 2007a, p. 589). I here posit that this problem is particularly pertinent in research within the field of language planning and policy (LPP). LPP practices are at heart language ideologically normative; that is, their stakes and interests center on the politics of a desired language situation (Canagarajah, 2005, p. 153). Hence, as Silverstein argues: Professional students of these transformative phenomena are, perforce, themselves engaging in a kind of explicit, necessarily ideological discourse about them. In its ideological aspects, to be sure, such discourse manifests a range of sociocultural positionalities of imagined linguistic projects within the global and national orders. (Silverstein, 1998, p. 421) From this vantage point, we can say that language planners and other intellectuals alike have language ideologies (e.g., Salö, 2014;Spotti, 2011), in the sense that they embody the values and beliefs of the social worlds where they have learned to think and act (Bourdieu & Wacquant, 1992, p. 136). The present contribution dwells on a metaprinciple apt for dealing with this issue -namely, epistemic reflexivity as a never-ending process of critical self-reflection which offers researchers a disposition for grasping these social worlds as an inroad to understanding the principles of their knowledge production (Brubaker, 1993). As a research tool, it is useful for any scholar who seeks to develop an epistemological take on his or her relationship to the object he or she has undertaken to study. Hence, it is also of value to scholars of language.

the SocioLogy of the SocioLogicAL eye
Bourdieu's stance on research and epistemology owes much of its foundations to the French philosopher of science Gaston Bachelard . Commonly labeled as an applied rationalist, Bachelard wrote extensively on historical epistemology and the foundations of the scientific mind (e.g., Bachelard, 2002Bachelard, [1938; Tiles, 1985 andBroady, 1991 offer overviews). As outlined by Broady (1991), Bachelard's philosophy of science can be summarized by three broad points. Firstly, it starts from the general proposition that, in essence, science composes a break with everyday mundane thinking, and the spontaneous representations of common sense. It follows from this premise that, secondly, the scientific object must be constructed and therefore not be taken for granted. Thirdly, the researcher's relation to the object should be analyzed as a dimension of the knowledge about that same object. Bachelard's key insights into these matters have had an impact on generations of scholars in France and elsewhere, not least of all in his view of critique as an essential means for overcoming the "epistemological obstacles" that hamper the progression of scientific thought (e.g., Ross & Ahmadi, 2006). Bachelard's insights also came to establish the basis of Bourdieu's take on reflexivity -epistemic reflexivity -that is, the integrated, systematic, and continuous device of the research practice whereby the analyst breaks with his or her own pre-given viewpoints, which are often found built into the research questions, theories, concepts, and analytical instruments that he or she has inherited (e.g., Wacquant, 1992, p. 36-46).
Frequently discussed under labels such as "socioanalysis", Bourdieu's take on reflexivity builds on the principles of Bachelardian thinking: Through rupture with the spontaneous thinking of common sense, "[t]he social fact is won, constructed, and confirmed" (Bourdieu et al., 1991, p. 57). The risk involved in constructing the object of inquiry, posits Bourdieu, is that the researcher naïvely imports into the research practice, as he puts it, "all that the view of the object owes to the point of view, that is, to the viewer's position in the social space and the scientific field" (Bourdieu, 1993a, p. 10). This is an issue fundamentally due to the fact that scientific knowledge can be obtained only by means of a break with common sense -the primary representations or "pre-notions" in Durkheim's vocabulary -in other words, the sort of mundane knowledge about the research object that the researcher has uncritically acquired elsewhere in the social world (Bourdieu, 1989, 15;Bourdieu & Wacquant, 1992, p. 235-238). In the language sciences, Cameron (1990, p. 81) similarly regrets what she sees as a bad habit in much sociolinguistic research: the acceptance and subsequent import of sociotheoretically naïve concepts (her examples being "norm" or "identity") -"used as a 'bottom line' though they stand in need of explication themselves." Bourdieu deplores this mistake; for one, in his own work the refusal to borrow common categories has implications down to the level of prose, where he is at pains to avoid the commonsensical understandings "embedded in common language" (Wacquant, 1989a, p. 31). Jenkins (1992), for example, has criticized this position for the reason that it makes Bourdieu's writings difficult to read and understand (also e.g., Burawoy, 2012, 20). Bourdieu, contrarily, sees this trait as an important technique of keeping science free from the everyday discourse on the social world, "the discourse of the semi-wise" (Bourdieu & Chartier, 2015, p. 29). For Bourdieu, then, the easy and readable style is thought of as dangerously manipulative, in that simplified discourse serves the end of oversimplifying knowledge about the social world, consequently found in the false clarity of dominant discourse (Bourdieu, 1990a, p. 52; see also Wacquant, 1993, p. 237, 247f. and note 5 there).
Allied to that, epistemic reflexivity is vital in cases in which analysts are a part of the group or "set of observers" whose apprehensions they aim at unraveling (e.g., Bourdieu, 1988). Clearly, this feeds into a well-known insider-outsider dilemma. On the one hand, argues Bourdieu, "one cannot grasp the most profound logic of the social world unless one becomes immersed in the specificity of an empirical reality, historically situated and dated' (Bourdieu, 1993b, p. 271). In this sense, indeed, being a member of the group that is investigated might well buy the researcher entrance into his or her key social worlds. The crux of the matter, however, is that the price paid for this insider's access is the overwhelming risk of not seeing the viewpoint from which it is stated, and thereby producing an account which says exactly what the researcher's position in the field allows him or her to say -and nothing else (Broady, 1991, p. 548; see also Bourdieu, 1990a, p. 183-184). Since the researcher, by this logic, is imprisoned by the field, reflexivity is brought up to date as a question of understanding and, subsequently, handling one's own position and dispositions, as handed down by one's field. Hence, one's only hope of producing scientific knowledge -rather than weapons to advance a particular class of specific interests -is to make explicit to oneself one's position in the sub-field of the producers of discourse […] and the contribution of this field to the very existence of the object of study. (Bourdieu, 1983, p. 317)

Science And poLiticS, And borderLAndS
To be sure, epistemic reflexivity in the research practice holds ramifications to social scientific research more generally, by virtue of the intricate ways in which its social objects of study tend to linger "between the scientific and the political registers" (Wacquant, 2009a, p. 125). However, it would seem that adopting a reflexive posture is a matter of particular significance in work that deals with saliently ideological and interest-laden topics of research. In what follows, I shall dwell on a poignant example of this sort, which occupies a central position in LPP research: representations of English as a language problem in the protection of national languages (e.g., Hultgren & Thøgersen, 2014). Representations, after all, are "performative statements which seek to bring about what they state" (Bourdieu, 1991, p. 225). Pielke (2007, p. 116) holds that it is "characteristic of the science and politics of the early twenty-first century to see scientists actively engaged in political debates." At least ideally, a "policy" is a decision; "politics" is bargaining, negotiation, and compromise in pursuit of desired ends; and "science" is the systematic pursuit of knowledge. (Pielke, 2007, p. 37) However, this conception seems overly innocent, since, as Pielke (2007, p. 124-125) acknowledges, the systematic pursuit of knowledge is often enacted as a part of the political pursuit of reaching desired policy. I propose that the case of Swedish LPP adds insight into this dilemma, since the field by some of its properties "follows the logic of the scientific field, but by others it follows the logic of the political field" (Bourdieu, cited in Wacquant, 1989b, p. 17). To speak with Foucault, we can rightly say that while research is one form of discursive practice, LPP is another. In reality, though, this is a difficult line to draw. To do so, as Lynch (2000, p. 31) notes, requires a clear understanding of the boundaries of knowledge-yielding practices. In viewing such discursive practices as fields in Bourdieu's sense, this assumption seems to beg for more attention, for, just as individuals can be a part of several fields, fields also overlap so that knowledge can be yielded out of the logic of one discursive practice while speaking, as it were, in the voice of another. This fact blurs the distinction between science and language politics and complicates the task of refraining from objects of knowledge which are not the products of research practice.
Ideally, at least, university research is a critical enterprise with a salient heirloom to the core values of "freedom in the autonomous pursuit of truth" (Krull, 2005, p. 99). As discursive practices, science and LPP follow different logics in this regard, and are likewise dictated by different terms.
Understanding sociolinguistic phenomena may certainly be one objective of LPP; yet, it cannot be the only -such a conclusion overlooks what Ricento (2000) calls the strategy component of LPP. As a practice, ultimately, LPP is about deliberate efforts to influence the behaviour of others (Cooper, 1989, p. 45). Characteristic of the knowledge production here is the creation of what Cibulka (1995, p. 118) calls "policy arguments", by which he means the use of research to fit a predetermined position, aligned with a desirable policy outcome. Such knowledge can both be imported from the outer fields and filtered by virtue of the field's internal logic or it can be produced from within the field. The difficulty here is that "the borderline between policy research and policy argument is razor thin" (Cibulka, 1995, p. 118). For the viewpoint of the critical scholar, therefore, it is vital to develop a sociological eye capable of seeing what the view of the research object owes to the point of view of the observer, in other words, to the scholar's past and present position in social space (Bourdieu 1993a, p. 10). "The progress of knowledge", Bourdieu (1990b, p. 1) holds, "presupposes progress in our knowledge of the conditions of knowledge." Reflexivity, here, is what differentiates habitus from scientific habitus, in that the latter includes a disposition to grasp its own principles of knowledge production (Brubaker, 1993, p. 225): As Grenfell notes, The 'empirical individual' is like everyman, he responds naïvely to what surrounds him. The 'epistemic individual', on the other hand, is the product of scientific training and experience (Grenfell, 2007, p. 118) the breAK: SituAting refLexivity in prActice It ought to be clear by now that the issue at hand in this essay pertains to the reflexivity of the researcher, not that of the groups studied. To be sure, although foci may vary, "being reflexive" is a watchword in many strands of sciences (e.g., Mauthner & Doucet, 2003;Lynch, 2000). In Wacquant's (1992) opinion, the most novel facet of Bourdieu's take on reflexivity is that it brings to the fore aspects that go beyond the individual researcher and instead emphasize the individual's position in the field that he or she embodies. In this vein, it is primarily the field's epistemological unconscious that needs to be unearthed, rather than that of the individual researcher (Wacquant, 1992, p. 41). Epistemic reflexivity thus differs from narcissist conceptions of "researcher positionality" in that it focuses neither on the individual person as a scientist, nor on the scientist's "privilege" or the violence potentially present in producing knowledge about other people. Rather, it attempts at unearthing what the scientist's vision of the object owes to his or her position in social space (Wacquant, 1989b, p. 19). Certainly, the point is not that the researcher should 'confess' things -political affiliation or sympathies, ethnic or religious membership, etc. -in the research product, but to deal with such matters if and when they may have a decisive impact on their research practice. As Wacquant puts it, "epistemic reflexivity is deployed, not at the end of the project, ex post, when it comes to drafting the final research report, but durante, at every stage in the investigation" (Wacquant, 2009a, p. 121-122). As a matter of fact, Bourdieu disapproves of pursuits in which reflexivity is added decoratively, often post festum, and customarily serving the therapeutic aim of self-understanding. Accordingly, Bourdieu stresses time and again that reflexivity should not be a self-absorbed return to the individual scientist. However, as noted by Maton (2003, p. 59), while epistemic reflexivity is designed to be a collective reflexivity, scholars often end up exemplifying enacted reflexivity in individualistic terms. Admittedly, this poses a dilemma in work seeking to outline and exemplify epistemic reflexivity, and there might be no way of by-passing the issue of self-centrism entirely. Yet, it is feasible to account for the position of the individual without adopting an overtly individualistic focus so long as focus is placed on the relationship to one's research object, and the value-laden social worlds previously encountered. In this sense, Bourdieu's brand of reflexivity pertains to a form of self-analysis that does not privilege the self (e.g., Bourdieu, 2007). Seeking to illustrate this position, the following account seems appropriate for presenting the relevant positions in social space that I have occupied that have had a bearing on my relationship to my object of study: English in Swedish academia.
Reflexivity is pivotal in relation to the question of moving from one distinct professional universe into another, each offering their particular point of viewspositions from which analysts see what they see (Bourdieu, 1989, p. 18-19). Having undertaken university training in applied Scandinavian linguistics, and later pursuing a professional career at the Swedish Language Council, I entered the research practice as a socialized agent of the LPP field, armed with pivotal preconceptions with important bearings on the object undertaken for investigation. This fact, then, does not merely pertain to matters of embodying "Swedishness" on the part of the analyst, but is also intertwined with and amplified by a set of professional dispositions with focal values attached to the significance of Swedish in Swedish society, enmeshed in the categories, objectives, and interests imbued in Swedish LPP. In relation to this point, at the Language Council, I had already made a contribution to the debate on English in Swedish academia, where the perceived risk of "domain loss" served as the key rationale of the enterprise (Salö, 2010). While this circumstance does not necessarily equate to being a deeply immersed insider with perspectives genuinely embedded into the value systems of LPP (see Josephson, 2014 for a true insider's account in that regard), it points to matters of working within a language ideological consensus. This involves reproducing accounts that one knows will be positively sanctioned by the field, after having acquired shared dispositions to a particular language problem.
The Swedish field of LPP referred to above pertains to the contexts encountered through my prior experiences, viz. a state-mandated body for language planning and its base of recruitment. Subsequently, upon entering the research practice, an inherited vision of the object was my evident point of entry, associated with anterior dispositions acquired across the life-span. With that follows the importing of pre-defined categories as well as a view of English as a problem which resembles language ideologies and "the epistemological unconscious" of the LPP field (Wacquant, 1992, p. 41). The construction of scientific knowledge, then, ought to begin with a break with the preconstructed object, as a form of "radical doubt" about the commonplace representations it brings to bear (Bourdieu & Wacquant, 1992, p. 235).
How does one go about doing this necessary labour? There are probably different ways, but in my case it was attempted by initially unravelling the state and structure of Swedish LPP as a particular intellectual social space (Brubaker, 1993, p. 221), which was realized by historicizing English as a sociolinguistic problem in Swedish LPP (published as Salö, 2012 andSalö, 2014). Reflexivity, thus, is here enacted by the very effort of undertaking an opening analysis of the field of which the researcher him-or herself is the product and where previous investments have already been placed (e.g., Salö, 2010). "The first and most pressing scientific priority", posits Bourdieu, is "to take as one's object the social work of construction of the pre-constructed object" (Bourdieu & Wacquant, 1992, p. 229, emphasis removed). Studying processes of problematization, then, helps recast LPP as "a field of cultural or ideological production, a space and a game in which the social scientist himself is caught" (Bourdieu, 1985, p. 210). Blommaert (1996) explores the international emergence of LPP as a field, which he dates to the mid-20th century. As such, holds Blommaert, LPP emerged as a new market for the application of modernist sociology of language and macro-sociolinguistic research in the global south, more often than not in the context of colonial dominance. Swedish LPP, on the other hand, has a somewhat different historical trajectory. What is known in Sweden as "språkvård" comprises a salient strand of functionalist corpus planning with a longstanding history in Sweden, pertaining as it does to the institutionalized standardization and "cultivation" of Swedish (Dahlstedt, 1976;Teleman, 2003Teleman, , 2005. By and large, however, the component commonly known as "status planning" (e.g., Cooper, 1989) had long been neglected, since, as many scholars have pointed out, the position of Swedish was taken for granted for such a long period of time (e.g., Hult, 2005;Milani, 2007;Oakes, 2001). One possible explanation for this circumstance is that Swedish was established as the national language of Sweden already before the 19th century and, unlike neighbouring countries such as Norway and Finland, has not had its national sovereignty challenged in modern times (e.g., Josephson, 2002, p. 80f.). In short, there were never reasons to give much thought to the position of Swedish in Sweden, as it had been unchallenged de facto.
In the 1990s, however, agents positioned in the Swedish field of LPP identified English as a threat to the position of Swedish in Sweden. In this discourse, the role of English in Sweden was grasped by virtue of a particular set of available apprehensions. During the same period of time, discourses about languages other than Swedish were established in the Swedish society. In fact, for a while, Sweden had five official minority languages but no official majority language. Highlighting this fact in the debate allowed the discussion to center on the status of Swedish by clinging to discourses attached to the other languages, which is to say that Swedish came to be piggy-backing on the position of minority languages in order to secure legal status at the top of the linguistic hierarchy of Sweden. Teleman (2003, p. 234), a central agent in the Swedish LPP field, acknowledges that grasping the situation in the perspective of the minority languages was inspired by ideological currents deriving from post-colonial contexts. Be that as it may, since 2009 Sweden has enforced a Language Act dictating that "Swedish is the principal language in Sweden" (Språklag 2009, section 4), which is a phrasing that owes much of its existence to the perceived impact from English (Salö 2012(Salö , 2014. De facto and de jure, thus, Swedish is the language of the Swedish state; yet, "[t]he existence of a language is always a discursive project rather than an established fact" (Woolard & Schieffelin, 1994, p. 64). It is axiomatic that national languages largely owe much of their existence to romanticist ideology and state formation (e.g., Hobsbawm, 1990). Ultimately, then, those who struggle for the unification of such markets likewise struggle for the upkeep of recognized domination (Bourdieu, 1977a, p. 652). By Bourdieu's logic, it is not an exaggeration to say that the maintenance and protection of Swedish is an object of inquiry "overladen with passions, emotions and interests" (Bourdieu, 1990a, p. 52). A reflexive posture concerning threats to national languages entails understanding a language problem, on the one hand, as a perceived social problem, one with bearings on people's investments, or their deep-seated feelings about their mother tongue, identity, and national belonging. On the other hand, it entails understanding it as a sociolinguistic problem in the sense of a scientifically legitimate problem (cf. Wacquant, 1989a, p. 55). In this context, Park and Wee state that [t]he characterization of a "language problem" usually reflects the apprehension of a social situation from the perspective of a particular observer or set of observers. In other words, what counts as a problem usually reflects the interests or ideological stances (even if subconsciously) of a particular group -and this is particularly so when language issues are involved. (Park & Wee, 2012, p. 23) Hence, grappling with these matters I came to realize that one can question the extent to which the threat against Swedish exists, as Foucault would put it, independent of the discourses about it. Particular representations of English as a sociolinguistic problem, moreover, were inherited from the context in which I had SALÖ © Salö and CMDR. 2018 dwelled -Swedish LPP -which had thus contributed extensively to the existence of the object of study. Reflexivity, in short, helped me see that LPP is politics on the battlefield of language, and as such it is about representing things, changing things, with words: The social world is the locus of struggles over words which owe their seriousness -and sometimes their violence -to the fact that words to a great extent make things, and that changing words, and, more generally representations [...] is already a way of changing things. Politics is, essentially, a matter of words. That's why the struggle to know reality scientifically almost always has to begin with a struggle against words. (Bourdieu, 1990a, p. 54) As can be seen, Bourdieu urges researchers to engage in a struggle against political discourse as part and parcel of developing a scientific -not scientistic -gaze. Does LPP research need straightforward distinctions between what is scientific and what is not? While Bachelard does not side with the positivist position that scientifically procured facts are value-free (e.g., Cameron et al. 1992, p. 6), he does share the view of figures like Popper that science should aim at objective knowledge, distinct from the knowing subject (Tiles, 1985, p. 43, 48ff.). Bachelard, however, preferred the notion "objectivation", thereby repositioning the urge for objective truths by emphasizing that the quest for objectivity is an activity, a line of work undertaken by the scientist (Broady, 1991, p. 347). My present-day position on this vast and deep-seated matter is that this aim is pivotal but immensely difficult, yet conceivable by adopting a reflexive posture (e.g., Bourdieu, 1983, p. 317).

interviewS And refLexivity
While many of the matters of epistemic reflexivity presented thus far might appear to the reader as pertaining mostly to the initial stages of research processes and practices, this is really not so. On the contrary, epistemic reflexivity has its place throughout; it is designed to be continuously and systematically implemented in every moment of the research practice and thus digs deep into the craft of the research practice. For example, it pertains to the formulating of one's interview questions, to interviews as situated and power-laden events in themselves (Bourdieu, 1996b;Briggs 1986;Slembrouck, 2004), as well as to transcribing (e.g., Bucholtz, 2000;Ochs, 1979;also Bourdieu, 1996b, p. 30ff.), etc. Maton (2003) has critiqued the foundations of Bourdieu's version of reflexivity on the premise that reflexive knowledge can and should also be subjected to reflexivity. In short, there is no way of knowing when to stop being reflexive. Both Bourdieu and Bachelard are aware of this fact; as Bachelard posits, "objective knowledge is never complete […] since new objects never cease to provide new topics of conversation in the dialogue between the mind and things" (2002, p. 243). Complying with this viewpoint, it would be inaccurate and indeed unrealistic to claim that every potential aspect of this exercise has been systematically implemented to the full extent in any scholarly work. Then again, it may be questioned whether a fully-fledged reflexive research trajectory is even possible, as there is always room for more reflexive thought (Maton, 2003, p. 59). As noted, in my work, reflexivity has had purport mostly in respect to the ways I have attempted to handle my dispositions and position in relation to the field where I previously dwelled and where I had therefore placed my investments. But it has also been a relevant instrument in the production of knowledge through interviews, and below I comment on some of the insights that were gained and difficulties encountered.
It goes without saying that all methodologies have their problems. In sociolinguistics, many studies are open to some of the manifold methodological problems that arise in studies where interviews are used. Scholars who have written critically on these topics recurrently point out that as a communicative event, the interview is skewed and situated, and accordingly yields data that should be thoughtfully interpreted (e.g., Briggs, 1986Briggs, , 2007bMertz 1993). In consequence, while in some respects it can be advisable to think of the interview as a conversation (e.g., Blommaert & Dong, 2010), analysts are often advised to keep in mind that, in actual fact, it is not an ordinary conversation. Rather, "[i]t is a deliberately created opportunity to talk about something that the interviewer is interested in and that may or may not be of interest to the respondent" (Dingwall, 1997, p. 59). Often, the questions asked by the analyst presuppose "certain sustainable metapragmatic starting-points", which may or may not correspond to informants' assumptions (Mertz, 1993, p. 160). One effect of this, naturally, is that the researcher can quite easily -oblivious to the fact or notproduce an account and thereafter pick some quotes "to illustrate a previously determined position on some personal or political issue" (Dingwall, 1997, p. 52). These issues should be acknowledged. However, as De Fina and Perrino (2011) note, much of the literature that seeks to critically scrutinize interviews as a source of bias in social scientific research seems strongly attuned to overcoming the perceived problem of interviews as "unnatural" contexts, which in itself is a problematic conceptualization. In my view, the issue resulting from using interviews is not primarily that the researchers carry out an analysis on a piece of data that they themselves have created -which is true, yet possible to overcome. Instead, the issue as I see it pertains to a point raised by Hymes (1981, p. 84), namely that "[s]ome social research seems incredibly to assume that what there is to find out can be found out by asking." Broadly, this critical comment aims at shedding light on the methodological pitfall of believing that people have more opinions about most things than what is regularly the case, which is a stance shared by Bourdieu (Blommaert & Dong, 2010, p. 3).
Bourdieu's position on interviews, thus, is somewhat similar: "It is the investigator who starts the game and who sets up its rules" (Bourdieu, 1996b, p. 19). However, Bourdieu goes further in arguing that interviews are problematic because they are linked to the inherent difficulties involved in having informants producing adequate accounts of their own practices.
Social agents do not have an innate knowledge of what they are and what they do: more precisely, they do not necessarily have access to the central causes of their discontent or their disquiet and the most spontaneous declarations can, without aiming to mislead, express quite the opposite of what they appear to say. (Bourdieu, 1996b, p. 29) In this quote, Bourdieu reveals his stance on reflexivity, which links to the general issues of agency raised in Bourdieu's framework. This pertains to what Ortner (2006, p. 111) sees as Bourdieu's "insistence on the inaccessibility to actors of the underlying logic of their practices." Indeed, while Bourdieu does not posit that agents are totally unaware of what goes on around them, he maintains that they grasp it differently. As he puts it, they do not "have in their heads the scientific truth of their practice which I am trying to extract from observation of their practice" (Bourdieu, 2003, p. 288). Consequently, as he notes elsewhere: "Workers know a lot: more than any intellectual, more than any sociologist. But in a sense they don't know it; they lack the instrument to grasp it, to speak about it" (Bourdieu & Eagleton, 1994, p. 273).
To many, this position is provocative. According to critics, by downplaying the informants' abilities to reflect upon their own practices, reflexivity becomes framed as the researcher's privilege only (Archer, 2007, p. 43;Lynch, 2000). It is clear that Bourdieu sees reflexivity as a key difference between habitus and a trained, scientific habitus (Brubaker, 1993). We can say that Bourdieu demands of the researcher to develop an eye capable of projecting an image that goes beyond what the people who are studied are capable of grasping. The problem is not necessarily that people will have nothing to say, but rather that they have not necessarily given much thought to the kind of matters that interest the sociolinguist. Thus, the question is how to deal with informants' accounts " [g]iven that one can ask anything of anyone and that almost anyone always has enough good will to give some sort of answer to any question" (Bourdieu et al., 1991, p. 42). For Bourdieu asking people about their point of view must be supplemented with an understanding of the point of view from which it is stated (Bourdieu, 1996b, p. 34).
One way of dealing with this intrinsic issue, as Dingwall (1997, p. 56) points out, is this: "If the interview is a social encounter, then, logically, it must be analysed in the same way as any other social encounter." Here reflexivity serves a device for understanding and mastering these distortions' (Bourdieu, 1996b, p. 18). Building on such insights in his work on reflexivity and sociolinguistic interviews, Slembrouck (2004) accordingly views the interview situation as a meeting between two habitus. By this logic, the research interview is also intrinsically linked to the linguistic market in which it unfolds and the particular notion of legitimate language that applies there (Slembrouck, 2004, p. 93). The power immanent in the interview events can be brought to bear in important ways. Indeed, informant accounts can be inclined to reproduce dominant conceptions of what is acceptable, conceivable, or normal, which in turn reflect the imperatives of power hierarchies beyond interpersonal relationships (Bourdieu, 1977b, p. 37). In analyzing interview accounts, therefore, it is important to add the social relation between the interviewer and the interviewee that censors discourse by making some opinions seem inexpressible or practices unacceptable (Bourdieu, 1996b, p. 25). For these reasons, it is easy to side with Briggs (1986), who holds that, generally, interviews should be complemented with other data sources. Interviews provide accounts of the practical experience of agents, and, as such, they are "situated performances in and of themselves" (Heller, 2011, p. 44).

concLuSion: the eye which SeeS itSeLf
Viewed through the prism of Bourdieu's reflexive sociology, research is a form of interested practice governed by a scientific habitus (Brubaker, 1993). To be sure, this holds ramifications for analysts, who are cultural producers with a stake in their own object, and who also bring their "spontaneous knowledge of the social world" (Bourdieu, cited in Wacquant, 1989a, p. 44) to their research practices (see Bourdieu, 1993a, p. 8ff.). Like other kinds of social research, language research is at times vested in ways which affect its eventual outcomes. As a case in point, struggles to achieve particular language conditions pertain to the backbone objectives of LPP practices, which, after all, are about "how things ought to be", not about what they are (Canagarajah, 2005, p. 153). Subsequently, LPP research produced in the borderland-like space between science and politics runs the evident risk of ending up showing and saying exactly what one would have expected it to show and say, based on the position -social, academic or otherwise -from which the research was produced. Often, this is because scholars embody the values of the group or object they investigate and, all too often, fail to create a rupture with their inherited view of the problem they investigate. However, as I have sought to signpost here, it is indeed possible to overcome this problem by adopting a reflexive posture. This work is vital in the process that Bachelard (2002) calls "the formation of the scientific mind", but which can more straightforwardly be understood as the acquisition of a professional habitus: a scientific habitus, incorporated as "a disposition to monitor its own productions and to grasp its own principles of production" (Brubaker, 1993, p. 216).
The key proposition of this paper is the impetus for implementing Bourdieu's notion of epistemic reflexivity in the research practice. Epistemic reflexivity offers the critical researcher the intellectual means to equip oneself with the necessary means to understand one's naïve view of the object of study (Bourdieu, 1996a, p. 207) and thereby "avoid being the toy of social forces in your practice" (Bourdieu & Wacquant, 1992, p. 183, emphasis removed). As I have attempted to show, for Bourdieu, the construction of scientific knowledge begins with a break with the preconstructed object, as a form of "radical doubt" about the commonplace representations it brings to bear (Bourdieu & Wacquant, 1992, p. 235). Criticism in this sense advances knowledge, and on this point, I concur with the Bachelardian standpoint that unfounded assumptions are epistemological obstacles (Broady, 1991, p. 365). Adopting this device, then, entails a rupture with previous viewpoints, ultimately with the goal of producing a better sociolinguistic understanding of the objects we endeavor to explore. For language scholars, this principle may prove to be useful in the efforts of producing a new gaze, a "sociological eye" (Bourdieu & Wacquant, 1992, p. 251), but also an eye that, as it were, is capable of seeing itself (Wacquant, 1989b, p. 20). This reflexive gaze, I hold, is a pivotal driver for yielding better research.

AcKnowLedgementS
This essay builds on thesis work, and I am therefore in intellectual debt to a number of colleagues who in different ways contributed to the work produced, including Kenneth Hyltenstam,