Epistemic Values and Information Management

Abstract

The philosophy of information is concerned with the nature, management, and use of information. Thus, it should be able to help us make better decisions about how to manage information (e.g., decisions about intellectual property laws, collection development policies, and Internet evaluation techniques). These decisions have knowledge acquisition as one of their principal goals. Thus, one way to improve these decisions is to clarify our epistemic values. In this article, we combine epistemology and decision analysis in an attempt to assist people in this important task.

Keywords:

• applied epistemology
• decision analysis
• epistemic value
• knowledge acquisition
• knowledge management
• information ethics
• information management
• information science
• philosophy of information
• value hierarchy

ACKNOWLEDGEMENTS

We thank audiences at the School of Library and Information Science, Indiana University, the Department of Library and Information Science, Rutgers University, and the School of Information Resources and Library Science, University of Arizona for their feedback. We also thank Terry Connolly, Sandy Goldberg, Richard Healey, Pen Maddy, and especially Kay Mathiesen for helpful comments on earlier versions of this research. We thank the anonymous reviewers for The Information Society for several helpful suggestions. This research was partially supported by a Professional Development Award from the Social and Behavioral Sciences Research Institute at the University of Arizona.

Notes

1. Many eminent philosophers of the past have discussed how we ought to manage information. For example, in the Republic, Plato argues that it is a good idea to misinform citizens on certain subjects. Also, in the Phaedrus, he argues that a speaker who intends to misinform someone must know the truth about her subject.

2. As Floridi (2002a) has pointed out, the philosophy of information is much broader than epistemology. But epistemology is still a central component of the philosophy of information (Fallis, 2006b, 508). In survey articles on the philosophy of information, epistemology typically features prominently (Herold, 2001; Floridi, 2004b; Greco et al., 2005).

3. Fred Dretske (1981, p. 86) claims that “K knows that s is F = K's belief that s is F is caused (or causally sustained) by the information that s is F.” Floridi uses results about communication protocols to argue that the Gettier problem is unsolvable. See Maruyama (1961), Popper (1972), von Foerster (1980), and Harms (1998) for other applications of information theory to standard debates in epistemology.

4. When epistemology has been applied to information science issues, it has usually gone under the label social epistemology (see, e.g., Shera, 1970; Goldman, 1999).

5. In fact, several of the early philosophers of information (e.g., Plato, 1961; Nietzsche, 1997) cited by Floridi were doing applied epistemology of just this sort.

6. In this article, we focus on how epistemology can help us make practical information management decisions. But just as work in applied ethics has influenced ethics more generally, work in applied epistemology can influence epistemology and the philosophy of information more generally (Floridi, 2002a, p. 38).

7. The normative structure of decision analysis is the same as any other articulation of decision theory. The techniques have simply been tailored to facilitate practical decision making.

8. Actually, this step is a bit more complicated when there is significant uncertainty about what the outcome is likely to be. See the later section on “How Epistemic Values are Distributed”.

9. One of the authors (Whitcomb) believes that knowledge always has at least some value for its own sake. The other author (Fallis) is agnostic on this point.

10. Of course, it is possible to measure knowledge acquisition (Hunt, 2003). In fact, we do so all the time. For example, we regularly give students tests to determine how much they know. But it is more difficult to measure the knowledge acquired by, for example, library patrons than the knowledge acquired by students in a classroom (Hamburg et al., 1972).

11. A number of other factors (e.g., educational institutions, family, community) have an effect on the amount of knowledge that people acquire (Hamburg et al., 1972). But this is a concern for almost any important decision (e.g., many factors beyond emissions laws have an effect on air pollution).

12. There are also a few philosophers (e.g., Popper, 1972) who discuss knowledge as something that exists independently of human beings.

13. For example, Alavi and Leidner (2001, p. 109) claim that “what is key to effectively distinguishing between information and knowledge is not found in the content, structure, accuracy, or utility of the supposed information or knowledge”.

14. At the moment, most work in epistemology is only comprehensible to, and only read by, other experts in epistemology.

15. Similarly, the accuracy of the information may sometimes be more important than the precision of the information (cf. Floridi, in press). But sometimes it will be the other way around.

16. Since work in epistemology has never been explicitly put into the framework of decision analysis, some questions about epistemic values are likely to arise that have not yet been addressed in epistemology (cf. Fallis, 2004a). Thus, an applied epistemology that will help people make better information management decisions may require extending, as well as leveraging, work in epistemology.

17. Information scientists (e.g., Wolkoff, 1996, p. 87; Dick, 1999, p. 315; Budd, 2004, p. 364) have correctly noted that information professionals are not in a position to be the “arbiters of truth” for their clients and patrons. But that is a question about what is the best means to achieving our goal (i.e., it is part of the second step of the decision making process) and not about what our goal is. Admittedly, some information scientists (e.g., Shera, 1970, p. 97) have claimed that information management decisions should not aim to promote true belief. But see Fallis (2006b, pp. 490–495) for a detailed critique of this claim.

18. In addition, the relative weights of these epistemic values may well vary across decisions. We discuss this possibility in the later section on “How Epistemic Values Are Weighted”.

19. Studies have found that people who have an inflated view of their abilities tend to be happier than people who have accurate self-images (Taylor & Brown, 1988).

20. A few epistemologists (e.g., Radford, 1966) deny that knowledge requires belief, but they are a small minority.

21. As Edmund Gettier (1963) showed, knowledge requires more than justified true belief. Some theorists claim that, whatever the additional conditions are, they have no effect on inquiry (Kaplan, 1985, p. 355). Others deny this (Williamson, 2000, pp. 60–93). If the former theorists are right, then perhaps those additional conditions should be left out of our hierarchies. We take no stand on these difficult issues.

22. Justification might also be valuable as a means to other epistemic goals. In fact, true belief and error avoidance might themselves be valuable as a means to further epistemic goals, such as understanding (Riggs, 2003, p. 350).

23. Even if justification does not appear in the epistemic value hierarchy for a particular information management decision, it should still appear in a means-ends diagram (Keeney, 1992, pp. 69–76). This diagram indicates exactly how those things that are valuable as means are connected to those things that are valuable as ends. The construction of this diagram is an important part of the second step of the decision making process.

24. For example, it is often suggested that an explanatory belief is only valuable if it is also a true belief (Hempel, 1965).

25. This is analogous to the sort of conflict that can arise between precision and recall in information retrieval.

26. Some philosophers (e.g., Jones, 1997, p. 435; Kvanvig, 1998) have argued that it is not always strictly greater.

27. Since having a true belief has positive value and having a false belief has negative value in this situation, these values seem to be epistemically permissible.

28. Since it is possible to lose knowledge, there are actually further complications here. For example, one might prefer to acquire the knowledge right before one needs it so that there will be less chance of forgetting it.

29. It might be suggested that legislators today are more interested in maximizing corporate profits (Vaidhanathan, 2001). But at least according to the original intent of the U.S. Constitution, legislators should be trying to come up with intellectual property laws that “promote the progress of science and useful arts”.

30. One of the authors (Fallis) counts himself among these epistemologists. The other author (Whitcomb) does not.

31. As noted earlier, each of these objectives might be further subdivided. For example, maximizing usage consists in maximizing circulation and maximizing in-house usage (Kraft & Boyce, 1991, p. 68). Also, minimizing financial costs consists of minimizing subscription costs, minimizing processing costs, and minimizing storage costs.

32. Many readers of fiction are primarily seeking entertainment rather than knowledge.| But even when selecting fiction, epistemic objectives can matter. For example, readers of historical fiction typically expect to acquire true beliefs about the larger historical context of a novel. Moreover, fiction can be a rich source of knowledge about the human condition. By reading a novel like Pride and Prejudice, we can deepen our grasp of the moral and emotional dynamics of interpersonal relationships (Elgin, 1996; Kivy, 1997).

33. Another common selection criterion is currency. Up-to-date information resources tend to bring about true belief because they are more likely to be based on our best evidence and because it is less likely that the facts have changed since they were published (Fallis, 2006b, pp. 501–502).

34. Public libraries frequently have to decide between acquiring information resources for which there is greater demand and acquiring information resources that have greater quality (Berry, 1990). More knowledge can be acquired from quality information resources, but fewer people will acquire this knowledge if there is less demand for these particular resources. Thus, this is another collection development decision that requires an epistemic value trade-off.

35. For example, even if we cannot precisely determine the relative importance of some of our values, we still may be able to reach a definitive conclusion about what to do by using a sensitivity analysis (Edwards & Newman, 1982, p. 81; Kirkwood, 1997, p. 82). That is, we may find that a particular alternative is best as long as our values fall within a certain range.