- 1 Researchers citing Egan and Shera (1952) as the first occurrence of the term SE include Budd 2002; (...)
- 2 Egan and Shera (1952, 133-4; italics in original) suggested a new discipline: “Thus the focus of at (...)
1Information science and knowledge organization (KO) can be proud that the term “social epistemology” (SE) was coined by library scientist Jesse H. Shera (1951, 82), in a paper about classification. A year later Egan and Shera (1952) published another paper using the term SE, but this time in another meaning and in the context of a theory of bibliography. Because Egan and Shera’ (1952) paper is by almost all subsequent researchers1 cited both as the first occurrence of the term SE and for discussing its meaning,2 the 1951 meaning has, unfortunately, been overlooked. Briefly explained the 1951 meaning is epistemological, while the 1952 meaning is sociological, a confusion also made by some later researchers. The term SE was not at the time develop into a research program.
2Outside library and information science (LIS) two schools of SE (founded by respectively Alvin Goldman and Steve Fuller in the 1980’s) developed and became dominant in relation to the term SE. These are discussed in Section 2. During this discussion arguments are put forward to conclusions made later in the article. Section 3 presents and discusses Shera’s (1951) approach to classification and his SE. It is argued that Shera’s understanding of SE is more fruitful compared to those of Goldman and Fuller.
3Epistemology is one among other fields studying knowledge (others being, e.g., sociology of knowledge and cognitive psychology). Epistemology is different from the other fields by being a normative field, meaning it tries to establish criteria about how to obtain knowledge, its purpose is to suggest methodological prescriptions. This is important, as we have already stated that the 1951 understanding was epistemological, but the 1952 understanding was sociological, and that this confusion also has been made by later researchers.
4Different theories of epistemology are further introduced in Section 4, but already in Section 2.1 we discuss the difference between individual and social epistemology.
5It should be said that the word “social” itself has more senses, a broad meaning that includes the socio-cultural context and social norms of people, and a narrow meaning that just understands “social” as individuals in the plural. The narrow conception is associated with positivism and the attempt to study social phenomena by empiricist methods. These two meanings are also reflected in the literature about SE, and, as we shall see later, Shera (1951) and Fuller’s position reflect the broad view, while Goldman’s position reflets the narrow view. Discussions about social knowledge involves several dichotomies, which are often confused, such as methodological individualism versus methodological collectivism, first-hand knowledge versus second-hand knowledge, private versus public and subjective versus objective. A more detailed discussion of the word “social” and these dichotomies are left out in the present article because of space restrictions but can be found in Hjørland (2023).
6In this section, two contemporary schools are presented and discussed. One of the contemporary schools of SE was, as already stated in the introduction, founded by Alvin Goldman. It is often called “analytic social epistemology”, e.g., by Quinton (2004) and Collin (2020, 21). The other was founded by Steve Fuller and by Quinton and Collin referred to as “critical social epistemology.” Collin (ibid.) described these two schools as “the twin roots and branches of social epistemology,” and found that both positions were founded about the 1980s.
7Goldman and O’Connor (2021; electronic resource, no pagination) emphasized that SE must be understood in opposition to individual/individualistic epistemology. As example of individualist epistemologies, they presented René Descartes and John Locke, writing:
According to the most influential tradition in (Western) epistemology, illustrated vividly by René Descartes (1637), standard epistemology has taken the form of individual epistemology, in which the object of study is how epistemic agents, using their personal cognitive devices, can soundly investigate assorted questions. Descartes contended that the most promising way to pursue truth is by one’s own reasoning. The remaining question was how, exactly, truth was to be found by suitable individualistic maneuvers, starting from one’s own introspected mental contents. Another major figure in the history of the field was John Locke (1690), who insisted that knowledge be acquired through intellectual self-reliance. As he put it, ‘other men’s opinions floating in one’s brain’ do not constitute genuine knowledge.
8Goldman and O’Connor (2021) presented their alternative to individual epistemology this way:
By contrast social epistemology is, in the first instance, an enterprise concerned with how people can best pursue the truth (whichever truth is in question) with the help of, or in the face of, others. It is also concerned with truth acquisition by groups, or collective agents.
9As Goldman and O’Connor (2021) mentioned, Descartes and Locke may be understood as two main representatives of individualist epistemology, Descartes representing rationalism, while Locke represents empiricism. Rationalism emphasizes the individual’s reason and the deductive method, whereas empiricism emphasized the individual’s sensory experience and the inductive method. Epistemological approaches are further addressed in Section 4.
10Goldman and O’Connor (2021; emphasis in original) also wrote:
In contrast with the individualistic orientations of Descartes and Locke, social epistemology proceeds on the commonsensical idea that information can often be acquired from others. To be sure, this step cannot be taken unless the primary investigator has already determined that there are such people, a determination that presumably requires the use of individual resources (hearing, seeing, language, etc.) Social epistemology should thus not be understood as a wholly distinct and independent form of epistemology, but one that rests on individual epistemology.
11This quote reveals an understanding of the difference between individual epistemology and SE that has been criticized by other researchers. Kusch (2001, 188), referring to Goldman (1999) emphasized that Goldman’s and others’ idea of individual knowledge as being primary, and social knowledge as being secondary is a problematic assumption. Perhaps the reader is questioning Kusch’s view and considering it to be nonsense. If so, the following thought experiment may help understanding Kusch’s idea: Think of a person (e.g., yourself) in an isolated position (e.g., on the toilet). Whatever you are thinking about in that situation, you use concepts, acquired during your upbringing. It is therefore not a mad idea to consider individual knowledge as secondary to the social and cultural context in which you have been socialized.
- 3 Kusch (2002, 2-3) calling his own position “communitarian epistemology” and Goldman’s approach “the (...)
12The implication of Kusch’s view is that SE cannot just be considered a supplement, itself based on individual epistemology, as claimed by Goldman and followers. Supporters of authors such as Kusch3 (2002), will argue contrary: that it is rather the individual epistemology, that rests on SE.
13To illustrate the social nature of knowledge we may consider a quote by Albert Einstein: “It is the theory which decides what we can observe” (cited from Heisenberg 1989, p. 40). This quote is a key to understanding SE. It is opposed to the empiricists (incl. logical positivists) belief in the objectivity of knowledge derived from individual observational reports. In the positivist view, theory is derived from and tested by observations, and theory and knowledge are assumed to be free of cultural and social factors, and from the observers’ theoretical assumptions. SE is different, as already Fleck ([1935] 1979, 38) realized:
[C]ognition must not be construed as only a dual relationship between the knowing subject and the object to be known. The existing fund of knowledge must be a third partner in this relation as a basic factor of all new knowledge. […]
14Cognition is therefore not an individual process of any theoretical ‘particular consciousness.’ Rather it is the result of a social activity, since the existing stock of knowledge exceeds the range available to any one individual.
15What Fleck here said is extremely simple and trivial, and it is strange that it has not always laid the foundation of epistemology, that mainstream epistemology has ignored this important dimension. Therefore, one way to understand the contrast between individual epistemology and SE can be formulated this way:
-
Individual epistemology: The individual’s observations are pure and direct (i.e., representing reality free of social, cultural, and theoretical influences).
-
Social epistemology: The individual’s observations are influenced by social issues including the individual’s orientations and views. In Fuller’s (2017, 4197) words: “Accordingly, individual cognitive orientations can be understood as either normal or deviant expressions of preexistent culturally entrenched dispositions.”
- 4 Library and information scientist Patrick Wilson’s book (1983): Second-Hand Knowledge, is about how (...)
16As emphasized by Kusch (2001, 188) Goldman and follower’s approach to SE does not recognize this problem related to individual epistemology. According to the definitions above, Goldman’s view represents an individualist rather than a social epistemology, but it is a research program focusing on second-hand knowledge (i.e., knowledge obtained from others, such as testimony). It is of course a legitimate area of research to study forms of biases in knowledge obtained from different kinds of testimony4.
17As explained in Hjørland (2023) there exists an “impoverished,” narrow understanding of the term “social,” which correspond to Goldman’s approach, not very unlike the meaning it got in experimental social psychology, where “social” is the opposite of being an isolated individual (e.g., on the toilet). The word “social” has in this tradition been understood as being together with other human beings (either physically or imaginarily).
- 5 It has always been a mark of scholarship to read the relevant literature, and thereby to urge peopl (...)
18Goldman and O’Connor (2021) also wrote: “Surprisingly, social epistemology does not have a very long, or rich, history.” Yes, it would indeed be surprising if this were the case5 . Gelfert (2010) traced SE back to Kant and the German Enlightenment tradition, even in the sense as understood by Goldman’s school. As here understood it at least goes back to Hegel (1770–1831), and have been influential for example, in pragmatism and critical theories but is has not been understood or considered in positions influenced by logical positivism and analytical philosophy (which may explain Goldman’s view as he is considered part of analytical philosophy). Fuller (1987), from his perspective, also traced SE further back in time, and Kusch (2011, 873), in line with the understanding by the present author, wrote: “Many contributors to Pragmatism, Marxism, Critical Theory or Hermeneutics also qualify [as being parts of SE].”
19Despite this criticism of Goldman’s school, a valuable research program into the problems connected with “second-hand knowledge” is possible. However, this is not entirely a new field, as claimed by Goldman. Historians, for example, have since Leopold von Ranke developed “source criticism” as the methodology for prioritizing kinds of sources, and medical researchers have developed “evidence-based medicine” as a criterion for which second-hand reports should be considered most valid. Also, in information science a well-known problem has been to select books, documents and information resources based on their reliability (see, e.g., Hjørland 2012). These three examples are all much related to problems discussed by SE, for example “which experts should you trust?” (Goldman 2011 and 2021). The question is whether Goldman’s SE has provided new understanding of such problems? There seems to be three problems in Goldberg’s approach:
-
Goldman’s articles are unrelated to research like that presented in these examples; therefore, it seems not to build on previous knowledge done in different disciplines. It is also strange that Goldman does not consider issues like consensus and disagreements among experts and its importance for selecting experts, as a high degree of consensus should imply a higher degree of trust. What was expected of Goldman was an epistemological analysis of how researchers and practitioners in different disciplines have dealt with second-hand knowledge, and his suggest methodological prescriptions on how to improve research and practice in different domains.
-
Goldman’s articles seem to be less concrete and more common-sense prescriptions compared to, for example, studies of the reliability of peer-reviews in information science (and suggestions made in other disciplines such as history and medicine).
-
- 6 Laudan (1984, 5-6) wrote: “The Leibnizian ideal holds that all disputes about matters of fact can b (...)
The articles discussion of novices is a bit strange. An alternative strategy would be to say how, theoretically, expert claims can be evaluated6, and how they are evaluated in practice, and then discuss how (or if) such metaknowledge can be, or has been, transferred to and used by other groups, including novices. To put up normative criteria for being experts is an epistemological task. To investigate whether different groups know about these criteria is an empirical task.
20Quinton (2004, 8) writes how metaknowledge about the authority of knowledge claims are something we learn about during our lifetime. Such metaknowledge should not just be based on common-sense and individual experiences, but as a far as possible on research into what Wilson (1983) termed “second-hand knowledge” and “cognitive authority.” This is a legitimate field for SE, but as said, such research has for a long time taken place outside Goldman’s SE. In the eyes of the present author, Goldman’s results developing methodological prescriptions are disappointing.
21Steven Fuller’s SE (e.g., 2016, 2017) is somewhat difficult to describe, as Collin (2020, 27) wrote: “Fuller’s work resists simple summary because of his somewhat unsystematic style of writing.” Fuller (2017, 4197) however, provided an important statement about his position:
- 7 Remark that the American school of pragmatism generally falls outside this dichotomy.
‘Social epistemology’ literally means the social theory or social science of knowledge. That simple definition already says a lot. It implies that knowledge is not normally seen as intrinsically social; hence, ‘social’ needs to be added to specify the field of inquiry. This point is worth noting because the image of knowledge as primarily acquired by individuals through their mental faculties (as perceptions, beliefs), who then combine with other such individuals, to construct more elaborate and durable knowledge products (such as theories, sciences), rests on a particular reading of the history of philosophy that is dominant only in the English-speaking world. For philosophers more influenced by French and German developments, knowledge is ‘always already’ social in both its constitution and import. In the United States, this distinction is marked as ‘analytic’ versus ‘continental’ schools of philosophy. 7
22We see, that in opposition to Goldman’s position, Fuller expresses a view of knowledge, that is genuine social: “knowledge is ‘always already’ social in both its constitution and import.” This is an important insight, but it does not say what Fuller’s epistemological position is. What are his normative guidelines for inquiry? In addition, it is a strange claim that the basis for constructing such normative guidelines is the same as “social science of knowledge”? Social sciences are empirical sciences influenced by conflicting epistemologies. They need epistemological clarifications; they do not themselves constitute an epistemology.
- 8 Bird (2003) discussed three claims about Kuhn’s conservatism, including Fuller’. He concluded (p.13 (...)
- 9 Rowbottom (2011) provided the interesting view that both Popper’s critical attitude and Kuhn’s cons (...)
23It seems difficult to find a clear answer about Fuller’s epistemology, although Fuller (2016, 2) indicates an answer: “An important precursor of social epistemology is ‘critical rationalism,’ a philosophy associated with Karl Popper, in which the two words that constitute the name should be taken equally seriously.” However, Popper’s “critical rationalism” is not a social epistemology. To demonstrate this point, we shall contrast it with the philosophy of Thomas Kuhn. Fuller (2003) expressed critical views about Kuhn, as he found him to be conservative, while he found Popper to be critical. In the eyes of Fuller, the effect of Kuhn’s (1962) The Structure of Scientific Revolutions was that philosophers of science gave up their critical attitudes towards science8. Popper, on the other hand, suggested that scientist should be free to suggest bold theories, but should focus on falsifying all theories.9 In one way, therefore, Fuller is right. The motto “be critical” is closer related to Popper’s philosophy than to Kuhn’s. But the motto “be critical” is insufficient as a methodological prescription if not followed by further guidelines about how to be critical, which means how to examine the theories. Popper’s well-known answer to this problem is “falsificationism:” researchers shall try to falsify theories. By implication, a criterion for being scientific for Popper is that a theory is formulated in such a way that it is possible to deduce its implications and thereby to test it. According to Popper, the theory “all swans are white” can never be confirmed no matter how many white swans have been observed. It can be falsified, however, if just on black swan is observed.
- 10 Criticism of falsificationism is further presented in Shea (n.d.; Section 3): https://iep.utm.edu/p (...)
24Popper’s view is, however, based on some problematic assumptions. The most important one is that whereas Popper accepts that observations are theory-dependent, and acknowledges that this makes confirmation of theories impossible, he ignores this when it comes to falsification—and therefore he fails to see that falsification is also inconclusive.10 Another problem is that research does not start with clear concepts, but concepts are typically made more precise as science progress. Therefore, the demand of clearly formulated concepts is in some instances problematic and harmful. When claiming that one observation of a black swan falsifies the theory that all swans are black, it is assumed that “swan” is an unproblematic concept, that can be identified with certainty. Recent research based on DNA analysis have, however, problematized many former definitions and classification of species of birds (see Fjeldså 2013). Thus, concepts are not just “given”, but are developed and modified by research, again implicating a social perspective. The last point to be taken up here is that Popper considered the Marxist theory of history and psychoanalysis as pseudo-sciences because he did not consider them falsifiable (which is a claim that has been denied by other philosophers.) But this makes Fuller’s preference for Popper strange, since Fuller himself is closer related to “critical” that to analytic and positivist schools but seems to prefer an epistemology that rejects “critical” theories. Despite Popper’s argument for the opposite, most philosophers consider his philosophy to be related to logical positivism, which is not an epistemology, that can be characterized as “social.”
25Kuhn’s philosophy, contrary to Popper’s, is genuine “social.” Scientists are trained under the influence of a paradigm, often in an indirect way, influenced not just by textbooks and theories, but also by the apparatus used, the kind of research questions raised etc. Theories are not just falsified, but they may be left when newer generations of researchers take over. It is widely recognized that Kuhn was a primary force in the historicist turn in the philosophy of science in the 20th century, and it is almost definitional that historicism implies a social point of view, as it implies the historicity of knowledge and cognition.
26Thus, when Fuller claimed that Kuhn was not critical, the answer here is that, opposed to Popper’s falsificationism, Kuhn’s view implies that the evaluation of theories cannot be limited to issues of logic and observation but must consider the broader socio-historical context of the theories (as will be demonstrated in Section 3 about Shera’s view on classification). See also further about Kuhn’s theory in Section 4 below.
27Fuller’s SE developed comments on more traditional epistemological issues such as the Popper-Kuhn controversy but is not a position that he clearly developed and defended. If anything, Fuller’s discussion of Popper and Kuhn seems to contradict his ambition of providing a social epistemology.
28Kusch (2002, 2) suggested to name Fuller’s school “the science policy programme.” Probably it is in this suggestion we should search for its core contributions. Kusch (ibid.) wrote:
‘Social epistemology’ has come to refer to two rather different programmes. I shall call them the ‘science policy programme’ [Fuller’s] and the ‘complementary programme’ [Goldman’s]. The science policy programme seeks to determine ways of making science more democratic and accountable to the public. It also hopes to increase our ability to choose between the development of different kinds of knowledge. This hope is based on the assumption that one can influence the collective production of scientific knowledge by manipulating the social organization of scientific communities. Changing social organization leads to a different type of knowledge [cf., Fuller 1988].
29Fuller (2002) is a book divided into four parts of which two are of interest in this connection: part three: “Issues in the Social Organization of Knowledge” and a part four: “Issues in Knowledge Policy-Making.” Part three consists of four chapters + an appendix, discussing respectively the demarcation of science, disciplinary boundaries, consensus in science, the Forman thesis (i.e., the thesis that physicist adopted quantum indeterminism in response to the cultural milieu of Weimar Germany), policy simulations of objectivity. Part four consists of two chapters + an appendix, discussing respectively normativity on the sociology of knowledge, cognitive authoritarianism with the politicization and depoliticization of expertise and the design of a core curriculum for a graduate program in knowledge policy studies. This content of a main book by Fuller seems to justify Kusch’s name “the science policy program” for Fuller’s version of SE. But does it represent an epistemology?
30From the mentioned chapters it is difficult to get an impression of how Fuller imagines SE can provide (or has provided) normative criteria for organizing science in ways which to their development of different forms of knowledge. What seems most important is that the point of department is not an analysis of which problems in the present sciences should be solved, followed by suggestions for how SE might contribute changing science in the right direction. The clearest example seems to be the appendix about a suggested graduate program in knowledge policy studies. Fuller found (2002, 289):
[G]iven their overarching administrative perspective, bureaucrats are potentially in a better position than any individual scientists to understand the collective product of epistemic pursuit. A goal of knowledge policy studies, then, would be to train bureaucrats in the kind of decision making that would instill confidence in scientists that government can have something epistemically interesting to say about how knowledge should be pursued.
- 11 Fuller (2002, 291-2) described a course in “the art of transideological policymaking, where “studen (...)
31Much in this quote can be discussed. For example, Goodall (2006) made a bibliometric study which showed that the best universities are led by the best researchers, which seems to conflict with Fuller’s idea about the role of bureaucrats. This leads to the question of Fuller’s epistemological basis for his claims about research policy?11 Clearly such claims need to be research based, and this research needs to be based on epistemology. In other words, science policy is a research field, but it is not epistemology and therefore it is not SE either.
32Another example is Fuller (2017), an article in Encyclopedia of Library and Information Sciences, which is impressing for the knowledge about and interest in library and information science (LIS), including the interest in providing an important role for this field. However, the point here is again that, in the eyes of the present author, is does not contribute epistemologically to this field, it does not provide normative rules for how to provide, for example, optimal selection, description, indexing, classification, etc. of documents. It is, as Fuller defined (2017, 4197) a “social theory or social science of knowledge,” which, although important, does not constitute an epistemology.
33In Section 2 it has been argued, that in the two positions mostly considered dominant in SE, Goldman’s is not genuine social, and Fuller’s is not genuine epistemological. In the view of the present author neither of these two schools seem helpful for addressing the research problems related to classification, which was the problem that original made Shera introduce this concept—and probably neither school provides useful methodological prescriptions for other areas of research either. (If they have contributed textbook relevant knowledge about research methodology, this has escaped the present author). SE is, however, more than these two schools, and, among others, feminist standpoint theory, and Kuhn’s (1962) theory of paradigms, seem better alternatives. However, the two chosen theories are those which have dominated the literature, which use the term SE, and which therefore have been found necessary to consider in this article. Now we turn to Shera’s original suggestion about SE, which, it is argued, provides a better foundation for SE.
34Shera (1951) coined the term SE in the context of classification research. He found that formerly all theory of the organization of knowledge, from Plato to Henry Bliss has been founded on four basic assumptions (1951, 72-3; italics in original; bulleted listing added):
-
“First that there exists a universal ‘order of nature’ that, when discovered, will reveal a permanent conceptual framework of the entirety of human knowledge;
-
second, that the schematization of that order is a hierarchy of genus and species, class and sub-class, that progresses downward from general to specific, from terms of maximal extension to those of maximum intension;
-
third, that the principle of differentiation that operates throughout the hierarchy is derived from the likeness or unlikeness of the properties or attributes of the component units of the classification; and
-
fourth, that these properties or attributes partake of the substantive nature or physical properties of the units being classified: an intrinsic part of the unit itself, permanent and unchanging, an essence, an essence that resists alteration by the external environment, and denies all consideration of the fortuitous or accidental.”
35Shera provided a criticism of these assumptions and suggested alternatives. It was in this context, he introduced the term SE (1951, 77; italics added):
Even a cursory examination of the history of classification of the sciences emphasized the extent to which any attempt to organize knowledge is conditioned by the social epistemology of the age in which it was produced. This dependency of classification theory upon the state of the sociology of knowledge will doubtless be even more strongly confirmed in the future. Here, then, is an implicit denial of Bliss’ faith in the existence of a ‘fundamental order of nature,’ a rejection of the belief that there is a single, universal, logically divided classification of knowledge.
36Shera’s claims are of course in need for further investigations, e.g., how some specific classifications in specific ages were influenced by that age. Perhaps there is even a contradiction by claiming on the one hand that all prior classifications were based on the four basic assumptions, and on the other hand claiming that all former classification “is conditioned by the social epistemology of its time.” One could say that Shera claims that what he suggests is a “paradigm shift” in classification, introducing a new epistemology against all former ones. However, this contradiction may only be apparent in that former classifications were intended to follow these ideals, but that that they in reality were conditioned by the social epistemology of the age in which they were produced (which also indicates that the former ideals did not work in practice, that they may have been falsified by history). The meaning of “social epistemology of the age” is —at least partly—revealed in the following quote (Shera, 1951, 77):
If one may learn anything from such a cursory examination of the history of classification it is that every scheme is conditioned by the intellectual environment of its age or time; that there is not, and can never be, a universal and permanent classification that will be all things to all men; and that each generation may build upon the work of its predecessors, but must create its own classification from the materials that it has at hand and in accordance with its own peculiar needs.
37This quote states:
-
that a classification cannot be universal, serving all purposes for everybody
-
that a classification cannot be permanent
-
that classificationists build on their predecessors, making classifications developmental and dynamic
-
classifications are based on “the materials at hand,” i.e., based on the knowledge and concepts of its time, and
-
that classifications are designed to serve specific needs.
38Shera’s expression “the intellectual environment of its age” may be translated to the dominant worldview, paradigm, epistemology, or metatheory. Probably, we should not take “generation” too literally. It may well be that some classifications have a longer and some have a shorter lifetime and that not all fields of knowledge necessarily develop in a synchronize way. What is important is that the classifier (and the resulting classification) is influenced by views represented in a broader social, cultural, and domain-specific context. This is a clear social epistemological position that denies the possibility of constructing classifications based on the isolated individual’s observation and cognition.
- 12 Shera (1951, 77) wrote: “The early systems of library classification may here be dismissed briefly (...)
39The positions in library classification that Shera argue against include that of Henry Bliss, but although Shera does mention the Dewey Decimal Classification (DDC) he presents no specific criticism of it (just implicitly since he dissociates himself from universal and hierarchical classifications). But the DDC, to an even greater extent, deserve to be criticized12. Rather than dismiss the systems from consideration, which Shera (1951, 77) termed “purely utilitarian and without philosophical foundation,” Shera should rather consider them as one among other approaches to classification and studied the relative merits and demerits of this approach more closely, and also considered Mill’s (1872, 498-9) distinction between artificial and scientific classifications, where the first are narrowly utilitarian while last suggests that scientific classifications should provide basis for general conclusions.
40Shera expressed the pragmatic approach clearly (citing the classical pragmatic philosopher William James), and thereby demonstrated the close connection between pragmatism and social epistemology (1951, 83-84; italics in original):
The pragmatic approach to classification through meaningful units of knowledge must be based on recognition of the obvious truth that any single unit may be meaningful in any number of different relationships depending on the immediate purpose. Thus, it is the external relations, the environment, of the concept that are all-important in the act of classifying. A tree is an organism to the botanist, an esthetic entity to the landscape architect, a manifestation of Divine benevolence to the theologian, a source of potential income to the lumberman. Pragmatic classification, then, denies the existence of the “essence” of tree, for each of these relationships owes its existence to different properties of the tree. Relationship is not a universal, but a specific fact unique to the things related, and just as these relations reveal the nature of the relata, so the relata determine the character of the relationship.
41Somebody might claim that Shera’s position is more metaphysical than it is epistemological, and that it comes close to an idealist view according to which boundaries are without an objective basis. There is, however, a realist position according to which the complexity of reality allows different classifications, according to different needs, e.g., Dupré’s (1993) “promiscuous realism”. These issues are still debated, but naïve realism is probably an untenable position which Shera opposed, and he also argued for the anti-essentialist viewpoint, which, although still discussed, have got many supporters today. Shera (1951) emphasized how different domains may consider the same term (e.g., tree) differently and thus describe different attributes and put it into different conceptual structures, thereby predating Hjørland’s domain analytic point of view (cf. Hjørland, 2017) for about half a century. This is also clear in Shera’s emphasis on the necessity of subject knowledge of the classifier, and by his priority of subject-specific classifications rather than universal ones.
42Given this description of social epistemology, it seems not to be as unclear, as others and Shera (1971, 79) himself later assumed. The position is epistemological in that it claims that different paradigmatic views (by the classifier) produce different classifications. The position is social in that different groups of people with differing perspectives and interests produces different classifications. Shera failed, however, to provide more specific methodological principles for classification based on SE. He realized that subject knowledge is important, but as Wilson (1973, 248) critically remarked: “This seems to imply an extensive knowledge of ‘academic disciplines,’ but Shera does not tell us how extensive”. Wilson’s question should probably rather have been: “what kind of subject knowledge is of special interest to the classifiers of a knowledge domain?” Shera’s problem was that in 1951 the world had yet to encounter a revolution in philosophy best know from Kuhn (1962), which introduced the concepts “paradigm” and “paradigm shift,” and which itself is associated with both SE (cf., Wray 2011) and pragmatism (cf., Mladenović 2017).
43Shera – and the proposals he made – did outline important methodological prescriptions for classification (much better than the two schools presented in Section 2.) He did not, however, concretize in the design of a domain-specific classification, or provide more specific guidelines for doing so. In relation to SE this first came with introduction of the domain-analytic approach, with the claim that different “paradigms” implies different classifications, which indicate that the design of a classification presupposes the deciding of the “paradigm” on which should it be based. Ørom (2003) may be considered a model.
44Today, there seems to be an unmanageable number of theoretical positions in epistemology, both in elementary textbooks and in the research literature. The situation seems overall not to be a healthy one.
45Some of the many suggested positions in epistemology seem to this author to be sound, but of a too narrow application. This is the case with, for example, the feminist “standpoint theory” and “critical race theory,” which probably are very fruitful for tackling problems related to equality in, respectively, gender and race/ethnicity. However more general courses in research methodology need to be based on broader epistemological principles, which can be taught in courses and applied to a field (e.g., LIS) as a whole. We should ask ourselves: What are the main epistemologies in use (explicitly or implicitly), how should they be evaluated, and what is the general learning for us? As a first step it may be suggested that “standpoint theory” and “critical race theory” form parts of a more general “critical theory”, which opposes individualist epistemologies like empiricism, rationalism, and positivism by denying the neutrality of research and knowledge. A general epistemological lesson from feminist epistemology has been, according to Code (1998, 597), to move the question “whose knowledge are we talking about?” to a central place in epistemological inquiry. A related lessen from critical theories is that engagement is important, “objectivity is not neutrality,” as Haskell (1998) put it. Based on such a need for a generalized classification of epistemologies, the present author has suggested the following classification, in, for example, Hjørland (2021):
-
Rationalism: emphasis on logical and rational principles, a priory knowledge, and the deductive method. Rationalism consider itself an ahistorical and neutral epistemology, not influenced by social issues. (This view is considered impossible by historicism and pragmatism, but nevertheless has a strong position in parts of the literature).
-
Empiricism: emphasis on observations, which are not influenced by the observers social-cultural or theoretical context or orientation. (This view is also considered impossible by historicism and pragmatism, but nevertheless also has a strong position in parts of the literature).
-
Historicism: emphasis on the social, cultural, and paradigmatic context of the observers. Any knowledge claim must defend the broader theoretical basis on which it is based.
-
Pragmatism: emphasis on the non-neutrality of knowledge claims. Any knowledge claim must be state which interests are served and provide arguments that the claim supports the stated interests.
46People with philosophical knowledge may provide arguments against this classification. It is well-known that the established contradiction between rationalism and empiricism is problematic, and that, for example, no philosopher has ever been 100% empiricist or rationalist. However, still this classification seems to work very well in classifying, for example, approaches to knowledge organization:
-
Some approaches to KO are based on rationalism, for example those described in the Bliss Bibliographic Classification (Mills and Broughton 1977), which emphasizes methods like logical division and a priori reasoning (in practice it also uses empirical materials, but the selection and use of this is not a part of the described methodology.)
-
Statistical approaches such as numerical taxonomy (Sokal and Sneath 1963) are examples of classifications that clearly are based on empiricism. (But as discussed by Richards 2016, 124ff, despite empiricism’s declared avoidance of subjectivity, this nevertheless necessarily influence classifications based on numerical taxonomy)
-
Examples of historicism are (1) Darwin’s (1859) approach to classification can be used because its criteria for which properties of organisms are important for classification are derived from evolutionary theory (see Richards 2016, 113ff); (2) Ørom’s (2003) demonstration that classification of art (both in museums, in comprehensive works and in library classifications) reflects paradigms in art studies. Historicism also uses empirical data and rational methods but consider these to be based on background theories, which makes it a SE.
-
Examples of pragmatist approaches to classification are those which emphasizes the analysis of the purposes, functions, consequences, interests, and political goals, which the classification is intended to support. Historicist and pragmatic approaches are often very related, but historicist approaches need not as pragmatism be based on explicit interests. Pragmatism also uses empirical data and rational methods but, like historicism, considers these based on background theories, which makes it a SE.
47As stated, the two last positions represent forms of SE. We shall not discuss this classification further, as it has been used in several publications already.
48Rationalism and empiricism became most influential by their combination in logical positivism (or logical empiricism), a position that by most philosophers of science is claimed dead, but the death of which was discussed by Bentz and Shapiro 1998, 26-31) under the heading “the mysterious death and afterlife of positivism,” stating (p. 30): “Postpositivism can become, like positivism did, an excuse for not reflecting on the grounds of one's beliefs and practices about knowledge and about one's social and historical context.”
49The central point of departure for alternatives to positivism is individualist versus social epistemology, as we defined it in Section 2.1. We take the point of departure in Kuhn (1962), and his introduction of the concept “paradigm” and “paradigm shift.” There are unclarities related to the term “paradigm,” but it is now widely used in a wider sense than the one suggested by Kuhn. When a scientist is trained and works in a field, he or she learns about the theories of this field, often in an implicit and subtle way, e.g., in the kinds of instruments used and the kind of questions asked, that does not reveal itself as “as theory” and therefore does not reveal itself as something one can question. This socialization of scientists is an important determinant for how questions are put and approached, and how observations and experiments are interpreted. The paradigm provides the researcher with a set of glasses through which s/he sees and acts in the world. Individual characteristics of the researcher (such as his/her talents, motivations, and interests) are also important, but the social dimension is more important from an epistemological point of view. What here is written about researchers can be generalized to all kinds of human acts, as it is done in certain psychological approaches such as activity-theory (e.g., Engeström 2015).
50Kuhn’s theory is not without problems and Fuller (1987, 150) rightly opposed Kuhn’s view that normal science is dominated by a single paradigm at a time. This is something that many critics have stressed: the historical incorrectness of Kuhn’s notion that normal science is characterized by a period of the sole existence of one dominating paradigm. Biologist Ernst Mayr (1997, 98–99), for example, found that Kuhn’s view “reflects the essentialistic-saltationistic thinking so widespread among physicists.” This article cannot provide a deep discussion about the concept “paradigm,” but it is assumed that paradigms often are competing phenomena, which primarily are characterized by conflicting philosophical assumptions. It is obvious, that researchers in different paradigms produces different theories and classifications. This does not imply a total relativism, as different paradigms may not be equally fruitful, and, as Kuhn (1970, 263) wrote:
[N]ature cannot be forced into an arbitrary set of conceptual boxes. On the contrary. The history of developed science shows that nature will not indefinitely be confined in any set which scientists have constructed so far.
51In other words: The world provides “resistance” to our conceptualizations in the form of anomalies, i.e., situations in which it becomes clear that something is wrong with the structures given to the world by our concepts. Therefore, Kuhn’s position, and thereby also SE, may be interpreted as a realist philosophy.
52Epistemology is important for all science because defending and arguing properly for any knowledge-claim in the end involves arguments about epistemology. In this article it has also been argued, contrary to Goldman, that individualist epistemologies are not tenable. The realization of this is mainly due to the collapse of logical positivism and the flourishing of historically, socially, and pragmatically oriented epistemologies such as Kuhn’s theory of scientific paradigms.
53Our focus has been the importance of SE for classification research, which was the context in which Shera (1951) first used this term. It is also the field of research which is the core interest of the present author. In many fields of knowledge, including biological taxonomy, principles of SE seem to be justified. SE has also been used outside classification research. The field of scholarly communication, with its primary, secondary, and tertiary document types and information services, i.e., the chain from publishing in a journal, to indexing the article in a database, and to aggregate knowledge from individual papers, for example, in a review article. All these activities are done by different agent informed by some kinds of epistemological views, which may be in conflict or in agreement. The identification and evaluating of these views are important because they influence the validity of (1) the knowledge claims in the primary literature (2) the relevance of the documents retrieved by secondary services, and (3) the bias in the aggregation in the tertiary literature. This is a broader perspective for SE, which, for example, has been investigated by Andersen (2002).
54SE is therefore extremely important for information science in general, as well as for KO with classification research. The term SE is used by different schools of research, and this paper has argued that the perspective introduced by Shera (1951) is the most fruitful one.