3.2    Ontological and Epistemological Foundations

Enterprise modeling is a process that involves human inquiry into some domain of interest. Different views on the nature of this domain and the relationship to the inquiring actor is the theme of section 3.2.

3.2.1 What is meant by ontology and epistemology?

The term ontology concerns what is said to exist in some world -- that which potentially can be talked about. Wand and Weber (1993:220) refer to ontology as
"a branch of philosophy concerned with articulating the nature and structure of the world."
By ontology is sometimes also meant a set of terms and their associated definitions intended to describe the world in question (e.g., Uschold, 1995:1).

The term epistemology denotes (Hirschheim et al., 1995:20)

"the nature of human knowledge and understanding that can possibly be acquired through different types of inquiry and alternative methods of investigation."
Guba and Lincoln (1994:108) categorize alternative inquiry paradigms according to their stance on the following three questions: Hirschheim et al. (1995:21) seem to collapse the epistemological and the methodological questions, and so will be done in the following discussion. Ontological and epistemological issues become related in the sense that the latter concerns how human actors may go about inquiring about and making sense of the former.

When speaking of the phenomena that are of interest in the "world", the term Universe of Discourse (UoD) is used. This term is well established within conceptual modeling (e.g., Sølvberg and Kung, 1993:173).

Ontological and epistemological questions concern what is commonly referred to as a person's Weltanschauung or worldview. Weltanschauung can be described as (Merriam-Webster, 1997)

"a comprehensive conception or apprehension of the world especially from a specific standpoint"
Two different worldviews will be considered: An objectivistic (section ) and a constructivistic (section ). As will be argued in proceeding discussions, an actor's worldview has profound influence on the perceived relative importance of aspects of enterprise modeling.

The two strands presented here are intended as stereotyped accounts of possible worldviews. One may find actors enacting one of them or both of them. Weick (1995:35), in his advice to students of sense-making, even advocates what he refers to as ontological oscillation, "changing" view depending on the situation.

Again, as for discussions of alternative research strategies in section 1.3.1, none of the worldviews are considered to be unconditionally superior to the other. Both may be appropriate for some purposes and insufficient or overly complex for other purposes. Boland and Tenkasi (1995) provide a theory of knowledge production and use that allow for incorporation of ideas from both worldviews. Their theory will be presented in section and used for analysis of enterprise modeling in the sequel.

3.2.2 An objectivistic worldview

From an objectivistic point of view, the Universe of Discourse is comprised of distinct objects with properties independent of the inquiring observer (Hirschheim et al., 1995:58). If two observers do not understand a phenomenon in the same way, it is due to human imperfection, e.g., lack of training, errors of judgement, illusions or plain misunderstanding (Lakoff and Johnson, 1980:187). Disputes over the true nature of the UoD can be resolved through additional inquiry into the UoD.

Epistemologically, observers get their knowledge about the world by experiencing it. It is generally possible for humans to investigate the world without influencing it (Guba and Lincoln, 1994:110). The ideal approach to investigation of the UoD is the scientific method of research, and replicable observations are considered to represent the truth.

With an objectivistic worldview, a mountain is a mountain for everyone, a product is a product for everyone, and a work process is a work process for everyone. The meaning of a phenomenon is inherent to the phenomenon and can be experienced by interacting with it.
A characteristic of an objectivistic worldview is the existence of objective, absolute and unconditional truths (Lakoff and Johnson, 1980:159). The objective meaning of a statement is given from a set of conditions of truth or falsity, and human understanding is a matter of knowing these conditions (ibid.:198), requiring precise and unambiguous definitions and rational explanations relying on deductive logic. Hence, sense-making from an objectivist point of view is considered as rational analysis of data in a mental problem space and construction of deductive arguments of cause-and-effect (Boland and Tenkasi, 1995:353)

A widely accepted objectivist view of communication is as sending of messages between actors in accordance with Shannon's conduit model of communication (e.g., Shannon, 1987) or variants thereof. Shannon (ibid.:223) states explicitly that

"Frequently the messages have meaning; that is they refer to or are correlated according to some system with certain physical or conceptual entities. These semantic aspects of communication are irrelevant to the engineering problem."
Thus, the model concerns the transfer of a message from point A to point B, and is not developed for semantic aspects of communication.

3.2.3 A constructivistic worldview

Schwandt (1994:125) argues that what constitutes a constructivistic worldview is shaped by the use and users of the term, and there is no widely agreed upon view of the paradigm (this again is consistent with a constructivistic worldview). Still, some traits seem to be central (from Guba and Lincoln, 1994; Schwandt, 1994; Gjersvik, 1993b; Hirschheim et al., 1995; Floyd et al., 1992; and Weick, 1995).

Firstly, realities are local and specific in the sense that they vary between groups of individuals (Guba and Lincoln, 1994:110). Constructions, being ontological elements of realities, are not absolutely true or correct in any sense, only more or less informed and sophisticated (Schwandt, 1994:129). However, even if all constructions are meaningful, some may rather be termed malconstructions, as they obviously are too simplistic or inconsistent. Whether a construction is malformed depends upon the paradigm the constructor operates within.

Secondly, reality is actively constructed, i.e., not merely discovered. Hence, the distinction between ontology and epistemology is blurred, as what constitutes reality depends on a particular actor and his values (Guba and Lincoln, 1994:111). The objective, value-free actor does not exist.

Thirdly, reality is socially constructed, i.e., the constructions are not personal or technical (Dahlbom, 1992:101). Although perception and thinking necessarily is individual, the construction process involves other social and cultural artifacts and therefore inevitably becomes social.

To formulate a more powerful analytical framework, the work of Gjersvik (1993b) is investigated more closely. Gjersvik has adapted the original ideas of Berger and Luckmann (1967) to better fit organizations. The main elements of his model of social construction of reality in organizations are illustrated in figure 3.1.

Figure 3.1: The social construction of reality in organizations (Gjersvik, 1993b:37)

The process in can be read as follows, starting with local reality: Actors in an organization have their private local realities, that is, subjective beliefs, views and values. Local realities are systems of meaning that are perceived by the actor to be valid -- they are reality. Groups of actors can also be considered to have local realities (from an analytical point of view). Local realities are not explicated theories of reality, although elements may be reflected upon.

Actors make their local realities more or less accessible to other actors through a process of externalization, i.e., enactment of their local reality. Actors communicate through their actions and act through their communication (speaking is also acting). Gjersvik (1993b:47) considers the most common ways to externalize local realities as through construction of language, objects and artifacts, and institutions.

Externalization implies construction of an organizational reality through objectivation. The organizational reality is inter-subjective and may include elements like institutions, roles, language, knowledge, objects, routines, technology, commitments and history (Gjersvik, 1993b:51). Organizational reality structures and restricts human action and thereby makes actions predictable and to some degree also controllable.

Organizational reality is interpreted by individual actors and made sense of through a process of internalization. Internalization is interpretation of elements of organizational reality (constructions) in terms of the actor's local reality.

Externalization and internalization of realities occur both continuously and simultaneously, despite the neat sequence depicted in . The process is an ongoing part of all human activity.

When the actors in an organization are reproducing the same organizational reality, in the sense that the same questions, problems and solutions keep reappearing, one may talk of closure in the organization. Gjersvik (1993a:20) states that

"the term 'closure' may easily be interpreted as negative. However, closure is something that all human societies and organizations need in order to have a common orientation, so that joint actions are possible."
The negative aspects of closure arise when it occurs prematurely or is dominated and controlled by a few actors' realities. Premature closure implies that the organizational reality might not be optimal, and closure dominated by a few actors' local realities are candidates for conflict at a later stage (in case of breakdowns).
Agreement upon a method for enterprise modeling might illustrate the concept of closure. Assume that a project leader have prepared an enterprise modeling method for her project in advance. If the organizational reality enforces that "the boss is always right", the project participants might accept the method and work according to it, reproducing the organizational reality. However, if breakdowns in the process occur due to problems with the method, the premature closure of the method might strike back on the project leader, as her local reality dominated the closure. 
Sense-making from a constructivist point of view is a process of attributing meaning to constructions according to the actor's local reality and simultaneously influencing the local reality. How an actor interprets a phenomenon and whether it makes sense or not depends upon the local reality of the actor. This view is consistent with ideas associated with hermeneutics (Guba and Lincoln, 1994:109; Klein and Lyytinen, 1992:214ff). In the vocabulary of hermeneutics, the local reality can be viewed as a horizon, being the pre-understanding that all perceived phenomena are interpreted and made sense of against. An actor may access his horizon through a process of bracketing, being an attempt at isolating and investigating one's own presumptions.

Communication from a constructivist point of view is also a process of social construction, as externalization can be viewed as a generalization of speaking, and internalization is to perceive, interpret and make sense of the externalized constructions.

3.2.4 Perspective making and perspective taking in communities of knowing

Attention is now turned to a theory proposed by Boland and Tenkasi (1995) concerning knowledge production and use in organizations. The theory incorporates ideas associated with both objectivistic and constructivistic worldviews, integrating the discussions above into an overall framework for analysis of enterprise modeling.

Communities of knowing

Organizations doing knowledge-intensive work can be seen as comprised of multiple communities of knowing, being actors with similar expertise or perspective. A perspective in this context is comprised of shared vocabularies, methods, theories, values and accepted logic. Boland and Tenkasi (ibid.:351) refer to "thought collective", "community of practice" and "interpretive community" as similar concepts. Kuhn's (1970:10) paradigm concept is also discussed as reasonably consistent with a perspective. A strong perspective is required to perform knowledge-intensive work, and multiple communities are usually required or more effective when business domains are so complex that no single actor or community can be an expert in all sub-domains.

A community of knowing may be comprised of a single actor or a group of actors, but the main characteristic is that different perspectives in some sense are incommensurable (otherwise, they would not belong to different communities of knowing). As Boland and Tenkasi (ibid.:351) claim,

"Thought worlds with different funds of knowledge and systems of meaning cannot easily share ideas, and may view one another's central issues as esoteric, if not meaningless."
The construction of a perspective within a community of knowing is referred to as perspective making, and attempting to understand the perspectives of other communities as perspective taking.

Perspective making and perspective taking

Perspective making is a process in which the perspective of a community complexifies. Complexification implies the construction of a more detailed vocabulary and sensitivity towards more subtle nuances in the knowledge domain. As a perspective complexifies, the community becomes more able to do work that is covered by the perspective, i.e., within their area of expertise.

Perspective taking is claimed to be a foundation of coordinated action in organizations, enabling actors to make assumptions about other actors and their behavior. Boland and Tenkasi (ibid.:358) state that

"Much of social behavior is predicated upon assumptions an actor makes about the knowledge, beliefs and motives of others. This is the beginning of perspective taking, and is fundamental to communication."
A prerequisite of perspective taking is the ability and willingness of actors to be reflective. Reflectivity is (ibid.:362)
"the ability to periodically suspend our natural attitude and notice the matter-of-course, taken-for-granted ways in which our communities of knowing are constructed and interpreted, which can open possibilities to change them."
Without reflection, actors are not able to change their own perspectives. Hence, perspective taking is not simply an act of sharing information -- it is an act of changing one's conviction of what is true and correct and, at the extreme, questioning of one's own worldview.

Another prerequisite for perspective taking to be possible is that the perspectives of other communities are externalized in some way, e.g., in terms of artifacts. These artifacts are referred to as boundary objects, stressing the role they play in the communication of perspectives across the border between communities of knowing. Examples of boundary objects are narratives, cognitive maps and models.

Models of cognition and communication

The concepts community of knowing, perspective, perspective making, perspective taking and boundary object have different qualities and different meanings depending on the underlying theories of human cognition and communication (Boland and Tenkasi, ibid.:352). Boland and Tenkasi discuss several alternative theories and their approach is to allow for a multitude of theories to coexist within the same framework. The rationale for this is that all theories have strengths and weaknesses, making them more or less powerful for explanatory or predictive purposes. By allowing multiple models to coexist, the strengths of one model may counter the weaknesses of another model.

The theories of human cognition discussed by Boland and Tenkasi are a rational information processing mode (Simon, 1977) referred to as the paradigmatic mode, and a narrative mode (Bruner, 1990). For communication, the conduit model (Shannon and Weaver, 1949) is contrasted to a language game model (Wittgenstein, 1974). The details of these theories will not be discussed here (as they do not play a decisive role in the sequel). The essential observation is rather their ontological and epistemological assumptions.

Boland and Tenkasi (ibid.:354) provide an overview of key assumptions underlying on one hand the paradigmatic mode and the conduit model (e.g., reality is given, knowledge about it is universally true and realization of objective knowledge is a rational process) and on the other hand the narrative mode and the language game model (reality is socially constructed and knowledge about it results from narration and playing of language games). The underlying assumptions listed in (ibid.:354) match very well what has been discussed in sections and as objectivistic and constructivistic worldviews, respectively.

Hence, the theory is a promising basis for developing a framework for analyzing enterprise modeling as a means to support human sense-making and communication: Both existing frameworks for enterprise modeling and observations from empirical studies can be analyzed to see whether they fit most comfortably with the assumptions associated with an objectivistic worldview or a constructivistic worldview. In this way, a comparison of alternatives is enabled within the same analytical framework.

3.2.5 Concluding remarks on worldviews and theory

Section 3.2 has outlined the theories that will be used for analytical purposes throughout the thesis. The variety of ideas and references may seem somewhat chaotic, and hence, a brief recapitulation and clarification of terms is attempted. Table 3.1 summarizes the concepts of sense-making and communication from both an objectivist and a constructivist point of view. The term meaning is also included, as it is central to the thesis.
Concept  Objectivistic  Constructivistic 
Meaning  Inherent to the phenomenon, can be experienced or defined.  How the phenomenon is related to phenomena already a part of the local reality. Meaning is local to community. 
Sense-making  Rational information processing (identification of causal relationships).  Never-ending interpretation against changing horizon (hermeneutics), and construction of stories that are plausible and interesting. 
Communication  Exchanging messages (messages have fixed meaning).  Communication is externalization (action, both creation of artifacts and speech) and internalization (interpretation, sense-making). 
Table 3.1: Meaning, sense-making and communication according to alternative worldviews

The different qualities of sense-making described in are consistent with Weick's (1995:95) discussion of occasions for sense-making. In case of perceived uncertainty (in the sense lack of interpretations), sense-making involves scanning, discovery and retrieval of information that enables the actors to attribute meaning to the phenomenon in question. In case of ambiguity (in the sense too many interpretations), Weick calls for social construction and invention of coherent wholes (in terms of stories) that make various interpretations meaningful.

Table 3.2 provides an overview of core concepts, each described from either objectivist or constructivist point of view.
Concept  Objectivistic  Constructivistic 
Community of knowing  Group of more or less knowledgeable actors.  Group of actors with shared vocabulary, values, beliefs, etc. 
Perspective  Shared, objective knowledge. Focus on causal relationships, formal logic and precise definitions.  Local reality meaningful to a particular community of knowing, manifest as narratives, artifacts with local meaning, etc. 
Perspective making  Development of more accurate causal laws and more precise definitions. The scientific method is ideal.  Construction of a community's local reality through externalization and internalization. 
Perspective taking  Scanning, retrieval and processing of information.  Internalization of local reality of other community of knowing. 
Boundary object  Explicit and more or less accurate representation of an actor's knowledge.  Externalized construction being inherently ambiguous. Meaning depends on interpreting actor. 

Table 3.2: Core analytical concepts according to alternative worldviews

Considering the appropriateness of the competing models of sense-making and communication, an objectivistic worldview can be defended when the focus is on distribution of information with a meaning that is reasonably well known to the actors (e.g., within a community that have developed a stable and mature perspective). A constructivistic worldview may be more appropriate for discussion of communication between actors with very different backgrounds (different local realities, e.g., between different perspectives or within perspective that is immature and evolving).