Aravali Institute of Management Knowledge Management Term Paper Topic: Knowledge Management and Organization Behaviour (Tracing the interrelationship) Submitted To Mr. Prithwi T. Banerjee Faculty, AIM Submitted by Group No. 4 Harshad Vyas Om Prakash Suthar Bhawani SinghRathore Amit Mathur Gourav Rathi Abstract Organizations are collections of interacting and inter related human and non-human resources working toward a common goal or set of goals within the framework of structured relationships.
Organizational behaviour is concerned with all aspects of how organizations influence the behaviour of individuals and how individuals in turn influence organizations. Organizational behaviour is an inter-disciplinary field that draws freely from a number of the behavioural sciences, including anthropology, psychology, sociology, and many others. The unique mission of organizational behaviour is to apply the concepts of behavioural sciences to the pressing problems of management, and, more generally, to administrative theory and practice.
The quest for technologies with strategic value for the organization but also with empowering strengths for the work context of the firm has persistently occupied the landscape of information systems. Knowledge Management is the latest techno-managerial buzzword earmarked for improving the work processes and creating value for a firm’s operations. Knowledge Management comprises a multiplicity of technological offerings for potential applications. Nevertheless, there is scarce empirical evidence on phenomena, conditions and factors related to the organizational adoption of these offerings.
The research treats Knowledge Management systems as IS innovations by explaining organizational situations and phenomena related to its adoption. This research seeks a deeper understanding of organizational phenomena taking place during the adoption and implementation of KM technical solutions. Knowledge Management is being considered for adoption as a practice that could facilitate the sustainable development of new products and services and beyond that the transition to a radically different set of operational arrangements. Introduction K nowledge M anagement:
A Framework for Building a Knowledge Sharing Culture Organizational Knowledge, Cognitively Plausible Actors, and Multi-Actor Systems Organizational knowledge and knowledge management can only be studied successfully if two basic requirements are fulfilled: (1) determination of what knowledge is about and which carriers of knowledge are allowed and (2) the mechanisms that provide the interaction between the carriers (actors and software agents). We therefore have to step down to a lower level of aggregation, which is to say, to actors, to (shared) mental models, to agents and to the interaction between them.
In order to guide the study of these constituting elements we formulate two questions. (A) What is the difference between information and knowledge and what consequences does this difference have for corporate and organizational issues? (B) If the human individual is one kind of actor, what other kind of actors (or agents) can we discern, what characteristics do the various actors have and what mechanisms are used to collaborate in Multi-Actor Systems (MAS)? Insights from cognitive science, artificial intelligence and knowledge technology are used to answer the questions.
We see knowledge as interpreted information. For the time being only human actors can entertain knowledge, because they have representations. Taking into account other components of (intelligent) actors, such as perception and interaction, other kinds of actors (and agents) can be defined. Various kinds of actors (and agents) may work together in an organization, which we call a Multi-Actor System. The “glue” that keeps such a system together is called: coordination mechanism. Various kinds of coordination mechanisms exist such as standardization, authority, and mutual adjustment.
This also depends on the characteristics of the involved actors. The perspective of cognitive science combined with the assumption that organizations are a MAS make “organizational knowledge” and “organization” operational, measurable and quantifiable. Especially the focus on actor characteristics and as a result the actor/agent taxonomy being combined in a multi-actor system with various coordination mechanisms, makes it a better framework for an easy and smooth inclusion of and integration with (software) agents. More Details…
In many situations, the term “organizational knowledge” is very useful as a short description of what organizations know. “Organizational knowledge” is also the basis for “organizational learning” (OL). One can only learn if one knows already something. In this article, we argue that this organizational perspective can only be studied fruitfully if two basic requirements are fulfilled: (1) a determination of what knowledge is and is about and which carriers of knowledge should be taken into account and (2) a determination of the mechanisms that provide the interaction between actors and possibly software agents.
We first state that the term “organizational knowledge” is a metaphor, a way of speaking. Organizations literally do not have knowledge. Human individuals, or to be more precise, the brains/minds of humans have knowledge. With this knowledge, humans work with each other and with other kinds of actors (software agents), such as (advanced) information systems. For reasons of clarity, we prefer to use the term actor for humans and the term agent for software entities. Similarly, “organizational memory” and OL are metaphors.
They are useful, but bounded and limited. Their usefulness lies in the fact that with these terms we can describe complex artifacts and constructs in a ready-made and short-handed way. However, their limitations are also clear, namely that you easily borrow properties or attributes from the one field and apply or ascribe them to the other. For example, talking about a diploma given to employees for following courses (an indicator of individual knowledge) and stating that the whole group has “corporate knowledge” is strange.
Alternatively, talking about the speed of “corporate memory”, which characteristic is relevant in computers and human memory, is at least quite beside the intention of using the metaphor. Metaphors are therefore inspiring, but on the other hand persuasive and occasionally wrong. Although discussions about knowledge (and learning) are very prominent in management and organization studies (Dalkir, 2005; McElroy, 2003), we also believe that often knowledge (and learning) are not made operational, quantifiable and measurable.
As we said, a term like “organizational knowledge” is a metaphor and “measuring” properties of metaphorical entities is often misleading. Two directions can then be chosen. The first is that one forgets about making the concepts operational, that is to say that one is satisfied with qualitative instead of quantitative observations, often at an abstract level. The second is that one tries to redefine the concept of e. g. , organizational knowledge into concepts at lower levels of aggregation and thereby tries to change a metaphorical into a more literal description.
This requires assessing the constituting elements of organizations, that is to say human individuals (actors), computer systems (agents) and the adaptive processes of them, individually as well as collectively. By presenting a framework and its constituting elements, we show that the second route is rewarding. To make various aspects of knowledge operational and quantifiable, we will step down to lower levels of aggregation, which is to say to characteristics of actors, to (shared) mental models, to agents and the interaction between actors and agents.
This does not imply that organizations are not relevant any more. On the contrary. Organizations are combinations of actors, more and more integrated with agents. We call them Multi-Actor Systems (MAS), in which the “multitude” exists in the various kinds of actors (and agents) with knowledge involved and in the coordination mechanisms that make the multi-actor system function. In our view, knowledge – and consequently learning, which we only discus as far as it is relevant for our focus on knowledge – is something human actors have, respectively do. This knowledge is based on information.
However, information is different from knowledge and a human actor is different from an organization or from complicated (software) agents. Although literature discusses the levels distinction in organization and individual, we believe that a real cognitively plausible perspective on human actors as information processing systems is missing and that consequently the treatment of actors (and agents) in KM literature is superficial and not very operational. What often remains is “lip service” to real humans in organizations, where in fact one discusses empty, or at best, very simple actors.
We claim to present a framework taking into account intra- and inter-individual mechanisms based on cognitive science, artificial intelligence and organizational semiotics (Helmhout, 2006) . To structure our analysis of the observable, empirical and measurable constituting elements, we formulate two questions: 1. What is the difference between information and knowledge and what consequences does this difference have for corporate and organizational knowledge issues? The determination of knowledge implies a further distinction in the content and the form or presentation of knowledge. 2.
If a human individual is one kind of actor, what other kind of actors (or agents) can we discern, what characteristics do the various actors have and what mechanisms are used to collaborate in a MAS? We start our discussion about knowledge and actors in Sect. 2 with a review of the literature within knowledge management (KM) (and organizational knowledge). Issues in Organizational Knowledge and Knowledge Management 1 Carriers of Knowledge and Coordination Mechanisms According to McElroy (2003), the discipline of KM evolved from the fields of OL and ICT (see for instance Firestone and McElroy, 2003).
According to McElroy (2003, p. 82), the field of information technology unjustly claims that KM is nothing more than the application of information technology. ICT forms the technological side of most KM approaches (Alvesson and Karreman, 2001). The field of OL concerns how learning by organizations takes shape (Argyris and Schon, 1978), and sets out the human-side of KM approaches. OL is described in terms of single-loop – learning within the boundaries of assumptions that apply within the organizational context -, and double-loop learning – learning by questioning the boundaries of the assumptions and changing them.
For long, learning by individuals has not been addressed within OL (e. g. , Kim, 1993). Kim (1993) bridges the gap between learning at the organizational and individual levels. He presents the OADI-SMM model, combining insight from OL and Kofman’s (1992) view on individual learning. OL and individual learning are linked through the notion of shared mental models (SMM), a notion that is similar to Helmhout’s concept of social construct. Although Kim (1993) presents a comprehensive model that convincingly links the two levels of learning, we argue that two issues are missing in his approach.
The first concerns the identification of the carrier of knowledge. Because knowledge cannot exist without someone keeping it in existence, we argue that a theory that deals with knowledge (and learning) should specify a cognitively plausible carrier. Although Kim (1993) touches upon this topic, in his model he does not refer to humans as cognitive, information processing entities. He starts from the position that the carriers of knowledge are human beings, but other carriers are storage facilities within an organizational context (e. g. paper, or computer systems), initially perceiving these carriers as equal. When discussing the likelihood for an organization to survive a total disintegration of all computer 1 See http://en. wikipedia. org/wiki/Knowledge_management http://en. wikipedia. org/wiki/Securitization. http://www. citywire. co. uk/adviser/-/news/property-and-mortgages/content. aspx? ID=293217. systems and paper material versus a complete replacement of employees, Kim favors the former situation. In other words, computer systems and human beings are not equal carriers of knowledge.
This is true, but what do the differences consist of? The second element that lacks from Kim’s model is the notion of coordination mechanisms between individuals who learn. Kim identifies organizational structure and size as the distinctive element between individual and OL. He argues that within small groups the two levels of learning are indistinguishable because of little structure (Kim, 1993, p. 40). As the group or organization grows, structure becomes more apparent, and organization and individual learning become distinctive realms.
In essence, we agree with this position. What Kim, however, hardly discusses is that there is no learning without knowledge and that this knowledge implies knowing how to function and act organizationally in a firm. Knowledge about the organization in which an actor operates implies a discussion about coordination mechanisms. As we will discuss later, the use of a coordination mechanisms stands apart from an organizational structure. Additionally, the concept of coordination mechanism links directly to the levels of individual and OL and therefore knowledge.
First-Generation Knowledge Management: Knowledge Distribution and Use2 KM is a relatively young discipline in both research and practice (Dalkir, 2005). From the start, the main objective of KM has been to get the right information to the right people at the right time in the right quality, in the right shape, against the lowest costs (Schreiber et al. , 2000; McElroy, 2003). According to McElroy (2003) , KM was highly technocratic in the beginning. Information technological applications dominated KM practice from the start (Ruggles, 1998).
This is understandable, for KM has been perceived as nothing more than the distribution, delivery and transformation of information. McElroy calls this kind of KM, first generation KM. Now, this approach within KM does not suffice anymore. KM, in its present second generation, has become much more than just delivering and transforming information and using information technology. It is now also about knowledge creation and production (Sect. 2. 3). Based on McElroy (2003), we first give a general picture of knowledge and KM (first generation) in organizations (see Fig. 1). In the figure, real world or environment is external.
Within the environment, there is a system or organization, with business processes (BP), with subjective and objective, coded and theoretical knowledge in what we call the distributed organization knowledge base (DOKB), governed partly by KM. It should be noted that we are describing the system here at a higher level of aggregation than just the human individual (the actor). 2 The Guardian Newspaper, 28 March 2008, 36 pp. Also in the perspective of McElroy, knowledge is not integrated with cognitively plausible actors. Furthermore, the system or organization consists of a multitude of various kinds of actors, what we call MAS.
McElroy (2003) states that KM starts from two assumptions that, he says, do not align with reality. The first assumption is that knowledge that is needed by individuals to perform a certain task is already there. It exists. From this assumption, the need arises to capture and codify knowledge to make it easy to handle, to transform it into manageable pieces of information that can be easily transported. The choice for information technology as an instrument for KM is logical, for it brings forth tools that realize the fast transport of information and in this way; it contributes to the realization of organizational efficiency.
From the first assumption, McElroy derives a second assumption that underlies KM: the equivalence of information and knowledge. Knowledge not solely depends on humans as carriers or social processes as distributors. McElroy concludes that within an organization one presupposes that just feeding humans with the information they need, will result in desired outcomes, and will make organizations perform better. The two assumptions lead to two consequences for KM. First, because knowledge exists already by definition, one only has to focus on the transfer of knowledge.
Just making sure that the right information reaches an individual at the right time, in the right shape and quality is sufficient from this KM perspective. Second, information technology is the key technology within KM. Therefore, KM is all about information transfer against the lowest costs. Information technology enables a fast transfer and transformation of all kinds of information against relatively low costs. In other words, information technology is KM’s “silver bullet. ” The codified perspective of information management resulting in KM, presupposes that human actors are controllable and predictable.
It, therefore, suffices to prescribe behavior they are allowed to display in rules and policies. Furthermore, the controllability of humans enforces the use of information technology. Because humans will behave as predicted, the inflexibility that often characterizes information technology does not stand in the way of efficient organizational behavior. If the two assumptions hold, KM should pursue the mentioned objective making strong use of information technology. Meeting the objective then should result in the functioning of organizations as if they were well-oiled machines (another etaphor). Problems of course arise when one or both of the assumptions do not hold. What should be done in case not all knowledge already exists and that therefore knowledge transfer is insufficient? Alternatively, what should be done if human interactive behavior as such does matter regarding knowledge and KM? Several authors have abandoned the assumption that knowledge already exists. They suggested that KM theories need a broader basis (Nonaka and Takeuchi, 1995; McElroy, 2003; Jorna, 2007). If all knowledge would exist, innovations would not be necessary anymore.
The focus on information systems in first generation KM has resulted in a neglect of knowledge creation and production within KM. McElroy and others suggested broadening and transforming the “old KM school” with structures to cover the creation of knowledge. The second assumption underlying KM – the similarity of information and knowledge – contrasts with our notion of knowledge in which only humans are carriers. Knowledge is something human actors posse and use, also in social interactions. McElroy (2003) concludes that, in addition to including a structure for knowledge creation, a theory of KM should focus on humans and social processes.
Second-Generation Knowledge Management: Knowledge Creation Dalkir (2005) provides an overview of KM theories that all describe some structures of knowledge creation. She presents theories by Meyer and Zack (1996), Bukowitz and Williams (1999), McElroy (2003), and Wiig (1993). Here, we adopt McElroy’s (2003) approach towards the creation of knowledge, a process he labels as “knowledge production”. We already mentioned that he labels the earlier theories of KM, in which information technology prevails and knowledge is assumed to exist as “first generation KM”.
In contrast, when KM theory incorporates knowledge production – leading to knowledge processes – McElroy speaks of “second generation KM”. Humans and social processes form the centre of his notion of second generation KM (see Fig. 2). The difference between Figs. 1 and 2 concerns the emphasized aspect of knowledge processing that is closely related to creative learning. KM should provide the policy to enable or endorse these processes. Regarding knowledge creation, McElroy (2003) speaks in terms of knowledge claims, a concept that links to the well-known concept of hypothesis in scientific research.
The process of knowledge production builds upon these knowledge claims. Knowledge production roughly consists of the sub-processes of knowledge claim formulation and knowledge claim evaluation. In the former, individuals collaboratively formulate a testable knowledge claim. Through information gathering and discussion about this information, knowledge claims are formulated. Both information gathering and discussion rely on human involvement and social interaction. Formulated knowledge claims are subsequently tested in the sub-process of knowledge claim evaluation. Peters et al. 2008) identify various types of evaluative mechanism that can be used to evaluate knowledge claims. An evaluated knowledge claim will become part of the overall “organizational” knowledge and will be integrated and so the knowledge processing continues. In summary, the incorporation of knowledge production processes in KM theory emphasizes the role of humans concerning knowledge processes and the importance of “inside knowledge”, i. e. , the knowledge held by human actors within the organization (Fearon and Cavaleri, 2006). The next step is to determine the ways to manage the process of knowledge production.
First generation KM relied strongly on the use of information technology. As indicated, this was possible because the role of human actors was largely ignored. In contrast, human and social processes reside at the core of second generation KM, and the use of information technology is not that straightforward anymore. We are talking here about information technology from the late 90s. However, the role of information technology has changed since then. The developments and possibilities of new software tools have dramatically increased the last ten years.
Until now, we humans are the only syntactic, semantic and pragmatic intelligent machines. Our digital “companions” (agents) are mainly syntactic machines and as far as they are semantically and pragmatically sound, it is because we humans provide the interface(s). We belief this might change within the next twenty years (see also Harper et al, 2008) . This brings us to the significant part concerning actors. Currently, our artificial actors are mainly assistants and companions. What characteristics – also with regard to knowledge – are required to be an actor and what are its implications for MAS?
We, therefore, change our focus. We direct our attention to actors, the basic elements within any organization. Information, Knowledge, Cognition, and Human Actors We see data, information and knowledge as parts of a three-stage rocket (Jorna and Simons, 1992; Schreiber et al. , 2000). Knowledge assumes information, and information in turn assumes data. The bottom layer is formed by data. Data are the noises, scratches, images and other unstructured elements, out there in reality. If the data are interpreted explicitly, we speak of information.
If this information is used by people in reasoning or in performing actions – i. e. , if it is interpreted – we have knowledge. This means that going from data, via information to knowledge, degrees of freedom increase. The same data can be interpreted in many different ways to serve as information. In a similar way, information can be interpreted in many different ways to serve as knowledge. Someone receives data and information, and with the aid of knowledge the person already possesses, information becomes knowledge, which, in turn, can consequently complement or change a person’s current knowledge.
The crucial difference between information and knowledge is interpretation and this interpretation is done with the human mind. It implies a cognitive perspective on knowledge. Presently, humans are the only carriers of knowledge. They are goal-oriented sign or symbol processing systems (Jorna, 1990; Newell, 1990). What is exclusive for humans as carriers of knowledge, here, is different for information and data. In that situation, not only humans, but also other kinds of actors (software agents) are involved. With respect to the knowledge actors have, various divisions can be made.
The most important one is the distinction in knowledge content and knowledge type. Knowledge content concerns what knowledge is about: about cars, about physics, about making coffee, about computers or about coordination mechanisms? Domains, fields and disciplines are examples of knowledge content. Postrel (2002) calls a knowledge domain a “singularly-linked cluster”, also named “discipline”. Scientific fields are good examples of knowledge domains, for example medical science, biology, chemistry or sociology. Knowledge according to content can often easily be covered by the question “what” and occasionally by the question “how”.
After determining knowledge content, it is very important to take into account the various forms of knowledge; the knowledge types (Polanyi, 1967; Pylyshyn, 1984; Boisot, 1995, 1998; Jorna, 2007). Leaving aside the many distinctions that can be found in literature (Cijsouw and Jorna, 2003), we suggest three knowledge types, classified along three non-orthogonal axes: sensory (ranging from rough to detailed), coded (from weak to strong), and theoretical (from concrete to abstract) knowledge. Together the types of knowledge can be depicted in a knowledge space (see Figs. a, b). Sensory knowledge forms the first dimension in the knowledge space. Sensory knowledge is the knowledge a person obtains using sensory organs. The knowledge is as concrete as the event that is interpreted. It is behavior. Examples of such knowledge are the knowledge of somebody’s face, the knowledge of bird songs or the skills and procedures one demonstrates in labor and performance. This dimension ranges from rough to detailed sensory knowledge. In detailed sensory knowledge, more fine-grained and specific sensory or behavioral aspects are present and relevant. Fig. : (a) The knowledge space: Five individuals (I1 to 5) on knowledge domain D at time T. (b) One individual’s knowledge types on knowledge domain D moving through time from T 1 to T 3 Coded knowledge is the second dimension in the knowledge space. Coded knowledge is the category that is formed on top of the knowledge of a concrete event (sensory knowledge). Coded knowledge means using signs. Words, diagrams, formulas and pictograms are all examples of codes. Coded knowledge forms a dimension that ranges from weak (icon or picture) to strong (mathematical formula) (Goodman, 1981/1968; Jorna, 1990).
The dimension from weak to strong is indicated based on decreasing ambiguity: the stronger the code, the less ambiguous the transferred knowledge is. Theoretical knowledge is the structure that can be formed on top of sensory and coded knowledge. All knowledge that reflects a structure, method, or pattern is theoretical. For example, natural laws and behavioral norms are theoretical knowledge, but ideological or religious coherent structures are theoretical knowledge as well. Theoretical knowledge can be made visible in asking and answering “why” questions.
This third dimension in the knowledge space ranges from concrete to abstract theoretical knowledge; concrete theoretical knowledge consists of small “why chains”, whereas abstract theoretical knowledge consists of long and complex chains. Note that theoretical knowledge is not used before coded knowledge has been acquired and that coded knowledge builds upon sensory knowledge in this structuring of knowledge types. Just as information and data can be made operational and quantifiable in terms of databases or in conceptual mathematical structures (i. e. Shannon and Weaver, 1963) , so can knowledge in this way be made operational (by using questionnaires to assess the types) in a knowledge space. Figure 3a depicts an example of the use of a knowledge space. It is an example of a static situation. In this knowledge space – a snapshot at time T – the individuals (I 1 to I 5), who are involved in a similar knowledge domain D, are positioned according to the knowledge types they use. Individual I 1 has knowledge type: highly theoretical, strongly coded and roughly sensory, whereas I 5 has the knowledge type: low theoretical, weakly coded and detailed sensory.
Accordingly, “corporate knowledge” can be depicted (not depicted in Fig. 3) involving a higher level of aggregation combining individual actors, knowledge content and knowledge types. In this way, given knowledge content, we can make a snapshot of the organizational knowledge of company A. It is also possible to depict individual (or organizational) development within the knowledge space, provided one makes various consecutive snapshots in time (see Fig. 3b). One individual develops knowledge from low theoretical and weakly coded at T 1 into highly theoretical and strongly coded at T 3.
Determining a specific knowledge content domain (e. g. , glassblowing, car production or childcare), the knowledge space makes it possible to assess and compare which individuals have which type of knowledge. It is also possible to accumulate the individual knowledge into a dominant knowledge type of an organization. The knowledge space can therefore be used as a measuring tool to assess the knowledge “state” of an individual, but also of a group or organization. In this way, we can operationalize organizational knowledge.
The assessment may then result in an organizational debate about knowledge use, knowledge distribution, and knowledge storage and knowledge accessibility. In combination with ICT, this is what McElroy (2003) called: first generation KM. Nevertheless, we still have to go into more details of the carriers of knowledge: what does our cognitively oriented view on actors (and agents) mean? Knowledge Carriers, Various Kinds of Actors, and Coordination Mechanisms 3 Characteristics of Cognitive and Other Kinds of Actors (Agents) 3 W. R. King (ed. ), Knowledge Management and Organizational Learning, 163 Annals of Information Systems 4, DOI 10. 007/978-1-4419-0011-1_11, © Springer Science+Business Media, LLC 2009 We already said that “actor” is the general term to talk about carriers of knowledge. Our focus is first on the individual cognitive actor. Within cognitive science, cognition of an actor consists of three essential characteristics: (a) a cognitive architecture, (b) representations and (c) computations or operations on representations. Newell and Simon (1972) stated that human thinking and reasoning consist of the manipulation of (internal) symbols (physical symbol system hypothesis).
Symbols are the basic constituents of our thoughts. They are the functional constituents of representations. They have a material carrier. For humans: their brain. A cognitive architecture implements the design and organization of the human mind. It presents various functions and properties of the mind and their interrelations. This concerns characteristics of functional components, such as memory, processing capacity, perception, motor systems and various kinds of central processors (Posner, 1989). Architecture without content is empty.
Representations are the content, the substantial knowledge in our cognitive system. They refer to books we have read, movies we have seen or experiences with our relatives and friends. All this knowledge and information exists in our memory in the form of representations: stories, icons, images, propositions, semantic nets and scripts (Jorna, 1990) . Operations on these representations are activities of symbol manipulation, such as combining them, forgetting, abstracting from and restructuring them. This general cognitive perspective of actors is implicitly present in our perspective on KM.
Putting aside for a moment the general cognitive structure, we consider an actor to be a coherent whole, consisting of several components. Within cognitive science, Posner (1989) formulated an extensive list of (cognitive) actor components. They include (a) perception, (b) interaction (including learning in the sense of habit formation), (c) representation and interpretation (including learning in the sense of (mental) knowledge formation and integration) and (d) autonomy and self-consciousness (Gazendam, 1993; Gazendam and Jorna, 1998). With perception, a system must be able to accept input in a general sense.
This input may include visible, audible and tangible stimuli and the accepting system may vary from a lobster, a human being to a software agent. Interaction is the process by which a system has contact with its environment. Stimuli as input in the system lead to output in the sense of responses. The reaction patterns of the system may result in learned behavior, that is to say that habits are formed. Representations and interpretation are necessary for a system that internally symbolizes the environment (Newell and Simon, 1972; Jorna, 1990). Based on representations and interpretations the system also learns.
Examples of representations are words, pictures, semantic nets, propositions and temporal strings (Kosslyn, 1980; Anderson, 1983). A system is autonomous, self-organized or self-conscious if it is able to have a representation of its own (physical and conceptual) position in the environment. This means that the system has self-representation. An autonomous system has reconstructing representational interaction patterns. In Fig. 4, taken over from the work of Helmhout (2006) , we depict a situation where two actors (having all the components) are interacting. Actor X and actor Y are examples of human actors.
Various combinations of aspects result in an actor hierarchy. An actor that only has perception is at the lowest level and cannot be called an intelligent actor, whereas an actor with selforganization, including perception, interaction and representations, is at the highest level. This last form is what we regularly call an actor that is reflective, intelligent and thoughtful. Human beings are presently the only instantiations of intelligent actors. It is debatable whether software agents at this moment have representations, but they certainly do not have self-organization, at least not the next years. We are aware of the fact that we are mixing up our earlier distinction in actor and (software) agent, but the fields of AI, cognitive science, computer science and economy are not consistent and constantly developing. In the remainder, we use as general term: actor. Environment Fig. 4: Actors, properties and multi-actor systems (from Helmhout, 2006) The above described classification in perception, interaction, representation and autonomy can be used to qualify various kinds of actors. We start with a cohesive, structured and organized entity.
In a sense this entity is an actor, because it is self-contained, strives toward continuation and, looking at the actor characteristics, it has perception and interaction including the possibility of learning in the sense of habit formation. We emphasize that this actor does not have internal representations. Its cognitive domain is absent or empty. We call this actor a Response Function system (RF-system), or Actor I (First Square in Fig. 5). We can compare it with the ant in the sand (Simon, 1998/1969).
In discussing complex behavior of systems, Simon stated that the behaviour of an ant on the sand could be called complex, not intelligent, because its behavior is a function of the complexity of the irregular environment that the ant has to cross. In the second place, we can define an actor that we call a Representational system (R-system). This actor has representations and is able to depict external events internally into its cognitive domain. We call this Actor II (Second Square in Fig. 5). This representational system has representation, to a certain extent autonomy and perception.
Interaction as we humans use it, is absent, that is to say that there is no device that semantically interprets causal inputs and outputs. Most present work in Artificial Intelligence, Knowledge Technology and Decision Support Fig. 5: Three kinds of actors Systems concern Actor II implementations, for example, ACT-R or SOAR. The focus here is mainly on the internal functioning of an “intelligent” system and very little on the interaction between this system and its environment. The third possible interpretation of an actor is the Representational Response Function actor (RRF-system).
This actor incorporates a really intelligent, interactive and cognitive system. We call this Actor III (Third Square in Fig. 5). This actor is able to perceive, to interact, to represent and to be autonomous. RRF-systems behave on the knowledge level, as Newell called it. “There exists a distinct (computer) systems level, lying immediately above the symbol level, which is characterized by knowledge as the medium and the principle of rationality as the law of behavior. ” (Newell, 1982, p. 99) Newell is proposing this knowledge level for natural (humans) as well as in the future for artificial (computers) intelligent systems.
Actors equipped with the integration of representations and responses have knowledge. “Knowledge”, says Newell “is whatever can be ascribed to an actor, such that its behavior can be computed according to the principle of rationality. ” (Newell, 1982, p. 105). We believe that it will not take a century before actors and agents are equivalent. Hopefully, this explains our confounded use of the terms actor and agent. Multi-Actor Systems and Coordination Mechanisms 4 The hierarchy of single actors returns in the composition of MAS. In a first multi-actor system, all actors are RF-systems (“empty actors”; see Fig. 6).
All actors have perception and interaction. To take up the example of Simon’s ant we are talking here about a group of ants 4 Larsson, R. , L. Bengtsson, K. Henriksson, and J. Sparks. 1998. The interorganizational learning dilemma: Lorenzoni, G. , and A. Lipparini. 1999. The leveraging of interfirm relationships as a distinctive organizational Capability perceiving and interacting with each other. Coordination is only defined in terms of reactions to the behavior of other actors. Fig. 6: RFS (multi-actor system) In the second place, we have a multi-actor system consisting of only representational systems (Actor II).
Every actor has internal representations in the sense of symbol structures and operations. Interaction is nearly absent for this kind of actors and if it exists it is of course not semantically or pragmatically meaningful (Fig. 7). In the third place, we may have representational response function systems in a multi-actor situation. The actors perceive each other and react to each other in a semantically and pragmatically rich and intelligent way (see Fig. 8). Each actor has perception, interaction, representation and autonomy and manages to integrate this into the organization as a multi-actor system.
A collection of human cognitive systems in a multi-actor perspective is an example of multiple representational response function systems. This is the situation of (human) organizations in practice. They consist of actors in the sense of actor III. In the fourth place, a combination of several kinds of actors is possible. Various MAS’s can be considered consisting of actors I and III, of actors II and III, and actors I, II and III. A combination of actors I and II seems difficult because we believe that at least one of the actors in a multi-actor system should have autonomy and self-organization.
Under the influence of developments in ICT more and more artificial actors (agents) will behave as actors III (see Fig. 9), which, to many people now seems horrifying. A multi-actor system necessarily requires a coordination mechanism to align, combine or integrate the various actors. We already indicated that for the moment we believe that for any such intelligently behaving MA-system at least one RRF-entity, an actor III, is necessary. This actor is the incorporation of a normal intelligent cognitive system. That being said, many MAS realizations can be made, depending on the used coordination mechanism.
Before discussing coordination mechanisms, we have to make a distinction between this mechanism and an organizational form. Fig. 7: RS (multi-actor system) Fig. 8: RRFS (multi-actor system) Fig. 9: Homogeneous and heterogeneous multi-actor systems An organizational form is a consistent structure of the cooperation (togetherness) of various actors (and agents). Examples of these forms are bureaucracies (professional or machine), networks (including webs and markets), clans and fiefs (Boisot, 1995). The reason why these forms exist and work is the coordination mechanism within these forms that makes the organization function.
In any organizational form, the actors have to understand the coordination mechanism, that is to say they need shared representations. For example in a bureaucracy, everybody has to understand and follow rules, procedures or norms. Moreover, in case of a clan, the members have to understand that there is one very important person who is in charge and has the authority; in most clans, based on family ties. However, family ties are not necessary. Clans are also found in R&D departments, where the authority is based on someone’s expertise and creativity.
Thompson (1967) distinguishes three major coordination mechanisms in MAS: standardization, planning and mutual adjustment. In using these mechanisms, a balance has to be struck between autonomous action and concerted action in order to gain an optimal performance (risk control, flexibility, learning) of the multi-actor system. Mintzberg (1983) distinguishes five coordination mechanisms that are used in five different types of organizations, i. e. simple structure, machine bureaucracy, professional bureaucracy, divisional structure, and adhocracy.
In a simple structure, direct supervision is used as coordination mechanism, whereas in the machine bureaucracy standardization (rules or procedures) of work processes is the coordination mechanism. Standardization of skills is the coordination mechanism used in professional bureaucracies. In a divisional structure, standardization of output is the main coordination mechanism used, and finally, people working in an adhocracy use mutual adjustment as coordination mechanism. If we look at Thompson’s, Mintzberg’s and many other distinctions in terms of knowledge, we can discern three basic coordination mechanisms.
The first is standardization, which presupposes that every actor knows the rule, procedure or norm and is able or obliged to act accordingly. The second is authority that can be based on age, expertise, official position or on family structures, like being a (god) father. In terms of knowledge, this requires that if one actor is the expert, the other actors acknowledge this and see themselves as having less expertise or authority. The third is mutual adjustment, meaning that beforehand no rule or authority does exist and that the actors have to negotiate or adjust in order to make the MAS function.
This requires knowledge of goals, preferences and constraints of the various actors. All these coordination mechanism and organizational forms are constructs. They are human made artifacts; they are social constructs that exist only because of our knowledge. Looking at the basic coordination mechanisms, it is evident that knowledge of what actors do and understand is essential in any multi-actor system. We also argued that knowledge can only be used meaningfully if at least one actor is a cognitively plausible system. This has consequences for the allowed combination of actors in MAS.
If knowledge is involved, we can think of two basic structures that are possible and two others that are impossible. At the moment, the most frequent MAS is one consisting of only humans (actor III). We will not repeat the various coordination mechanisms and organizational forms (also see Sorge and Warner, 2001). They are obvious. The second MAS that already occurs very often, is a combination of humans (actor III) and (software) agents (actor I or actor II). In the case of classical information systems, they function at the level of actor I and in combination with the authority of humans, we see them all around us.
In the case of actor II, we are talking about software agents that may reason, analyze or solve problems, but under the guidance or control of humans, at least in principle. Examples of such systems are expert and knowledge systems, advanced or dedicated decision support systems or other implementations of artificial intelligence systems. These MAS’s in whatever organizational form, use coordination mechanism such as rules, procedures and authority based on expertise, function or power. Whether the coordination mechanism of mutual adjustment also works here, if many actors are software agents, is debatable.
If software agents have self-representation and autonomy, it might be possible in the future. At the moment, it does not work. We also have two impossible basic structures of MAS’s, now. The first one only consisting of actors I and the second one consisting of actors II, in which some actors I may be involved. There is a very simple reason for this impossibility. A MAS consisting of actors I is just a behavioural system, without reasoning and representations, whereas a MAS consisting of actors II is lacking interaction possibilities and autonomy.
Actors II, consisting of AI-systems, knowledge technology and robots, require the presence of an actor III to be fully “cognitively plausible” and realistic. Without the inclusion of humans, the composition of such MAS does not function. The various kinds of actors (and agents) and various organizational forms and coordination mechanisms all presuppose the availability and accessibility of knowledge. As cognitive science already demonstrated, intelligence and rationality being present in reasoning, thinking, decision making and problem solving, presuppose that we humans are cognitive systems.
We have and manipulate representations consisting of signs and symbols. We do this internally, but we also use signs and symbols in our external interaction. Intra-individually as well as inter-individually, we are sign and symbol systems. In our minds, we use and work with sign and symbols, but also in our interaction, cooperation and communication we work with signs and symbols. The interesting point is that agents, software entities, are also sign and symbol using and interpreting systems (Newell and Simon, 1972).
Without starting a new conceptual discussion, we can say that covering the field of information systems, organizational structures and cognitive plausible actors, the term organizational semiotics was coined (Stamper, 1973; Liu, 2000). Semiotics, as the study of sign systems and sign interpretation in general, is thereby used to apply an existing conceptual framework to study MAS (Helmhout, 2006 , Helmhout et al. (2009) Sun, P. Y. T. , J. L. Scott, and D. McKie. 2005. Reframing and engaging with organizational learning constraints. In Current topics in management, Tsoukas, H. 003. Do we really understand tacit knowledge? In The Blackwell handbook of organizational learning and knowledge management , ed. L. Easterby-Smith, Cambridge, MA: Blackwell. Tsoukas, H. , and E. Vladimirou. 2001. What is organizational knowledge? Journal of Management Studies Conclusions and Future Directions The concept of “organizational knowledge” is important. The concept is used in intra- and interorganizational analyses and studies. By using the concept, an organization values its own and other organizations’ possibilities for development, innovation and cooperation.
However, because it is so important we wondered why so few quantifiable and measurable indicators of “organizational knowledge” have been developed. The same can be said for the KM issues in organizations. We believe that two important reasons exist for this situation. The first is that too much a top-down approach, from departments and processes to tasks and actors in KM, is practiced. The second is that as for actors and agents too little use is made of the availability of concepts and operational tools within cognitive science and artificial intelligence. For both shortcomings, we opened up the literature and suggested alternatives.
Instead of a top-down approach, we sketched a bottomup approach, starting with actors having various kinds of knowledge in their minds and we end with processes and organizations as MAS. Within this perspective, it is possible to measure and to quantify knowledge. Because knowledge is something until now only human actors have, we can study this knowledge by using mental maps, knowledge types in “knowledge spaces” and reasoning patterns of actors. Mental maps of the content of actors show what similarities and differences in knowledge various actors in organizations have.
Use of the knowledge space makes explicit whether actors use sensory, coded and theoretical knowledge and how the distribution of the various types is for the various actors with regard to specific domains. Studying reasoning patterns of actors makes accessible what they do in decision making, problem solving or planning. These tools and instruments can not only be used with regard to the primary or main processes in organizations (the processes expressing why organizations exist (making cars, teaching or curing), but also with regard to the organizational, executive or secondary processes.
We showed the availability of concepts and tools of cognitive science (CS) and artificial intelligence (AI) in the description of various characteristics of actors and as a consequence the variation in kinds of actors. Within CS and AI, we find extended discussions of components of actors, varying from perception and interaction to mental representations and reasoning. If actors do not have mental worlds at their disposal, they are empty actors. They do behave, but do not think or reason.
We showed that the sophisticated concepts and tools of CS and AI can be used when we discuss actors as the building stones of organizations. Again, we argue that in this way we are able to measure and quantify the carriers of information and especially knowledge. Organizations without human actors do not exist, as much as one human does not exist without the interaction with another human. This situation requires coordination and organization, not only as a leading principle, but also as an abstract entity and as a social construct. “Organization” as a structuring principle can be found in many oordination mechanism that we use. “Organization”, as an abstract entity, embodies a social construct and emphasizes the fact that as such an organization can only exist if we humans think about it. When the university as an organization closes down during the night and opens again in the morning, it does not mean that it did not exist in the night. It exists as buildings and other artifacts, but especially because during the night it exists in the minds of their members, whether they have it unconsciously, consciously or in their nightmares.
The perspective of cognitive science combined with the assumption that organizations are MAS’s make “organizational knowledge” and “organization” operational, measurable and quantifiable. Especially the focus on actor characteristics and as a result the actor/agent taxonomy being combined in a multi-actor system with various coordination mechanisms, makes it a suitable framework for the inclusion of (software) agents. Our future organizations in terms of hardware will consist more and more of combinations of brains, neural nets and electronic circuits, separated in different physical entities.
In terms of functional structures, or if one wants to call it that way: of software, we will more and more be interwoven and connected. That this functional structure does not have to be a vague, imprecise and abstract notion, we hope we showed in this article. In any case, artificial intelligent systems (agents) are here to stay and they will be more and more integrated with human cognitive systems. The big challenge for the future is to make them semantically and pragmatically more adaptable to us, so they can help us in sustaining a human future (Harper et al, 2008) . References Alvesson M. & D. Karreman. 2001.
Odd Couple. Coming to terms with knowledge management. Anderson, J. R. 1983. The architecture of cognition . Cambridge, MA: Harvard University Press. Antonacopoulou, E. P. 2006. The relationship between individual and organizational learning: New evidence from managerial learning practices. Argyris, C. , and D. Schon. 1978. Organizational learning: A theory of action perspective . Reading, Boisot, M. H. 1995. Information space: A framework for learning in organizations, institutions, and culture . London: Routledge. Boisot, M. H. 1998. Knowledge assets: Securing competitive advantage in the information economy .
New York: Oxford University Press. Bukowitz, W. , and R. Williams. 1999. The knowledge management fieldbook . London: Prentice Hall. Cijsouw, R. S. , and R. J. Jorna. 2003. Measuring and mapping knowledge types. In Dynamics and change in organizations: Studies in organizational semiotics , ed. H. W. M. Gazendam, R. S. Cijsouw, and R. J. Jorna Dalkir, K. 2005. Knowledge management in theory and practice . Amsterdam: Butterworth-Heinemann. Fearon, D. S. , and S. A. Cavaleri. 2006. Inside knowledge. Rediscovering the source of performance improvement . Milwaukee: Quality press. Firestone, J. and M. McElroy. 2003. Key issues in the new knowledge management . Burlington: Butterworth Heinemann. Gazendam, H. W. M. 1993. Variety controls variety: On the use of organizational theories in information management . Groningen: Wolters-Noordhoff. Gazendam, H. W. M. , and R. J. Jorna. 1998. Semiotics, multi-agent systems and organizations. Meyer, M. , and M. Zack. 1996. The design and implementation of information products. Sloan Mintzberg, H. 1983. Structures in fives: Designing effective organizations . Englewood Cliffs: Prentice Hall. Newell, A. 1982. The knowledge level.
Artificial Intelligence Nonaka, I. , and H. Takeuchi. 1995. The knowledge-creating company: How Japanese companies create the dynamics of innovation . New York: Oxford University Press. Peters, K. , L. Maruster, and R. J. Jorna. 2009. Knowledge evaluation in organization (accepted for publication). Simon, H. A. 1998/1969. The sciences of the artificial . Cambridge, MA: MIT Press. Sorge, A. and M. Warner. 2001. The IEBM Handbook of Organizational Behavior . London: The End Group No. 4 (PGDM – G) Name Group Members Harshad Vyas Om Prakash Suthar Bhawani Singh Rathore Amit Mathur Gourav Rathi