Foucault and Research Ethics

Foucault’s Ethics

A fruitful way of approaching the whole orientation of Foucault’s work is to note that he rejects one version of historicism while reconstituting another. There are at least two senses that have traditionally been given to this term in 20th century Western thought (see Mandelbaum, 1967; Scruton 1996, pp. 236-237 ). For Karl Popper, “historicism” refers to a teleological view of history according to which the future of human societies is laid out by some process of historical development, frequently one that is supposedly directed toward the realization of human ideals (Popper, 1957; see also Hayek, 1952, p. 1, chapter VII). Hegel and (at least early) Marx adopted a historicist perspective of this kind, and Berlin (1954, p. 11) gave it the label “metaphysical historicism.” The other sense in which the term “historicism” has been used, itself open to some variation in interpretation (see Beiser, 2011; Rand, 1964; Lee & Beck, 1953-4), refers to the view that different periods in history are characterized by distinctive, we might say incommensurable, systems of belief, and sharply divergent ways of life; and that these cannot be understood, even less evaluated, except in their own terms. From this point of view, there is no overall perspective in terms of which different historical cultures can be explained or judged, not even one that is immanent in the process of historical development in the manner that is assumed by metaphysical historicism. ⁵

Foucault clearly rejects the first kind of historicism. Like many of his intellectual generation in France, he reacted against the teleological meta-narratives built into both Orthodox and Western Marxism. For him, this was not so much a matter of playing down determinism so as to recognize human agency, but more radically of both insisting on the presence of historical discontinuities and rejecting any tendency to view the process of historical development in terms of a collective subject. This is part of a broader antihumanism that can be found across poststructuralist philosophy, and also in the historical epistemology of science developed by Bachelard and Canguilhem, which was very influential on Foucault (Gutting, 1989; Davies, 1997; Dews, 1995).

Foucault adopts a version of the second kind of historicism, even though he distances himself strongly from earlier forms of this which had been organized around ideas like “the spirit of the age” or typical modes of consciousness (Foucault, 1972, Introduction). Much of his work was concerned with identifying the governing “épistémès” or “discursive formations,” what might be thought of as historical a prioris of possible experience, operating at particular historical times, and the network of practices they entailed, these practices being bodies of rules to which people are “subjected” rather than products of their conscious decisions. ⁶ And he treated these as sui generis, rather than as part of any grand narrative or as representing fixed universal types. At the same time, a distinctive feature of his historical research was that it was aimed at understanding present-day modes of thinking and ways of life through documenting how they had come about. His aim was a “history of the present” (Roth, 1981) or more significantly, as Miller (2007, p. 17) puts it, to “historicize the present.” Referring to Foucault’s first book, Gutting (1989) writes:

Foucault’s history of madness was not written simply to provide an accurate overall picture of how people’s views of and treatment of the mad have changed through the years. It also had the critical intent of discrediting the idea that our contemporary conception of mental illness is [ . . . ] the objective scientific truth about what madness really is (p.105).

Moreover, Foucault presents his attempt to understand the present through studying the past as of value for its role in facilitating action aimed at bringing about a different future. His model here is, to a large extent, Nietzsche, who had criticized (under the heading of “antiquarian” history) what we listed earlier as the second kind of historicism for only understanding “how to preserve life and not how to create it,” complaining that it hindered “the mighty impulse to a new deed” (Nietzsche, 1893, para. 28). Foucault’s historical work was motivated by, if not precisely aimed at, facilitating creative new deeds. Thus, the conclusion he drew from historicism was that what we believe and how we act today are arbitrary, in the specific sense that they come out of a contingent history and are therefore themselves open to change. This contrasts with the more usual tendency for historicists to assume that those living within some particular historical culture will, and perhaps should, accept the legitimacy of that culture. ⁷

Thus, in the first volume of his History of Sexuality, where he starts to address the issue of ethics, Foucault (1990a) challenges the idea that we have liberated ourselves in our attitude toward diverse sexual practices as compared with the repression of the 19th and early 20th centuries. He emphasizes continuity here, how what occurred previously was not so much repression as a “discursive explosion” as regards sex—a preoccupation with differentiating forms of deviance. This still continues today, albeit perhaps with a different inflection from before, and it constrains, as well as enables, what we can be and do—perhaps to much the same extent as in the past. This line of argument parallels his claims elsewhere that the change from what we now regard as barbarous forms of punishment to therapeutic treatment amounts to a shift in mode of control rather than to a relaxation of it, and perhaps to one that is more insidious (Foucault, 1977a). The demand for self-examination, from religious confession through to modern counseling and occupational appraisal, is associated with the authority of various sorts of experts and of the bodies of knowledge on which they rely, and it structures our experience of who we are and our ideas about what we can and should do. In broad terms, his interest is in how our lives today have come to be directed by a particular set of ethical norms that give our existence a specific meaning and purpose without our being aware of their historical and contingent character.

So, much of Foucault’s work would suggest that we are ourselves constituted in and through the operation of sociohistorical discourses and practices. As has often been argued, this seems to leave little room for any kind of active role on our part, in other words for personal autonomy. There is, however, an apparent break with this position in the two later volumes of the History of Sexuality (Foucault, 1990b, 1992). ⁸ Here, instead of focusing on developments in 19th century medicine and psychopathology and how these shaped present-day attitudes and practices, the focus is on ancient Greek and Roman ethical practices. ⁹ Furthermore, rather than analyzing the contingent character of these practices, and how they were constituted in and through the power relations characteristic of Greek and Roman society (notably those around both slavery and patriarchy), we find Foucault preoccupied with the way in which ancient ethics prescribes “care of the self” (ϵπιμέλϵια ϵαυτOύ/epimeleia heautou) and the techniques it offers for this.

From this new point of view, “[ethics is the] process in which the individual delimits that part of himself [sic] that will form the object of his moral practice, defines his position relative to the precept he will follow, and decides on a certain mode of being that will serve as his moral goal. And this requires him to act upon himself, to monitor, test, improve, and transform himself” (Foucault, 1992, p. 28). In other words, the demand here is for “a certain way of attending to what we think and what takes place in our thought” (Foucault, 2005, p. 11), with a view to engaging in “actions by which one takes responsibility for oneself and by which one changes, purifies, transforms, and transfigures oneself’ (Foucault, 2005, p. 11). ¹⁰

Another model for Foucault in thinking about such “action on the self” was the cultural invention characteristic of modernist art. Faubion (2001, p. 87) notes that Foucault follows Baudelaire in adopting a modern attitude toward the value of the present that involves “a desperate eagerness to imagine it, to imagine it otherwise than it is, and to transform it not by destroying it but by grasping in it what it is” (Foucault, 1997, p. 311). An exemplar here is Baudelaire’s dandy, who “makes of his body, his behavior, his feelings and passions, his very existence a work of art” (p. 312). Rabinow (1997, p. xxxix) refers to this as “a form of continual self-bricolage,” one that involves not just resistance to prevalent practices but also active transgression. Faubion goes on to argue for the significance of Blanchot’s influence on Foucault in this respect, as one “who willfully transgresses the stylistic criteria that distinguish one genre from another, who conflates fiction with philosophical reflection, critique with narration” (Faubion, 2001, p. 87; on Blanchot see Haase & Large, 2001).

One way to understand Foucault’s position here is as a radicalization of Kant’s earlier radicalization of the notion of autonomy, in which all reference to nature (and therefore all external constraint) is excluded in favor of an idea of selfhood as rationality willing itself entirely from its own resources (see Pippin, 1991). However, for Foucault, as for Nietzsche and modernist art theory, imagination is substituted for rationality—in light of the allegedly oppressive and groundless character of any assertion of rationality, whether the instrumental form of reason associated with Western capitalist institutions or the dialectical reason of Hegel and Marx.

Implications Drawn for Research Ethics

In relation to social research, Foucault’s ethics can be read, first of all, as demanding reflexive interrogation of ethical norms, and indeed of methodological rules and established conventions more generally. This fits with the model that he provides in his own work. He operated in what might be called a postdisciplinary fashion, in the sense that his work conformed neither to the rules of philosophy nor to those of historiography—the two disciplines with which it is most obviously connected. In this way, like Blanchot, he sought actively to challenge existing “genres” in the development of “the history of thought” (see, for example, Foucault, 1989, p. 294). ¹¹

So, a first ethical injunction that has been derived from Foucault’s work is that researchers must subject to continual scrutiny prevailing ethical and methodological ideas—both those ingrained in institutional norms and practices and their own intuitions about what is good or bad, right, or wrong—thereby opening the way for new modes of research and forms of life. After all, Foucault sees research knowledge as always implicated in the operation of power. Indeed, many of his analyses were precisely concerned with this aspect of “the human sciences,” and the institutional practices associated with them. It could even be that the primary goal of research from a Foucaultian perspective should be to unsettle the systems of thought that underpin current social institutions and practices, including those of social science, to show that alternatives are possible. In these terms, much conventional discussion of research ethics, and of methodology more generally, would be seen as “an inhibition to creative, liberating and significant social research” (Pels, 2010, p. 1).

An equally important implication that has been drawn from Foucault’s work for research ethics is the idea that through empirical study of ethical ideas and practice in diverse societies we can come to understand what must be challenged and what should be accepted (Faubion, 2001). For example, Caduff (2011) argues that Foucault’s later work provides a model for how the anthropological study of morals could inform the ethics of anthropology. Building on the discussions of Fassin (2008) and Stoczkowski (2008), he suggests that this can provide a via media between “moral positionalism,” on the one hand, and a “cultural relativism” that amounts to a form of nihilism, on the other. He sees the kind of “critical” anthropology characteristic of, say, Nancy Scheper-Hughes (see Scheper-Hughes, 1995) as representing a version of moral positionism in which evaluations are made of people, institutions, and so forth, on the basis of the researcher’s own values, these simply being taken for granted. This contrasts with a more traditional cultural relativism that effectively undercuts any such evaluation because this would amount to an imposition of alien values on the culture being studied. ¹² What Caduff proposes, instead, is that a moral framework for evaluation can be developed internally to the discipline of anthropology through the study of ethical ideas and practices within societies, in a similar manner to the way that Foucault seeks to draw lessons from ancient ethics. ¹³

Needless to say, Foucault’s work has been subjected to considerable criticism, and, before reflecting further on what can be learned from it in thinking about research ethics, it is necessary to consider the problems it involves.

Problems With Foucault’s Position

Like many other philosophical perspectives on ethics, that of Foucault implies a critical attitude toward both prevailing moral codes and our own ethical intuitions. This is a feature not just of those views that are commonly associated with critique—such as the philosophies of Kant, Hegel, and Marx—but even those, most notably utilitarianism, that have come to be seen as constituting the prevalent ethical culture. ¹⁴ From Foucault’s point of view, as from these others, we must not treat acting ethically as a matter of obeying the dominant moral code or, even more subtly, of finding reasonable compromises among the values to which we and others happen to feel committed. Instead, there must be continual reflexive assessment not just of the moral norms that others seek to impose upon us but also of our own ethical commitments and beliefs. In other words, Foucault argues that we must seek to understand and challenge the sociohistorically constituted character of what passes for morality—objectively and subjectively.

At the same time, there is an important difference between what is offered by Foucault and other critical philosophies. The latter generally provide some means for deciding what should and should not be subjected to critique and in what terms, or at least they offer general guidance about the sort of evaluation required. For example, in the case of Kant, genuine duties are to be determined by using a principle that (he claims) any rational agent would adopt simply by virtue of being a rational agent. And he supplies arguments specifying what these duties include. By contrast, Foucault’s position does not provide any substantive guidance: there is an injunction to be critical, but no suggestion of principles that should guide the critique. ¹⁵ Similarly, while, drawing on ancient ethics, he emphasizes the use of techniques to craft the self in light of an “aesthetics of existence,” he does not tell us how we are to decide what sort of person to become: in other words, which aesthetic criteria are most appropriate or how to determine this.

Nor is this a matter of simple omission. Any such specification would be at odds with his interest in opening up the possibility of some quite new form of life, since the specification would necessarily depend upon current assumptions. As we saw, he is keen to emphasize historical discontinuity. Such a specification would also violate his principled rejection of general principles. Thus, along with many of his French philosophical peers (see Gutting 2011), Foucault rejects all of the various foundations that have been used to ground some alternative to existing moral codes or ethical concepts: whether appeals to human nature—either as manifested at all times, or as represented in some genuine form to be found in the historical past (Rousseau) or to be realized in the future (Marx)—or rationality—whether an a priori mode of reasoning (Kant) or some naturalistic “felicific calculus” (Bentham).

In the case of Foucault, then, we are faced with a theoretical problem, and one that has serious practical implications. While he insists that we must recognize the arbitrariness of whatever presents itself as an ethical demand, externally or internally, he provides us with no grounds on which to decide what should be resisted and what ought to be accepted, or how this could be determined; no indication even of toward what human ideals our efforts ought to be directed. Yet, from a commonsense point of view at least, it seems clear that not all new possibilities would be desirable. In effect, Foucault resists such evaluations on historicist grounds: we cannot know what the future will find of value or what it will denounce. But the result is that his position has no capacity to evaluate ethical demands, internal or external, in either theoretical or practical terms (Fraser, 1989, chapter 1; Ramazanoğlu, 1993, p. 11); though this certainly does not stop Foucault engaging in practical evaluations, as we shall see.

The problem can be illustrated by considering a legal case discussed by Foucault in volume 1 of The History of Sexuality (Foucault, 1990a). He describes Jouy as a “simple-minded” farmhand, who was arrested and imprisoned “till the end of his life” for the purposes of medical study (p. 32) after he had “obtained a few caresses from a little girl” (p. 31): he “would give a few pennies to the little girls for favors the older ones refused him” (p. 32). Foucault asks “What is the significant thing about this story?” and his answer is: “The pettiness of it all; the fact that this everyday occurrence in the life of village sexuality, these inconsequential bucolic pleasures, could become, from a certain time, the object not only of a collective intolerance but of a judicial action, a medical intervention, a careful clinical examination, and an entire theoretical elaboration” (p. 31). ¹⁶

What is interesting about this case, for our purposes here, is that while one could certainly view it in the way that Foucault does, on the evidence he provides it is at least as plausible to argue that it amounts to an example of child abuse, so that Jouy perhaps deserves the treatment he received. More importantly, the judgment Foucault provides in this case is not, and cannot be, derived from his philosophical principles but relies instead upon notions of what is (and is not) “consequential” and the idea that because something is common (an “everyday occurrence”) within a particular culture it is therefore legitimate—arguments that are very questionable.

In fact, the absence of philosophical guidance effectively leaves open the way for the exercise of dogmatism or whim. In terms of Foucault’s framework, his own arguments in this case cannot be exonerated from such a charge any more than could others. Gutting (1989) comments:

Foucault’s [ . . . ] concrete judgments about what is oppressive and what is not conform fairly well with the standard views of leftist intellectuals [at the time]. For him the major threats to human freedom are such things as bourgeois morality and the exploitation of workers. But why should we prefer these judgments to those of rightists who think labor unions, pornography, abortion and international Communism are real problems? (p. 283)

Foucault does not have any answer to these questions, other than by appealing to his personal preferences, along with his sense that the present is “intolerable” (Foucault, 1977b, p. 216, see also Glucksmann, 1992), and that something new is required—though we should note that, ironically, in the case of Jouy he is actually arguing in favor of village tradition.

So, the key issue for us here is not which of the two interpretations of the case is most persuasive, but rather that while Foucault takes a stand he does not do this on the basis of assumptions that are intrinsic to his philosophical position, and he could not have done so because his philosophy offers no guidance one way or the other. It could equally have justified dismissal of “village sexuality” as a normalizing patriarchal force that should be challenged rather than validated.

The conclusion we must reach is that Foucault’s position results in a kind of decisionism, whereby all evaluations are unavoidably arbitrary. Furthermore, as we have seen, it is not just that Foucault’s philosophy is incapable of providing crucial guidance, but that it specifically rejects any possibility of such guidance. At the same time, Foucault puts forward his own evaluations as obviously correct, denouncing opposing views.

Gutting writes about the case of Jouy: “Many of us today will be shocked at Foucault’s insouciance over what we might well judge sexual molestation, but [he] would no doubt see our reaction as itself a sign of the effects of the modern power/knowledge system” (Gutting, 1989, p. 95). While this might have been his response, he would have had no grounds for it. How are we to determine what does and does not derive from the modern power/knowledge system—if, in fact, anything does not—or to decide why some alternative framework, in which the activities of Jouy would be tolerated, or even applauded should be regarded as more desirable? That this is undecidable in terms of Foucault’s philosophy reflects the relativism long associated with historicism. If each historical era has a culture that is incommensurable with all others, we have no basis for judging them outside of themselves. The paradoxes that this generates become particularly obvious when historicism is used to “historicize the present.” ¹⁷

It appears that for Foucault, even more than Kant, the only important value is autonomy, formulated in this case as detaching ourselves from the effects of the power-knowledge system. Yet Foucault’s philosophical position cannot justify even his valorizing of resistance and transgression, or the positive valuation of novelty and difference that lies behind it. On what grounds should resistance or transgression be regarded as a virtue, and why should novelty be treated as the most important ideal? He simply takes it as a given that the present is intolerable, and that alternatives would be better. Moreover, we should note that what he champions here are, in effect, the characteristic values of modernism: the valuing of change and novelty has long been a prevalent feature of the dominant culture in Western society (see Oakeshott, 1962). This surely undermines any claim that Foucault has managed to “escape himself” or the culture that had formed him, and perhaps highlights the futility of this aim.

This connects to another problem, mentioned earlier: that any injunction “to resist” assumes that there is some part of us that is not constituted by what is to be resisted. In other words, the concept of resistance assumes some “I,” some agency, that is not constituted by the power-knowledge system. Yet it is not clear what grounds Foucault could have for believing that any such agency exists. Gutting (2005) writes “No doubt the reason [Foucault] so resisted any fixed identity was his realization that even what might seem to be his own autonomous choice of identity could be just an internalization of social norms” (p. 101). Within the framework that Foucault has developed we not only have no basis for deciding what to resist but also no grounds for believing that we have the autonomy required for doing this anyway. Our resistance could, without our being aware of it, be just as much a product of normalizing forces as our conformity.

Indeed, the general thrust of Foucault’s work as a whole is to suggest that there can be no escape from those forces. He writes at one point:

these “power-knowledge” relations are to be analysed, therefore, not on the basis of a subject of knowledge who is or is not free in relation to the power system, but, on the contrary, the subject who knows, the objects to be known, and the modalities of knowledge must be regarded as so many effects of these fundamental implications of power-knowledge and their historical transformations. In short, it is not the activity of the subject of knowledge that produces a corpus of knowledge, useful or resistant to power, but power-knowledge, the processes and struggles that traverse it and of which it is made up, that determines the form and the possible domains of knowledge. (Foucault, 1977b, pp. 27-28)

On this basis it seems that only the power-knowledge system itself could exercise any autonomy or agency. There is a paradox, then, that while, on the one hand, Foucault seems to treat autonomy as the most important ethical principle—interpreting it in an absolute way as the shaking off of all constraints—on the other hand his philosophy denies the possibility of autonomy, or at least that we can know whether and when we are exercising it. We are in the grip of a process that produces both who we are, and what we can be, as well as any knowledge we believe we have about this. Here, the very possibility of an ethics seems to be denied.

There has, of course, been a perennial tension within social science generally, including within Marxism, between recognition that we cannot be completely autonomous—that our thoughts and actions are always constrained in various ways—and an insistence that we must have (or believe we have) at least some autonomy or agency. In line with this, in his later writings on ethics Foucault seems to shift to a less radical view of the nature of autonomy, and to suggest how the problem might be solved. Faubion (2001) argues that:

[In] the first volume of The History of Sexuality, Foucault often seems of the opinion that our modern ‘‘liberal’’ polities diminish ethical possibilities nearly to a zero-degree. [However,] during the course of the long interruption between the first and second volumes of the History, he undergoes a change of mind. He never comes to presume that liberalism had led to anything approaching a true ethical commons, but does come to acknowledge in the present a greater array of ethical interstices than he had previously recognized. Such interstices are not removed from power. They are not, however, spaces or places of [ . . . ] political micromanagement or psychosocial bondage [ . . . ]. (p. 89)

According to this interpretation, then, in his later work Foucault recognizes variation in degree of autonomy rather than defining it in absolute terms; in other words, as either simply present or absent. And, it is suggested, he seeks to deal with the issue through the notion of “problematization.” Gutting (2005) explains what is involved as follows:

Problematizations formulate fundamental issues and choices through which individuals confront their existence. The fact that my existence is problematized in a specific way is no doubt determined by the social power relations in which I am embedded. But, given this problematization, I am able to respond to the issues it raises in my own way, or, more precisely, in a way by which I will define what I, as a self, am in my historical context. (p. 103)

Gutting (2005) presents Foucault as contrasting problematization with marginalization:

Marginalisation corresponds to the strongest constraints that a society exercises on individuals. Even the marginalized are not entirely determined by a society’s power structures, since they are capable of engaging (and succeeding) in revolutionary movements against what dominates them. But they can define themselves only through their struggle with power. The “mainstream” members of a society, those who are not marginalized, are less constrained. The power network defines them in a preliminary way but allows for a significant range of further self-definition. Unlike the marginalized, they have available “niches” within the society that provides them room for self-formation in their own terms. (pp. 103-104) ¹⁸

While treating autonomy as a matter of degree may be a more defensible position, it remains unclear how it can be reconciled with Foucault’s broader philosophy. For example, how can the claimed difference between problematization and marginalization operate when, according to Foucault, power is dispersed throughout the whole of society rather than centralized (Foucault, 1990a, pp. 92-97)? Given this dispersal, why would those at the centre have any more scope for autonomy than those at the margins? Furthermore, the practical question of what is an unavoidable constraint determined by the current problematization and what would amount to the exercise of freedom is left unanswered. And, here again, it seems as if the exercise of autonomy itself is all that matters, even if this is now more of a possibility than it was in the case of absolute autonomy. ¹⁹

A final problem is that it is not clear how any genealogical analysis we carry out, in the manner of Foucault, on the prevailing moral code—or, more broadly, on the institutional arrangements within which we live—or on our own ethical intuitions, can itself escape the skeptical questioning that Foucault unleashes. Much the same is true of any history or anthropology of ethics; at most all that these could do would be to loosen the grip of current preconceptions and commitments. But to what purpose?

Discussion: Lessons for Research Ethics

In the previous section, we have discussed some very serious problems with Foucault’s arguments. There may be a temptation to conclude from this that little can be learned from him in thinking about research ethics. We believe that this would be a mistake: there are both positive and negative lessons.

One positive lesson is that research ethics should not be entirely preoccupied with how researchers ought to deal with other people, notably the people they are studying—as it currently tends to be. Foucault’s ethical writings are concerned with “action on the self” (Foucault, 1992, p. 28), a focus that derives from the character of much ancient ethics, and that is inherited by modern “virtue ethics.” ²⁰ The idea behind this approach is that only if we acquire relevant virtues will we be able to act well in relation to other people. This is because acting well is not a matter of enacting principles—of following rules in a mechanical way—but rather of making wise judgments in particular situations about what is the best course of action. The implication of this for research ethics is that it should start from the question of what virtues are required of researchers. And, to us, this suggests that what is and is not ethical in dealing with other people must be judged in the context of the distinctive obligations built into the role of researcher, rather than simply being assessed in terms of general ethical principles. Moreover, the central obligation here is to pursue knowledge (Hammersley & Traianou, 2012). ²¹

A second positive lesson from Foucault is that we must be fully prepared to question institutional pronouncements about what is and is not ethically acceptable, as well as our own initial judgments and feelings about this. The implication is that we must adopt a highly reflexive attitude and seek to make our own independent assessments. ²² As researchers we necessarily act in contexts where other agents make declarations about ethical principles and put forward ethical judgments, ones which they sometimes seek to enforce upon us—whether this is university authorities, ethics committees, funders, gatekeepers, or people in the field. We must take account of these judgments, at the very least prudentially, but Foucault would demand that, however plausible and justified they may seem, we should not simply treat them as valid, nor do this on the grounds of respecting others’ views. Equally important, he insists that our own initial judgments about what would and would not be ethically justified ought to be questioned, since these too may be the product of normalizing sociocultural forces. ²³

A third lesson from Foucault’s ethics is that while the individual has a personal responsibility to examine critically both external ethical demands and felt commitments, this cannot be done on the basis of some absolute foundation, whether a set of principles derived from nature or a fixed, determinate method. Rather, it must be done in ways that recognize the endemic uncertainty surrounding what can and cannot be relied upon, in which the “origins” of ideas and feelings neither count unequivocally for nor against them. Thus, Foucault rejects the notion that we can achieve a single system of knowledge, a general theory, that will encapsulate everything and serve all purposes; insisting instead that we rely upon more local, situated, and acknowledged-to-be-fallible, forms of understanding.

While, as we have seen, he provides us with no grounds for justifying these forms of understanding, this problem can be addressed by drawing on contextualist arguments in analytic epistemology. A starting point would be Elgin’s (1996) notion of “considered judgment,” where the aim is to achieve “reflective equilibrium.” Here, the idea that we can engage in rational judgment is preserved despite recognition that all judgments are fallible, that there can be no absolute principles on which we can depend without fear of being misled. Furthermore, it might be argued that reflection is not only fallible but is also always situated, in the sense that it involves assumptions that, within the contexts concerned, must necessarily be treated as fixed, at least for the moment. And it must be stressed that this is a defensible stance, not one that needs to be corrected or even defended in the face of challenges from foundationalism or radical epistemological skepticism. It is true, as skeptics argue, that there can be no knowledge of the world that is logically founded upon empirical givens, nor any other means of establishing its demonstrable validity, but they are wrong to conclude from this that no knowledge is possible. Moreover, not even skeptics can stand entirely outside the web of existing belief to ask their challenging questions, and there are potential costs associated with doubt, just as there are with belief, so that the pertinence of those skeptical questions is always open to dispute (see Williams, 2001).

What might be concluded from this more contextualist epistemology is that the false binary according to which ethical requirements are either absolute or entirely arbitrary should be rejected. Rather, following Ross and some other philosophical writers on ethics (Ross, 1930), we need to recognize that all value principles and their implications are defeasible, so that what we give priority to should depend upon judgments about what is relevant, and what ought to be treated as a priority, in the circumstances. The implication here seems to be, rather against the grain of Foucault, that what is usually required is not total rejection of current demands, and of what exists, but rather a judicious assessment of these matters as against other concerns—in ways that are not unduly influenced by prevailing opinions or personal feelings.

Another important lesson we can take from Foucault trades on one of the main points that he seems to derive from ancient ethics. This is that we should neither demonize those desires or personal interests that are conventionally judged to be immoral, or of low status, nor at the same time believe that acting upon them is a form of liberation. Most obviously, the pursuit of sexual gratification is neither to be rejected, as in some influential interpretations of Christianity, nor to be freed from all restraint, in the name of personal fulfillment or sociocultural transformation, as some have argued (see, for example, Marcuse, 1969). A rather less seductive example of demonization, this time in the context of research ethics, is the commonly expressed denunciation of researchers as pursuing their own careers at the expense of the people they study. Not only does this, as a general criticism, rely upon an implicit set of questionable assumptions about the necessarily inequitable nature of the “exchange” involved in the researcher-researched relationship (Hammersley & Atkinson, 2007, chapter 10), but it also appears to deny that pursuing a career is a legitimate aspiration. This stems from the fact that the notion of career is effectively being interpreted as the pursuit of personal gain. Yet this is a very narrow and reductive assumption—one that neglects any notion of vocation or even a more mundane commitment to the value of one’s work. At the same time, we should not assume that pursuing a career is always justified: whether or not it is, at what cost, and what this entails, are matters for situational judgment.

There are also negative lessons to be drawn from Foucault’s work. Above all, we can learn that great care needs to be exercised in handling the concept of “autonomy.” If we think of it in absolute terms, as Foucault does most of the time, then we are locked into a framework which is likely to lead to the conclusion that autonomy is impossible, that there can be no agent who is not entirely constituted in and through sociocultural forces of one kind or another, and that (so the argument goes) we are therefore entirely constrained. ²⁴

Furthermore, it is not even clear that autonomy in this absolute sense could be a coherently thinkable concept: there must always be some “I” exercising autonomy, and this must necessarily involve, at any point in time, at least some fixed substance, however thin, which will by its nature involve potential constraints. Moreover, in practice if an agent is going to be able to act effectively in the world then this substance will have to be relatively “thick,” in other words to include all manner of capabilities and tacit knowledge. A newborn baby is a relatively thin agent and is, as a result, relatively incapable and dependent. So, a more effective agent would have to be “thicker” and thereby also, ironically, more constrained internally. This is the other side of the idea that power is enabling not just constraining. Foucault recognized this, of course, but in our view did not follow through on its implications. In effect, most of the time, his philosophy is the obverse of Sartre’s: it is locked within the same, unsatisfactory, framework whereby, in principle, we must either be absolutely free or totally constrained. What is required instead is recognition that autonomy and constraint are normative more than ontological matters—they are evaluations; and situationally relative ones at that.

There is at least one other lesson we might learn from Foucault. If we draw from him the conclusion that research ethics should focus on identifying the virtues researchers ought to have, we can note that at least some of these will be epistemic in character; in other words, they will be concerned with what is required on the part of researchers if knowledge is to be pursued effectively—and as we saw earlier this links up with Foucault’s later interest in ancient ideas about “truth-telling.” Included amongst these epistemic virtues would be, for instance, being explicit about changes in one’s own views and the reasons for these. Equally important would be a commitment to trying to be as clear and precise as possible about what one is claiming and why, and about the appropriate degree of confidence there should be in the likely validity of any conclusions offered.

Now, it seems to us that, in practice, Foucault did not match up well to these requirements. It was noted earlier that he shifted his position significantly over the course of his work, and particularly in writing the last two volumes of his History of Sexuality. Yet he does not explicitly acknowledge this. ²⁵ Indeed, he famously rejected any requirement placed upon him as regards consistency: “Do not ask who I am and do not ask me to remain the same: leave it to our bureaucrats and our police to see that our papers are in order. At least spare us their morality when we write.” (Foucault 1972, p. 17). However, a concern with consistency, and honesty about changing one’s mind, is not simply the morality of bureaucrats or the police; and it is of particular importance in the context of intellectual work—if collective discussion, so important for the production of research knowledge, is to operate effectively (Hammersley, 2012). Of course, Foucault might well have argued that consistency is of little importance by contrast with continual transgression in search of new possibilities. But, as we saw, in fact his philosophical position can offer no justification for that futile project; and, in practice, his work was clearly aimed at contributing to academic knowledge, where the epistemic virtues we have identified are of obvious importance.

Similarly, despite Foucault’s recognition of the difficulties and uncertainties involved in producing knowledge, he took positions and denounced other views in a manner that implied a much stronger basis for his own conclusions than he could justify. His judgment in the case of Jouy, discussed earlier, is a case in point. Gutting (1989)—a very sympathetic commentator—writes as follows, referring to one of Foucault’s early books, Madness and Civilisation:

the rhetorical force of Foucault’s prose often goes far beyond the evidentiary force of his arguments in condemning modern psychiatry. [ . . . ] Here his view is particularly distorted by his prejudice against bourgeois morality and by his romantic desire to see madness as an infrarational source of fundamental truth. Foucault simply assumes that a moral system built around respect for property and family relations strikes at the root of the individual’s autonomy and happiness. His overwhelming emotional aversion to such a morality leads him to reject it in toto, with no analysis of its precise values and deficiencies. [ . . . ] Similarly, his romanticization of madness as a source of deep truth blinds Foucault to the fact that, quite apart from any exclusion or exploitation by society, the mad are often cripplingly afflicted by the distortions and terrors of madness itself. (pp. 108-109)

Summarizing, Gutting comments: “[ . . . ] alternatively befogged by the tortuous opacities of his prose and dazzled by the seeming gratuitousness of his audacious claims, [readers] may well ask if there is anything at all here worth their while’ (Gutting, 1989, p. 261). His answer to this question is that there is, and we agree. But the deficiencies of Foucault’s work must also be recognized.

There are, then, both positive and negative lessons to be learned from Foucault’s later writings, and they carry quite profound implications for our thinking about social research ethics.

Conclusion

Our aim in this article has not been simply to apply Foucault’s ideas to research ethics but to use critical engagement with them to facilitate thought about this topic and also about those ideas. We started from the importance of researcher autonomy, and we have explored Foucault’s views on ethics to gain some illumination of its character and importance. We noted that the most obvious injunction that follows from his views requires that prevailing moral codes and ethical intuitions must be subjected to reflexive assessment. In our view, this emphasis is of considerable value in the context of research ethics today, not least because there seems to be great pressure to believe that acting ethically is exhausted by conformity to codes and regulations, or to the decisions of ethics committees. Meanwhile, there are those who want to insist that principles of social justice or human rights lay down absolute requirements on how others must be treated, and that these should be placed at the centre of the research process (Mertens & Ginsberg, 2009). In the spirit of Foucault we reject both of these contemporary tendencies.