ASIST 2005-10-04
"Epistemologies as (normative) theories of information seeking and use"
Birger Hjørland and Jeppe Nicolaisen
Abstract
How do people determine what information to consider for their different tasks? They will, of course, tend to use the information sources and the information that they consider beneficial and worthwhile. More generally, how does beliefs of this kind develop in human beings? Most research in information science has focused on the study of the users without considering the collective development of knowledge. Humankind in general and different cultures and domains have made some common experiences, which are communicated to the single individuals, directly or indirectly. Most of us have learned to trust medical science for information about health issues (and some mistrust medical science and uses alternative sources). The general theories about how valid knowledge is produced (or should be produced) are epistemological theories. Classical epistemological theories were empiricism and rationalism. More recent theories are related to historicism and pragmatism, respectively. Any epistemological theory will (if it is clearly formulated) provide normative criteria for, which information sources should be preferred, not only in science, but also in everyday life. The other way round: The information sources preferred in any context are a reflection of the dominant epistemological view. In every domain different epistemologies struggle to gain influence. In medical science, for example, the movement known as “evidence based medicine” may be seen as one epistemological view which emphasis certain kind of information (such as clinical trials) at the expense of other views (such as more clinical experience). The claim of this presentation is that explanations in people’s use of information should basically be found by examining the (mostly implicit) epistemological views that people have. In other words: The most stable patterns in people’s use of information is an expression of en implicit epistemological theory. The explication of such epistemologies may is necessary if library and information science is going to help people identify the information they need.
1) Introduction. What are "epistemologies"?
One of our basic assumptions is that in any domain of knowledge exist competing theories. The more fundamental theories share properties with theories in other fields. For example, cognitive information scientist share fundamental assumptions with cognitive psychologists, feminist sociologists share fundamental assumptions with feminist art historians and so on. Such basic theories are visible in many fields of knowledge. They are concerned with worldviews and with the nature of knowledge. They are philosophical theories or positions although they often have developed in different domains relatively independent of philosophy as a particular field. They are, by the way, not just scientific or scholarly theories, but also theories influencing "ordinary people's" "everyday information use": some people tend to think more like feminist scholars, other tend to think more like empiricists and so on.
In "The epistemological lifeboat" (Hjørland & Nicolaisen, 2005) we try to document the existence of different epistemologies in all fields of knowledge. In the section termed "Positions" we try to list and describe the different positions (and provide links to further information and discussion concerning their nature and status). In the section termed "Domains" we try to document the existence of different epistemologies in all domains.
Sometimes, e.g. in sociology, this seems easy. Almost all "positions" can be identified in writings within sociology. We refer to papers about, for example:
· Functionalism in Sociology
· Gender and Feminist Studies in Sociology
· Marxism in Contemporary Sociology
· Phenomenology in Sociology
· Positivism: Sociological
· Postmodernism in Sociology
· etc.
In other cases, such as chemistry, it is more difficult to trace the influence of different epistemologies. (A feminist theory of chemistry may be a joke - or it may not). Still, our working hypothesis is that different epistemological positions can be traced in all domains - and that they are important in order to understand the information needs and relevance criteria of a given domain.
We are aware of different kinds of dangers. Both in mathematics and in law, there are positions termed "realism". There is the obvious danger, for example, that the word is not used in the same sense, that legal realism and mathematical realism do not share fundamental assumptions. This is a problem that we can only solve by considering the domains one by one. Our working hypothesis is that it is possible to identify similar epistemologies in all fields of knowledge, not in pure forms, but as ideal types.
A given epistemological position may exist in a domain, whether or not this is recognized by people in that domain. Abrahamsen (2003), for example, asked music-librarians about their attitudes towards different views on music, including different kinds of music. He found that the librarians could not or would not take a standpoint; they claimed to be "neutral". In studying how they selected and organized music he found, however, that certain kinds of music were given higher priority, why he claimed that librarian in their behaviour revealed an epistemological view of music. We believe this example to be representative for most fields: researchers, professionals as well as ordinary people normally do not explicitly claim certain epistemological positions, but nonetheless, a study of what they are doing, what kind of information they prefer, reveals a certain pattern, which form an epistemological position. In order to uncover the epistemologies in a given domain it is not the best way to ask the average person in that domain. The best way is to ask the people who have studied the theories in the domain. Such people may be historians of science, historians of ideas, philosophers, bibliometricans, and other kind of science studies (both from within the domain and from outside the domain). [This view implies a criticism of user studies].
Sometimes are there explicit positions within a domain, with dedicated journals and other formal criteria. In such cases it easier to make bibliometrical studies and to demonstrate how such epistemologies are deeply influencing what is seen as relevant (cf. Hjørland, 2002; Nicolaisen, 2004). Given a mapping of epistemologies in a domain, this can be used to evaluate information systems. Ørom (2003), for example, has demonstrated how different views of art has influenced classification systems in LIS and thus pointed to the importance of considering epistemology in relation to knowledge organizing systems.
2. Empiricism as an epistemology
Empiricism is the view that experiences, observations or sense data are the only or the most important way of acquiring knowledge. It has a long history in science and philosophy, and it also has a long history of criticism and alternatives (see, e.g., Hjørland, 2005).
One common misunderstanding is the confusion of empiricism with empirical studies. Empiricism does not have a monopoly on empirical studies but represent specific views on how studies should be done. The empiricist attitude it to let data speak for themselves, not to look at them from a theoretical point of view. The empirist is a collector of data rather than a builder of logical systems or a critical person questioning given data as valid. A pure empiricist is impossible in reality, but as an ideal type, the empiricist is a fruitful concept. Each paper published in any domain is more or less reflecting the ideals of empiricism. Each paper may be characterized as being more or less empiricist (or more or less some other ideal type such as rationalism, historicism or pragmaticism). This way it is possible to identify an empiricist "paradigm" in almost all domains as well as criticisms and alternatives.
The concept of empiricism is important for LIS in two ways: It is important for how library and information researchers approach their objects of research (e.g. by preferring quantitative or qualitative research methods). LIS is one domain among all domains in which empiricist thoughts can be traced, discussed and evaluated. Much more important is, however, its importance for how the literatures that LIS study is constituted and organized. When information scientists are designing, evaluating, using or managing information systems in a given domain (e.g. Arts or medicine) may empiricist norms be traced and confronted with other epistemological norms. This provides in our opinion a foundation for information science which has been missing hitherto. We claim that epistemologies are the most important explanatory variables in IS and that a focus on these variables is the only way to establish a proper theoretical basis.
Empiricism as a theoretical position has a paradoxical relation to recorded knowledge that runs throughout its long history. Its motto was formulated by Francis Bacon: “Read Nature not Books” (cf., Rayward, 2004, 119). This motto is of course paradoxical in itself because empiricists publish a lot of papers themselves and use libraries and other systems of organized knowledge. Perhaps of its obvious problematic nature, it has not been discussed very much. But if it is a core assumption in the empiricist philosophy, and if the empiricist philosophy is a dominant view, then of course, this paradox should be examined. It is also obvious that the field of LIS has a core interest in determining fundamental issues concerning the relation between knowledge and its representation in literatures, libraries and information systems. So, our thesis is that a theory about the relation between researchers and information systems is primarily an epistemological theory (and not, for example, a psychological theory, as has mostly been assumed in LIS).
Other examples from the empiricist tradition:
· Francis Bacon also wrote Atlantis, a vision of a research based future filled with laboratories. But there are no libraries in Atlantis!
· David Hume wrote: Burn all books that do not contain numbers or experiments
· Ludwig Wittgenstein and the logical positivists dammed all positivist books, but Wittgenstein admitted that his influential book "Tractatus" was itself a metaphysical book.
· The psychologist B.F. Skinner declared that he followed Hume's principle and was poorly read in psychology. He was interested in experiments, not in subject literature.
These examples, we believe, prove the close, paradoxical relation between empiricism as a movement and a normative theory about how to use literature: don't.
We may go a little deeper into this question by asking: Given that it is unavoidable to use literature as a source of knowledge, in which way will the empiricist ideals influence the way literature is used? The answer to this question will, by the way, also reveal the nature of an empiricist view of relevance, citations and information needs.
In our opinion, the evidence based approach to medicine, the Cochrane reviews as well as the approach known as meta-analysis are examples of how modern empiricists use recorded knowledge. In accordance with Hume, modern empiricists tend to disregard all studies that are not of a quantitative nature. They tend to regard published data as given units, which -if they meet some given methodological standards - may be processed in rather mechanical ways.
The empiricist view on the use of recorded knowledge is challenged by other epistemological positions. Rationalism shares some ideals with empiricism: That pieces of published knowledge are given and may be cumulated. Rationalism disagrees, however, about the nature of those units. A more radical alternative is, for example Kuhn's theory of scientific revolutions. This is a view that acknowledges the role of textbooks in forming the basic paradigms, which governs the scientists use of information. We cannot go further into this issue at this place, but we hope to have demonstrated the extremely important relations between epistemological theories and theories of IS. As information scientists it seems important to subscribe to an epistemological view that is opposed to Bacon's ideals. We should subscribe to a view that emphasizes the role of information systems in the development of knowledge.
3. Medicine as a case-study.
In medicine is a relatively new movement termed "evidence based medicine" (EBM). According to the Centre for Evidence-Based Medicine, "Evidence-based medicine is the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients."
We will use this example, because we believe it is important to demonstrate our point of view outside the humanities and social science. So, we have chosen a perhaps difficult example in order to try to make our claims more general.
In this case we are lucky that evidence based medicine is a self-conscious movement: It is easy to identify researchers and documents subscribing to this view.
Different positions tend to describe themselves in more positive terms and competing views in more negative terms. This was, for example, the case when American behaviorists established themselves as a movement in the beginning of the 20.Century. From our point of view it is important to try to make a description that emphases the relative strength and weaknesses in each position.
Here is one description of what EBM is an alternative to:
“The former paradigm was based on the following assumptions about the knowledge required to guide clinical practice.
· Unsystematic observations from clinical experience are a valid way of building and maintaining one's knowledge about patient prognosis, the value of diagnostic tests, and the efficacy of treatment.
· The study and understanding of basic mechanisms of disease and pathophysiologic principles is a sufficient guide for clinical practice.
· A combination of thorough traditional medical training and common sense is sufficient to allow one to evaluate new tests and treatment.
· Content expertise and clinical experience are a sufficient base from which to generate valid guidelines for clinical practice.” (Evidence-Based Medicine Working Group, 1992).
Another way to formulate the difference between different views have been given by Tonelli (2001): . Clinicians, then, need to incorporate knowledge from 5 distinct areas into each medical decision: (1) empirical evidence, (2) experiential evidence, (3) physiologic principles, (4) patient and professional values, and (5) system features. The relative weight given to each of these areas is not predetermined, but varies from case to case.
The last quote is a much more positive way to formulate the information needs of clinical doctors. What EBM term "evidence" is only one kind of information needed. It is not possible a priory to say what kind of knowledge is most important, it varies from case to case.