2.1. The Domain Knowledge/Activity Dimension

On the most basic level ‘conceptual domain knowledge’ (a) is internalised knowledge of a discipline’s research objects (logical entities and concepts, philosophical arguments, proofs, etc.), as found in, for example, books, articles or other media from the domain (Goldsmith, Johnson, & Acton, 1991).

‘Domain activities’ (b) are the communicative and noncommunicative acts (‘doings,’ cf. Schatzki, 1996) of logicians and philosophers within their domain. These activities can serve intellectual or coordinating functions. Where they serve an intellectual function they are directed at joint knowledge building (Cress & Kimmerle, 2008) and addressing the research questions of the domain or project, and are therefore closely related to conceptual domain-specific knowledge (a). The coordinating functions include all the practical, research-related questions in a domain that require communication, for example, project coordination, research dissemination, funding and cooperation over publishing projects, and organisation of meetings and conferences (Rolland & Potter, 2017). Coordinating labour is only indirectly linked to research questions and domain knowledge, and there is a clear hierarchy of research activities, with intellectual labour regarded as proper research and coordinating labour as merely facilitative.

‘Domain community knowledge’ (c) highlights that a domain researcher is part of a disciplinary community of researchers, with a collective history, norms, quality standards, criteria for good arguments and good research, and academic motivational systems (Elsayed et al., 2011; Tompkins, Perry, & Lippincott, 1998). This set of norms, criteria, and standards guides researchers’ activities in the domain (b), whether intellectual (e.g., discussion about a research paper) or coordinating (e.g., department meetings, discussing funding possibilities, etc.).

Domain-specific procedural knowledge typically consists of sequences of activities which ‘make sense’ (Dervin, Foreman-Wernet, & Lauterbach, 2003) in the context of the domain concerned. Thus, domain community knowledge determines the order in which tasks are executed and the nature of communicative activities in the context of the domain concerned. One example would be the structuring of processes a domain uses for information seeking (Case, 2012; Engerer & Gudiksen, 2016), from seeking a reference and accessing the full text document to checking its relevance to the researcher’s interests, downloading the reference into the researcher’s reference-managing program and relating it to other references therein, downloading the full-text document to a target destination with accepted ordering principles (for example a list in a Dropbox folder that is ordered alphabetically by author), and so on. This also demonstrates that not all activities in the domain are necessarily communicative; an activity such as downloading a document makes sense for the agent alone by virtue of being preceded by saving the corresponding reference (a prerequisite for subsequent citation) and succeeded by the naming of the document file according to the ordering principle of the researcher’s private repository (to ensure it is easily retrievable) (Østerlund, Snyder, Sawyer, Sharma, & Willis, 2015). These three aspects of domain research are linked as Fig. 1.

새창으로 보기

Fig. 1.

Relationships between conceptual domain knowledge (a), activities in the domain (b), and a community’s tacit, practical knowledge how to do things in the domain (c).

A community’s tacit, practical domain knowledge (c) organises the activity within the domain (b) and structures activities having directly to do with the domain’s disciplinary objects of knowledge (intellectual labour) and other more practical research-related communications (coordinating labour). At the same time, these communicative activities instantiate and externalise the tacit, procedural knowledge of the research community (c). There is also a two-way relationship between intellectual labour-related communication and cognitive knowledge-building (a). This connection highlights that research is not an isolated, individual process of cognitive knowledge-building, but, like all learning (Wenger, 1998), is a social and communicative activity as well.

2.2. Information Science Perspectives on the Domain

These three knowledge/activity aspects of researchers in a domain yield three distinct perspectives on information science knowledge and how it is brought into play in an interdisciplinary project where information science is the supporting discipline.

Perspective 1, which relates to conceptual domain knowledge (a), entails interdisciplinary interaction in which information science concepts such as ‘information’ and ‘knowledge’ are integrated into logical and philosophical reasoning and research (Floridi, 2011). Today interdisciplinarity is a research discipline in its own right (Frodeman, Klein, & Mitcham, 2010; Klein, 2010; Krohn, 2010) and it has already provided fruitful concepts and a theoretical background for the analysis of information science’s relationships with other disciplines (for a study exploring the interdisciplinary relationships between information science and linguistics see Engerer, 2017b). As this paper is concerned with information science in a supportive, rather than constitutive, role in collaborations the interdisciplinarity perspective will not be discussed further.

Perspective 2 is connected with the activities in a domain (b) and relates information science knowledge to that domain’s digitally mediated activities in a project context. From perspective 2, digital communication tools are typically conceptualised as information systems (Boell & Cecez-Kecmanovic, 2015; Urquhart, 2018), distinguished from other system components by their high degree of functional autonomy, input/output features, and knowledge-organising properties (Hjørland, 2003, 2008, 2013) and interactivity (Borlund, 2013; Kiousis, 2002; Ruthven & Kelly, 2011), as well as by their typically dichotomous functional structure. In this structure ‘content’ is linked to (or ‘mediated by’) an interpreting receiver, either a human user or a machine (Engerer, 2017a). Under perspective 2 information systems are often regarded as analytically closed systems which can be studied in isolation. In the following section this line of thought is elaborated and illustrated by an analysis of the Prior website’s information systems.

Under perspective 2 the concept of self-contained information systems relates to both intellectual and coordinating labour. Expertise for coordinating functions includes informing the development of project platforms, customising of wikis for information-sharing, integrating conference programmes into a researcher’s work desk, embedding tools for announcing project activities into a project website, and developing and maintaining other information systems that directly or indirectly facilitate coordination of a project.

An intellectual labour orientation implies a focus on the knowledge aspect of digital information systems, which serve as media for domain researchers in their cognitive and joint knowledge-building. Information science has a much richer tradition of studying systems for intellectual labour rather than coordinating labour and emphasises that distributed and equal access to domain-relevant information resources and joint terminology are important to the sharing of domain knowledge. One of the characteristics of these systems is a special type of dichotomous functional structure, a nexus linking indexing/metadata and user queries. In the context of intellectual labour information systems deal with both the metadata and representational properties of items of information (Chowdhury, 2010, p. 1; Frohmann, 1990; Lancaster, 2003; Mai, 1999; Svenonius, 2000), typically documents and their retrieval (Baeza-Yates & Ribeiro-Neto, 2011; Pandey, 2003; Ruthven & Kelly, 2011; Warner, 2010). Examples of this kind of dichotomous information system are digital full text repositories, bibliographies, library catalogues, directories of domain-relevant sources, and other digital aids which help researchers to access the information they need directly.

Perspective 3 relates information science knowledge to another kind of knowledge, i.e. practical domain community knowledge (c) about the domain activities (both intellectual and coordinating) of logic and philosophy specialists (b). Practical knowledge of the type c) specifies how things are done in the domain, why they are done this way, and which digital tools are typically used (Wenger, 1998). The first step in connecting information science knowledge with a domain community’s practical knowledge is to elicit and externalise tacit domain knowledge, and the second is to map it onto complexes of related information systems in the domain. Hence the question which information science knowledge can be drawn upon has a less straightforward answer than under perspective 2. It is not simply a matter of information systems typology. Faced with the task of eliciting tacit practical knowledge (Nonaka, Toyama, & Konno, 2000), information science can fall back on methods such as domain analysis (Hjørland, 2002), ontology building, user studies, and other qualitative, ethnomethodological techniques (Boaduo, 2011; Daniel, 2011; Pickard, 2013), which can be used to help establish the practical understanding that domain researchers have of their own and their peers’ activities. From these approaches we will in the following paragraph elaborate in more detail on domain analysis and ontologies.

As highlighted above, a perspective 3 approach to designing, evaluating, and improving scholarly digital resources builds basically on a correspondence of research community features on the one hand and the structure and design of the information systems used in that community on the other. The more general postulate of a close relationship between user features (e.g., linguistic, social, situational, professional) and the structural makeup of the knowledge systems for these users is in information scientific terms described by the notion of the “domain” (here used in a somewhat narrower sense than in the remainder of this article). A domain from an information science view captures the dependency between knowledge of a specific subject field (such as time logic) and skills in managing and organizing the information resources specifically of that field (such as the Prior Internet Resources [PIR]). Domain analysis takes the view that managing information resources and information systems (databases and websites) in a specific field demands knowledge of this field including its traditions, terminologies, norms, and practices (Albrechtsen, 2015; Bawden & Robinson, 2012; Hjørland, 2002, 2017; López-Huertas, 2015; Robinson, 2009; Tennis, 2003). Birger Hjørland devised a rude but useful and pragmatic methodology (better: systematics) for producing domain-specific knowledge necessary for information scientists in solving their tasks. This methodology includes approaches such as subject gateways, specialist classifications and thesauri, disciplinary peculiars of indexing and retrieving practices, user studies, bibliometrical studies, document and genre studies, terminological studies, historical studies, and more (Hjørland, 2002). Domain-analysis is a practical approach and thus appears as a good starting point for information specialists to systematically collect knowledge about the practices, modes of information seeking, language and communication conventions, and the more in the domains of other disciplines.

A further refinement from a perspective 3 standpoint is ontologies. The information scientific concept of an ontology encompasses the sphere of indexing terms and related search terminology at the same time, and therefore regards index terms as closely linked to (if not identical with) the vocabulary used by specialists in their domain. The step from traditional thesauri and classification schemes to ontologies of knowledge domains demarcates not only the integration of semantic web principles into the description of data (Berners-Lee, Hendler, & Lassila, 2001), but is foremost a move from barely developing a search terminology towards a controlled language for knowledge representation (Engerer et al., 2017). This step from the lexical-terminological component to a whole language with a built-in logic, a syntax, and inference rules makes it possible to derive information which is not explicitly contained in the descriptive terms themselves (Antoniou, Groth, van Harmelen, & Hoekstra, 2012, p. 4). Knowledge is therefore no longer just named, as it is the case in traditional controlled vocabularies, but it can be described and “confirmed” via ontology languages through constructing sentence-like complex formulas operating with linguistically-informed components such as subjects, verbs (relations), and objects.

Ontologies reflect a common understanding of a domain (Antoniou et al., 2012, p. 11) by expanding the restricted repertoire of thesaural relations between terms (e.g., broader/narrower terms, related terms) to an unrestricted range of semantic relationships realized and acknowledged in the domain idiom. Ontologies must therefore be constructed for each specific domain in order to reflect the language use practiced in the domain in question (Stuart, 2015). Generally, an ontology models the expert user’s view on information in his/her domain. More practically, in the case of PIR, the information scientist collects terms, their definitions, and mutual semantic relationships and builds a formal vocabulary system, including syntactical and inference rules. The advantages of ontologies for specialist users are, among others, improved possibilities for exploring data, “semantic search” (King & Reinold, 2008, p. 22), enhanced serendipity, and optimized search results by using ontology-based search techniques including Natural Language Processing (King & Reinold, 2008, p. 12).

Not unlike domain analysis, a complete methodology is linked to the creation of ontologies in specific domains, including steps such as collecting the vocabulary, defining and classifying the vocabulary terms, and indicating the semantic relationships between the established classes (King & Reinold, 2008, ch. 3; Stuart, 2015). Building an ontology is therefore in a way similar to doing a domain analysis as described in Hjørland (2002); both methodologies aim at transferring expert knowledge, often in tacit form, into the realm of explicit knowledge organization, and both respect the linguistic form of this knowledge when modeling it in a knowledge system. Domain analysis and the development of expert ontologies can therefore be regarded as two sides of the same coin, though ontology building is a more specialized activity and part of the broader theoretical endeavors and coverage of domain analysis. The three perspectives are mapped to their respective domain categories a) to c) in Fig. 2.

새창으로 보기

Fig. 2.

Three information science perspectives assigned to three knowledge/activity modes of domain researchers.

The illustration highlights two opposite, yet complementary and necessary strands of the process of bringing information science knowledge together with research in other domains (marked with grey). Setting the interdisciplinary perspective (1) aside, we can see that perspective 2 takes an analytical, isolationist approach to information systems, viewing them as self-contained units that are studied by information science in a rather straightforward way, in particular with respect to information systems for intellectual labour. In contrast perspective 3 takes a synthesising, holistic, and integrative approach to information systems, prompted by the tacit, procedural nature of domain communities’ knowledge, which means that existing practices and workflows must be reflected in the interplay between different information systems (Nicolini, 2013).

Below these two perspectives on information systems, the analytic-isolationist and the synthesising-holistic, are elaborated. It is demonstrated that information scientists need to use both approaches in order to access relevant professional knowledge and respond to domain researchers’ two essential ways of constructing their disciplinary domain, namely in terms of what they do (b) and what they know ([a] and [c]).

3.1. From Websites to Domain Knowledge

The path from websites to knowledge via information systems and knowledge types is illustrated in Fig. 3. Below the individual steps are justified and discussed in greater detail.

Inspecting the illustration from left to right we can see that the relationships between formal websites and information systems are not necessarily one-to-one; an information system (in this example information system 3) can formally be distributed over two or more web domains (indicated by a circle overlapping with two websites) and, perhaps trivially, one website can contain more than one information system, as it is the case with website 1, in which information systems 1 and 2 are embedded. The relationships between websites and information systems connect the digital, formal organisational level with the information systems level, marking the transition from a rather non-specific domain of communication to a functionally more specific domain, in our case the informational domain.

Looking further to the right of the illustration we reach the knowledge representation level (Blair, 1990; Stock & Stock, 2013), where the type of information system is specified. Again, a non-unique relationship connects knowledge types with information systems. One information system can be related to just one type of knowledge (in this example information system 1 is connected only with knowledge type X), or two information systems (here, 2 and 3) may be instantiations of the same knowledge type (here, type Z). Note that under perspective 2 the knowledge types (and hence also the domain knowledge) are assigned to the information systems regardless of 1) how many instantiations they possess on the level of information systems and 2) their formal organisational properties, which is the level at which the context of these systems is determined (in other words, the context in which an information system is embedded is irrelevant to the assignment of knowledge types). Clearly, therefore, perspective 2 is unsatisfactory, and perspective 3, which is complementary, compensates for this, as shown below.

Moving from information systems to knowledge representations is a step towards domain specialisation. This specialisation is achieved by mapping non-specific information systems onto an information scientific nomenclature for domain terms. This acknowledgement of the disciplinary terminology of a knowledge-organising system (Temmerman, 2000), not unlike a thesaurus system (Broughton, 2006; Foskett, 1994; Lykke Nielsen, 2001), is the key to the conceptual domain-specific knowledge.

The right side of Fig. 3 shows the linking of conceptual domain-specific knowledge (more commonly termed ‘expert knowledge’ or ‘expertise’) and information system types. Relationships between knowledge-type terms and the knowledge system itself are also not unambiguous (although this is not indicated in the illustration). In practice an information system type can map to more than one knowledge system or concept; similarly, two or more different information system types can relate to the same kind of knowledge system or concept. Theories of knowledge organisation treat these relationships as ambiguous descriptor terms in a thesaurus system (our knowledge types) with several, often incompatible ‘meanings,’ ‘definitions,’ ‘referents,’ ‘semantics,’ etc. (our conceptual domain-specific knowledge). In traditional information organisation, ambiguities like the ones mentioned above are treated as mismatches and undesirable drawbacks (Svenonius, 2000). In our very different context we might simply conclude that the information professional faces the challenge of identifying the most appropriate domain knowledge for the information system type under inquiry.

Once the relevant information science knowledge is identified, it must be projected into the realm of information systems where it guides analysis and, eventually, prompts interventions. In our illustration, the transfer of partial information science knowledge into information systems is indicated by backwards arrows.

3.2. Illustration of Moves Using the PIR

In this section the abstract moves presented in the previous section are illustrated using cases taken from the PIR. The numbers 1-6 designate the six information systems in the PIR, as illustrated in Fig. 4. Methodologically, the identification of information systems (3.2.1) differs from the identification of knowledge types (3.2.2) and conceptual knowledge (3.2.3). While the status of an information system is mainly a matter of design affordances, web site architecture and, most importantly, the way user practices apply to these affordances, the identification of knowledge types and concepts in relation to these systems is guided by disciplinary, information science meta knowledge regarding how to assign a knowledge type to an existing information system and connect this type with the relevant knowledge domains of the discipline. Thus, on the one hand, whether a website facility functions as an information system or not is to a high degree dependent on the user practices, here the uses that time logicians (logicians working on time as a philosophical subject), philosophers, historians, and more make of a website. On the other hand, assigning information science knowledge types to digital systems and accessing the corresponding information science concepts is solely based on information science practice and method.

새창으로 보기

Fig. 4.

General structure of Prior Internet Resources (as of summer 2017): six information systems representing five distinct information system types, implemented on three Internet domains functioning in three areas.

We can therefore methodologically ground the move from websites to information systems (3.2.1) by addressing design features, website architecture characteristics, and the information practices of the website’s user group. Decisions concerning types and concepts (3.2.2 and 3.2.3), however, must methodologically be based on information science practice, i.e. the ways how information specialists draw on their professional knowledge in developing information systems. With regard to the former, knowledge from interaction design (for example, Preece, Sharp, & Rogers, 2015) and information architecture (for instance, Ding, Lin, Zarro, & Marchionini, 2017) is coupled with informal assessments of the website and its information systems by members of the domain group (collected in informal conversations). With regard to the latter case, in which information professionals access the relevant disciplinary knowledge they need to apply to a digital system, only introspective reflection into their own professional practices reveals their competencies and practical skills. These reflections have been discussed in the information science group of the project and have been discussed by the information science group in several publications (Engerer et al., 2017; Engerer & Albretsen, 2017; Engerer & Sabir, 2018). What we do here, in a nutshell, is to make explicit a posteriori our own tacit model of accessing and applying information science disciplinary knowledge in our concrete practice of website development. In this sense, other researchers may arrive at other ‘models’ in their own practice, but it is finally this awareness and explicitness that is the basis for development and improvement in professional environments.

3.3. Information Systems in the PIR

The analytic approach will now be illustrated using the PIR, which comprise the Internet resources on Arthur Prior that are associated with the Prior project. The term PIR encompasses both formal digital elements such as websites, knowledge-organising units such as bibliographies, and other information systems, which will be presented in further detail below. The overall structure of the PIR, depicted in Fig. 4, consists of three main components, i.e. ‘content areas’ (these do not strictly coincide with particular Internet domains). ‘Foundations of Temporal Logic—The WWW-site for Prior-studies’ (Hasle & Øhrstrøm, 2016), hereafter ‘Priorstudies,’ is the main entry point (research portal) for scholars interested in Arthur Prior’s work and life. The related ‘Virtual Lab for Prior Studies’ (Albretsen, 2016), hereafter Prior Virtual Lab (PVL), is the virtual platform used by researchers transcribing handwritten documents by Prior. Finally, we have the so-called ‘Nachlass,’ a full-text archive of transcribed and published Prior manuscripts (Nachlass area). As already mentioned, our analysis refers to the Prior sites as they were until 2017, not taking into account subsequent changes and design modifications. This older version is accessible and archived at http://web.archive.org/web/20070609124540/ http://www.kommunikation.aau.dk/prior/index2.htm.

As indicated in the preceding section, the information systems have to be identified first. The term ‘information system,’ which has its roots in the world of management and business (Burton Swanson, 2009), refers by default to IT-based support to enable organisations to accomplish specific tasks (cf. Wallace, 2015), but definitions and conceptions of information systems vary considerably according to whether the perspective is technological, social, sociotechnical, or process-oriented, as a thorough review of information systems’ definitions (Boell & Cecez-Kecmanovic, 2015) demonstrated. Perhaps the broadest definition of an information system—but still meaningful in the context of this paper—is the one that comes from Wikipedia; it is widely cited in textbooks (for example, Bourgeois, 2014) and on conference websites (in the Wikipedia article itself no references are given). According to this definition “[a]n information system […] is an organised system for the collection, organisation, storage and communication of information. More specifically, it is the study of complementary networks that people and organisations use to collect, filter, process, create and distribute data.” (Wikipedia, 21 April 2017, https://en.wikipedia.org/wiki/Information_system, link marking and bold type removed).

This broad definition of information systems is practical and allows the approximate identification of six information systems embedded in the PIR.4 Each information system in the PIR has been glossed in the list below with the specific information-related action associated with it, taken from the definition above:

In information science, information systems are of several types—most prominently documentary languages implemented in knowledge organisation systems such as classification systems, thesauri, and ontologies (Hjørland, 2003, 2013; Stock & Stock, 2013, sect. L); information services such as bibliographies, retrievable databases, and text repositories; and, last but not least, research portals and collaborative academic platforms in general. In order to identify the information science knowledge relevant to the six PIR-embedded information systems, these systems have to be mapped onto specific types of information systems, such as the ones mentioned. The goal of this exercise is to enable systematic access to relevant and useful scientific disciplinary knowledge, which improves our understanding of the PIR and can be a starting point for professionals seeking to develop and improve the existing digital resources.

The PIR, defined as the virtual space delimited by the three abovementioned content areas and Internet domains (Priorstudies, PVL, and Nachlass), contains, on first inspection,5 six information systems of five distinct types. All information systems types are well-known and acknowledged in the information science research tradition, and disciplinary knowledge relevant to each type is readily accessible:

At the 2017 stage of the Prior project the digital information structure of the PIR consists of four theoretical levels: formal organisation (three Internet domains), information systems (six partial systems), knowledge representation (five information science subtypes), and conceptual domain-specific knowledge (five partial knowledge domains, corresponding to five information science subtypes). At the time of writing a complete restructuring of the website is in progress; some preliminary results are presented in Engerer and Albretsen (2017). Fig. 4 sketches the general structure of PIR in summer 2017.