Philip Kitcher’s Science in a Democratic Society treats science and values and particularly the value of science to a democratic society—as did his earlier Science, Truth, and Democracy (New York: Oxford University Press, 2001). However, the book under review takes the debate further than its predecessor in many respects. For one, it marks a much clearer break with the traditional conception that science’s role in society could be described as providing knowledge in a value-free fashion. Second, Kitcher’s new proposals on how science should ideally be integrated within a well-ordered democratic society are far more concrete and geared toward applicability. The new book also undertakes to integrate Kitcher’s own conception of a naturalized value theory, published in the same year under the title The Ethical Project (Cambridge, MA: Harvard University Press, 2011), with his political philosophy of science.

The main topic of the book is best outlined by introducing Kitcher’s concept of “division of epistemic labor.” This is not to be confused with the division of cognitive labor within a scientific research community, a subject to which Kitcher had devoted himself in earlier work (Philip Kitcher, The Advancement of Science: Science without Legend, Objectivity without Illusions [New York: Oxford University Press, 1993], esp. chap. 8). In contrast, the division of epistemic labor refers to the simple fact that in modern societies, knowledge and the capacity to generate and process it are as much subject to specialization as everything else. Whatever subject you choose, there are some people who know more about it than the average citizen, who accordingly relies on their expertise. As Kitcher observes, deference to expertise is not in itself inimical to the democratic ideal of a self-determined citizen—to the contrary, it is required. More reliable information than any citizen could generate all by herself is needed to ensure her capacity to formulate and pursue projects and political preferences that adequately reflect her interests and values (21; cf. Torsten Wilholt, “Scientific Freedom: Its Grounds and Their Limitations,” Studies in History and Philosophy of Science 41, no. 2 [2010]: 174–81, 177–78). Nonetheless, there is a tension between the democratic ideal of equality and “the plausible idea that, with respect to some issues, some people know more than others” (21). This is by no means an academic problem. Political conflicts often involve controversy over the proper division of epistemic labor—Kitcher discusses the political controversies over evolutionary theory, stem cell research, genetically modified organisms, and climate change. But how exactly should the division of epistemic labor in a democratic society work? It is this normative question that the book sets out to answer. “We urgently need a theory of the place of Science in a democratic society—or, if you like, of the ways in which a system of public knowledge should be shaped to promote democratic ideals. The central task of this book—and, as I shall suggest below, the central question for the general philosophy of science (if not for epistemology) today—is to present a theory of this sort” (25–26).

Before Kitcher begins to develop his positive proposals on this topic (chaps. 5–9), he attempts an analysis of what has been going wrong so far (chaps. 1–4). According to this analysis, our collective difficulties to come to a shared understanding of the right division of epistemic labor are rooted in at least four major problems. First, the phantom of value-free science still haunts public expectations (chap. 1). In his 2001 book, Kitcher had already stressed the importance of the insight that there can be no value-free decision on which research questions to pursue and which to abandon or postpone. Now his argument goes considerably further. Even when it comes to the procedures that determine which hypotheses are ultimately settled on as the accepted results of scientific inquiry, the influence of values cannot be excluded. The reason for this is ultimately that methods in the inductive sciences are never fail-safe and that each and every methodological decision involves weighing the question of how reliable the result of the particular investigation needs to be. The foreseeable consequences of getting it right or wrong must bear on this consideration in a way that inevitably involves value judgments. This line of reasoning, which has been used since the 1950s to show that scientific methodology is inevitably value infused (Richard Rudner, “The Scientist qua Scientist Makes Value Judgments,” Philosophy of Science 20, no. 1 [1953]: 1–6) and which has recently been revived by Heather Douglas (Science, Policy, and the Value-Free Ideal [Pittsburgh: University of Pittsburgh Press, 2009]) and others, is now clearly also endorsed by Kitcher (see esp. 34–35). In Kitcher’s view, the failure to admit that values play a great role in science, and the lingering myth that genuine science is value free, is the “deepest source of the current erosion of scientific authority” (40).

It is also connected to the second large problem he identifies, the widespread conviction that questions of value are inaccessible to reasonable debate (chap. 2), which he regards as the main cause why many scientists and commentators on science shy away from acknowledging any role for values in science. Kitcher’s effort to counter this prejudice builds on his own naturalistic genealogy of ethics (cf. Kitcher, Ethical Project). He views ethics as a social technology that has evolved from simple practices of avoiding social trouble into a refined system of values and norms. Two important corollaries to his account are that there are no experts with regard to questions of value but only the authority of the ongoing societal conversation and that this conversation can in fact consist of rational contributions intended to convince others that a certain constellation of values is more progressive than another—and not just of expressions of interests or tastes. This allows Kitcher to impose a constraint on the kinds of values that can play a legitimate role in scientific practice: they need to be “sustainable in an ideal conversation” (60).

As a third problem, Kitcher maintains that a superficial understanding of democracy often hinders our appreciation of the exact role of a system of public knowledge. He thus devotes chapter 3 to a discussion of democracy that highlights the importance of public knowledge for forestalling oppression—especially oppression of the kind that is not carried out by one easily identifiable oppressor. The fourth and last large obstacle on the way to a proper division of epistemic labor is then identified (in chap. 4) in our failure to understand that our current system of public knowledge, including the social arrangements that govern the generation and use of scientific information, has evolved by happenstance rather than by design. “Not much time, if any, has been devoted to wondering about how public knowledge might be shaped so as to be good for democracy” (100). This sets the stage for Kitcher to do just that in the remaining five chapters of the book.

The positive conception that Kitcher develops in chapters 5–9 is based on his ideal of “well-ordered science” (chap. 5), which is familiar from Science, Truth, and Democracy: if problems find their way onto the agenda of scientific research in accordance with their respective significance, Kitcher argues that the relevant notion of significance should ideally follow those judgments that a collective of ideal deliberators would agree on at the end of a conversation under mutual engagement in the course of which multiple points of view were considered and weighed by well-informed discussants. In Science in a Democratic Society, Kitcher extends this basic idea in two directions. First of all, while well-ordered science remains an ideal and the idealized kind of discussion imagined in its defining scenario is clearly impossible to realize in practice, as Kitcher acknowledges (115), he now offers some ideas on first steps that might actually take us toward approximating the ideal. Second, his ideas are extended and adapted to cover not only the procedures by which the agenda for inquiry is determined but also the submission of scientific claims to the system of public knowledge and their certification as well as their application. This wider focus of Kitcher’s new political philosophy of science is a consistent implementation of his expanded views on how deeply infused with value judgments the practices of science are.

A newly proposed ideal of “well-ordered certification,” for instance, requests that the levels of probability of getting the truth (and of proximity to the truth) that are aimed at in certifying pieces of scientific information and admitting them to the corpus of public knowledge are those that would be chosen in a process of ideal deliberation (149). It is complemented by an ideal of transparency: the method of certification should ideally be understood and accepted by everyone (151).

Kitcher’s method, which he applies consistently, is to first argue in the abstract for a certain normative ideal for one aspect or another of the proper division of epistemic labor, and only then to discuss possibilities of approximating the ideal in practice. In the limited space of this review, only some of his suggestions can be discussed. One recurring idea is the possibility to “create groups of citizen representatives, drawn from diverse segments of different societies,” who undergo tutoring in order to attain some understanding of the goings-on of contemporary scientific research and are “[led] behind the scenes” of cutting-edge research activities (129). This idea is first introduced in chapter 5, where it is envisaged that such tutored outsiders, if they subsequently engaged in discussion with wider communities of lay citizens, could improve the quality of public input for debates about the agenda for scientific inquiry.

Kitcher takes up the same idea when it comes to approaching the ideals of transparency and well-ordered certification. In actual practice, what is accepted as canonical public knowledge is subjected to distorting influences, such as the workings of people (labeled “resisters” by Kitcher) who are desperately trying to undermine the acceptance of truths they deem too bothersome to face (about global warming, say, or about the effects of secondhand smoke). Resisters often assume sinister and opaque value influences at work in the respective research. Tutored outsiders going behind the scenes, Kitcher suggests, could help to make all value judgments involved less opaque and to enable fruitful deliberation on whether, given what is at stake, the data are good enough to warrant the conclusions (164–65). Groups of tutored outsiders, Kitcher suggests, might be better placed to resolve disputes over the proper division of labor—such as the question of whether maverick dissenters on anthropogenic climate change ought to be taken seriously in public decision making—than a debate in a fully public forum, due to the difficulties involved in fully understanding all the relevant information (178).

Tutored outsiders are even taken into consideration for the role of helping with the organization of the division of cognitive labor: together with the research community, they might decide which mix of strategies is best suited for tackling the problems on the research agenda (216). This is a problem that is usually left to be sorted out automatically through the competition among scientists, but Kitcher rightly points out disadvantages of this mechanism (such as disincentivizing sharing of information and other forms of cooperation). So this, too, is an aspect of the research process that might be improved by a more cooperative approach (including public participation), by which Kitcher also hopes to cultivate an ethos of team playing (217).

Quite generally, Kitcher envisages modes of public participation to take over roles that are commonly reserved for the research communities themselves. This reflects a deliberately chosen line: “Central to this book is an argument that the insistence on autonomy should go, to be replaced by a very different image”—an image that includes an ethos of team playing and a commitment of scientific research to the common good (217).

The traditional insistence on the autonomy of science serves several purposes (cf. Wilholt, “Scientific Freedom”). For one, individual scientists’ autonomy in pursuing research strategies of their own devising may be assumed to enable them to make optimal use of their individual and local resources and preconditions for creative scientific work, about which they themselves have a privileged kind of knowledge. Would any team of biologists and tutored outsiders have had the wisdom, in the 1940s, to recommend a close study of the details of the developmental biology of maize in connection with its genetics—let alone to assign Barbara McClintock to the task? Even an only slightly centralized mode of decision making about the cognitive division of labor is bound to incur a loss in terms of the efficient utilization of the personal and local knowledge that is scattered over the community. Perhaps this loss would be compensated for by a gain in significance of the knowledge produced. But whether this is the case is far from obvious or easy to determine, so the possible consequences of curbing autonomy on the level of the cognitive division of labor remain precarious.

Another, perhaps more decisive, purpose by which the autonomy of science has been defended is to ensure its political independence. If the knowledge generated in science serves as an important basis for the development of well-informed political preferences among citizens, then political powers must not have control over the generation of knowledge on pain of undermining the democratic process. Two things ought to be granted at once: first, it is far from clear that the way scientific research is currently organized warrants its political independence. Second, Kitcher nowhere recommends to grant the government or the legislature direct control over scientific research. Entrusting not the established political institutions but groups of tutored outsiders with the task of providing public input to the system of public knowledge may in fact be one way to circumvent the problem—if it is possible to organize these new mechanisms in ways that ensure their independence from those whose offices, careers, and political agendas are influenced by the scientific results that find their way into the corpus of public knowledge. Again, I do not mean to dismiss the possibility of organizing public participation in such a way as to ensure the required independence. But I submit that the importance of political independence grounds a presumption in favor of the autonomy of science that requires a solid consideration before we settle on alternative modes of organization.

One aspect of the idea of tutored outsiders that particularly calls for discussion is the tutoring. On the one hand, it is a necessary ingredient of the idea, as citizen groups are supposed to serve as a remedy for the lack of specialized information and competence among the general public. On the other hand, tutoring seems to presuppose the possibility of giving a neutral or objective primer on the subject matter at issue. Kitcher himself considers the possibility that in contexts laden with controversy citizens who have received tutoring might be regarded by those opposed to the scientific consensus as having been “brainwashed” (223). He proposes a three-stage model of tutoring to counteract this problem, in which the citizens are first introduced to the basic concepts and standards of evidence, then confronted with challenges to the scientific consensus, and finally familiarized with the community response to such counterarguments. However, will the alienated dissenters not complain that one side has the last word in the process by definition? The assumption that the community consensus is epistemically privileged is built into the procedure. In some cases (such as climate change) there is ample justification for this assumption. (But still, what is the added value of groups of tutored outsiders if this assumption is already built into the tutoring process?) In other cases of controversy about the proper division of epistemic labor, it is less clear what the seminal community consensus even is (when it comes to effects of genetically modified organisms on the ecosystem, e.g., it can matter whether you ask the ecology community or the molecular genetics community; cf. Daniel Sarewitz, “How Science Makes Environmental Controversies Worse,” Environmental Science and Policy 7 [2004]: 385–403). Accordingly, it will be much harder even to define what would constitute a “neutral” primer for the tutoring of citizen representatives.

The tutoring is bound to aggravate problems familiar to practitioners of participatory science governance from the notorious framing issues: outcomes of public discussions are strongly influenced by which questions are asked, how exactly they are articulated and ordered, and so on. Framing often itself becomes a subject of controversy (cf. Rob Hagendijk, Peter Healey, Maja Horst, and Alan Irwin, “Science, Technology and Governance in Europe: Challenges of Public Engagement,” STAGE [HPSE-CT2001-50003], final report, February 2005, http://www.stage-research.net/STAGE/documents/STAGE_Final_Report_final.pdf). Whoever organizes groups of citizen representatives and “leads them” behind the scenes is invested with considerable power and at the same time confronted with an enormous challenge: to frame the process in such a way that it indeed enables an engagement with public concerns (and not just predefined issues that suit the policy needs of political actors) and at the same time inspires trust on all sides of the controversy.

So far, experiments on deliberative science governance have more or less been exercised as an end in itself, with little immediate impact on actual policy. If similar practices were to be employed in order to approximate the ideal of well-ordered science, and groups of tutored citizens actually had a say in science policy, that would be a major change in the game of deliberation itself. It is not difficult to imagine the kinds of pressure—the lobbying, bullying, and bribery—that citizen representatives would be subjected to if the outcomes of their consultation had actual policy consequences. The same pressures would be experienced by those involved with tutoring and other framing activities. Kitcher’s desire for an arena for deliberation on the ends and means of science that is exempt from the forces of misinformation and manipulation that distort public debate as we know it is probably shared by almost every thoughtful observer. But the question of whether and how it might be realized remains as open as ever.

The real achievement of Kitcher’s new book is that it provides a framework for a normative discussion on the place of science in a democracy. A short review can only hint at the detail and extent of this accomplishment—the convincing development of important basic concepts such as of the division of epistemic labor and the system of public knowledge, the careful separate discussion of the contexts of investigation, submission, certification, and application. And while some of Kitcher’s particular suggestions on how the ideals he derives might be approximated in practice require further discussion, they are each an important step toward a truly political philosophy of science. Most of all, this book will make philosophers of science aware that current systems of public knowledge should engage our critical creativity no less than the peculiarities of current political systems absorb political philosophers’ attention.